Datasets:

health360

/

Healix-Shot

like 5

Follow

Health 360 7

loose or watery stools in the past 24 hours), upper respiratory tract infection (URTI) and fast breathing suggestive of pneumonia. Any current episode of upper respiratory tract infection (URTI) and acute watery diarrhoea was also recorded. The grading of protein energy malnutrition (PEM) was done using the Indian Academy of Pediatrics (IAP) classification [13] with the help of anganwari personals. The IAP classification is based on expected weight of the child for that age. Severe PEM is defined as grade III (expected weight for age between 51 to 60%) and grade IV (expected weight for age less than 50%), with or without oedema. Ethical considerations The project officer, Ministry of Women and Child Development, Ujjain approved the study. The ethics committee of RD Gardi Medical College granted ethical approval for the study (approval number 42/2007). Sample collection The child's head was tilted back gently and steady from the chin. Sterile cotton swabs pre-wetted with sterile saline were rotated against the turbinate of both anterior nares of each participating child. Both swabs were inserted into a tube of Amies transport media with charcoal (HiMedia, Mumbai, India) and transported to the microbiology laboratory at Madhav Science College, Ujjain. A range of temperature between 4 to 8°C was maintained during transport. Microbiological laboratory method In the microbiology laboratory swabs were plated on 5% sheep blood agar. Colonies of S. aureus identified by the typical colony morphology, Gram's staining, biochemical tests for anaerobic utilization of glucose and mannitol, catalase production and tube coagulase test [14]. Antibiotic susceptibility tests were performed using Kirby-Bauer's disc diffusion method according to performance standards of CLSI [15]. Screening for methicillin resistance was done using cefoxitin disk screen test and 6 μg/ ml of oxacillin in Mueller-Hinton agar supplemented with NaCl (4% w/v; 0.68 mol/L) according to Clinical and Laboratory Standard Institute (CLSI) guidelines [15]. S. aureus ATCC 25923 was used as control strain. Due to limitation of resources the following restricted panel of antibiotics was selected: ampicillin, ceftriaxone, ciprofloxacin, levofloxacin, ofloxacin, gentamicin, doxycycline, tetracycline, co-trimoxazole and vancomycin. For both methicillin-sensitive S. aureus (MSSA) and MRSA we defined multi-drug resistant (MDR) isolates as those resistant to three different antibiotics groups [16]. Statistical analysis The data was entered in an Excel spreadsheet and then transferred to Stata 10.0 (Stata Corp. College Station, Texas, USA) software for statistical analysis. Prevalence of S. aureus and MRSA were estimated with 95% confidence intervals. The relationship between each variable and the outcome (nasal carriage of S. aureus) was explored using odds ratios (OR). Crude OR's were calculated from two by two tables. A given variable was entered in the final multiple logistic model if the bivariate analysis yielded a P value less than 0.1. All the variables were adjusted for age and sex. Results In the study 1002 children were enrolled from 100 preschools. Out of them 51% (n = 514) were boys and the remaining 49% were girls. Most children (37%, n = 374) belonged to the age group of 2 to 4 years of age, and to a family size of between 4 to 10 members (58%, n = 585) ( Table 1). Thirty-six percent (n = 365) of the children's parents or caretaker of the children gave a history of any acute illness (like LM, URTI and fast breathing suggestive of pneumonia) in the past two weeks. Twenty-seven percent (n = 268) children had PEM grade III or IV (severe malnutrition). Nasal carriage of S. aureus and MRSA Out of 1002 children included in the study a total of 351 children were culture positive for S. aureus. Thus, the prevalence of S. aureus nasal carriage was 35% (95% confidence interval CI 32.07 to 37.98). Out of the 351 S. aureus isolates, 102 isolates were methicillin resistant S. aureus (MRSA). Thus, the prevalence of MRSA nasal carriage was 29% (95% CI 24.28 to 33.88). Factors associated with nasal carriage of S. aureus The factors that were independently associated with nasal carriage of S. aureus were: "age-group" i.e. as the age increased beyond the age of 2 years the OR of nasal carriage decreased, "family size of more than 10 members" OR 2.59 (95% CI 1.53-4.37; P < 0.001), and PEM Grade 3 or 4 (OR 1.40, 95% CI 1.04-1.90; P = 0.026) ( Table 1). Antibiotic susceptibility pattern of S. aureus isolates The antibiotic susceptibility pattern of S. aureus to individual antibiotics is shown in Figure 1. Their resistance pattern showed that co-resistance was highest for combination of tetracycline and gentamicin (30%), followed by doxycycline and gentamicin (26%) and cotrimoxazole and gentamicin (18%) ( Table 2). The MDR pattern of S. aureus isolates is shown in Table 3. The commonest pattern of multi-drug resistance was found for a combination of gentamicin, tetracycline and cotrimoxazole (n = 62, 18%). Antibiotic susceptibility pattern of MRSA isolates The results of antibiotic susceptibility patterns for MRSA isolates for ciprofloxacin, ofloxacin, levofloxacin, gentamicin, doxycycline, tetracycline, co-trimoxazole and vancomycin are shown in Figure 2. The commonest pattern of co-resistance was for tetracycline and gentamicin (n = 62, 61%). Table 4 shows the MDR patterns for MRSA. The commonest MDR pattern for MRSA was for a combination of doxycycline, gentamicin and cotrimoxazole (n = 44, 43%). Discussion In this study 1002 children aged between one to six years were enrolled. The prevalence of S. aureus nasal carriage was 35% and that of MRSA 29%. The nasal carriage decreased as the aged increased beyond two years. Children living in larger families and those having severe PEM had higher nasal carriage of S. aureus. The resistance pattern of S. aureus and MRSA showed resistance not only to single antibiotic class but co-resistance and multi-drug resistance was also common. A study done in the same geographical location among 1562 healthy children from 1 to 59 months of age, 6% of the children tested positive for nasal carriage of S. aureus [11]. The factors associated with nasal carriage were "child attending preschool" (OR 4.26, 95% CI 2.25-8.03; P = 0.007) or "school" (OR 3.02, 95% CI 1.27-7.18; P < 0.001) and "family size more than 10 members" (OR 2.76 95% CI 1.06-7.15; P = 0.03) [11]. The OR of nasal carriage in a child attending preschool were 4 times and the present study was done among the "preschool children" the prevalence is expected to be higher. In a community-based study done in Mangalore, southern India on 250 patients of pyoderma the nasal carriage rate of S. aureus and MRSA was 54% and 12% respectively. However, only half of the patients were children less than 10 years of age [17]. In another study done in north India among children of age group 5 to 15 years living in slums, the nasal carriage rate of S. aureus and MRSA was 52% and 4% respectively [18]. A statistically significant correlation of age with nasal carriage of S. aureus was noticed in our study which is similar to that reported in other low-middle income settings. In a study from Taiwan [19] the carriage rate of MRSA was higher among 2 to 6 months old children. In a study done in an elementary school in Seoul, Korea the prevalence of nasal carriers was found to be higher in younger children (≤7 years) (mean 69%) than that in older children (mean 47%) [20]. In relation to nasal carriage of S. aureus and sex of the child are not consistent. Another study from Taiwan showed that nasal carriage was higher for female compared to males and also increased with age [21]. In contrast a study from Lebanon showed increased risk of carriage among males [22]. Other studies have however, not found statistically significant association between sex of the child and nasal carriage of S. aureus [11,23]. In our study it was observed that when the family size increased the prevalence of nasal carriage of S. aureus increased as well. Similarly, in a study from Taiwan [19] MRSA colonization was associated with the number of children in the family (adjusted odds ratio [aOR], 1.114) and day care attendance (aOR, 1.530). Association of family size with nasal carriage of S. aureus is probably due to overcrowding and greater sharing of nasal flora within a large family. Prevalence of nasal carriage of a The co-resistance shows the pattern of resistance for the given two antibiotic groups (one in the row and other in the column). The figures in the bold represent the common pattern of co-resistance involving different antibiotics. MRSA has been shown to be higher among household contacts of patients with community onset MRSA disease with significant strain relatedness among the index cases and contacts [24]. Likewise, in a study done on healthy postmenopausal women in Ujjain, India one of the factor significantly associated with carriage of MDR Escherichia coli was a family size of more than 10 members (OR 8.23, 95% CI 2.73-24.73; p < 0.001) [25]. To our knowledge, a relationship between URTI and nasal carriage has not been studied in any Indian study before. Our study did not find any statistically significant association, however studies from other settings have reported increased spread of S. aureus during an episode of URTI [10]. The effect of PEM on nasal carriage of S. aureus in children has not been reported before. In severe PEM, acquired immunity-i.e., lymphocyte functions-as well as innate host defense mechanisms-i.e., macrophages and granulocytes-are affected [26]. Malnutrition causes atrophy of the thymus [27]. Suppression of the delayed cutaneous hypersensitivity, decreased helper T cells, impaired secretory immunoglobulin A antibody response, decreased antibody affinity, reduced concentration and activity of complement components and phagocyte dysfunction [27]. Also, there is the appearance of immature T cells in the circulation [27][28][29]. Because of the mechanisms discussed above malnourished children suffer in greater proportion from respiratory infections, infectious diarrhea, measles and malaria [26,28]. The infections are also characterized by a protracted course and exacerbated disease [29]. Because of the low secretory IgA levels in malnourished children the mucosal response to pathogens such as rotavirus and E.coli in the intestine and measles virus in the nasopharynx are found to be impaired [28]. Therefore, impaired immune response might be responsible for increased nasal carriage of S. aureus observed in the present study. In the present study we found higher proportion of resistance for commonly used antibiotics then that reported in a previous study [11]. The percentage of resistance to commonly used antibiotics in our study was higher e.g. tetracycline (41%), doxycycline (39%), gentamicin (32%), ampicillin (29%) and cotrimoxazole (28%). In a study from Turkey done among 5 to 7 year-old healthy children attending day care center the resistance pattern was reported for erythromycin, clindamycin, fusidic acid, and tetracycline to be 16.6, 8.3, 5.6, and 8.3%, respectively [30]. The co-resistance pattern also demonstrates change in the resistance pattern over time in the same geographical location as the present study [11]. Antibiotic use is one of the most important determinants of antibiotic resistance [5]. A high antibiotic use rate has been reported both in the outpatients [6] and among admitted patients [31] in the same geographical area i.e. Ujjain. Antibiotic stewardship programmes that promote judicious use of antibiotic are thus urgently needed. Other strategies to prevent community spread of S. aureus include targeted screening among hospitalized patients based on risk factors, isolation of the carriers and decolonization. Our study has limitations; we selected a limited pool of antibiotics for susceptibility testing and did not do molecular studies to confirm MRSA isolates due to financial constrains. We could not evaluate drugs given especially antibiotics given to the children. The dynamics of transmission of S. aureus in a large family needs to be studied in future using molecular diagnostics and might shed new light on the subject. Conclusions The high rate of nasal carriage of S. aureus and MRSA and presence of resistance to commonly used antibiotics is disturbing. Antibiotic stewardship programmes that promote judicious use of antibiotic along with strategies to prevent community spread of resistant bacteria like the ReAct's Civil Society Organization project (http:// cso.reactgroup.org) are urgently needed. The Sociodemographic and Clinical Characteristics of Occupational Accident Victims Followed in Physical Therapy and Rehabilitation Department of a Tertiary Hospital ABS TRACT Objective: The aim of this study is to evaluate the so- ciodemographic, injury and employment characteristics of the patients included in a physical medicine and rehabilitation program after

an occupational accident (OA). Material and Methods: The study included 102 patients who were admitted to our clinic for a rehabilitation program after an OA between January 2012 and December 2018. Electronic media files in the hospital automation system were retrospec- tively scanned. We recorded the sociodemographic, injury and employment characteristics for each patient. Results: The mean age of 102 patients included in the study was 36.39±10.38 (min: 18, max: 59). One hundred (98%) of them were male, two (2%) were female. 18.6% of the patients are between the ages of 18-25; 42.2% were between the ages of 26-40 and 39.2% were over the age of 40. The mean number of physical therapy sessions was 42.89±25.53 (min: 15, max: 180). OAs mostly occurred in the construction industry, most fre- quently as a result of falling from a height, followed by sharp object injuries. Of all patients, 48% (n:49) were primary school graduates. It was found that most sequela was observed statistically significantly after upper limb injuries. The most common sequelae was contracture with 45.1% (n: 46). Conclusion: The results of our study showed that the primary school graduate young males working in the construction and metal industry sectors were mostly affected by OAs. It was found that many of the patients in need of physical therapy and rehabilitation programs recovered with sequelae despite all treatment applications. Anah tar Ke li me ler: İş kazaları; rehabilitasyon; komplikasyonlar DOI: 10.31609/jpmrs.2019-71471 ORİJİNAL ARAŞTIRMA ORIGINAL RESEARCH and 1405 in Social Security Institution of Turkey (SSIT) statistics, respectively. 2 Additionally, 374 million non-fatal OAs and WRDs occur each year, many of them resulting in extended absences from work. 1 In addition to substantial economic losses, WRDs and work-related injuries bring along significant individual and social burdens. 3 The injuries may cause serious disabilities, workforce losses and even mortalities. Many patients who experienced an OA face numerous problems in the long term period and return-to-work (RTW) may get complicated for them. Our priority is to carry out occupational safety and accident prevention activities systematically to prevent OAs and subsequent financial and emotional damage. In unpreventable situations, physical therapy and rehabilitation applications are important, especially in facilitating RTW and decreasing the morbidities. The longer an injured worker is away from work, the lower possibility of a successful RTW is. 4 Physical therapy and rehabilitation practices facilitate RTW for the injured workers and help them maintain their functional and cognitive capacity. 5 The purpose of this study is to increase data related to OAs through a retrospective analysis of the demographic and clinical characteristics of the patients included in a physical therapy and rehabilitation program. Thus, it will be possible to draw attention to the injuries and rehabilitation process after OAs and in order to start rehabilitation programs earlier, to increase the cooperation of the physical medicine and rehabilitation department with orthopedics, plastic and reconstructive surgery and neurosurgery. MATERIAL AND METHODS The study included 102 patients who were admitted to our clinic for a rehabilitation program after an OA between January 2012 and December 2018. The study was approved by the Local Ethics Committee of Tokat Gaziosmanpaşa University (20.02.2019/no. 19KAEK-035). It was performed in compliance with the principles of the Declaration of Helsinki. The inclusion criteria of the study were determined as patients who had an OA over the age of 18, whose accident was reported to the SSIT and who were included in a physical therapy and rehabilitation program after an OA. The electronic health records of the patients in the hospital database were retrospectively scanned. The sociodemographic, injury and employment characteristics of the patients were recorded, symptoms not related to OAs were ignored. Sociodemographic characteriSticS This group includes gender, age, marital status and educational levels of the patients. Gender is a categorical variable with two categories: female and male. The age of injured workers is categorized as follows: 18-25 years, 26-40 years, 41 years and above. For providing more reliable information, age groups were used rather than single-age values. Marital status is also a categorical variable with two categories: single and married. Educational status has been assessed in four categories: primary school, secondary school, high school and college. injury characteriSticS This group includes bodily location, injury type, cause of injury, need for surgery after injury, presence of sequelae, post-injury ambulation levels and the number of physical therapy treatment sessions. Injury type has five categories: amputations, tendon lacerations, non-vertebral fractures, vertebral fractures and soft-tissue injuries. Bodily location is categorized as follows: upper extremity, lower extremity and axial skeleton. The cause of injury has six categories: falling from a height, sharp object injuries, electric shocks, burning, traffic accidents and chronic mechanical loading. employment characteriSticS In this group, occupational sectors where the OAs occurred were evaluated. The sectors were divided into five categories: construction industry, textile sector, metal industry, transportation sector and other service industry. StatiStical analySiS The SPSS 19 (IBM Co., Somers, NY) was used for statistical analysis. Descriptive analysis were made to give information about the general characteristics of the study groups. The data of the continuous variables were given as mean±standard deviation and median [interquartile range]; data for categorical variables were given as n (%).The means of the quanti-tative variables among the groups were compared by using independent samples t test, One-Way ANOVA for normally distributing data and Kruskal Wallis test for non-normally distributing data. Cross tabulations and the Pearson chi-square Test were used. p <0.05 were considered to be statistically significant to evaluate the relationship between qualitative variables. RESULTS One hundred and two patients were involved in the study. The mean age of the patients was 36.39±10.38 years (min: 18, max: 59). One hundred of the patients (98%) were male, two (2%) of the patients were female. The patients were divided into three categories by ages: 19 (18.6%) of the patients were between 18 and 25 years, 43 (42.2%) were between 26 and 40 years, and 40 (39.2%) were 41 years and above. Table 1 shows the sociodemographic characteristics of the patients. The injury and employment characteristics of the patients were shown in Table 2 and Table 3, respectively. The sharp object injuries with a ratio of 47.1% (n:48), falling from a height with a ratio of 40.2% (n:41) were the most common causes of OAs. The other causes of OAs are shown in Table 2. Ten (9.8%) of the patients had an amputation and all of these patients had an injury with a sharp object in etiology. Eighty-three (81.4%) of the patients included in our study had one or more surgical procedures due to OAs. There was no statistically significant difference among the occupational groups in terms of surgical history (p: 0.170) ( Table 2, Table 3). All patients with tendon lacerations (n:27) and amputations were operated due to these injuries. Ten patients covered with no sequela left ( Table 2). The number of the patients fully independent and ambulatory without the need for any support was 93 (91.2%). However, three (2.9%) patients were ambulatory with a wheelchair. These three patients had experienced vertebral injuries ( Table 2). The mean number of the physical therapy sessions received by patients was 42.89±25.53 (min:15, max:180). There was no difference among different occupational sectors regarding the duration of the rehabilitation programs (p: 0,799). However, the patients who underwent surgery experienced a longer duration of rehabilitation (p: 0,003). The patients with amputation and vertebral fractures required the longest period of rehabilitation (Table 4). DISCUSSION The economic cost of OAs and WRDs reaches 3.94% of the world national income. Therefore, if an OA occurs despite all preventive measures, it is important to implement the most appropriate treatments and rehabilitation programs for the patients in the earliest period. 6 Multidisciplinary approach after an OA is of great importance in the success of treatment. Rehabilitation practices are an important component of this multidisciplinary team. Physical therapy plays a key role in restoring the ability of the affected person to perform daily living activities independently and providing RTW. This study evaluated sociodemographic and clinical characteristics of 102 patients who received physical therapy and rehabilitation after an OA retrospectively. In their retrospective study, Serinken et al. documented 746 patients who experienced a hand injury after an OA and the mean age of the patients was 27.8±6.1 years (ranged 16 to 46 years). The number of the male patients was 87.2% (n:213) and 57.0% (n:139) of them were between 25 and 34 years of age. Furthermore, the metal industry was the first with 41.4% (n:101), followed by the textile sector with 16.8% (n:41) and mining-construction industry with 14.7% (n:36). 7 Karakurt et al. attributed several reasons for the male predominance (96.6%) in OAs: 80% of all insured workers in Turkey are male. They work in more dangerous sectors hence they experience more OAs. 8 Çelik et al. reported that 92.4% of the patients with OA were male, while Erdemli et al. found the ratio to be 73%. 9,10 SSIT statistics stated that OAs mainly occurred in males (84.2%) in Turkey. 2 In our study, the mean age was 36.39±10.38 (min: 18, max: 59). Of 102 patients (100 male; 2 female); 18.6% were between 18 and 25 years of age, 42.2% were between 26 and 40 years and 39.2% were 41 years and above. The frequency of male workers who experienced an OA might stem from several reasons: men are more predominantly involved in working life and dangerous professions. Female workers behave more carefully at work and mainly work as unregistered day laborers. In the study by Serinken et al., the mean age was lower than our study group. 7 This may be due to the difference of the region where the study is conducted. Their study was carried out in an industrial region with plenty of job opportunities at earlier ages. Our findings on the mean age, the distribution by age groups, and the age range where OAs are the most prevalent were in close correlation with the literature. [11][12][13] Low educational levels lead to an increase in OAs because employees with a low level of education accept working in more heavy and dangerous works and they are more exposed to occupational risks. Low levels of education even prevent recognition of existing occupational risks. 14 In a study by Jafari et al., it was found that workers with a non-academic education were at an increased risk of OAs. 15 Because these individuals had jobs that required a lot of exercise and physical effort. Also they were generally less well prepared to learn safety practices in the workplace or to follow safe working procedures. [16][17][18] In our study, 48% (n: 49) of the patients were primary school graduates. Only 7.8% (n: 8) were college graduates. In the construction industry, where OAs occur most frequently, 58.2% (n: 32) of the employees were primary school graduates, while 37% (n: 10) of the employees in the industrial sector were primary school graduates. These sectors require specific attention, training, and knowledge. The low education levels of the employees may be the reason why OAs are most frequently seen in these two sectors. Increasing the frequency of vocational training and inspection may be a solution proposal to reduce OAs. In different studies, it was found that the incidence of OAs varies by months and industry. 8,9 However, in our study, it was found that the applications to our clinic due to OAs did not differ by month. This may be due to the fact that patients who have had an OA apply to the physical therapy clinics for rehabilitation, not immediately after the accident, but for later periods. Sayhan et al. reported that most of the employees who had an OA work in the manufacturing and construction sectors. 19 Ada et al. found that the employees most frequently injured were the unqualified ones. 20 Çelik et al. stated that OAs occur most frequently in the industrial sector (26%) and the construction industry (28.7%). 9 In our study, 53.9% (n:55) of the patients were in the construction industry. Almost one-quarter of the patients included in the study were employed in the metal industry. These results were consistent with the previous studies. The reason for this distribution may be that the construction industry and metal industry sectors involve dangerous jobs that require high attention. Schloenfisch et al. and Ulutaşdemir et

al. found that the most frequently injured anatomic regions were upper extremities with 40% and 61.75% (n:82) respectively. 21,22 In our study, 56.9% (n:58) of the injuries were upper extremity injuries, 29.4% (n:30) of them were lower extremity injuries and 13.7% (n: 14) were vertebral fractures. This frequency of upper extremity injuries was expected since hands are the most commonly used parts of the body. Çelik et al. found that lacerations were the most common injuries (36.4%). 9 Serinken et al. reported that the most common injuries were lacerations, penetrations, amputations and avulsions. 7 According to SSIT data, sharp object injuries take the first place with 15.2% and crash-related injuries rank number two with 10.5%. 2 In our study, the most common injury was non-vertebral fractures with 42.2% (n:43) followed by tendon lacerations with 26.5% (n:27). Akbarzadeh et al. reported that the duration of hospital stay served as a proxy indicator for the severity of the injury. 11 In our study, it was found that the patients who underwent surgery after OAs required a longer duration of rehabilitation. The type of injury was decisive in the duration of rehabilitation and vertebral fractures required the longest period of rehabilitation. The superiority of our study is that it is the first study in the field of physical therapy and rehabilitation analyzing the sociodemographic and clinical characteristics of the victims after an OA. LIMITATIONS There are a number of limitations of our study. First, it was a cross-sectional and retrospective study. Due to the design of the study, the precautions taken in the workplace where the patients had an OA and the knowledge levels of the patients about the OAs and work safety could not be evaluated. In addition, our study was single-centered and performed in a tertiary hospital, therefore our data did not cover the entire society. Only patients who needed physical therapy and rehabilitation after an OA were included in the study. Therefore, we do not have any data on overall morbidity and mortality of OAs. CONCLUSION The results of our study showed that the primary school graduate young males working in the construction and metal industry sectors were mostly affected by OAs. It was found that many of the patients in need of physical therapy and rehabilitation programs recovered with sequelae despite all treatment applications. For this reason, it is important to take measures to prevent OAs and to increase educations on occupational safety. Detection of primary sites in unknown primary tumors using FDG-PET or FDG-PET/CT Background Carcinoma of unknown primary tumors (CUP) is present in 0.5%-9% of all patients with malignant neoplasms; only 20%-27% of primary sites are identified before the patients die. Currently, 18F-fluorodeoxy-glucose positron-emission tomography (18F-FDG PET) or PET combined with computed tomography (PET/CT) is widely used for the diagnosis of CUP. However, the diagnostic yield of the primary site varies. The aim of this study was to determine whether PET or PET/CT has additional advantages over the conventional diagnostic workup in detecting the primary origin of CUP. Findings Twenty patients with unknown primary tumors that underwent PET or PET/CT were included in this study. For all patients, the conventional diagnostic workup was unsuccessful in detecting the primary sites. Among 20 patients, 11 had PET scans. The remaining nine patients had PET/CT. In all 20 patients, neither the PET nor PET/CT identified the primary site of the tumor, including six cases with cervical lymph node metastases. The PET and PET/CT revealed sites of FDG uptake other than those associated with known metastases in seven patients, but these findings did not influence patient management or therapy. Two patients had unnecessary invasive diagnostic procedures due to false positive results on the PET or PET/CT. Conclusions Although it is inconclusive because of small sample size of the study, the additional value of PET or PET/CT for the detection of primary sites in patients with CUP might be less than expected; especially in patients that have already had extensive conventional diagnostic workups. Further study is needed to confirm this finding. Introduction Carcinoma of unknown primary tumors (CUP) is a biopsy-proven malignancy in which the anatomical origin of the tumor cannot be identified from the patient history, physical examination, laboratory testing, chest radiographs, computed tomography of the chest, abdomen and pelvis, and (in women) mammography [1]. CUP is present in 0.5-9% of patients with malignant neoplasms; however, only 20-27% of primary sites are identified before the patients die [2]. Some studies have reported that although the median survival time of patients with CUP is less than 1 year, if the primary site is identified and specific therapy started, the survival time can be increased [3,4]. However, primary tumors are detected in less than 40% of patients by conventional diagnostic procedures, even when multiple examinations are performed [1]. Currently, positron-emission tomography (PET) with 18F-fluorodeoxyglucose (FDG) or PET combined with computed tomography (PET/CT) is widely used in the diagnostic evaluation of patients with CUP [5]. The rate of detection of the primary site varies; 24.5-41% for the FDG-PET [6,7] and 22-73% with the FDG-PET/CT [8]. These variable diagnostic yields might be due to different patient inclusion criteria and the extent of the diagnostic workup in different studies. Therefore, the efficacy of PET or PET/CT for the detection of primary sites in patients with CUP remains to be determined. Furthermore, there is concern about the false positive PET results in regions endemic for TB [9]. Therefore, the aim of this study was to determine whether PET or PET/CT had additional advantages over conventional Patients The medical records of patients with CUP that underwent PET or PET/CT imaging were reviewed retrospectively. All patients were admitted to the Seoul National University Hospital for further evaluation between January 2003 and September 2005. Carcinoma of unknown primary tumor was defined as a biopsy-proven malignancy whose anatomical origin could not be identified by a conventional diagnostic workup (history, physical examination, laboratory tests, chest radiography, CT of the chest, abdomen and pelvis, MRI of the suspected lesion, endoscopic examinations where indicated, and, in women, mammography). All patients had biopsy-proven malignancies and the results of conventional diagnostic examinations were negative. The workup performed was determined based on the histological results, and therefore, the procedures used to detect the primary sites of tumors differed among the patients. PET or PET/CT imaging All patients underwent whole-body 18F-fluorodeoxyglucose positron-emission tomography (18F-FDG PET) or PET/CT scans according to the following procedure. Patients were fasted for at least 8 h before receiving an intravenous injection of 555-740 MBq(15-20 mCi; 0.22 mCi/kg body weight) of 18F-FDG. The uptake period was 60-90 min. The PET was performed on a dedicated PET scanner with a 5-min emission acquisition per imaging level. Attenuation correction was performed using the CT technique (140 kV, 80 mA) in the case of the PET/CT. PET images were reconstructed with a 128 × 128 matrix, an ordered subset expectation maximum iterative reconstruction algorithm (six iterations, 16 subsets), a 2-mm Shepp filter and a 16.2-cm field of view. PET/CT images were reconstructed with a 144 × 144 matrix and a 3 D row action maximum likelihood algorithm (two iterations, 0.006 relaxations). The results of PET or PET/CT scans were evaluated by two experienced nuclear medicine physicians that were unaware of the histology of the metastatic sites. Evaluation 'Detection of the primary tumor using PET or PET/CT' was defined when additional information about the primary tumor was revealed by PET or PET/CT imaging. Although the suspected primary site was seen on the PET or PET/CT, it was not considered 'detection by PET or PET/CT' if the suspected primary site was seen on other imaging modalities such as the CT. When the FDG uptake site in the PET or PET/CT was confirmed as a benign lesion, this was defined as a 'false positive' PET or PET/CT result. Among the 20 patients, 11 underwent 18F-FDG PET scans, four of whom also underwent PET/CT scans several months after the initial PET scan was performed. The remaining nine out of the 20 patients only had a PET/CT. Data on the 20 patients are shown in Table 2. Neither the PET nor PET/CT detected the putative primary site of the metastatic tumor in any of the 20 patients. In six cases, the initial presentation was cervical lymph node metastasis. Neither the PET nor PET/CT detected the primary site of cervical lymph node metastases. A 54-year-old male (patient no. 9) presented with right cervical lymph node metastasis from an unknown primary tumor. The patient had a radical neck dissection, unilateral tonsillectomy, blind biopsy of the nasopharynx and the tongue base. The pathology revealed metastatic squamous cell carcinoma in one out of 20 cervical lymph nodes. However, there was no evidence of malignancy in the other tissues including tonsil, tongue base, parotid gland, salivary gland, and nasopharynx. The PET/CT showed a hypermetabolic lesion (SUV 13.0) in a right cervical lymph node, at level II. However, there was no additional FDG uptake suggesting a primary site (Figure 1). These findings contrast with those of previous studies [6,[10][11][12] that demonstrated the efficacy of PET for localizing primary sites of cervical lymph node metastases. In this study, not even the PET/CT was able to localize the primary site of the cervical lymph node metastases. The PET or PET/CT revealed FDG-uptake lesions other than the known metastases in seven out of 20 patients. Five lesions were confirmed as benign (false positive results), and two were pathologically confirmed as another metastatic lesion after biopsy. Three out of five false positive cases (patients no. 4, 7, and 17) also displayed FDG uptake by the thyroid or pharynx (standardized uptake value [SUV] = 2.9-7.1); in patients where the initial physical examinations showed normal thyroids and pharynxes. Clinically, these thyroid glands and pharynxes were not considered to be the primary tumors and did not exhibit malignant changes during the follow-up period. In two out of five false positive cases, the results of the scans were initially thought to show the primary tumors and a diagnostic workup of these patients was expanded to include invasive procedures. For example, one patient (patient no. 5) with a metastatic adenocarcinoma of the skin had additional FDG uptake around the mid-esophagus (SUV 8.9). To confirm this lesion, the patient underwent a repeat esophagogastroduodenoscopy (Initial result of endoscopy was negative.); however, there was no evidence of a malignancy in the esophagus. A mid-esophagus lesion was observed as subcarinal lymph node uptake on a subsequent chest CT. Because the follow up PET/CT scan showed decreased size and FDG uptake of the subcarinal lymph node, this lesion was confirmed to be a benign lesion. Another patient, with metastatic leiomyosarcoma of the brain (patient no. 10) had mild hypermetabolic findings in the lower left lung field. Because a malignancy could not be ruled out, the patient underwent bronchoscopy and a chest CT. There was no evidence of a malignancy. Because of the false positive PET result, the patient had unnecessary invasive procedures. In two patients (patient no. 13 and 20) out of the seven that had additional FDG uptake, other metastatic lesions were confirmed by pathological examination of biopsies. Because these metastatic sites were just additional, the management plans of these patients did not change. The results with additional FDG uptake did not positively influence the management and therapeutic plans of the patients. In four cases, the PET/CT was performed after the initial PET scan (patient no. 3, 5, 10 and 15). The PET/ CT, which is anatomically more accurate than the PET, did not confer any additional advantage in the detection of the primary sites of patients with CUP. Discussion Detection of the primary tumor can change the prognosis of patients with CUP by enabling targeted treatment. Previous studies have indicated that PET and PET/CT are useful for the detection of primary sites [7,8,[13][14][15]. In this study, neither the PET nor PET/CT imaging improved the detection of the putative primary sites in patients with CUP that already had thorough conventional diagnostic workups. Neither PET nor PET/ CT detected the primary sites in any of the 20 patients with CUP. Previous studies have reported that PET detects primary lesions in 24%-41% of patients with CUP [7]. However, these studies differ in the definition of patients with CUP and conventional workups. Lassen et al. [16] identified primary cancers in nine out

of 20 patients (45%) in their prospective study. Among them, eight patients had primary lung cancers and did not undergo chest CTs during the conventional workups. Bohuslavizki et al. [17] studied 53 patients with CUP, of whom primary tumor sites were detected in 20 (37.8%) using the PET. This study did not include CT or magnetic resonance imaging (MRI) in the conventional workup. Alberini et al. [18] investigated 41 patients with CUP. PET detected primary sites in 26 patients; however, 15 patients had the primary site revealed in the conventional workup. The results of previous studies differed from the findings of this study, which was limited to patients for whom a complete conventional workup, including CT or MRI of suspected lesions, failed to show a primary lesion. According to the previous literature [2], 20-27% of primary sites are identified before the patients with CUP die. The primary site was not initially identified in all of our study patients and was detected in 2 of 20 cases by immunohistochemical staining of biopsied metastatic lesions during the follow-up period. We included only patients whose primary sites were not identified by initial complete diagnostic workups including CT of chest, abdomen, pelvis and endoscopic examinations. Therefore, the patients in our study might have less chance to detect primary sites. The poor resolution of PET has been superseded by PET/CT, which identifies anatomical landmarks more accurately. The PET/CT detects the primary tumor in 22-73% of patients with CUP, according to a recent review article [8]. However, in our study PET/CT did not improve the detection rate of primary sites in patients with CUP. The findings are consistent with those reported by Gutzeit et al.[14] that the identification rate of primary cancers using PET/CT was 33%, but the diagnostic accuracy did not differ significantly from that of the other modalities even though it revealed more anatomical detail. The PET and PET/CT have gained widespread acceptance as useful methods for the management of cancer [19]. However, they do not appear to be effective in identifying a primary lesion after a thorough conventional workup fails to do so. This may be due to the biological characteristics of primary tumors. Primary tumors may disappear after seeding metastases because their angiogenetic incompetence leads to marked apoptosis and cell turnover [20]. Primary tumors that have regressed would not be detected by PET or PET/CT. In this study, neither the PET nor PET/CT detected primary sites in six patients with cervical lymph node metastases, contrary to the findings of other studies [6,12]. The 6 patients with cervical lymph node metastases in this study included 4 poorly differentiated carcinomas and 2 squamous cell carcinomas. The metastases of poorly differentiated carcinoma would have marked cell turnover and apoptosis and that leads to early regression of the primary site. A higher portion of poorly differentiated carcinoma would be one reason for a low detection rate of primary sites in cases with cervical lymph node metastases. The PET or PET/CT revealed FDG uptake lesions other than the known metastases in seven patients. These additional uptake lesions were of no value for detecting the primary sites of tumors, and false positive FDG uptake lesions complicated the diagnosis. Despite no additional value of the PET or PET/CT in the detection of the primary site, primary lesions were identified in two cases by immunohistochemical staining of biopsied metastatic lesions during the follow-up period ( Table 2). Various immunohistochemical markers were used to identify the primary site according to the pathology of the metastatic sites. In one patient (patient no. 6) with metastatic adenocarcinoma, immunohistochemical markers such as the estrogen-receptor, progesterone-receptor, C-erbB2, cytokeratin 7, cytokeratin 20, TTF-1(Thyroid Transcription Factor-1) and GCDFP-15 (Gross Cystic Disease Fluid Protein 15) were used to identify the primary site. Immunohistochemical staining for the estrogen-receptor, progesterone-receptor and C-erbB2 were positive, but cytokeratin 7, cytokeratin 20, TTF-1 and GCDFP-15 were negative. Therefore, the primary cancer was presumed to be a breast cancer. Cytokeratin 7, cytokeratin 20 and TTF-1 were used in a patient with cervical lymph node metastasis (patient no. 11). Results of immunohistochemical staining showed positive cytokeratin 7, negative cytokeratin 20 and focal positive TTF-1 in this patient. The primary cancer was presumed to be a non-small cell lung cancer. A careful conventional workup that includes immunohistochemistry would be helpful for cases in which the primary site cannot be successfully identified using PET or PET/CT. PET or PET/CT scans are easy to perform because of their non-invasiveness [5]; however, subsequent invasive procedures and biopsies are inevitable for pathology confirmation of the results of the PET and PET/CT. The limitations of this study included the following. First, the sample size was small and the study design was retrospective. Second, this study was performed in the early stages of PET and PET/CT, when the PET and PET/CT were not widely used. It is possible that the study results do not reflect current PET or PET/CT scanning. In conclusion, neither PET nor PET/CT improved the detection of primary sites in patients with CUP in our study. Although it is inconclusive because of small sample size of the study, the additional value of PET or PET/CT for the detection of primary sites in patients with CUP might be less than expected; especially in patients that have already had extensive conventional diagnostic workups. Further study is needed to validate this finding. Periorbital Dirofilaria repens imported to Denmark: A human case report Dirofilaria repens, a filarial nematode of dogs and other carnivores, can accidentally infect humans. The infection occurs widely throughout Europe. We report a case of D. repens in a Danish woman who had been traveling to Crete. A nematode was visualized on examination and ELISA was positive for antibodies against D. repens. Introduction Dirofilaria repens is a filarial nematode transmitted by mosquitoes that primarily affects dogs and other carnivores. The adult worms are found in the subcutaneous tissues of dogs and other animals and the adult filariae deposit microfilariae in the blood of the host. Mosquitoes ingest the first-stage larve with a blood meal from the infected host, which develop into the infective third-stage filariform larvae. When the third stage larva have matured and are found in the salivary glands of the mosquito, the infection can be transmitted to the next animal including humans from which the mosquito take the next blood meal. The incubation period in the vertebrate host is 6-8 months [1] (Fig. 1). Incidental human infection can manifest as a single subcutaneous nodule. Migration of the worm may result in local swellings with changing localization. The majority of cases are found on the upper half of the body, mainly around the eyes [1]. Severe clinical manifestations have also been reported and may affect organs including the brain and lungs [2]. Infections have been reported from various regions of the world, mainly Europe, Africa, Middle East and Asia. Areas in Europe where D. repens has been solidly established include countries of the Mediterranean region where the warmer climate facilitates the development of infectious larvae in mosquitoes [3]. Studies from Russia and northern Germany found that the geographical distribution is expanding [1,4]. So far there are no data from Denmark. Case report A 39-year-old Danish woman was seen in January 2014 referred by an ophthalmologist on suspicion of a larval infestation around the right eye. Symptoms began suddenly one week before referral. The patient had noticed a larvae moving under the skin around the right eye with redness and swelling of the skin. Periodically, the structure disappeared. Severe headache in the right temporal area occurred when the structure disappeared. The patient had visited Crete in August 2013 and had never traveled outside Europe. In September 2013 the patient had noticed a subcutaneous nodule on her left leg. The patient lives on a farm and has a dog. On examination a larva was seen moving across the right upper eyelid. MRI of the right orbit did not demonstrate any abnormalities. The MRI was performed four days after the patient were first seen and after start of albendazole treatment. There was no eosinophilia and IgE was normal. Toxocara antibody titer was negative, but there was a positive filarial antibody titer, using Dirofilaria immitis as antigen. Serum was sent to the Rostov Scientific Research Institute of Microbiology and Parasitology, Rostov-on-Don, Russia, for determination of specific antibodies against D. repens, which has developed an ELISA for the detection of Dirofilaria-specific antibodies based on purified somatic antigen of immature female D. repens removed from human cases [1]. The test was found positive for D. repens-specific antibodies. ELISA tests with echinococcus, toxocarosis and trichinellosis antigens were performed, and they were all negative. The patient's dog was tested negative for microfilariae in the blood. Antihelminthic treatment with albendazole 400 mg twice daily orally for 5 days was initiated and the patient recovered rapidly. On the day of admission we send the patient to an ophthalmologist for surgical removal of the worm. The diagnosis may be achieved parasitological when a living and intact worm can be extracted. Unfortunately it could not be removed. Discussion This is the first reported case of human infection with D. repens in Denmark, probably imported from Crete. In Greece the prevalence of dirofilariasis in dogs caused by D. repens infection has been estimated to range between 6.7% and 22% [5]. During the last years cases of human dirofilariasis have also been reported from countries farther north, including Austria, Poland and Germany [6][7][8]. However, with the increasing prevalence of D. repens in northern Germany infection in Denmark is another possibility as D. repens may be present in Danish canines. In February 2014 dogs and mosquitoes have been found positive with D. repens in northern Germany and a case of human D. repens was reported [3,6]. The climate change and increased translocation of dogs might be responsible for the spread to the north [9]. It is surprising that both the eosinophil count and IgE were normal, as these parameters are usually elevated in invasive nematode infections. However, we believe that it reflects that the patient only had a single nematode and that it caused modest local inflammation and thus failed to elicit eosinophilia and an increase IgE. Indeed in a large case series from Russia eosinophilia was found in only 38% of cases (IgE were not measured) [1]. Although this is the only case reported from Denmark, infection with D. repens probably occurs more frequently and is either unrecognized or misdiagnosed. In 2013 67.4% of the Danish population traveled by air outside Denmark [11,12]. Patients who traveling in Europe should consider limit their exposure to mosquitoes and protect themselves from bites. In conclusion a subcutaneous dirofilariasis should be considered as a differential diagnosis in patients who presents with a subcutaneous nodules or a migrating larva. Conflicts of interest None of the authors have any conflicts of interest to declare. Targeting HIV-1 Envelope Proteins Using a Fragment Discovery All-Atom Computational Algorithm Introduction: HIV viral envelope proteins are targets for small inhibitor molecules aimed at disrupting the cellular entry process. Potential peptide-class inhibitor molecules (rDNA drugs) have been previously identified, with mixed results, through biomimicry and phage display experimental methods. Here we describe a new approach based on computational fragment discovery. The method has the potential to not only optimize peptide binding affinity but also to rapidly produce alternative inhibitors against mutated strains. Methods: A comprehensive, all-atom implicit solvent method is used to bombard the C-heptad repeat unit of HIV-1 target envelope protein GP41 with single D-amino acid residues as they exist in their native state. A nascent peptide computational search process then identifies potential favorable sequences of attached ligands based on four peptide bond criteria. Finally, dynamic simulations of nascent peptides attached to host targets help refine potential peptide inhibitors for experimental HIV-1 challenge assays and testing. Results and Discussion: Initial testing of the method was done using 64,000 total ligands of D-amino acid residues at a total computational time of 0.05 microseconds per ligand, which resulted in several thousand attached ligands. Peptide bond criteria search employing three of the four bond constraints with a tolerance of 20 percent, resulted in four potential peptide inhibitors of 5 to 6 residues in length. Only one of the four peptides demonstrated IC50 values and partial viral inhibition based on cell challenge assays using CEM-SS host

cells. That peptide inhibitor also computationally demonstrated long-time attachment and stability to a helical groove in its C-heptad target. This initial testing of peptide fragment discovery against HIV-1 has helped us refine the protocols and identify key areas of improvement. Conclusion: Our methods demonstrate the potential to design efficient peptide inhibitors to viral target proteins based on an all-atom dynamic simulation and using a ligand library as fragments of potential nascent peptides. Our methods can be greatly improved through the use of higher numbers of ligands, increased time of bombardment, and tighter constraints on the peptide bond search step. Our method may be important in the need to rapidly respond to target mutations and to advance multiple targeting methods based multiple peptide inhibitors. INTRODUCTION Infectious disease from viral agents continues to represent one of the most significant health threats to society. The ability of these agents to mutate, transform, and develop across species makes them a formidable opponent to the development of therapeutics, diagnostics, and vaccines aimed at their debilitation. Inhibitors that can bind to viral surface proteins can significantly reduce infectivity rates and, thus, function as therapeutic agents. In this study, the use of advanced molecular computational algorithms for generating small peptide inhibitor compounds that are specific to HIV-1 viral surface protein antigens has been demonstrated. Peptide therapeutics, in general, represents one of the fastest growing segments of the pharmaceutical industry, known as rDNA drugs, and often includes efficient delivery platforms and attractive pharmacokinetic profiles [1]. Advances in ex-vivo production methods also allow production of proteolyticdefying D-peptide drugs and provide inexpensive and large production capability routes for protein and peptide drugs in general. GP41 is part of an envelope glycoprotein complex of HIV-1 that binds to target cell receptors CD4 and CCR-5 or CXCR-4 [2]. GP41 is a three-stranded coiled-coil structure that is exposed during the viral entry process (prefusion state). GP41, therefore, has been a target for the development of inhibitory compounds that bind to it and disrupt the viral entry process. Each subunit of GP41 consists of an N-heptad repeat unit from its N-terminal region (NHR) and C-heptad repeat unit from the C-terminal end (CHR) arranged in an antiparallel fashion. During fusion, the subunits fold to form a six bundle helix with three NHR regions in the core stabilized by interactions with three ectodomain CHR regions. The NHR and CHR interacting regions were synthesized and structurally determined [3], as illustrated in Fig. (1). Peptide sequences based on the CHR region (C peptides) potentially bind to the NHR region and vice versa [4]. Cpeptides have been experimentally shown to be potent inhibitors resulting in, for example, the successful drug Fuzeon (Roche) or Enfuvirtide (T-20) [5]. Previously, we used a static all-atom energy landscape mapping algorithm [6] that yielded a 32 residue peptide sequence from the CHR region (Residue Numbers C628-C659) as the dominant energy interaction region between NHR and CHR almost exactly overlying a number of known experimental nanomolar binding peptides (C34's and SJ-2176) from the CHR region [3]. In this study we describe the development of an alternative, ab initio method of identifying potential peptide binding partners based on a fragment discovery algorithm. We demonstrate the ability to design peptide molecules that bind to GP41 and inhibit viral entry based on experimental HIV-1 cell challenge assays. The fragment discovery algorithm has the potential to rapidly develop inhibitory peptides and, thus, may have broader impact to the need for a robust response to viral infections and outbreaks, in general, including the rise of mutational variations. All Atom Implicit Solvent Methods Biological macromolecular interactions, notably proteinprotein interactions, are fundamentally dynamic events involving the transport and diffusion of ligands to target molecules or sites, followed by attachment or physical adsorption of the ligand to the target molecule. Physical adsorption is, in general, reversible leading to desorption and diffusion and transport of ligands away from the target site. The host solvent (water and dissolved ions) plays a number of important roles, as both a transport media (fundamental fluid mechanics) and in modifications of ligand-receptor force interactions via dielectric and hydrophobic effects (micro-structure considerations). Fig. (2) illustrates a simple ligand-receptor process that takes place in a model biological system, where the target molecule is fixed in space. This particular process is modeled in detail in our methodology, including a full accounting of molecular interactions and their manifestation to observed macroscopic behavior. Due to its potential to significantly reduce the computational load, we have developed an all-atom implicit solvent method derived from first principles via the N-body Liouville equation -the fundamental equation of molecular statistical mechanics. Broadly speaking, coarse-grained methods, stochastic dynamics, Brownian dynamics, and implicit solvent methods, are closely related terms representing macromolecular dynamic methods that involve some type of averaging method for the solvent phase. Here we simply refer to our method as implicit solvent or Brownian dynamics (BD) method [7][8][9]. Our first-principles method is comprehensive in that it includes all possible rotational, translational, and coupled rotational-translational modes of ligands (peptide amino acid residues in this study), quantitative limits on the Fig. (2). Illustration of the overall dynamic process of ligand receptor interactions. ĐŚĂŝŶ E ĐŚĂŝŶ use of BD methods via multiple time scale perturbation theory, inclusion of any external surfaces (protein molecular targets in this study), and formal, comprehensive prescription of all implicit solvent terms via time force autocorrelation expressions. In general, the BD algorithm involves a short-time averaging of the host solvent dynamics coupled to a relatively longer-time, macromolecule dynamics step. The macromolecular dynamics step requires Brownian particle diffusion terms and implicit solvent force terms that can be determined via the short-time behavior, or, alternatively, the diffusion and implicit solvent force terms can be approximated via separate analytical or computational studies for any given system, as demonstrated in proteins [8,9]. Thus, BD has demonstrated the potential to greatly reduce the computational load required for protein dynamic simulations by reducing both the total atom-atom force computations load and increasing the integration time steps required. Briefly, the Brownian dynamics method is in the form of dimensionless translational and rotational displacement equations describing the change in the position of the ligand center of mass, R l , and the orientation about the center of mass, l , for the lth ligand or B-particle as [7]. where i and j represent the Cartesian coordinates for coupled translation and rotation of the lth particle, is a particle Stokes' number, D is a 6 by 6 grand diffusion tensor for the B-particle, F and T are the forces and torques acting on the B-particle as specified below, t is the time step, and the superscript (o) indicates values at the beginning of the time step. The stochastic function C i l (D 0 ij l , t) is a multivariate, Gaussian random number with zero mean and variancecovariance given by The force and torque terms appearing in the Brownian displacement equations above consists of two parts: (a) the local equilibrium average force and torque of the solvent acting on each ligand or B-particle (also, called the implicit solvent force) and (b) the external field force or torque due to the interactions of the ligand with the receptor molecule atoms, or in symbols, respectively F l < F fl > eq + kl ks F kl-ks (4) T l < T fl > eq + kl ks T kl-ks (5) where k l represents the atom of the lth B-particle and k s represents the atom of the target receptor (s). These atom-atom interaction forces and associated torques are taken from the AMBER 03 force field model for all results shown here [10]. Atomic structural information (for ligand and receptor) is supplied by *.pdb or equivalent structure files. Solvent Considerations The local equilibrium average force of the solvent on the ligand consists of polar and apolar contributions as described below. Polar Implicit Solvent Forces Polar implicit solvent forces lead to the so-called dielectric behavior of the solvent. In the calculations given here, following the work of Ramstein and Lavery [11] as reviewed by Smith and Pettitt [12], we employ a distance dependent dielectric of sigmoidal shape with a decay constant of 0.5A 1 . Apolar Implicit Solvent Forces Following previous studies [9], the apolar implicit solvent potential was taken to be proportional to the solvent accessible surface area with the surface tension parameter set to 0.5kcal/molA 2 . The solvent accessible surface area was calculated at the beginning of each Brownian dynamics step using the method of Hasel et al. [13]. Diffusion Tensor for the Ligand For the purposes of calculating the diffusion tensor, each ligand is modeled as a sphere in a continuum with an effective diameter based on their respective van der Waals volume [9]. The hydrodynamic interaction of the ligand with the receptor is neglected in all calculations given here. We note that any of these assumptions can be relaxed through the incorporation of short-time force autocorrelation or analytical studies. Now, although the general algorithm can also include the flexibility of both ligand and receptor [9], the computational costs are currently prohibitive for this application, which necessarily involves the dynamic simulation of tens of thousands of ligands interacting with the target molecule. Consequently, all results given here are restricted to rigid ligands and receptors. Fig. (3) shows the basic simulation "box" where the "ligands" are, in general, peptide fragments from a compound library. Here we take the ligand library of compounds to be the 20 amino acid residues as they naturally appear in a protein, i.e., the residues contain one hydrogen and one oxygen attached to the N and C atoms, respectively (Fig. 4). These amino acid residue structure files are taken from random protein structure files from the protein data bank. The center of mass of a single target receptor is placed at the center of the box, and a fixed lab frame Cartesian coordinate system is assigned to it (Fig. 4). Ligands are initially placed randomly, including position and orientation, within the available volume of the box. The "available volume" excludes the box volume occupied by the receptor, and the center of mass of any ligand must reside within the box and not outside. Multiple Processing The BD method traces the trajectory of each ligand in real time as it interacts with the molecular target. The target receptor is "bombarded" with multiple copies of each type of ligand. Since ligands and their copies are not to interact with each other, each trajectory can be done independently and on a separate processor. Thus, the method is highly amenable to multiprocessor strategies [14], as graphically illustrated in Fig. (3). Fig. (4). Illustration of fragment passing over specified target. Peptide Discovery After a sufficient simulation time on the order of 0.1 microseconds here, a large number of amino acid residues will remain attached to the surface target by virtue of favorable atom-atom interaction energies. For these attached amino acid residues, we have developed an additional code to search for the possibility of patching residues to create a peptide of at least 6-12 attached residues in order to potentially obtain a strong binding, approximately linear peptide. In brief we search over attached residues to find potentially favorable peptide bonds based on the orientation and position of attached residues. These residues are then connected via peptide bonds to create the "birth" structure file of the peptide or peptide candidates (Fig. 5). Peptide Bond Criteria We have utilized four peptide bond criteria in order to determine potential peptide binding segments as given below: Where the subscript i is the residue number and C , N, and C, and O are the carbon atom, backbone nitrogen, backbone carbon, and backbone oxygen of the amino acid residue. The nascent peptide can then be studied dynamically using the flexible implicit solvent simulation [9] in order to obtain its more natural state and examine its stability or continued binding ability with the target surface. We note that all results given here are restricted to finding linear peptides; however, the general method could be extended to secondary structures, such as helices from some known ensemble of potential helical fragment candidates. RESULTS AND DISCUSSION We sought to initially demonstrate the potential feasibility of the method using a limited number of ligands and constraints. For our amino acid segment library we used

the Dform here, based on its enhanced pharmacokinetics, taken randomly from protein data bank structure files for D-amino acid proteins. However, the method is not restricted to this form and, for example; the natural L-form residue library is also easily generated. With the C-heptad repeat as the target, 64,000 total ligands (3200 per residue) bombarded it at a total time of 0.05 microseconds for each ligand using a Sun Ultra Sparc III microprocessor. This resulted in approximately 3000 attached residues. From these 3000 attached residues 4 linear peptide segments of 5 to 6 residues were Fig. (3). Illustration of the control volume and multiprocessor computational strategy for ligand-receptor interaction dynamics. [6]. However, due to the small number of peptides and the small number of residues, all four of the peptides were synthesized and tested. We note that currently we have been able to greatly expand the number of ligands to hundreds of thousands using multiprocessors, increased the total simulation time per residue, employ all four of the bond constraints in the peptide search, and reduce the tolerance levels to 10-15 percent. These improvements generally result in 8 to 12 residue segments and stronger binding affinities as compared to the initial study and results given here. The four D-peptides listed above were synthesized (Sigma Genosys) and sent to Southern Research Institute for HIV-1 Cytoprotection Assays using CEM-SS cells and HIV 1RF (Project 12209.01). Only one of the ligands (Decoy D) demonstrated IC50 inhibition as shown in Fig. (6). Fig. (7) shows the computer generated Decoy D peptide attached in one of the helical groves of the C-heptad helix as predicted from the attached residues and peptide bond constraint criteria. As shown in Fig. (8), we conducted further longer time dynamic simulations with the Decoy D peptide on the target surface in order to verify its superior binding strength and stability compared to the other three peptides. Decoy A, B, and C peptides exhibited less overall interaction energies compared to Decoy D and had less overall interactions with helical grooves of GP41. We note that a static energetic mapping of the binding strengths and subsequent ranking of the peptides are now formally available [6], which can provide a rapid ranking method. However, longer time (millisecond) dynamic algorithms may be necessary due to natural conformational changes associated with the nascent peptide and protein target. Formal off-rate screening is also a challenge due to the long simulation times required and alternative methods are necessary, such as temperature programmed desorption, which are under current study. In that regard, we note that the algorithm given here does not provide equilibrium binding data. This initial study points to the importance of carrying out longer time dynamic simulations with nascent peptides prior to any experimental testing in order to optimize the process of fragment peptide discovery; this is in addition to the number of ligands, constraints, and tolerance issues noted above. If possible, the addition of in-vitro kinetic binding studies, not done for the four peptides above, is also important in viral applications given the complexities and costs of cell challenge experimental methods [6]. CONCLUSION We have demonstrated the ability to conduct an ab initio design of peptide inhibitory molecules to an envelope glycoprotein target molecule GP41 of HIV-1 using a real-time fragment discovery algorithm. By passing on the order of one hundred thousand amino acid residues in their natural protein state over the target surface and then using general bond constraint criteria for peptides, we were able to obtain several potential peptide candidates in a docked state for further study and testing. Our approach includes the development of potentially multi-use rapid viral detection and therapeutic agents based on ab initio computationally generated small peptides with molecular weights less than 1000 Da. In general, peptides can be designed specifically from any validated viral protein target molecule by this method. Since our approach heavily relies on real-time computational methods, there is a potential to significantly reduce the time necessary for the identification of efficacious candidate compounds and also potentially respond in a very rapid manner to newly mutated forms [15]. Clearly, however, more bench studies are needed in order to validate the fragment discovery method beyond the limited results for HIV-1 GP41 inhibition given here, including other viral and nonviral protein targets. The challenges of this approach include the use of a greater number of residues or fragments; longer simulation times, tighter bond constraints on the discovery of nascent peptides, and longer time dynamic simulations of nascent peptides in the bound state of the target for the purposes of ranking and partial validation. The addition of flexibility of nascent peptides and target proteins, not considered here, would also be desirable using either implicit solvent or molecular dynamics methods [9], albeit at the expense of longer computational times. CONFLICT OF INTEREST The author confirms that this article content has no conflict of interest. Fig. (8). Total interaction energy, U, time profile for the attached state of Decoy D in helical groove. Note that this is a typical dynamic energetic signature for attached ligands. The total interaction energy is scaled by kT, where k is Boltzmann's constant and T is absolute temperature (371.15K) and time is scaled by to, 10 11 seconds. Analysis of the cytoskeleton organization and its possible functions in male earthworm germ-line cysts equipped with a cytophore We studied the organization of F-actin and the microtubular cytoskeleton in male germ-line cysts in the seminal vesicles of the earthworm Dendrobaena veneta using light, fluorescent and electron microscopy along with both chemically fixed tissue and life cell imaging. Additionally, in order to follow the functioning of the cytoskeleton, we incubated the cysts in colchicine, nocodazole, cytochalasin D and latrunculin A. The male germ-line cells of D. veneta are interconnected via stable intercellular bridges (IB), and form syncytial cysts. Each germ cell has only one IB that connects it to the anuclear central cytoplasmic mass, the cytophore. During the studies, we analyzed the cytoskeleton in spermatogonial, spermatocytic and spermatid cysts. F-actin was detected in the cortical cytoplasm and forms distinct rings in the IBs. The arrangement of the microtubules changed dynamically during spermatogenesis. The microtubules are distributed evenly in whole spermatogonial and spermatocytic cysts; however, they primarily accumulate within the IBs in spermatogonia. In early spermatids, microtubules pass through the IBs and are present in whole cysts. During spermatid elongation, the microtubules form a manchette while they are absent in the cytophore and in the IBs. Use of cytoskeletal drugs did not alter the general morphology of the cysts. Detectable effects—the occurrence of nuclei in the late spermatids and manchette fragments in the cytophore—were observed only after incubation in nocodazole. Our results suggest that the microtubules are responsible for cytoplasmic/organelle transfer between the germ cells and the cytophore during spermatogenesis and for the positioning of the spermatid nuclei. Electronic supplementary material The online version of this article (doi:10.1007/s00441-016-2398-6) contains supplementary material, which is available to authorized users. Introduction The formation of the syncytial groups of cells, which are usually called cysts, clusters, nests or isogenic groups, seems to be a conservative and widespread phase of animal gametogenesis (Pepling et al. 1999;Ventelä 2006;Haglund et al. 2011;Greenbaum et al. 2011). Within germ-line cysts, there is a cytoplasmic continuity between interconnected cells due to the presence of broad (up to 10 μm; Ventelä 2006;Haglund et al. 2011) cytoplasmic channels (stable intercellular bridges, cytoplasmic bridges, ring canals) that allow the cytoplasm to be shared. The intercellular bridges (IBs) are modified contractile rings that do not close during late cytokinesis (there is no abscission) and, as a result, the cytoplasm of sister cells is common (Greenbaum et al. 2011;Haglund et al. 2011). It has been experimentally demonstrated that in model species such as Caenorhabditis elegans, Drosophila melanogaster and Mus musculus the absence or severe disorders in the formation and functioning of the IBs are connected with infertility (e.g., Robinson et al. 1994;Brill et al. 2000;Maddox et al. 2005;Greenbaum et al. 2006;Green et al. 2011;Yamamoto et al. 2013;Lorès et al. 2014). It is widely accepted that the interconnections of germ cells into syncytial clusters regulate and synchronize germ cell development (Pepling et al. 1999;Guo and Zheng 2004;Ventelä 2006;Greenbaum et al. 2011;Haglund et al. 2011;Amini et al. 2014). However, it seems Electronic supplementary material The online version of this article (doi:10.1007/s00441-016-2398-6) contains supplementary material, which is available to authorized users. that the actual role of a cyst is sex-dependent. In males, cysts are formed similarly in both invertebrate and vertebrate animals (Roosen-Runge 1977;Adiyodi and Adiyodi 1983). Male germ cells develop and differentiate synchronously within the cysts and the gene products and even organelles may be shared between cells and therefore haploid spermatids remain phenotypically diploid in late spermatogenesis (Braun et al. 1989;Morales et al. 2002;Ventelä et al. 2003). On the other hand, in female gametogenesis, matters are more complicated. Germ-line cysts are completely absent in some taxa (e.g., Büning 1994;Tworzydło et al. 2014), whereas in vertebrate species such as Xenopus laevis or M. musculus the female cysts are transient and germ cells quickly spilt into individual cells and the role of cell clustering is not clear (Pepling and Spradling 1998;Pepling et al. 1999;Kloc et al. 2004Kloc et al. , 2008Greenbaum et al. 2009). Finally, in some invertebrate taxa including insects, only several (sometimes one) cells in a cyst pass meiosis, gather nutrients and become oocytes and the rest of the interconnected cells supply the growing oocytes with macromolecules and cell organelles-these cells die late in oogenesis and serve as auxiliary (nurse) cells (Telfer 1975;Büning 1994;Matova and Cooley 2001). Numerous analyses of developing male and female germ cells have shown that the spatial organization (architecture) of germ-line cysts varies among taxa. Taking only male germ cysts into account, it can be assumed that three types of cysts are predominant: linear, branched and cysts that have a central cytoplasmic mass. In linear clusters, the cells form chainseach germ cell has two IBs that connect them to their neighbors and only the terminal cells have one IB. Linear chains, which are composed of hundreds or even thousands of male germ cells, are widespread in vertebrates (Fawcett et al. 1959;Roosen-Runge 1977;Greenbaum et al. 2011). Some germ cells in branched cysts may have more than two IBs and, as a result, the Bbranchings^occur in these sites. The branched cysts of male germ cells are best known from some insects, especially D. melanogaster (Hime et al. 1996;Eikenes et al. 2013). In some invertebrates, e.g., nematodes, flat worms and annelids, germ cells are not directly interconnected to one another but rather, as a rule, each germ cell in a cyst has only one IB joining it to a common and anuclear cytoplasmic mass (the cytoplasmic core). In the gonads of some nematodes such as C. elegans, the cytoplasmic core has the form of a cylinder and is known as the rachis or gonad core (Hirsh et al. 1976;Wolf et al. 1978;L'Hernault 2009). In flat worms and annelids, the cytoplasmic core is traditionally called a cytophore and its shape depends on the developmental stage of the cyst (Davis and Roberts 1983;Olive 1983;Ferraguti 1983;Świątek et al. 2009). The formation, organization and functioning of male germline cysts is best known in the case of branched cysts. Numerous studies, mainly on D. melanogaster, have revealed many sophisticated molecular aspects of the biology of the cysts, e.g., the germ-line-specific cytoskeletal-rich cytoplasm called fusome is engaged in the correct segregation of IBs during the subsequent divisions of daughter cells (Hime et al. 1996;Eikenes et al. 2013). Fewer studies have been devoted to germ-line cysts that have a cytoplasmic core. Although we have a powerful model for such studies, i.e., C. elegans, the numerous aspects of cyst biology such as the mechanism that governs the formation or the organization of these cysts and the role of the cytoskeleton are far from elucidation. Recently, it has been experimentally shown that the actomyosin cytoskeleton plays an active role in cytoplasmic streaming during oogenesis in C. elegans and that C. elegans homologs of the protein anillin regulate the stability of IBs (Wolke et al. 2007;Amini et al. 2014). Moreover, morphological studies on clitellate annelids have revealed that the formation of germ-line cysts with a cytophore differs markedly from the one that is known from branched and linear cysts (Świątek et al. 2009). In

the latter cases, the IBs are always formed de novo as modified and stabilized contractile rings (Robinson et al. 1994;Robinson and Cooley 1996;Ong and Tan 2010;Greenbaum et al. 2011;Haglund et al. 2011). In clitellate annelids, the specific orientation of mitotic spindles causes the contractile ring of dividing germ cells to merge with a pre-existing bridge and splits it into two new bridges (Świątek et al. 2009). Clitellate annelids (Clitellata) seem to be a very attractive model for analyses of germ-line cysts that have cells that are clustered around the cytophore. The germ-line clusters that have a cytophore have been found in all the species of Clitellata that have been studied to date, in both male and female gametogenesis (Ferraguti 2000;Jamieson 2006;Świątek et al. 2009;Urbisz et al. 2015). Male germ-line cysts are exceptionally easy to obtain and manipulate. Although the first divisions of spermatogonia occur in the testis, which are usually tiny structures that are not easy to handle, the later spermatogonial divisions, meiosis and spermiogenesis take place in the specialized diverticula of the septa, the so-called seminal vesicles (Jamieson 1992). The seminal vesicles are easy to manipulate and, what is more important, there are hundreds or even thousands of germ cysts freely floating in the coelomic fluid in each seminal vesicle. Moreover, the clustered germ-line cells that are floating within the seminal vesicles are not associated with any somatic cells (Jamieson 1981(Jamieson , 1992, which greatly facilitates their analysis. The different aspects of spermatogenesis (especially spermiogenesis) and sperm structure in Clitellata are well described at the ultrastructural level in many taxa (see Ferraguti 1983Ferraguti , 2000Jamieson 1981Jamieson , 1992Jamieson , 2006 for a review), because the characters that are connected with sperm structure have been widely used in phylogenetic assessments (e.g., Marotta et al. 2008;Marotta and Ferraguti 2009). To date, ultrastructural studies on spermatogenesis and cyst structure in Clitellata have concentrated only on such aspects as the formation of the IBs and cytophore (Martinucci et al. 1977;Świątek et al. 2009), nuclear fragmentation and the multiplication of basal bodies during paraspermiogenesis (Boi et al. 2001;Ferraguti et al. 2002) or the distribution of cytoskeletal elements such as actin-and gelsolin-related proteins (Krüger et al. 2008). We decided to perform systematic studies on the development and functioning of male germ-line cysts in the commercially available earthworm Dendrobaena veneta using techniques such as light, fluorescence and electron microscopy along with both a chemically fixed tissue and life cell imaging. This paper is devoted to the analysis of F-actin and the microtubular cytoskeleton in the cysts that develop in the seminal vesicles. Additionally, experiments using cytoskeletal drugs such as colchicine, nocodazole, cytochalasin D and latrunculin A showed that the general architecture of cysts is not altered after drug treatment. The only detectable effects were the disorganization of the microtubular manchette and the presence of late spermatid nuclei within the cytophore. Animal material Commercially obtained, adult (marked as size no. 4) specimens of the earthworm Dendrobaena veneta, Rosa 1839 were used. The earthworms were kept in their original boxes in laboratory conditions for several days/weeks. Light and electron microscopy Dissected seminal vesicles were fixed in 2.5 % glutaraldehyde in a 0.1 M phosphate buffer (pH = 7.4) for 24 h at room temperature. After fixation, the material was rinsed several times with a mixture of 50 ml 0.1 M phosphate buffer and 50 ml of ddH 2 O to which 4.6 g saccharose was added. Afterwards, the material was postfixed for 2 h in a mixture of 1 % OsO 4 in a 0.1 M phosphate buffer, dehydrated in a graded series of ethanol replaced by acetone and then embedded in an Epoxy Embedding Medium Kit (Sigma, St. Louis, MO, USA). Semi-thin sections (1 μm thick) were stained with methylene blue and then examined under an Olympus BX60 microscope equipped with an XC50 digital camera and CellSense Standard software. After contrasting with uranyl acetate (15 min) and lead citrate (20 min), ultra-thin sections (70 nm thick) were examined using a Hitachi H500 transmission electron microscope. Visualization of microtubules on Steedman wax sections The dissected seminal vesicles were fixed in 4 % formaldehyde, which was freshly prepared from paraformaldehyde, in PBS for 30 min at room temperature for all the cytoskeletal staining. Fixed seminal vesicles were dehydrated in a graded series of ethanol: 30, 50, 70, 90 and 96 % for 15 min each and then 2× 1 h in 100 % ethanol. After that, the tissue was saturated in Steedman wax/ethanol solutions: 1:3 for 24 h, 1:1 for 24 h and 3:1 for another 24 h. For the final saturation, 100 % wax was used for 24 h. The seminal vesicles were then embedded in Steedman wax, left for polymerization and cut into 6-to 12-μm-thick sections on a microtome Zeiss HYRAX M40. Before staining, the sections were mounted onto microscope slides and de-waxed by using a reverse series of ethanol: 100 % for 2× 15 min and 90, 70 and 50 %, ddH 2 O for 15 min each. The de-waxed sections were then washed with 1 % Triton X-100 in TBS (Trisbuffered saline) for 20 min and washed with pure TBS for 5 min and afterwards incubated for 1 h in 1 % BSA (Bovine Serum Albumin) in TBS. Mouse anti β-tubulin antibody (T-4020; Sigma-Aldrich) was used to visualize the microtubules. Primary antibodies were diluted in 1 % BSA in TBS at a ratio of 1:100, applied onto slides and incubated for 24 h at 4°C. After washing in TBS (5 min), the anti-mouse secondary antibodies, which were conjugated with Alexa Fluor 488 (Sigma-Aldrich) that was diluted in 1 % BSA in TBS at a ratio of 1:50, were applied (1 h at room temperature). Finally, after staining with antibodies, the samples were additionally stained with DAPI (1 μg/ ml; Sigma-Aldrich) for 30 min in the dark. Visualization of F-actin and microtubules in the cyst suspension After fixation (as described above), the seminal vesicles were torn manually with tweezers under a stereomicroscope in order to obtain the cyst suspension in PBS. The suspension was stained with a rhodamine-conjugated phalloidin (2 μg/ml; Sigma-Aldrich) for 45 min, washed with PBS (5 min) and additionally stained with DAPI (1 μg/ml) for 30 min. To detect microtubules, the same protocol and anti-tubulin antibody was used as described above. The germ-line cysts were applied onto microscopic slides using a cytocentrifuge MPW 223c (MPW MED. INSTRUMENTS, Poland) and were examined under an Olympus BX60 fluorescence microscope or under an Olympus FV1000 confocal microscope with a ×60/ NA 1.30 sil objective. Z-stack images were generated using a 405-nm laser for DAPI, a 488-nm laser for Alexa Fluor 488 dye and 568-nm for rhodamine-conjugated phalloidin dye. 3D datasets were analyzed as volume-rendered datasets using Imaris (custom software developed by Bitplane Scientific Software, Zurich, Switzerland) and Fiji (freeware software). Staining of live cells Before sectioning, the animals were narcotized in 50 % ethanol for 5 min. TubulinTracker Green and Hoechst 33342 (Life Technologies) were used to visualize tubulins and cell nuclei in living cells. After dissection, the seminal vesicles were q u i c k l y t r a ns f er r ed t o a c e l l c u l t u r e m e d i u m (Dulbecco's Phosphate Buffered Saline, DPBS; Sigma-Aldrich). The seminal vesicles were then torn manually with tweezers under a stereomicroscope to obtain the cyst suspension in DPBS. Both of the above-mentioned live cell dyes were diluted in DPBS at a ratio of 1:1000 and the staining time oscillated between 10 and 20 min. After staining, the cyst suspension was immediately applied onto microscopic slides using a glass pipette. The microscopic slides were examined under an Olympus BX60 fluorescence microscope or an Olympus FV1000 confocal microscope with a ×60/ NA 1.30 sil objective. Z-stack images were generated using a 405-nm laser for Hoechst 33342 and a 488-nm laser for TubulinTracker Green. Experiments using cytoskeletal drugs In order to examine the role of the cytoskeleton during the development of male germ-cell cysts, a series of experiments were performed. After the earthworms were narcotized in 50 % ethanol for 5 min and, after they were dissected, the freshly obtained seminal vesicles were incubated in DPBS to which cytoskeletal drugs had been added. Nocodazole and colchicine (Sigma-Aldrich) were used to disrupt the microtubule cytoskeleton. Cytochalasin D and latrunculin A (Sigma-Aldrich) were used to destabilize the microfilaments. According to the manufacturer's protocol, the nocodazole and cytochalasin D and latrunculin A were initially dissolved in DMSO (dimethyl sulfoxide) and the colchicine in ddH 2 O. Then, all the drugs were dissolved in DPBS in order to obtain the working solutions: for nocodazole 20 μM and 100 μM; for colchicine 250 μM; for cytochalasin D 4 μM and 8 μM; and for latrunculin A 0.25 μM and 0.5 μM. For colchicine and nocodazole, the seminal vesicles were incubated for 6, 12, 24 and 48 h. and for cytochalasin D and latrunculin A for 6 h. Additionally, blind and control samples were prepared for each experiment. For the blind samples, the seminal vesicles were fixed (as described below) immediately after dissection without any incubation. For the standard control experiments, the seminal vesicles were incubated in a cell culture medium to which a cytoskeleton drug solvent was added only (DMSO and ddH 2 O, respectively) without the addition of any cytoskeleton drug. After all the experiments, the seminal vesicles were fixed in 2.5 % glutaraldehyde for light and electron microscopy or in 4 % formaldehyde, which was freshly prepared from paraformaldehyde, for immunofluorescence and cell suspension labeling. The experimental probes were then stained and analyzed as described above. General cyst organization The process of spermatogenesis in clitellate annelids has been intensively studied at the ultrastructural level in recent decades and detailed summaries have been provided by Jamieson (1981Jamieson ( , 1992Jamieson ( , 2006 and Ferraguti (2000). In this paper, we only describe the general aspects of spermatogenesis in D. veneta with the emphasis on the structure of the cysts and the cytoskeleton organization. The terminology was adopted after Jamieson (1981) with a few modifications such as using cysts instead of Bmorulae^and intercellular bridges (IBs) instead of Bzonulae collaris^. The seminal vesicles of Dendrobaena veneta are roughly oval structures that are located in segments IX-XI. We usually found three pairs of vesicles per specimen; however, in some cases, there were four pairs of vesicles or a number of vesicles were unpaired. There were hundreds of germ-line cysts within the vesicles (Fig. 1a). All the germ cells that were clustered in a single cyst were at the same phase of spermatogenesis; however, there was no synchrony between the cysts, i.e., cysts with a different number of clustered cells (i.e., cells in different phases of spermatogenesis) were observed in the same vesicle (Fig. 1a). Generally, in all the germ-line cysts that were observed, the center was occupied by an anuclear cytophore (Figs. 1, 2, 3, 4, 5). The shape and dimensions of the cytophore changed during the successive stages of spermatogenesis (see below). The gem cells were clustered around the cytophore and, as a rule, each germ cell was connected to the cytophore by one IB (Figs. 1,2,3,4,5). The first step of study was to identify the cysts that were in the consecutive stages of spermatogenesis and to determine the cell number in the cysts. In order to find out how many cells constituted each spermatogenic stage, we counted the nuclei in 10 DAPI-stained cysts that contained elongate spermatids (these cysts are the easiest to identify). The number of cell nuclei in the elongate spermatid cysts varied between 248 and 255 (Table 1). Taking into account the synchronic divisions of germ cells that were observed within a given cyst (Figs. 1, 2c), it may be assumed that the total number of cell divisions during spermatogenesis is eight, which would potentially produce 256 spermatids. As a result, cysts that contained 64 cells were composed of primary spermatocytes, whereas secondary spermatocytes were clustered in the 128 cell cysts. Additionally, spermatogonial cysts were also observed within the seminal vesicles; these were 8-, 16-and 32-celled cysts (Fig. 1a, b). Cysts that had fewer than eight cells were never observed in the seminal vesicles. The spermatogonia had large nuclei with dense chromatin that was embedded in

a transparent matrix and their cytoplasm contained numerous small mitochondria and Golgi complexes, whereas cisternae of ER were rare (Fig. 1b). The cytophore in the spermatogonial cysts were small and had the form of an irregular mass of cytoplasm (Figs. 2a, 3a-a''''') that contained numerous ribosomes, mitochondria, short ER cisternae and lysosomes (Fig. 1b). The primary spermatocytes were easy to identify at the electron microscopy level where synaptonemal complexes were visible within their nuclei (not shown). However, at the light microscopy level, the clear distinction between primary and secondary spermatocytes was usually difficult and therefore we generally only identified the spermatocytic cysts (Figs. 1a, 2b, c, 3b-b'''''). At the light microscopy level, cysts that had isodiametric (young) and elongate (late) spermatids were the easiest to identify (Figs. 1a, 2d, e, 3c-c''''', d-d''''', e-e'''''). The isodiametric spermatids were roughly rounded cells that were connected to the cytophore, which was irregular or rounded in shape (Figs. 1a, c, 2d, 3c-c'''''). The nuclei of these spermatids were oval and Vesicles are filled with hundreds of germ cysts. Germ cysts are at different stages of spermatogenesis and therefore spermatogonial cysts (sg), spermatocytic cysts (sc) and cysts with isodiametric (is) and elongate spermatids (es) can be observed. The center of a cyst is always occupied by a more or less voluminous anuclear cytophore (stars). Arrow marks the vesicle wall that is made up of somatic cells. d cyst with dividing cells. Light microscopy, methylene blue staining, scale bar 20 μm. b Ultrastructural detail of a spermatogonial cyst. cy cytophore, m mitochondria, sg spermatogonia, sgn spermatogonia nuclei. Ellipse the intercellular bridge; arrows the electron-dense bridge rim. Transmission electron microscopy (TEM), scale bar 1 μm. c A fragment of cysts with young (isodiametric) spermatids. Ellipse the intercellular bridge connecting the spermatid (is) with the cytophore (cy); c centriole, er endoplasmic reticulum, gc Golgi complex, m mitochondria, isn spermatid nuclei. TEM, scale bar 1 μm. d A portion of cysts with elongate spermatids (es) cross-sectioned at the level of their nuclei. Chromatin in elongate spermatids is tightly condensed and the nuclei (esn) are surrounded by a microtubular manchette (thin arrows). Thick arrows the electron-dense rim of the intercellular bridge; cy cytophore, gc Golgi complex. Inset higher magnification of nuclei enveloped by the manchette (thin arrow). TEM, scale bar 0.5 μm, inset scale bar 0.5 μm the chromatin was not condensed. Rough ER cisternae and small mitochondria were scattered within the cytoplasm, while the Golgi complex and centrioles were located close to the proximal (anterior) end of the cell, i.e., the pole of the cell connected to the cytophore via the IB (Fig. 1c). Numerous mitochondria and cisternae of endoplasmic reticulum were observed within the cytophore of the isodiametric spermatids (Fig. 1c). The process of cell transformation into spermatozoon occurred intensively in the elongate spermatids (Figs. 1d, 2e, 3d-d''''', e-e''''') (see Jamieson 1981;Ferraguti 2000 for details). The following structures along their proximal-distal axis could be observed within the transforming spermatids: the developing acrosome, an elongated and condensing nucleus, mitochondria, the centriolar complex and the axoneme Fig. 2 The germ-line cysts stained with rhodamineconjugated phalloidin (a-d) and DAPI (b-d) in the consecutive stages of spermatogenesis. Blue cell nuclei or metaphase plates (in c), red circles intercellular bridges, thin arrows cortical Factin, cy cytophore. a spermatogonial cyst, b spermatocytic cyst, c cyst with synchronously dividing cells, d cyst with isodiametric spermatids, e cyst with elongate spermatids. Single slices from confocal microscopy, scale bar 10 μm Fig. 3 The distribution of microtubules and F-actin within the germ-line cysts not incubated (blind samples) and treated with the cytoskeletal drugs (incubation time = 6 h., concentrations are marked on the panels). Note that the general cyst morphology is the same in both cases. a spermatogonial cysts (sg), b spermatocytic cysts (sc), c cysts with isodiametric spermatids (is), d cysts with early elongate spermatids (es), e cysts with late elongate spermatids (es'); blue cell nuclei visualized by DAPI staining (a, b, a''-e''), green microtubules visualized by TubulinTracker Green (e) or by antibody against β-tubulin (a-d, a'-e', a''-e''), red F-actin stained with rhodamine-phalloidin (a'-e', a'''-e''', a''''-e'''', a'''''-e'''''); ax axoneme, cy cytophore, mm microtubular manchette. Thick arrows microtubule concentrations in regions that correspond to the intercellular bridges, thin arrows intercellular bridges. a-d Steedman wax sections, the rest of the panels, whole mount preparations. Confocal (a'-e') and fluorescence microscopy (the rest of the panels). Scale bars 10 μm b (not shown). The acrosome and nucleus in the elongate spermatids were enveloped by a microtubular manchette (Fig. 1d). The cytophore in these cysts reached its maximal dimensions (Fig. 2e). It was spherical and contained numerous mitochondria, ER cisternae and organelles that had a dense content and that were determined to be lysosomes (not shown). The IBs that connected the germ cells during all of the stages of spermatogenesis had a similar ultrastructure, namely, the cytoplasm strand was surrounded by an electron-dense layer of fibrous material that lined the plasma membrane, which is the so-called bridge rim (Fig. 1b-d). Cell organelles such as mitochondria, ER and Golgi complexes were observed within the bridges (Fig. 1b-d). F-actin and microtubule distribution In order to visualize the F-actin, the cysts were labeled with rhodamine-coupled phalloidin ( The ultrastructure of a cyst incubated for 6 h (nocodazole concentration 100 μM). The cyst morphology is not altered; is isodiametric spermatids, cy cytophore, m mitochondria, nu cell nuclei. Ellipses the intercellular bridges. Scale bar 2 μm. b A fragment of a cyst with elongate spermatids that was incubated for 24 h (20 μM). The nuclei of spermatids (sn) inside the intercellular bridges (ellipse); the cytophore (cy) can be observed. Scale bar 1 μm. c) Fragments of two cysts with late elongate spermatids incubated for 48 h (20 μM). Numerous nuclei of the spermatids (sn) can be seen within the cytophore (cy). Scale bar 2 μm. d A cross-sectioned spermatid nucleus (sn) in the cytophore; note the remnants of the microtubular manchette (thin arrow) in close proximity of the nucleus. Scale bar 0.5 μm area where the germ cells were connected to the cytophore (Figs. 2a-e, 3a'-e', a'''-e''' , a'''''-e''''', a'''''-e'''''). The F-actinenriched rings corresponded to the electron-dense fibrous material that formed the IB rim (Fig. 1b). It should be stressed that we always observed just one F-actin ring corresponding to the IB between a given germ cell and the cytophore (Figs. 2, 3; see also Electronic Supplementary Material, Fig. S1). Multiple bridges between a given germ cell and the cytophore or between adjacent germ cells were never observed. Two methods were used to detect the microtubules (MTs)immunofluorescence using antibody against β-tubulin and live cell imaging using TubulinTracker Green. Both methods produced comparable results (Fig. 3a-e, a'-e'). The distribution of MTs changed during the consecutive stages of spermatogenesis. In the spermatogonial cysts, MTs were observed Fig. 5 a The scheme of F-actin (red) and microtubule (green) distribution in the consecutive stages of cyst development. All cell organelles except for the germ cell nuclei (grey) and axoneme (ax) are omitted for clarity; cy cytophore, intercellular bridges are marked by arrows (for more details, see text); sg spermatogonial cyst, sc spermatocytic cyst, is cyst with isodiametric spermatids, es cyst with early elongate spermatids, es′ cyst with late elongate spermatids. b Microtubule (green) and F-actin (red) distribution in cysts that were incubated in nocodazole and cytochalasin D, respectively. All cell organelles except for the germ cell nuclei (grey) and axoneme (ax) are omitted for clarity; cy cytophore, intercellular bridges are marked by arrows (for more details, see text); sg spermatogonial cyst, sc spermatocytic cyst, is cyst with isodiametric spermatids, es cyst with early elongate spermatids, es′ cyst with late elongate spermatids to be distributed within the germ cell cytoplasm and within the cytophore (Fig. 3a-a'). However, a high condensation of MTs was detected in the areas where the germ cells connected to the cytophore, i.e., within the IBs (Fig. 3a'; ESM, S1). On the other hand, in the dividing spermatogonial cysts, the MTs only formed mitotic spindles and no MTs were found in the bridges or within the cytophore (not shown). In the spermatocytic cysts, MTs were distributed both within the cytoplasm of the germ cells and within the cytophore, whereas no accumulations of MTs were observed in the areas that corresponded to the IBs (Fig. 3b-b'). In cysts that were composed of isodiametric spermatids, the MTs formed a dense network in the cortical cytoplasm of the germ cells and in the cytophore (Fig. 3c-c'). Specifically, prominent bundles of MTs passed from the cytoplasm of the isodiametric spermatids into the cytophore (Fig. 3c-c'). As was shown at the ultrastructural level, MTs form a manchette that tightly surrounds the transforming cell organelles such as acrosome and the nucleus in elongate spermatids (Fig. 1d); they also encompass the mitochondria and finally the MTs form the axoneme. Immunofluorescence and TubulinTracker Green labeling confirmed these observations (Fig. 3d-d', e-e'). MTs were not found in the cytophore and in the areas that corresponded to the IBs in either the late elongated spermatids or at the onset of the elongation of the spermatids (Fig. 3d-d', e-e'). Figure 5a presents the scheme that summarizes the changes in the F-actin and MTs distribution within germ-line cysts during the consecutive stages of spermatogenesis. Experiments using cytoskeletal drugs The first step in conducting the experiments using cytoskeletal drugs was to elaborate the methodology of the germ-line cyst culture or outstay in vitro. We found that the germ-line cysts together with coelomic fluid (cyst suspension), which were freshly obtained from the seminal vesicles, mixed with DPBS and kept in small plastic vials, were vital for at least 48 h. The ultrastructure of spermatogonia, spermatocytes and spermatids, which were kept in vitro for 6, 12, 24 and 48 h, respectively, did not differ from the one that was found in the non-cultured cysts (not shown). In order to disrupt the organization of MTs and microfilaments, seminal vesicles that had a developing germ-line cyst were incubated in media that contained nocodazole, colchicine, cytochalasin D or latrunculin A. The incubation times and the doses that were used varied from 6 to 48 h and from 5 to 250 μM, respectively (see BMaterials and methods^for details). The use of colchicine produced no detectable results (not shown). The cyst architecture after treatment with cytochalasin D and latrunculin A also remained unaltered (Fig. 3a'''-e''', a''''-e'''', a'''''-e'''''); however, numerous granular rhodamine-phalloidin positive accumulations were found within the germ cells and the cytophore, especially after incubation in cytochalasin D (Fig. 3a'''-e'''). The general organization of the cysts that were incubated with nocodazole also did not changethe morphology of the cysts appeared to be unchanged (Fig. 4a). However, when it was visualized using immunofluorescence, the microtubular cytoskeleton was found to be severely damaged (Fig. 3a''-e''). In the spermatogonial and spermatocytic cysts and also in the cysts that were composed of isodiametric spermatids that were incubated in nocodazole, the detected tubulin was mainly globular in form and MTs were observed only occasionally (Fig. 3a''-c''). No bundles of MTs that passed the intercellular bridges were observed ( Fig. 3a''-e''). MT manchettes were detected in the cysts that contained spermatids at the onset of their elongation that were treated with nocodazole; however, the immunofluorescence visualization suggested that the manchettes did not fully surround the spermatid organelles ( Fig. 3d''). These results were confirmed by ultrastructural analysis (Fig. 4c, d). Manchettes that enveloped organelles were not detected in the cysts composed of elongate spermatids by the immunofluorescence methods (Fig. 3e''). A weak signal from tubulin was observed within the cytophore of these cysts only (Fig. 3e''). As was demonstrated by the ultrastructural analysis, these signals may have come from the manchettes that were partially destroyed and passed through IBs to the cytophore (Fig. 4d). The ultrastructural observations and immunofluorescence methods revealed that, within the cysts with the elongate spermatids that were treated with nocodazole, the nuclei of the spermatids had a tendency to pass through the intercellular bridges into the cytophore (Fig. 4b). In such cysts, numerous nuclei of the spermatids were detected within the cytophore (Fig. 4c). In the blind and control experiments, no nuclei of the spermatids were ever observed within the cytophore Table 1 The number of cells (cell nuclei) that were counted in the ten cysts that contained elongate spermatids; the cysts were stained with DAPI for counting

and analyzed using confocal microscopy (not shown). The distribution of tubulin in the cysts after nocodazole treatment is schematized in Fig. 5b. Discussion The general and ultrastructural aspects of sperm formation that were found for D. veneta in this study are in accordance with the general description of spermatogenesis in clitellate annelids that was summarized by Jamieson (1981Jamieson ( , 1992Jamieson ( , 2006 and Ferraguti (2000). One aspect of spermatogenesis in D. veneta is especially noteworthy. Usually, there are 128 spermatids per cyst in Lumbricida and other clitellate annelids (Jamieson 1981). We found that there are 256 spermatids per cyst in the species studied here. This means that there is one extra mitotic division of spermatogonia in D. veneta in comparison with other earthworms such as L. terrestris (Walsh 1954;Krüger et al. 2008). Among Oligochaeta, 256 spermatids that were clustered in a cyst have also been reported in two lumbriculid species (Ferraguti 2000). The organization and role of the cytoskeletal elements in developing male gametes have been intensively studied, especially in vertebrates (mainly mammals) and in model invertebrates such as D. melanogaster (recently reviewed in Lie et al. 2010;Sun et al. 2011;Sperry 2012;O'Donnell and O'Bryan 2014). Usually, these descriptions concentrate on the mitotic and meiotic divisions of the germ cells and on late spermatogenesis, i.e., spermiogenesis when the haploid spermatids undergo rapid changes in cell organization such as cell elongation, chromatin condensation, the formation of the acrosome and flagellum. It is well known that the cytoskeleton plays an active role in all these processes (Sperry 2012). It has also been demonstrated that the correct organization and regulation of the cytoskeleton dynamics is essential for male fertility (see Lie et al. 2010;O'Donnell and O'Bryan 2014 for details). To the best of our knowledge, the results presented here are the first comprehensive description of the F-actin and microtubular cytoskeleton in male germ-line cysts that have a cytophore. F-actin Using rhodamine-coupled phalloidin, we found strong F-actin positive signals from the IBs in the form of ring-like structures. The presence of F-actin delimiting the IBs interconnecting germ cells appears to be a widespread phenomenon of gametogenesis, especially in oogenesis (Haglund et al. 2011). F-actin has been detected in the IBs in the male cysts of mammals such as rat, squirrel and mouse (Haglund et al. 2011), whereas F-actin is absent in fully formed bridges in fruit fly males (Hime et al. 1996;Eikenes et al. 2013). Using rhodamine-phalloidin staining, F-actin rings within the IBs have been found in clitellate annelids in both female (e.g., the fish leech Piscicola geometra, Spałek-Wołczyńska et al. 2008; the sludge worm Tubifex tubifex, Urbisz et al. 2015) and male cysts (the earthworm Lumbricus terrestris - Krüger et al. 2008; the sludge worm, Boi et al. 2001). F-actin rings appear to stabilize the bridges and keep them permeable, which is, in turn, a prerequisite to the efficient transfer of cytoplasm between germ cells. While the presence of F-actin in the cortical cytoplasm and in the IBs in cysts that are composed of spermatogonia, spermatocytes and spermatids is not unusual, surprisingly, we were unable to detect F-actin by rhodaminephalloidin staining in any other localizations within the cysts that clustered around the spermatids. Krüger et al. (2008) analyzed the distribution of the gelsolin-related protein (EWAM-P1) and both G-and F-actin using rhodamine-phalloidin and immunofluorescence in the developing male cysts of L. terrestris. Like our studies, they found actin in the IBs but they also found actin in the proximal and distal parts of the spermatocytes, in the cytophore of spermatid cysts and in the distal part of the spermatids. Additionally, the pattern in which the anti-actin antibodies marked the periphery of the cell nucleus depended on the stage of spermatogenesis (Krüger et al. 2008). According to the authors, the actin concentrations in the distal poles of spermatocytes and spermatids may be connected with the activity of the Golgi apparatus and the actin that is localized in close proximity to the spermatid nuclei seems to Bflow^from the spermatids towards the cytophore (Krüger et al. 2008). Additional studies using the anti-actin antibodies should clarify whether the localization of G-and Factin in the developing male cysts of D. veneta is similar to that found in L. terrestris or whether the actin distribution is species-specific. Generally, it may be stated that the actin cytoskeleton in the male cysts of earthworms is not especially complex; no actin-rich structures such as, e.g., acroplaxome, actin cones or acroframosome, which are known from developing sperm of the other animals (see Sun et al. 2011 for a review), have been found. It is worth mentioning that a complex F-actin cytoskeleton has been found in female germ-line cysts that have a central cytoplasmic core in both annelids (T. tubifex; Urbisz et al. 2015) and nematodes (C. elegans; Wolke et al. 2007). Microtubules The dynamics of MTs and MAPs (microtubule-associated proteins) are indispensable for the correct sperm formation; the role of MTs and MAPs in Sertoli cells, during germ cell divisions, in nuclear elongation, cytoplasmic redistribution and flagellar formation in spermatids are well known, especially in rodents (see Lie et al. 2010;Sperry 2012;O'Donnell and O'Bryan 2014 for details). Much less is known about the dynamics of MTs in invertebrate spermatogenesis, in which MTs are primarily recognized at the ultrastructural level and special attention is usually focused on the manchette, which is a transient microtubule-rich structure that tightly envelops the elongating spermatid nucleus. Our studies showed that MTs not only form the manchette in elongating spermatids but also reorganize dynamically during the subsequent stages of spermatogenesis. In spermatogenic cysts, MTs form a dense network in the germ cells and within the cytophore, in which they are highly concentrated in the areas that correspond to the IBs. During the meiotic phase, the network of MTs is also well developed in both the germ cells and the cytophore; however, no specific concentrations of MTs were found in the bridges. At the onset of spermiogenesis, the prominent microtubular bundles pass through the IBs and radiate into the cytoplasm of the cytophore (Fig. 5a). Later in spermiogenesis, MTs were only found in the manchette and the axoneme; no MTs were detected in the bridges or in the cytophore (Fig. 5a). Such a distribution of MTs is certainly connected with the transfer of cytoplasm and the formation and functioning of the cytophore. It is widely accepted that the cytophore is formed due to an accumulation of the cytoplasm of germ cells in the cyst center (Jamieson 1981;Ferraguti 2000;Świątek et al. 2009). The cytophore enlarges in the subsequent stages of spermatogenesis, e.g., the volume of the cytophore that interconnects the secondary spermatocytes increased at least 22fold in E. foetida in comparison with spermatogonial cysts (Martinucci et al. 1977). The presented studies, together with the ultrastructural observation of MTs within the intercellular bridges that interconnect early spermatogonia in the testis of E. foetida (Martinucci and Felluga 1975), suggest that the microtubular cytoskeleton is responsible for the transfer of organelles and macromolecules. It should be stressed that the transfer between germ cells and cytophore is most probably bidirectional. The cytoplasm with the cell organelles, e.g., mitochondria, flows from the developing germ cells towards the cytophore. On the other hand, it has been suggested that some macromolecules such as proteins (e.g., tubulins) or glycogen are produced in the cytophore and then transferred to the germ cells (Martinucci et al. 1977). The presented observations, which show that MTs were not detectable during late spermiogenesis either within the IBs or the cytophore, suggest that there is no transfer (or that the transfer is weak and/or not MTs dependent) between the elongated spermatids and the cytophore and vice versa. At this time, the spermatozoa are almost fully formed a n d a r e r e a d y t o d e t a c h f r o m t h e c y t o p h o r e (Martinucci et al. 1977;own unpublished results). Drug experiments In order to study the role of microfilaments and MTs, we incubated the germ-line cysts in media containing cytoskeletal drugs such as cytochalasin D and latrunculin A, which affect the microfilaments and colchicine and nocodazole, both of which affect the dynamics of the MTs (see Lie et al. 2010 for literature and details). Surprisingly, although we used several different working concentrations and the time of incubation was as long as 48 h (the time was limited due to the vitality of the cysts in an artificial media), none of the drugs that were used altered the general geometry of the cysts (Fig. 5b). The cyst morphology in the blind samples (without any incubation) under the control conditions (cysts incubated without drugs) and in the experimental probes was very similar; the shape of both the germ cells and the cytophore was still stage-specific and not altered. Weak cytochalasin D and latrunculin A incubation effects were found within the germ cells and the cytophore in the form of numerous granular rhodamine-phalloidin-positive accumulations. More prominent effects were found after the incubation of the cysts in nocodazole: in this case, numerous spermatid nuclei were found within the cytophore in late spermiogenesis. It is known that the efficacy of cytoskeletal drugs, especially those that alter MTs, is organism-specific, i.e., in some organisms the drugs alter the cytoskeleton organization whereas other organisms remain resistant (Dostál and Libusová 2014). The mechanisms of resistance are especially well known in some cancer lines (Pellegrini and Budman 2005). Why cytochalasin D, latrunculin A and colchicine did not affect the cytoskeleton in D. veneta germ-line cysts remains unresolved. On the one hand, it is known that the MTs in spermatids are modified by acylation, polyglutamylation or detyrosination and are more resistant to depolymerization (see Sperry 2012 for details). On the other hand, the effects of colchicine treatment on spermiogenesis have been described in, e.g., molluscs via the disruption of chromatin condensation (Maxwell 1983), whereas the use of cytochalasin D in some flatworms (Stitt et al. 1991) and rodents (Russell et al. 1987(Russell et al. , 1991Ventelä et al. 2003) caused the formation of multinucleate spermatozoa or significantly inhibited the granule movement between neighboring cells. The next explanation is connected with the timing of the germ cell divisions. We do not know whether germ cells divide at least once during the maximal treatment period (48 h). If the cells do not divide, this might explain the absence of some drug effects. Further analyses with an extended period of incubation and using drugs that have different mechanisms of cytoskeleton disruption than those that were used in the presented study such as, e.g., colcemid or taxol, should explain whether the F-actin and MT in the male germ-line cysts of D. veneta are especially resistant to cytoskeletal drugs. Using immunofluorescence, we showed that nocodazole caused the depolymerization of MTs; however, the cyst morphology was untouched, which shows that the MTs are not responsible for keeping the shape of the germ cells and the cytophore. Interestingly, while the nuclei of spermatogonia and spermatocytes did not change their positions, the nuclei of the elongated spermatids had a tendency to move away from the cells towards the cytophore in the nocodazoleincubated cysts. This suggests that not only the IBs prevent the spermatid nuclei from passing from the cells into the cytophore as was suggested based only on the ultrastructural observations (Martinucci et al. 1977) but also that the microtubular cytoskeleton is engaged in keeping them in a place. Most likely a manchette, a structure surrounding the spermatid nuclei in a diverse group of animals that is rich in MTs (e.g., annelids, molluscs, mammals; see Maxwell 1983;Ferraguti 2000;O'Donnell and O'Bryan 2014), is responsible for the positioning of the spermatid nuclei during late spermiogenesis. The structure and functions of the manchettes are best known in rodents, in which the manchette contains both MTs and F-actin and is engaged in, e.g., shaping the sperm head (Russell et al. 1991) and cytoplasmic redistribution (Kierszenbaum 2002). The microtubular manchette is also well developed in clitellate annelids (Jamieson 1981(Jamieson , 1992Ferraguti 2000) and some ultrastructural observations suggest that there is a relationship between the MTs, chromatin condensation and nuclear morphogenesis (Ferraguti and Lanzavecchia 1971). Our experiments showed that nocodazole disrupts the integrity of the manchette and that some fragments of the manchette, together with spermatid nuclei, also move

from the spermatids into the cytophore. It may be supposed that, in natural conditions, the manchette prevents the spermatid nuclei from moving while the residual cytoplasm is transferred towards the cytophore; when the manchette is disrupted, it flows (in an actin-dependent manner?) into the cytophore together with the nucleus. A functionally similar mechanism that prevents the germ cell nuclei from moving towards the IBs during mass cytoplasm transfer is known from the ovaries of some insects; e.g., in some hymenopterans, MTs form cages around the nurse cell nuclei and after colchicine treatment, the nuclei move towards the IBs (for details, see Adamska and Biliński 1997;Biliński and Jaglarz 1999). Finally, it should be added that there are some ultrastructural reports that show that, in natural conditions, the nuclei of the spermatids in some clitellate annelids may extend through the intercellular bridges into the cytophore (e.g., early spermatids in the earthworm Microchaetus pentheri, Hodgson and Jamieson 1992 and in E. foetida, Martinucci et al. 1977) or that even spermatid nuclei have been observed within the degenerating cytophore (Martinucci et al. 1977). The nuclei of the female germ cells have also occasionally been found within the cytophore in late oogenesis in several clitellate annelids (Świątek 2008;Urbisz and Świątek 2013;Urbisz et al. 2014). It has been suggested that such a phenomenon is connected with defective mechanisms of spermatid detachment that is related to the relaxation of an IB (Martinucci et al. 1977) or with the degeneration of germ cells and/or the disturbance of the cytoplasm transfer from the nurse cells to the oocytes via the cytophore (Urbisz and Świątek 2013). Conclusions The results that were obtained show that: 1) F-actin is enriched in the cortical cytoplasm and forms distinct rings in the IB rims in all the stages of cyst development that were studied (Fig. 5a); 2) the distribution of the MTs changes dynamically in the consecutive stages of spermatogenesis (Fig. 5a); 3) cytoskeletal drugs such as colchicine, nocodazole, cytochalasin D and latrunculin A do not alter the general morphology of cysts; 4) after treatment of the germ-line cysts with nocodazole, spermatid nuclei and manchette fragments were found within the cytophore (Fig. 5b); 5) it appears that the MTs play the main role in cytoplasm/organelle transfer between the cells and the cytophore and prevent the nuclei of spermatids from passing through the IBs. Beyond the Core-Deficit Hypothesis in Developmental Disorders Developmental disorders and childhood learning difficulties encompass complex constellations of relative strengths and weaknesses across multiple aspects of learning, cognition, and behavior. Historically, debate in developmental psychology has been focused largely on the existence and nature of core deficits—the shared mechanistic origin from which all observed profiles within a diagnostic category emerge. The pitfalls of this theoretical approach have been articulated multiple times, but reductionist, core-deficit accounts remain remarkably prevalent. They persist because developmental science still follows the methodological template that accompanies core-deficit theories—highly selective samples, case-control designs, and voxel-wise neuroimaging methods. Fully moving beyond “core-deficit” thinking will require more than identifying its theoretical flaws. It will require a wholesale rethink about the way we design, collect, and analyze developmental data. these theories, a complex array of observable characteristics are frequently categorized according to a single defining neurocognitive deficit. As understanding of a particular set of diagnostic features evolves, most such theories are gradually pruned from the field because of insufficient evidence, counterevidence, or the emergence of betterspecified theories. These theoretical accounts fail, but that is their purpose. Choose any clinical categorization of developmental differences, and there is a graveyard of once-popular theoretical accounts, from the magnocellular theory of dyslexia (Stein, 2001) to the mirror-neuron theory of autism (Williams, Whiten, Suddendorf, & Perrett, 2001). Over and above specific problems with any single theory, this general class of theory is problematic because of the reliance on the very notion of a "core deficit"something that has been repeatedly debunked both within specific (Happé, Ronald, & Plomin, 2006) and across multiple diagnostic categories (e.g., Pennington, 2006). But in practice, developmental psychologists and developmental-cognitive neuroscientists frequently return to core-deficit theories, either implicitly or explicitly, if not to motivate studies, then to interpret them. Here, we provide a working definition of a core-deficit account, describe the inherent weaknesses in the theory's accompanying methods and methodology, and argue that building the empirical foundations for more complex (and accurate) theoretical frameworks will require a wholesale rethink in the way we design, collect, and analyze developmental data. Identifying a Core-Deficit Hypothesis A core-deficit hypothesis pins a multiplicity of cognitive, behavioral, and neurobiological phenomena onto a single mechanistic impairment and is assumed to have the power to explain all observed profiles within a particular diagnostic category. To provide one example, in autism, the dominant core-deficit hypothesis since the mid-1980s has been the theory-of-mind model. This model proposes that autistic people uniquely lack the ability to detect, interpret, or understand the mental states of others. Versions of this model vary in the range of behaviors and mental states they attempt to encompass. At its broadest, the theory-of-mind deficit can draw in differences in play, language development, and all types of processing of emotions, including basic emotions, as well as desires, beliefs, and higher-order complex mental states such as suspicion. On a narrower scale, theory-of-mind impairments in autism are understood to apply only to the automatic and easy processing of complex mental states. Until recently, evidence for this latter position has looked fairly robust, but now even this pillar of autism theory is under threat, as innovative research has revealed that the apparent "deficits" in mental-state understanding exhibited by autistic people may apply only to understanding the mental states of nonautistic people (e.g., Crompton et al., 2020). At the same time, there is a growing body of evidence showing that nonautistic people show impairments in detecting and interpreting the mental states of autistic people (e.g., Edey et al., 2016). This reframes the characteristic "deficit" of autism as a typical manifestation of failed communication across different sociocultural groups (Milton, Heasman, & Sheppard, 2018). More importantly for our argument here, the model has never been adequately able to explain the sleep disturbances, sensory sensitivities, intense and consuming interests, or executive difficulties that are equally prevalent among autistic people. The Origins of Core-Deficit Theories and Their Persistent Appeal Across multiple categories, and despite well-articulated perspective articles highlighting their limitations (Happé et al., 2006;Pennington, 2006), core-deficit accounts remain a recurring theme in developmental-disorders research (e.g., Kucian & von Aster, 2015;Mayes, Reilly, & Morgan, 2015). Part of their intuitive appeal is their relative simplicity. Publishing in higher-tier journals is easier when the story is simple, and doing so can quickly turn into a citation gold mine if the field invests in challenging your theory. If a series of findings can be woven together, this provides an ideal way of combining decades of research and cementing its contribution to the field. However, such tapestries are riddled with loose threads that, if pulled, quickly reveal the flaws in their construction. Examples of these construction flaws are found in the design choices, participant selection, sample sizes, statistical methods, and restrictive range of measures that are hallmarks of core-deficit thinking. These flaws are still prevalent in the wider developmental literature because despite being mindful of the problems inherent in core-deficit theories, we have yet to change the methodological template inherited with them. Highly selective samples Studies of developmental disorders typically use strict exclusion criteria (e.g., Toplak, Jain, & Tannock, 2005;Willcutt et al., 2001), including the removal of children with co-occurring difficulties. But in reality, co-occurring difficulties are the norm rather than the exception (Gillberg, 2010). For example, it is rare to have selective reading or math impairments; the majority of children with one difficulty will also have the other (e.g., Landerl & Moll, 2010). The same is true within clinical samples: 44% of children who receive an ADHD diagnosis would also meet the diagnostic criteria for a learning difficulty, and a similar percentage who meet the latter criteria would also meet the criteria for an ADHD diagnosis (Pastor & Reuben, 2008). Because most children with the difficulties of interest are excluded, the literature overstates the "purity" of developmental differences. In turn, basing models on highly selective samples biases theory toward simpler core-deficit accounts. Where studies do include children with different diagnoses or with co-occurring difficulties, the purpose is usually to identify what is unique to each diagnosis rather than to establish which dimensions are important for understanding individual outcomes, irrespective of the diagnostic category applied. Study designs do not capture heterogeneity Most studies use a case-control design, with children grouped according to strict inclusion criteria (see above) and then matched to one or more control groups. Significant differences in group-level statistics are then taken as evidence for a specific deficit in the "case" group. Variability in performance within groups is rarely studied, in part because few studies have sufficient power but also because of reliance on univariate statistics. This issue becomes more and more problematic as diagnostic criteria broaden. Regarding heterogeneity as noise to be minimized removes a crucial signal that could lead to a more complex and accurate theoretical conclusion. Although we endorse the application of Occam's razor to interpretation of findings, we must be wary of parsimony achieved via flawed means. Circular logic of measurement selection Theoretical conclusions about underpinning mechanisms are constrained by our choice of measures. Tasks are often included because they are regarded as the gold standard for distinguishing a particular group from controls, even though the conceptual underpinnings of the task are poorly understood. The logic can become circular: "We always include a theory-of-mind task when studying autistic people, and autism is characterized by theory-of-mind task performance." The task has become an implicit requirement within the field without any real mechanistic understanding of why different children find it hard. Moreover, the dominance of a specific core-deficit theory with an associated goldstandard measure eliminates consideration of other possibilities: In the case of autism, why not consider an executive-planning or sensory-profiling measure? If the same domains of measurement are selected in every study, to the exclusion of alternatives, then the supposed cardinality of this profile will inevitably be reinforced. But little has been explained, and alternative possible profiles are not documented. Neuroimaging methods inherited from the adult literature Finding a shared neural substrate that purportedly causes the difference is typically taken as strong evidence for a core-deficit theory. But this is largely an artifact of the analytical approach we take to neuroimaging. Most canonical neuroimaging methods assume a direct correspondence between spatially overlapping brain differences (structural or functional) and surfacelevel cognitive profiles. A voxel-wise approach to analysis will always produce peaks. Despite its dominance, this approach yields remarkably inconsistent results. For example, ADHD has been associated with differences in gray matter within the anterior cingulate cortex (Amico, Stauber, Koutsouleris, & Frodl, 2011), caudate nucleus (Onnink et al., 2014), pallidum (Frodl & Skokauskas, 2012), striatum (Greven et al., 2015), cerebellum (Mackie et al., 2007), prefrontal cortex (Dirlikov et al., 2015), premotor cortex (Mahone et al., 2011), and most parts of the parietal lobe (Shaw et al., 2006). Our contention is that the assumption of spatial correspondence is not valid for understanding brain-cognition or brain-behavior links in childhood, especially in children with developmental disorders (see also Johnson, 2011). Difficulties that emerge over developmental time can be arrived at via multiple different underlying neural routes-a phenomenon called equifinality (Bishop, 1997). There may also be multifinality: The same apparent underlying neural effects can have different consequences for behavior and cognition across children (Siugzdaite et al., 2020). The canonical voxel-wise neuroimaging methods that dominate developmental cognitive neuroscience instead create the false impression that the neural underpinnings of developmental differences are akin to those resulting from acquired focal brain damage, and in turn this implicitly leads us back to "core deficits" in our theoretical interpretation. Moving Beyond the Core-Deficit Hypothesis There is no shortcut to a comprehensive empirical basis for more complex theory. Nonetheless, in this section, we make some nonprescriptive suggestions as a position from which to move forward. Well-powered studies with inclusive samples If our samples are more reflective of the children we are seeking to understand, then our theories, though necessarily more complex, will likely have greater practical value. It is necessary to include participants with different or multiple diagnoses. Emerging

first in adult psychiatry (Cuthbert & Insel, 2013;Morris & Cuthbert, 2012), transdiagnostic approaches focus on identifying underlying symptom dimensions that likely span multiple supposed categories (e.g., Astle Zhao & Castellanos, 2016). A review of a transdiagnostic approach is well beyond the scope of the current article, but suffice it to say, contemporary developmental science needs larger and more diverse samples. Embracing methods that capture complexity There is now a growing array of analysis methods allowing researchers to move beyond univariate group comparisons or pairwise associations between variables. A well-developed tool is structural equation modeling (SEM; e.g., Kline, 2015). SEM combines the strengths of path-modeling and latent-variable approaches to allow researchers, for instance, to specify how latent factors may explain the continuous variability on a set of observed measures (e.g., task performance) and the potential causes and consequences of such factors. SEM offers the tools to identify continuous dimensions that cut across diagnostic boundaries or to directly compare competing causal accounts. More advanced variants of SEM, such as latent-growth-curve models and growthmixture models, can address more sophisticated questions about how changes in one latent construct relate to changes in another (e.g., Hjetland et al., 2019;Kievit, Hofman, & Nation, 2019). Other variants-for example, hierarchical mimic modeling-can be used to identify multiple routes to particular outcomes (equifinality), such as the role that multiple different brain structures might play in developmental changes to a particular aspect of cognition (e.g., Kievit et al., 2016;Ritchie et al., 2015). Whereas SEM methods are ideal for testing complex theories or pitting theories against one another, other methods are able to handle complexity in a different way. A more recent development from data science that aims to capture complex interrelationships in an exploratory way is network analysis (e.g., Bathelt, Holmes, et al., 2018;Epskamp, Borsboom, & Fried, 2018;Mareva, the CALM team, & Holmes, 2019). The resulting network can open the door to a new toolbox of analytical techniques, such as graph theory. Rather than inferring the presence of singular latent factors, these approaches capture various different ways in which individual measures can be related over developmental time. For example, does phonological processing act like a hub for verbal short-term memory and literacy development, or is there a dynamic relationship between these constructs over time? Network analysis can also capture heterogeneity within a sample and provide a relatively theory-free means of exploring the underlying structure and composition of a data set. With a network analysis, it is possible to identify subgroups of individuals within which the task interrelationships are different, and a strong advantage of this approach is that a network analysis incorporates metrics for identifying these clusters (e.g., Bathelt, Holmes, et al., 2018). Unsupervised machine-learning techniques are also capable of capturing the multidimensional space in which children may differ (e.g., Siugzdaite et al., 2020). But relative to SEM-based approaches and network analysis, machine-learning applications remain underdeveloped. Although popular in other areas of science with similar challenges, these methods have yet to gain much traction within the study of developmental differences (but see Astle et al., 2019). These algorithms are highly flexible, and the resulting models can easily accommodate nonlinear relationships, make predictions about unseen data, be combined with simulations, incorporate different data types, and open the way to tools for testing generalization, such as cross-validation. For the data shown in Figure 1, a simple artificial neural network was used to map different profiles across multiple measures of short-term memory, working memory, fluid reasoning, vocabulary, and phonological awareness in a large group of struggling learners. Once the Fig. 1. Distributions of children within a simple artificial neural network trained on data from 530 children taken from the Centre for Attention, Learning and Memory (CALM) sample (Holmes, Bryant, the CALM Team, & Gathercole, 2019). Each node represents a profile learned by the algorithm, with spatially nearby nodes having more similar profiles. Therefore, the maps represent the multidimensional spaces that reflect the performance differences across the children. The left-most panel shows the best-matching unit for all children, and the subsequent panels show those for children with different diagnoses. The training data set included measures of fluid reasoning, vocabulary, verbal and spatial short-term and working memory, and phonological awareness (see also Astle, Bathelt, The CALM Team, & Holmes, 2019). ADHD = attention-deficit/hyperactivity disorder; ASD = autism spectrum disorder. individual nodes of the network were trained to represent the different profiles within the data set, the locations of children with different diagnoses could be identified. If common diagnoses are predictive of the cognitive profiles learned by the network, then those children's performance profiles should group togetherbut they did not. This highlights the potential utility of this approach to mapping the heterogeneity of a data set and exploring its composition. This is something largely untapped within the field to date. Beyond voxel-wise measures of brain structure and function Just as methods that capture complexity in cognitive and behavioral data are needed, the same is true for neuroimaging. In theory, the methods outlined above can be used alongside neuroimaging, although in practice there are very few examples. SEM techniques could allow researchers to identify many-to-one mappings (e.g., Kievit et al., 2016;Kievit et al., 2014), allowing the possibility that a particular set of symptoms could be associated with multiple different neurobiological effects. And network science has enabled the subfield of brain "connectomics," enabling researchers to demonstrate that apparently disparate neural effects could have very similar effects on brain organization, providing a meaningful endophenotype to bridge complex causal interrelations and shared surface-level profiles (e.g., Bathelt, Gathercole, Butterfield, the CALM Team, & Astle, 2018;Bathelt, Scerif, Nobre, & Astle, 2019). Conclusions Core-deficit models long held promise as optimistic researchers aimed to interpret behavioral complexity via cognitive or neurological simplicity. However, as more and more of these attempts fall by the wayside, many researchers have come to question the validity of the principle of the core deficit. At the same time, increased pooling and sharing of data, as well as better diagnostic ascertainment, has improved our capacity to gather substantial samples for well-powered complex analyses. New technologies provide opportunities for creative, data-driven analysis of such samples, which can provide us with an empirical basis for the development of new theories. We must embrace complexity within and between diagnostic boundaries in such theoretical models. In doing so, we will unlock our potential to understand the cross-cutting issues that frequently have the biggest impact on people's lives rather than dwell on the narrow selection of domains that seem to be unique to a specific population. Importantly, a move away from the concept of "core" should also entail a movement away from the concept of "deficit"-there is no objective reason why a condition should be defined by its disadvantageous elements instead of its advantageous elements. Although the former may be needful of intervention, the latter may be essential to delivering that intervention. Examples of successful application of this principle come from formal evaluations-for example, technological supports for autistic children (Grynszpan, Weiss, Perez-Diaz, & Gal, 2014;Kasari et al., 2014)-but are also evident in practitioner-focused guides-for example, image-based rather than text-based learning materials for children with dyslexia (Mortimore & Dupree, 2008). In this way, the current direction of research in neurodevelopmental diversity is at a potential tipping point. The issues we outline above, and the recent developments we highlight, could have a beneficial impact not only on research innovation and knowledge generation but also on policy, practice, and societal understanding. (See References). Uses an artificial neural network to map between cognition and brain structure. Comparison of Reconstituted, Acidified Reconstituted Milk or Acidified Fresh Milk on Growth Performance, Diarrhea Rate, and Hematological Parameters in Preweaning Dairy Calves Simple Summary The preweaning phase is the period for the rapid growth and development of dairy calves. During this period, dairy calves receive their nutrients through milk. Feeding hygienic milk is of great benefit to optimum growth rate and health status of dairy calves. Previous studies focused on the effects of hygienic milk by acidification on dairy calves’ health and growth. Reconstituted milk, as the common source of milk, is being used in dairy calves feeding. However, no previous studies reported the effects of feeding acidified reconstituted milk on dairy calves’ health and growth. Our study will provide the evidence that the acidification of reconstituted milk had positive effects on growth performance and health status of preweaning dairy calves. Abstract The present experiment was carried out to assess the effects of reconstituted milk (RM), acidified reconstituted milk (ARM), and acidified fresh milk (AFM) on growth performance, diarrhea rate, and hematological parameters of preweaning dairy calves. For this purpose, a total of 27 Holstein female calves (one month of age) with initial body weight of (67.46 ± 4.08) kg were divided into three groups in such a way that each group contained nine calves. Calves were housed individually, and starter was offered ad libitum to each calf. The dietary treatments were RM, ARM, and AFM. The highest milk intake was observed in calves receiving AFM as compared to other treatments (p < 0.01). Calves fed AFM had more feed intake than those fed ARM and RM (p < 0.01). Feed efficiency was significantly lower for calves offered ARM than those offered RM and AFM (p < 0.01). A lower withers height growth was found for calves fed RM than those fed ARM and AFM (p <0.05). Diarrhea rate and white blood cell (WBC) and lymphocytes (LYM) counts were greater for calves fed RM than those fed ARM and AFM (p < 0.05). These findings suggested that ARM and AFM had positive effects on growth performance and health status of the preweaning dairy calves. Introduction The preweaning phase is the period for the rapid growth and development of female dairy calves [1]. Providing calves with sufficient nutrients to achieve the optimum growth rate in preweaning phase is beneficial for promoting growth performance after adulthood, which positively associated with increased productivity in later life [2]. On large scale dairy farms, pre-weaning dairy calves are usually reared by milk or milk replacer [3][4][5]. Delay of feeding fresh milk to calves results in the multiplication of bacteria in the milk due to favorable conditions. Therefore, preservation of milk is recommended at large-scale dairy farms to overcome the consequences of bacterial load in fresh milk. It has been reported that acidification of milk to a pH value between 4.0 and 4.5 can inhibit pathogenic bacterial growth and enable milk to be preserved without refrigeration for a short time prior to feeding to calves [6,7]. It has also been reported that acidified milk protects calves from diseases caused by bacteria growth and reduces the incidence of diarrhea [8]. Reconstituted milk is resulted from the addition of water to dried or condensed milk powder in the amount necessary to reestablish the specified water:solids ratio. Reconstituted milk is also being used on the dairy farms for preweaning dairy calves [9]. However, to our knowledge, the acidification of reconstituted milk in preweaning growing calves has not been investigated. Previous studies only focused on acidified milk or acidified milk replacer in dairy calves [10][11][12]. Therefore, a comparative study was undertaken to compare the growth performance, diarrhea rate, and hematological parameters of preweaning dairy calves fed reconstituted milk, acidified reconstituted milk, or acidified fresh milk. Calves, Feeds and Management The animal care and experimental procedures were conducted by the Animal Welfare and Ethics Committee of Heilongjiang Bayi Agriculture University. Animal care and handling were followed the guidelines by the regulations for the Administration of Affairs Concerning Experimental Animals (The State Science and Technology Commission of China, 1988). Twenty-seven Holstein female calves procured from large dairy herd of same age with initial body weight of (67.46 ± 4.08) kg were selected and randomly divided into three groups in such a way that each group contained nine calves. Three types of milk were applied to different treatment groups-(1) reconstituted milk (RM), which was made by mixing milk powder with warm water (35-40 • C) in a weight ratio of 1:7; (2) acidified reconstituted milk (ARM), which was obtained by adding 30 mL formic acid (food grade) of 8.5% concentration in 1L reconstituted milk at temperature 5-10 • C; and (3) acidified milk made by using fresh milk (AFM), which was obtained by adding 30 mL formic acid (food grade) of 8.5% concentration in

1L fresh milk at temperature 5-10 • C. The fresh milk was procured from general and healthy herd and not pasteurized before feeding. Milk samples of 50 mL were collected every 10 days to detect the compositions. A milk analyzer (Foss Milkoscan 4000, Foss, Hillerod, Danmark) was used to determine the composition of milk, and milk composition are shown in Table 1. The pH value of acidified milk was determined by a pH meter (FE20, Mettler Toledo, Zurich, Switzerland). Milk samples of 50 mL were successively diluted (10 −1 to 10 −4 ) using distilled water, and a volume of 1 mL of each dilution was inoculated onto the surface of plate count agar (PCA), violet red bile agar (VRBA), and MRS agar for the cultivation of total bacteria, lactobacillus, and Escherichia coli, respectively. Total bacteria and Escherichia coli were incubated at 37 • C for 48 and 24 h. Lactobacillus was incubated anaerobically at 37 • C for 72 h. After incubation, the numbers of total bacteria, lactobacillus and Escherichia coli colony-forming units (cfu) were calculated. Bacterial numbers of different types of milk are shown in Table 2. Prior to onset of trial, all the calves were offered the same basal ration from day 1 to 21. Experimental treatments were applied from day 22 onward. Calves were given a nine-day adaptation period before the start of data collection. No abnormal behaviors were observed during the adaptation period, and the initial data were collected on day 30. Calves were housed individually in calf hutches, and 6 L milk were offered twice daily at 5:00 and 16:30. All calves had free-choice access to clean, fresh water and starter throughout the experiment. The chemical compositions of the starter diets are shown in Table 3. The experiment period lasted for 40 days until weaning. Standard management and environmental conditions were ensured to avoid any stress as described in recent researches [13,14]. Growth Performance Calves were weighed before the morning feeding for three consecutive days at the start and end of the experiment to calculate the initial body weight, final body weight, and average daily gain (ADG). Milk and starter consumption were recorded daily. Feed efficiency was determined from total feed consumption and ADG. Feed efficiency = Total feed intake(kg/day DM) ADG(kg/day) In addition, body measurements of each calf including body length, withers height, heart girth, and shin circumference were taken at the beginning and end of the trial according to the study made by Li et al. (2014) [15]. Fecal Score and Diarrhea Rate Fecal consistency was scored everyday on a scale of 1 to 4, where 1 = normal-firm but not hard, 2 = soft-does not hold form, 3 = runny-spreads easily, and 4 = devoid of solid matter. Fecal score of 3 or 4 were considered as diarrhea. The diarrhea rate was calculated according to the procedure described in the recent study [16] and formula is given below. Diarrhea rate = number of calves with diarrhea × days of diarrhea total number of calves × examined days × 100 (2) Hematological Parameters Blood samples (about 10 mL) were taken from the jugular vein in the morning, approximately 3h post feeding on the first and final day of the experiment. Blood was anticoagulated with EDTA for hematological analysis within three hours after collection. Hematological parameters values including white blood cell (WBC), lymphocytes (LYM), monocytes (MO), red blood cell (RBC), hemoglobin (Hb), and hematocrit (HCT) were detected by using automatic hematology analyzer (XN9000, Sysmex, Kobe, Japan). Statistical Analysis All the data were analyzed by PROC MIXED of SAS version 9.4 (SAS Institute Inc., Cary, NC, USA), the statistical model was where Y ij is the observed variable, µ is the overall mean, Di is the fixed effect of diet treatment, C j is the random effect of calves, and ε ij denotes the residual error. Significance was declared for p < 0.05, and trends were reported at 0.05 < p < 0.10. When a significant effect of treatment was detected (p < 0.05), differences between the means were tested using Bonferroni multiple comparison test. Growth Performance The results of growth performance of dairy calves are shown in Table 4. The initial BW, final BW, ADG, and starter intake were not different among treatments. Milk fed amount in liters were not different among different treatments. However dry matter intake from milk differed among treatments. Lowest dry matter intake from milk was observed in calves fed ARM, while the highest dry matter intake from milk was observed in calves fed AFM (p < 0.01). Calves fed AFM had increased total feed intake compared to those fed ARM and RM (p < 0.01). The value of feed efficiency was significantly lower for calves offered ARM than those offered RM and AFM (p < 0.01), which suggested that the feed efficiency was better for calves fed ARM than RM and AFM. Body measurements data of different treatments are presented in Table 5. Body measurements were not different at the beginning of the experiment. Smaller heart girth was observed in the end of the experiment for calves fed AFM than those fed ARM and RM (p < 0.05). Compared with calves fed ARM and AFM, a lower withers height growth could be found for calves fed RM (p < 0.05). Similarly, final withers height was lower for calves fed RM as compare to ARM and AFM (p < 0.05). Fecal Score and Diarrhea Rate Fecal score and diarrhea rate of dairy calves were presented in Table 6. Average fecal consistency score and diarrhea rate during 30-50 day, 51-60 day, and 61-70 day were greater for calves fed RM than those fed ARM and AFM (p < 0.01). Hematological Parameters Hematological parameters of different treatments are presented in Table 7. Hematological parameters values of WBC, LYM, MO, RBC, HB, and HCT were not different among the treatments at the start of the experiment. At the end of the trial, higher white blood cell and lymphocytes counts were found in calves offered RM (p < 0.05). Growth Performance Preweaning growth of female dairy calves is considered an important factor of milk production in their later stage of life. Feed intake, ADG, and feed efficiency are important indicators for growth performance of female dairy calves. Feed intake affects the level of nutrient intake, thus affecting ADG and final BW. ADG and final BW play an important role in assessing the feed efficiency and overall health status of dairy female calves [17]. Akins reported that female diary calves should gain 0.75 kg per day to reach optimum weight of weaning at eight weeks of age [18]. In the present study, ADG of calves in different treatment groups exceeded 0.75 kg/day. Although ADG were in favor of calves fed ARM and AFM compared to those fed RM (0.92, 0.96 vs. 0.88 kg/day), however the differences were not significant. Eren found that the calves offered acidified milk had increased ADG than those offered whole milk [19]. Zhang et al. reported that acidified milk replacers contribute to the improvement of growth rate of calves [12]. Dry matter intakes (DMI) from milk was the highest for calves fed AFM. This was expected as the highest total solid concentration of AFM which resulted in the higher consumption of DM from milk by calves. DMI from starter were not different for calves fed ARM, RM, and AFM, which were 0.43, 0.40, and 0.42 kg/day DM, respectively. The results are consistent with other research [10][11][12]20]. However, Sun et al. found a reduction in the intake of starter for the calves fed acidified milk with butyric acid addition [16]. Interestingly, in the current study, starter feed DMI was similar among all the group, which could be explained by different acidifier and experimental treatments as reported in the previous studies. Overall, ARM, RM, and AFM behave similarly on the intake of starter feed in the current study. Body measurements usually reflect body growth and development, which can be used to estimate liveweight [21]. Calves assigned to the ARM and AFM treatment groups exhibited greater withers height gain in the body measurements than the RM calves. The results suggested that calves reared using ARM and AFM had better skeletal growth. Todd et al. reported the calves fed acidified milk replacer were associated with greater preweaning structural growth [5]. However, some studies found that body measurements of calves were not affected by feeding acidified milk or acidified milk replacer [3,10,16]. Fecal Score and Diarrhea Rate Diarrhea is the most important cause of disease of calves, with high morbidity and mortality, which threatens the health and growth performance of calves and is a major cause of economic loss for cattle raisers [22]. Results in the present study showed that calves fed ARM and AFM exhibited lower fecal consistency scores and diarrhea rate. It has been reported that lower pH values inhibit the growth of total bacteria and Escherichia coli [23], and the current study provided the evidence that ARM and AFM contained lower number of total bacteria and Escherichia coli ( Table 2). The lower number of total bacteria and Escherichia coli reduces the growth and reproduction of saprophytic bacteria that harm the health of gastro-intestinal tract and cause diarrhea. On the other hand, feeding acidified milk can increase the acidity in the intestinal tract of calves, as reported by Woodford et al. [24]. Relatively higher acidity of intestinal tract is likely to exhibit a bacterial influence, therefore reducing the incidence of diarrhea among calves. Therefore, the lower incidence of diarrhea in ARM and AFM could be explained by the lower pH of intestine caused by ARM and AFM and by the lower number of total bacteria and Escherichia coli. Hematological Parameters Hematological parameters such as WBC, LYM, and RBC counts and Hb concentration are considered as important clinical indicators that are widely used to reflect health and disease status [25]. Diseases of the preweaning calves is one of the major cause of economic loss in dairy production [26]. Thus, the measurement of hematological parameters could be used to evaluate the health and physiological status of the calves [27]. WBC are the major part of the immune system that helps fight infection and defend the body against other foreign materials. Different types of WBC are involved in recognizing intruders, killing harmful bacteria, and creating antibodies to protect the body against future exposure to some bacteria and viruses [28,29]. LYM are one of several different types of white blood cells that are also important in the immune system, with T cells being responsible for directly killing many foreign invaders and B cells producing antibodies to guard the body against infection [30]. Some research reported that mean value of WBC and LYM were 8.08-12.75 (10 9 /L) and 2.67-4.77 (10 9 /L) of weaning Hanwoo and Holstein calves [31][32][33]. In the present study, mean value of WBC was 28.55 (10 9 /L) and LYM was 4.83 (10 9 /L) for calves fed RM, which were much higher than that of the research mentioned above. However, mean value of WBC and LYM for calves offered ARM and AFM were very close to the values of the research we have reviewed. Therefore, it can be concluded that the calves offered RM were susceptibility to infection, which may result in detrimental effects on calf health. Conclusions Results from this research indicated that calves fed acidified reconstituted milk or acidified fresh milk had greater average daily gain and withers height growth, lower diarrhea rate, and white blood cell and lymphocytes counts. We concluded that calves offered reconstituted milk were susceptible to infection, which may have a negative effect on calf health. Acidification of reconstituted milk as well as fresh milk had positive effects on growth performance and health status of the preweaning calves. A study of maternal mortality in Government Maternity Hospital, Sri Venkateswara Medical College, Tirupati Introduction: Maternal mortality is a key indicator of health services provided to population and reflects the health status of community. Aim: To study the factors responsible for maternal mortality. Materials and Methods: Maternal deaths that occurred in one year in Government Maternity Hospital, Tirupati were studied. Present study was conducted from January 2017 to December 2017 in Department of Obstetrics & Gynaecology, Government Maternity Hospital, S.V. Medical College, Tirupati. Results: 21

maternal deaths had occurred during the study period. 57.1% were unbooked cases, 61.9% were primiparous women and 61.9% were belonging to the age group of 21-25 years. Leading cause of death was hypertensive disorders of pregnancy (47.6%) followed by postpartum haemorrhage (28.4%). Maternal deaths due to direct causes were 85.7% and indirect causes were 14.3%. type 2 and 3 delays contributed to 57.1% and 38.1% maternal deaths respectively. Most of the cases were referred to the institute in late stages. Conclusion: Health education, regular antenatal check ups, early identification of high risk pregnancies and timely intervention are needed to reduce maternal mortality. Introduction Maternal death is defined as death that can occur as a direct result of obstetric complications or indirectly due to pregnancy-induced aggravation of pre-existing medical conditions, but not due to incidental or accidental causes. 1 The same was defined and classified according to WHO international classification of diseases, 10th revision. 2,3 Direct maternal deaths are caused by pregnancy, labour and puerperium. Indirect deaths are caused by previously existing disease or disease that developed and/or increased during pregnancy. Maternal mortality ratio is defined as the number of maternal deaths that occurred per 1 lakh live births up to 42 days after the termination of pregnancy. 1,4 Inspite of marked fall in maternal mortality in recent years, every day about 830 women die from pregnancy or childbirth-related causes. These deaths are preventable. Goal of United Nations is to reduce MMR by 75% by 2015. 5 There is a wide variation in MMR in India and variations in the regional states. Multiple factors like social factors, economic status, literacy, patient care, hospital care, infrastructure facilities in the referral units etc has its impact on maternal death. The United Nations Population Fund (UNFPA), 2017 report is as follows: one woman dies every two minutes and 20 to 30 women have suffer with severe morbidity for every one woman who dies. 6 Global ratio of maternal mortality is 400 per one lakh live births. India has made significant progress in reducing maternal mortality. MMR ratio has declined from 556 per one lakh live births in the year 1990 to 167 in 2011-13 and 130 per one lakh live births in the year 2016. 6 Approximately 44000 maternal deaths occur in India. Millennium Development Goals mandate a further reduction in MMR to 103. Maternal death review process initiated by the Government of India in 2010 by analysing and identifying lacunae in the healthcare systems to improve the quality of obstetric services. 7 Approximately 1.4 lakh women die every year. The target of MMR set for India was 139 per one lakh live births. This had to be achieved by the year 2015 and India has achieved MMR of 130 per one lakh live births by the year 2016. 7 As per sustainable developmental goals India has aimed to reducing MMR to less than 70 per one lakh live births. 7 The various states in India have a stark contrast in MMR. Among the south Indian states, Kerala has the least MMR with 46 per 1 lakh live births for the year 2016. Andhra Pradesh has reduced MMR from 92 per one lakh live births in year 2013 to 74 per one lakh live births in year 2016. Among all the states Assam has highest MMR of 237 per one lakh live births in the year 2016. 8 Main causes of maternal mortality is postpartum haemorrhage (27.1%) worldwide. Indirect causes of maternal deaths are anemia, malaria, heart disease (27.5%), infection (10.7%), unsafe abortions (7.9%), hypertensive disorders (14%), obstructed labour, ectopic pregnancy, embolism (3.2%) and anaesthesia complications. 45% of postpartum haemorrhage occur within first 24 hours. 66% occur during first week. 9 In developing countries MMR has been attributed to three delays-type 1. Delay in deciding to seeking care, type 2. Delay in time taken to reach the health care facility Type 3. Delay in receiving care at the health facility. 10 Government Maternity Hospital, Tirupati is a tertiary care centre receiving high risk and moribund patients. It is a referral centre for four districts and prestigiously known as Institute for Pregnant women. In Government Maternity Hospital, S.V. Medical College, Tirupati, 22 deaths occurred during the year 2015 and 13 deaths during the year 2016. The present study has been taken up to analyse causes with an aim to reduce preventable deaths. Aim To analyse and assess the direct and indirect obstetric factors involved in maternal mortality aimed to prevent maternal deaths and to provide safety measures to reduce maternal morbidity. Medical College, Tirupati were followed and studied. An analysis was done in the departmental MMR reviews with the involvement of professors, associate professors and assistant professors, department of anesthesiology, department of medicine and transfusion medicine. All cases were followed and analysed on demographic factors, social, obstetric factors, therapeutic factors and analysis of specific causes and leading cause of maternal mortality and identification of the level of delay done. Data was subsequently recorded in Microsoft Excel 2013 and analysed using descriptive statistics. Results Government Maternity Hospital of S.V. Medical College, Tirupati provided antenatal services both OP and IP to 64,299 antenatal women during the period of January 1, 2017 to December 31, 2017. During this period a total of 12,276 deliveries were conducted. Total number of live births were 12,134 during the study period. 21 maternal deaths occurred during the study period. 1 case of death remained inconclusive due to the failure of diagnosis of pregnancy in that specific case. The MMR in this institute is 173 per 1,00,000 live births. According to Kuppuswamy scale of socio-economic classification based on education, occupation and income the socio economic status of women were categorised. 11 90.4% of them belonged to class V and 19.6% belonged to class IV. 61.9% belonged to 20-25 years age group and were primipara. Direct causes are responsible for 85.7% of maternal mortality. 54.6% maternal deaths were due to hypertensive disorders and it is the leading cause of death in this institute. The other common cause of maternal death is haemorrhage. Most of the deaths in the institute occurred due to late referral. Majority of the women 13(61.9%) were in the age group of 21-25 years. 5(23.8%) were below 20 years. Youngest of them was a mere 15-year-old teenage girl. 13(61.9%) were primiparous and 8(38.1%) multiparous. 5 women were second gravida, 2 women were third gravida and one woman was fourth gravida. 13(61.9%) cases where deaths occurred were delivered by caesarean section. Indications for caesarean sections being performed were post caesarean pregnancies and causes like eclampsia where immediate termination of pregnancy was warranted. 6 cases were delivered by the vaginal route. 2(9.5%) were delivered by normal vaginal delivery, 2(9.5%) were delivered by abnormal vaginal delivery owing to prematurity of the fetus and 2(9.5%) were delivered by assisted vaginal delivery. 1(4.8%) case was delivered by hysterotomy. 1(4.8%) case died undelivered. Discussion All the maternal deaths that occurred were of women who belonged to the low socio-economic group. 20 of the 21 maternal deaths occurred after delivery and one case died during the antenatal period. Most of the cases were unbooked cases (57.1%) which showed the importance of antenatal care in preventing maternal death. In the present study 90.4% women belonged to lower socio-economic group according to Kuppuswamy scale and 19.6% were of the upper lower category. Most of the maternal deaths were among women belonging to poor class (76.4%). 11 Literacy, parity and socioeconomic status are some of the determinants of maternal mortality and can act as risk factors if are on the unfavourable side. 1 Primiparous (61.9%) women are at more risk of maternal mortality. 85.7% of maternal mortality is due to direct causes which include pre-eclampsia, eclampsia, PPH and antepartum haemorrhage. Deaths due to hypertensive disorders of pregnancy were the leading cause with 10(47.6%) of the 21 maternal deaths due to pre-eclampsia, eclampsia and HELLP syndrome. HELLP syndrome complicated 3 cases of severe pre-eclampsia and eclampsia. WHO and Government of India have listed haemorrhage as number one cause of death but in the present institute, hypertensive disorders formed a major share. This may be due to the present institute being a referral centre and receiving a large number of cases of hypertensive disorders of pregnancy in late stages. Younger age women are vulnerable to develop hypertensive disorders and maternal mortality. Postpartum haemorrhage, the leading cause of maternal deaths worldwide was a direct cause of death in 6 mothers during the study period. Introduction of SR cannula for prevention and control of atonic PPH during later end of the year had a significant impact in reducing morbidity and mortality due to PPH. 4 cases of atonic PPH were recorded, 1 death was due to traumatic PPH and another due to coagulation abnormalities causing uncontrolled haemorrhage. Sepsis as the cause of death was observed in 3 women and antepartum haemorrhage in 1. Postoperative complications were observed in 2 women who died. Pulmonary embolism, amniotic fluid embolism and medical complications were the cause in 5, 1 and 9 maternal deaths. Adverse blood transfusion reaction was responsible for 1 maternal death which had a renal failure due to the same and was referred to the present institute. 18 of the 21 mothers had live babies at delivery. 2 were delivered stillborn and 1 women died undelivered. Anaemia the leading and precipitating indirect cause of a number of obstetric complications was identified in 16(76.2%) women. Anaemia is an easily modifiable risk factor which if corrected can avert a number of maternal deaths. A large number of programmes are launched and implemented by Government of India like weekly iron and folic acid prophylaxis programme, parenteral iron therapy at PHC and CHC level, provision of diet to antenatal mothers and iron prophylaxis with IFA tablets with an aim of early detection and correction of anaemia. In the majority of the cases (16), death occurred due to multiple obstetric and medical complications one leading to another. A large number of women belonged to the age group of 21-25 years (61.9%). However other studies like Oladapo et al reported a higher maternal mortality in elderly women. 12 The disparity may be due to cultural differences regarding age at marriage and conception in India which is at a relatively younger age compared to other countries. Majority of the women in the present study were primiparous at death (61.9%). The results are higher when compared to study by Oladapo et al. where one third women were primiparous and similar to results in study by Paul B et al 12,13 (55.8%). The leading cause of death in the present study is hypertensive disorders of pregnancy i.e. 47.6% in the year 2017. The findings are similar to study by Paul B et al which reported first cause of maternal mortality as hypertensive disorders(30%). 13 However Government of India and WHO report haemorrhage as first cause of maternal mortality. The results can also be compared to study by Oladapo et al who reported leading cause of maternal deaths as hypertensive disorders of pregnancy (28%). 12 Direct cause of death is 85.7% in the present study. This is compared to study by Paul B et al who reported 76.7% direct causes of maternal mortality in 2006-07. 13 Haemorrhage, hypertension and sepsis were responsible for more than half of the deaths worldwide. Direct causes of death are postpartum haemorrhage (6), infection (3), embolism (6) and eclampsia (3). In the present institute 5.9% of cases of postpartum haemorrhage resulted in death out of the recorded 102 cases of major postpartum haemorrhage during 2017. In the present study type 1, 2 and 3 delays had role in 4.8%, 57.1% and 38.1% maternal deaths respectively. This is similar to the study by Paul B et al where type 1 and 2 contributed to 58% maternal deaths and type 3 contributed 46.5% maternal deaths. Regular monitoring of blood pressure and detection of albuminuria at primary health care centres with early detection of hypertensive disorders of pregnancy can prevent maternal deaths at facility level. Early referral must be done to tertiary care centres where blood transfusion, ventilator facilities and dialysis units are available. Interventions to reduce MMR are by increasing quality care services. Interventions are 1. Target Government of India has provided number of schemes to improve maternal health in the

form of PMSMA, JSSY for drugs, blood, transportation and early identification of risk factors for direct and indirect causes of maternal mortality at PHC and CHC hospitals for better care of antenatal mother. 14 Government of India has also provided infrastructure, obstetric ICU, emergency obstetric units and medical personnel at tertiary level to reduce maternal morbidity and mortality. There is also nil tolerance towards home deliveries. Conclusion The present study had aimed to analyse community factors and institutional factors contributing to maternal mortality. Hypertensive diseases of pregnancy presented as the most common cause of maternal mortality. Early antenatal care, health education, identification of hypertension at an early gestational age and timely intervention can reduce maternal morbidity and mortality. Postpartum haemorrhage can be controlled by active management of third stage of labour and employing newer techniques like SR cannula. 15 Lack of life support equipment, ventilator and dialysis units with non-availability of blood products have contributed to maternal mortality. Services of experienced medical personnel where essential services are needed to identify high risk factors to reduce maternal mortality. The Use of Small-Sided Games as an Aerobic Fitness Assessment Supplement Within Elite Level Professional Soccer Abstract The purpose of this investigation was to quantify the association between 5 vs. 5 small sided games (SSG) running performance and physiological performance during the Yo-YoIR1 test to ascertain the utility of SSGs as a potential fitness test modality within elite professional soccer players. Twenty-three (n = 23) elite male professional soccer players (mean ± SD age 25.3 ± 3.1 yrs, mass: 76 ± 9 kg, height: 176 ± 9 cm) were assessed. Players completed an intermittent aerobic fitness test (Yo-YoIR1) and a 5 vs. 5 SSGs protocol for the purpose of the study. During all SSGs players wore GPS (Statsports 10-Hz, Viper Pod, Newry, Northern Ireland) and HR monitors (Polar, Oy Kemple, Finland) with these measures related to Yo-YoIR1 running performance. Results revealed SSGs running performance (TD; m) and physiological performance (HR) showed the lowest CV% (< 5%), with high speed movements, accelerations and decelerations highlighting higher CV% during SSGs. Possibly small to possibly very large associations were observed for running performance during 5 vs. 5 SSGs and Yo-YoIR1 performance, with negative associations observed between physiological performance during SSG and YoYoIR1 running performance. To conclude, the current study observed how running performance during a standardised 5 vs. 5 SSG protocol within elite soccer cohorts is associated with the Yo-YoIR1 running performance. Given the low CV%, repeatability and large association of global running performance and internal load measures during a 5 vs. 5 SSG with Yo-YoIR1 performance, this particular soccer specific SSG protocol potentially supplements traditional non-sport specific testing assessments. Introduction The demands of soccer match-play have been reported (Casamichana et al., 2012;Malone et al., 2016;Taylor et al., 2018). During official games players complete between 9 and 14 km (Manzi et al., 2014;Osgnach et al., 2010), with a high proportion of distance covered jogging and walking (Manzi et al., 2014). However, the outcome of a soccer match is heavily influenced by the high-speed components of play despite the game primarily taxing the aerobic system (Manzi et al., 2014). Notably, ~300 acceleration and deceleration efforts (when categorized as changes in movement above 0.5 m·s-2) are performed per half (30) and ~18% of the total distance covered during a soccer match is completed while accelerating or decelerating (Akenhead et al., 2016;Manzi et al., 2014). Given the need for coaches to train technical, tactical and physiological stimuli during training, the majority of soccer-based training is completed through the use of small-sided games (SSGs). These games represent a concomitant form of training allowing for many physical qualities to be trained within shortened pitch dimensions across specific player numbers and game formats that also allow players to engage in sport-specific decision making (Fleay et al., 2018;Owen et al., 2012). Indeed, within many team sports SSG are widely utilised by coaches to develop aerobic and anaerobic components of fitness within players, with these games shown to provide similar physiological stimuli as interval training (Delall et al., 2012). However, when organizing SSGs, coaches who want to achieve physiological and physical performances that allow for the development of aerobic endurance need to consider a number of different factors which may affect exercise intensity. These factors include the number of players and the pitch size (Djaoui 2017;Owen et al., 2014), the game rules (Halouani et al., 2014;, coach encouragement (Halouani et al., 2014;Kolu et al., 2015;, the absence or presence of goalkeepers (Kolu et al., 2015), and training regime (Mallo et al., 2008). Additionally, bout duration (Kolu et al., 2015), goal size and work to rest ratios should be considered. Furthermore, the variability of SSGs with a standardised format has been found to be low when assessing both the heart rate and total distance (coefficient of variation (CV) < 5%) (Hill-Haas et al., 2011;Mallo et al., 2008;Owen et al., 2014). This is despite larger variability being reported in higher-speed distances (Hill-Haas et al., 2011;Mallo et al., 2008;Rampinini et al., 2007). As a result of the recent findings surrounding SSGs, it may be suggested that through the use of smaller standardised training games, coaches have the possibility to utilise these games to detect worthwhile changes in running and physiological performance that may be linked to increased fitness capacities. Unlike match play SSG performance if standardised, is less position, score-line and opposition dependant. Therefore, the physical and physiological performance during these games could potentially be used to highlight aerobic fitness changes within players, removing the need for maximal fitness assessments such as the time-trials, linear maximal running testing and shuttle testing. Previously Stevens et al. (2016) have shown that a 6 vs. 6 SSG showed moderate to large correlation with YoYoIR2 performance for total distance (r = 0.45; 90%CI: 0.31-0.56; moderate), high speed distance (r = 0.70; 90%CI: 0.61-0.77; large) and high accelerations (r = 0.59; large) showing that SSG performance can relate to testing performance for soccer cohorts. However, the above study placed caution with regard to the application of SSGs as fitness tests due to these games having poor validity measures within the tested cohort. In order to best ascertain sport specific training adaptations, coaches will engage in specific testing batteries at given time points during the competitive season, however, no such soccer specific assessments have been reported which allow technical and physical staff to test within training situations in order to maximise training content. Several test protocols are used and well-reported within the literature to assess specific physical qualities, such as maximal (repeated) sprinting and aerobic and anaerobic endurance tests, e.g. the Yo-Yo tests, however, none are fully integrated with a sport-specific technical involvement. Furthermore, recent analysis by Malone et al. (2018) has shown that improved intermittent aerobic fitness capacities result in lower odds risk of injury within soccer cohorts, additionally improved aerobic capacity increases coaches' ability to increase the amount of the prescribed training load, high speed and sprint running across training cycles. However, most of these intermittent aerobic testing batteries require players to complete maximal efforts over specific time duration depending on the test selected. Due to the competitive schedule within elite soccer and periods of fixture congestion these maximal testing protocols may not be feasible for coaches to run due to tim constraints and increased levels of fatigue amongst players (Halson, 2014). Given that standardised SSGs have shown a low CV for physical and physiological variables and that SSGs are a regular component within elite soccer training, the aim of the current investigation was to determine whether a 5 vs. 5 SSG protocol could be utilised as a surrogate of aerobic fitness within an elite soccer cohort. To determine it, we first sought to understand the validity and reliability of the 5 vs. 5 SSG and secondly, sought to understand the association between running performance measures within SSGs and YoYoIR1 performance. As a result of the unique focus of this current investigation, the potential to develop and propose a soccer specific assessment protocol amongst elite level professional soccer players as a supplement to the traditional current methods © Editorial Committee of Journal of Human Kinetics already well-documented, led to the development of the research topic. It was hypothesised that the 5 vs. 5 SSG would have limited variation for total distance and HR measures and that there would be a strong positive association between the SSG protocol performance and Yo-YoIR1 performance. Participants The current investigation was an observation study of elite level professional soccer players competing for a team that at the time of data collection were competing within the top tier of their respective league competition. Additionally, 80% of the players assessed were representing their respective European national teams at the time of investigation. Data were collected for 23 players (Mean ± SD, age: 25.3 ± 3.1 years; body height: 183 ± 7 cm; body mass: 72 ± 7 kg) over one season. The study was approved by the local institute's research ethics committee (Technological University Dublin, Tallaght, Ireland) and written informed consent was obtained from each participant. The study period involved all training sessions during the 2017/2018 season. Small-Sided Game (5 vs. 5) During the investigation period, a total of 1560 individual SSG data points and metrics were assessed with a median of 28 observations per player. The SSG consisted of free play with the focus of the SSG to keep possession in a 5 vs. 5 method within a 25 x 25 m grid resulting in a relative player area of 62.5 m 2 . All SSGs were performed with a continuous rhythm, under the supervision and motivation of several coaches in order to keep running performance of players high (Manzi et al., 2014). During all SSGs free play was allowed with maximal touches, in all cases multiple replacement balls were available for prompt replacement when hit out of play (Djaoui et al., 2017). Before the study period, these SSGs were frequently performed to ensure familiarisation before the experimental period. All sessions were performed on the same natural grass pitch. In addition, all exercise games were performed at the same time of the day (10:30 -11:00 am) to limit to effects of circadian variations on measured variables (Kolu et al., 2015). All of the SSG assessment protocols were completed after a standardised warm up of 15 min. All games were standardised for time (3 min in length) with players performing 3 repetitions of 3 min games within a single session and a 2 min passive recovery period between repetitions of the protocol. All games were coach driven through verbal encouragement to ensure maximum effort. Yo-Yo Intermittent Recovery Test Level 1 The Yo-YoIR1 consists of 2 x 20 m shuttle runs performed at increasing speeds, with 10 s of active recovery between runs (intermittent recovery version). The Yo-YoIR1 has been reported to assess a player's endurance capacity with a high aerobic energy contribution. During the current investigation testing was completed in line with the procedures described in a previous study . This included participants completing a 15-min dynamic warmup involving multi-joint and running activities of progressive intensity. Failure to complete a shuttle resulted in a verbal warning with participants being withdrawn on a second failure. Total distance and corresponding maximum speed at the final completed shuttle were recorded at the end of the test. The heart rate was measured at 1 s intervals during the test using the Polar Team 2 System (Polar Electro Oy, Kemple, Finland). Stored data were then exported out of the dedicated Polar software (Polar Team 2 Software, Kemple, Finland) to a bespoke data base (Microsoft Excel, Redmond, USA). The YoYoIR1 test has been previously shown valid and reliable with the distance covered during a Yo-YoIR1 being directly related to match play running performance within soccer in addition to being sensitive to changes in fitness across acute training periods in team sports (Lacome et al., 2017;. Running Performance Monitoring During SSGs players' running performance was monitored using a portable nondifferential 10-Hz GPS device integrated with a 100 Hz, 3-dimensional accelerometer, 3dimensional gyroscope, and a 3-dimensional digital compass (STATSports, 10-Hz, Viper Pod, Northern Ireland). This type of a system has previously been shown to provide valid and reliable estimates of instantaneous velocity during acceleration, deceleration, and constant-velocity movements during linear, multidirectional, and soccer-specific activities

(Beato et al., 2016). Each player was assigned a GPS vest that was tightly Journal of Human Kinetics -volume 71/2020 http://www.johk.pl fitted to their upper torso, holding the receiver between the scapulae. All devices were always activated 15 minutes before the data collection to allow acquisition of satellite signals in accordance with the manufacturer's instructions. In addition, to avoid inter-unit error, each player wore the same GPS device for each training session (Malone et al., 2016) After recording, the data were downloaded to a computer and analysed using the software package Viper version (STATSports, Viper, System). Based on GPS data, total, high-speed (>19.8 km·h -1 ) and sprint distance (>25.5 km·h -1 ), as well as average metabolic power (W·kg -1 ), high metabolic power distance (m; ≥ 20 W·kg -1 ), accelerations (n; ≥ 3.3 m·s -2 ), decelerations (n; ≥ -3.3 m·s -2 ) and the dynamic stress load (AU) were calculated during each training session. The dynamic stress load represents the weighted total of all accelerometery above 2G based on convex curved G-force ratings. These measures are reflective of standardised measures of training loads regularly reported within soccer cohorts Russell et al., 2016). Statistical Analysis Pearson's correlation (r) was used to assess the association between Yo-YoIR1 distance and selected 5 vs. 5-SSG variables. Magnitude of the correlation (r) was considered trivial (< 0.1), small (> 0.1-0.3), moderate (> 0.3-0.5), large (> 0.5-0.7), very large (> 0.7-0.9), nearly perfect (> 0.9-1.0), and perfect (1.0) (Hopkins, 2011). Analyses were performed using IBM SPSS Statistics for Windows (Version 22.0). The intraclass correlation coefficient (ICC), typical error (TE) and TE as a coefficient of variation (CV) of selected 5 vs. 5-SSG physical and physiological variables were calculated within a specific spreadsheet for the assessment of ICC created by Hopkins (2000). The reproducibility of 5 vs. 5 SSGs was determined using Bland and Altman (1995) limits of agreement. Reference lines were determined as the mean difference ± 1.96 standard deviations. All analyses were conducted with the statistical significance set at p < .05. Validity and Reliability of 5 vs. 5 SSGs The ICC (90% CI), bias ± random error and CV% (90% CI) for running performance measures during SSGs are shown in Table 1. During these 5 vs. 5 SSGs total distance (0.94; 0.76-0.98), the dynamic stress load (0.94; 0.76-0.98), average metabolic power (0.82; 0.76-0.91), time above 85% of the HRmax (0.86; 0.65-0.91) and the percentage of the HRmax (0.87; 0.53-0.94) showed very large ICCs. Global running performance measures showed the lowest CV% (< 5%) during SSGs. Objective internal load measures (Time above 85% HRmax, average HRmax, HRmax, percentage HRmax) also showed low CV%. Higher speed movement as well as acceleration and deceleration measures presented higher CV%. Table 2 shows typical running and physiological performance measures during 5 vs. 5 SSGs within elite soccer players. Table 3 shows the observed association between running measures and explained variance during SSG and Yo-YoIR1 performance measures. Possibly small to possibly very large associations were observed for running performance during Yo-YoIR1 and 5 vs. 5 SSG running performance. Total distance (0.88; 0.67-0.91; possibly very large) during SSGs had a possibly very large association with YoYoIR1 running performance in addition to the dynamic stress load (0.80; 0.67-0.91; possibly very large) that showed similar possibly very large associations. Likely large associations were observed between high speed running and average metabolic power during 5 vs. 5 SSG and Yo-YoIR1 running performance, with negative associations observed between objective internal measures during SSG and YoYoIR1 running performance. Discussion The current investigation sought to explore the possibility of utilising SSGs as an assessment of aerobic fitness within elite soccer players. We firstly assessed the reliability and repeatability of these SSGs, with the main findings highlighting the lowest CV% (< 5%) showed by global running performance measures (total distance, average metabolic power, dynamic stress load) during SSGs. Furthermore, physiological measures (time above 85% HRmax, average HRmax, HRmax, percentage HRmax) also showed low CV% during 5 vs. 5 SSGs. When the association between SSG and Yo-YoIR1 running performance was considered, the movement © Editorial Committee of Journal of Human Kinetics metrics recorded by athletes within these SSGs (5 vs. 5) were related to the aerobic fitness profile of players. These data may suggest that the standardised SSG protocol employed within the study could potentially be utilised by coaches as a fitness assessment tool within the soccer ergonomic process. Interestingly, total distance (m) and the dynamic stress load (AU) recorded the highest associations with Yo-YoIR1 performance with most movement metrics recording moderate to large associations with Yo-YoIR1 performance. Based on these findings, it may be suggested that running performance within the standardised SSG protocol may be utilised to assess aerobic fitness within elite soccer cohorts. Figure 1 Scatter-Plot for the association between total distance (m) covered during SSG and YoYoIR1 (m) performance in elite soccer players. © Editorial Committee of Journal of Human Kinetics Figure 2 Scatter plot for the high metabolic power distance (m; ≥ 20 W·kg -1 ) covered during SSG and YoYoIR1 (m) performance in elite soccer players. Recently within the soccer training process, coaches have evolved towards integrated physical training with the aim to maximise training time during which players are in possession of the ball (Lacome et al., 2017). It has been shown across many investigations that SSGs can improve aerobic and anaerobic physical qualities within team sport athletes in addition to these games improving soccer-specific fitness Owen et al., 2012Owen et al., , 2014 and match winning-related factors. Within the current investigation, the 5 vs. 5 SSG protocol resulted in similar percentages of HRmax responses that were previously published in the literature within soccer cohorts. Interestingly, players spent on average 4.3 ± 0.8 min above 85% of the HRmax. As such, it may be suggested that the application of these games over a training period would result in improved aerobic fitness capacities of soccer players Owen et al., 2011Owen et al., , 2012. Coaches are often reluctant to utilise SSGs as fitness assessments given the open loop nature of these games and potentially high variability (%CV) of specific running performance measures from set to set. Furthermore, coaches may understand the many factors such as pitch size, work to rest ratios, player numbers, player area, game rules, goal size, potentially impact SSG running performance. Therefore, the present investigation aimed to first assess the variability and repeatability of a 5 vs. 5 SSG protocol within an elite soccer cohort for commonly used GPS and HR variables, including the dynamic stress load (AU) and acceleration motion measures such as high metabolic loading distance (m). Interestingly, results from the current study found a low CV% for global running performance measures and physiological measures. Specifically, total distance, average metabolic power and dynamic stress load presented low CV% (> 5 %). However, higher speed movements (8.1%), accelerations (14.1%), decelerations (16.2%) and sprinting movements (16.3%) were shown to have high Journal of Human Kinetics -volume 71/2020 http://www.johk.pl variability within SSGs. These findings are in line with previous literature that has observed that high speed movement variables, acceleration and deceleration movement in addition to sprintbased variables present a high CV% during SSGs (Hill-Haas et al., 2008;Rampinini et al., 2007;Stevens et al., 2016). During fitness assessment, most coaches are concerned with specific variables that within the current investigation presented a low CV% such as total distance covered or physiological performance. Indeed, coaches will often utilise physiological data and GPS data such as total distance and time spent above the percentage of the HRmax across a week to monitor player's readiness and injury risk. Interestingly, given the low CV% and high associations between the specific SSG running metrics and Yo-YoIR1 scores, it may be suggested that coaches could regress their own data during SSGs against Yo-YoIR1 data and utilise the equation to understand how each SSG completed has impacted Yo-YoIR1 running performance, given the nuances between different teams' training philosophies. As a result of the information extracted within this investigation, coaches could potentially utilise the 5 vs. 5 SSG protocol as a fitness assessment of players during specific periods of the season given the known time constraints associated with fitness assessments in soccer cohorts. When considering SSG running and physiological performance and association of these variables with Yo-YoIR1 performance, it was observed that many variables had moderate to very large associations with Yo-YIR1 test performance. The current investigation is not the first study to relate test performance to an element of soccer specific performance. Indeed, previous studies (Buchheit et al., 2010;Casamichana et al., 2012;Castagna et al., 2010) have found similar correlations between aerobic fitness tests and match play running performance. Moreover, Manzi et al. (2014) observed moderate to large correlations (r = 0.52-0.83) between several aerobic fitness variables (MAS, VO2max) and distance covered in high-power categories (> 20 W·kg-1) during matches. Our findings are in line with the study by Stevens et al. (2016) who showed moderate to large associations between running performance during 6 vs. 6 small-sided games and Yo-YoIR2 running performance within professional soccer players. Specifically, it was shown that the total distance covered during 6 vs. 6 SSGs had a moderate association with Yo-YoIR2 performance (r = 0.45; 90%CI: 0.31-0.56); furthermore, HR measures showed similar trends within the current investigation when compared to Stevens et al. (2016) with negative associations observed between HR measures during SSGs and Yo-YoIR2 performance. However, despite the reported associations, these differed across populations from elite male to female cohorts resulting in the authors recommending caution in the thought process of the application of SSGs as a potential sole fitness assessment (Stevens et al., 2016). However, given the low CV% for total distance and HR measures within the current and previous investigations (Stevens et al., 2016), it may be suggested that these games provide an ecologically valid measurement of aerobic fitness within elite soccer player cohorts. The advantage of applying a 5 vs. 5 SSG protocol from a testing perspective is that this is an already frequently used methodology of training within all levels and age groups of soccer cohorts in addition to the ecological validity of SSGs being higher than traditional testing methodologies. Furthermore, the multiple use of this SSG assessment protocol can help objectively quantify changes or reductions in intermittent endurance capacities of individual players across a season. However, limitations of the current study include the fact that these data are taken amongst elite level professional players and may not be comparable across a range of playing cohorts. Additionally, not all coaches have access to the expensive equipment used within the study to provide the data capture required to assess the test. Furthermore, consideration with regard to a number of co-founding factors such as player motivation, player skill level and player pacing within SSGs may impact results of the assessment protocol. These suggested factors have previously been shown to influence locomotor performance as well as the physiological load of individual players. Therefore, given these potential cofounding factors, it may be suggested that within practice, coaches may utilise integrated training load ratios, i.e. the integration of the individualised heart rate (internal load) with distance measure (external load) during SSGs. In line with recent literature in this area, an © Editorial Committee of Journal of Human Kinetics additional strength of this particular investigation is the fact that this protocol may aid the identification of fitness changes and possibly counteract the potential impact of player pacing on external load variables (Akubat et al., 2014;Malone et al., 2016;Taylor et al., 2018). Given that improved aerobic fitness has been shown to reduce the odd risk of lower limb non-contact injury within soccer cohorts, and the observed associations between 5 vs. 5 SSGs and the Yo-YoIR1, future research may aim to best ascertain the number of exposures to 5 vs. 5 SSGs that result in improved aerobic fitness and whether this improvement is associated with reduced likelihood of players sustaining lower limb noncontact injuries. Furthermore, future research should target the use of multiple 5 vs. 5 SSG protocol assessments to view seasonal-variation in both group and individual cohorts of professional players, and even compare across various age groups or levels of professional soccer, or players of different training backgrounds or cultures. Additionally, future research may direct its interest towards the association of the 5

vs. 5 SSG assessment protocol with more laboratory physiological testing assessments. Conclusion To conclude, the current investigation confirms how specific measures of running and physiological performance have a low CV% combined with good repeatability during the 5 vs. 5 SSG assessment protocol within elite soccer cohorts. Furthermore, these measures showed moderate to very large associations with the wellreported Yo-YoIR1 performance and as a result, promote the fact that a soccer specific 5 vs. 5 SSG assessment protocol can be utilised more regularly by coaches as a supplement to traditional aerobic fitness testing to assess the intermittent aerobic capacity of elite soccer players. Within the competitive season, coaches who find it difficult to provide time for maximal or submaximal aerobic fitness assessments due to fixture constraints within elite soccer may view this as a feasible assessment option. As such these coaches could potentially utilise these 5 vs. 5 SSGs within a typical weekly training cycle to best ascertain players' soccer specific aerobic fitness and performance capacity. The SSG protocol utilised within this investigation ultimately represents a proven reliable, time-efficient method of soccer specific assessment of intermittent aerobic fitness within elite level professional soccer. Practical Implications • A standardised 5 vs. 5 SSG protocol presented validity and repeatability with regard to specific metric measurements of movement within an elite soccer cohort. • Given the low CV% shown with specific physical and physiological variables with 5 vs. 5 SSGs, it is suggested that these games could be utilised as a time efficient method of assessing the intermittent aerobic capacities of elite soccer players. • Consistent associations observed between distances covered in SSGs and YoYoIR1 performance suggest that 5 vs. 5 SSGs may be used as a soccer specific assessment of intermittent aerobic fitness. • The SSG protocol utilised within this investigation ultimately represents a proven reliable, time-efficient method of soccer specific assessment of intermittent aerobic fitness within elite soccer cohorts. • Given the fixture schedule within elite professional soccer, SSGs with a low CV% may be placed within training sessions by coaches and used as a tool in the assessment of intermittent aerobic capacities to guide a future training prescription. Isolation and Selection of Potentially Beneficial Autochthonous Bacteria for Piaractus mesopotamicus Aquaculture Activities Marcos Gabriel Guidoli1,3, Juan José Santinón1, Sergio Enrique Pasteris2, Sebastián Sánchez1 and María Elena Fátima Nader-Macías3* 1Department of Veterinary Sciences, Institute of Northeast Ichthyology (INICNE), National University of the Northeast, Corrientes, Argentina 2Superior Institute of Biological Research (INSIBIO-CONICET), San Miguel de Tucuman, Argentina 3Reference Center Lactobacilli (CERELA-CONICET), San Miguel de Tucuman, Argentina Introduction Piaractus mesopotamicus, commonly known in South America as Pacú, Pez chato, Mirabí or Pirarí is a serrasalmid fish endemic to the Paraguay-Paraná river basin that has been incorporated to the aquaculture production system in the northern region of Argentina. The resistance of this autochthonous fish to the breeding conditions has led to a high increase in its production, turning it into the main produced fish in Argentina with 1345.32 tons, equivalent to 44.45% of the total aquaculture production in Argentina in 2012 [1]. However, the poor knowledge on the behavior of this specie and the conditions of the intensive and super intensive breeding systems have created a deficit in the number of animals at different stages of their biological cycle, mainly for the obtaining of larvae, juveniles and sexually mature adults. The most worldwide proposed technique to increase the production in aquaculture facilities is the use of antibiotics. Besides acting as antiinfectious agents, antibiotics are associated with decreases in animal gut mass and increases in intestinal absorption and energy sparing; allowing a higher amount of consumed nutrients to be used for growth and production [2]. However, nowadays the use of antibiotics in aquaculture represents a human health risk characterized by three items: resistance transference [3], toxicity of antibiotic residues [4], allergies and effects of the antibiotics on human intestinal microbiota [2] and environmental risks [5]. A novel and emerging alternative to antibiotics are probiotics, defined for aquaculture as "a live microbial adjunct which has a beneficial effect on the host by modifying the host associated or ambient microbial community, by ensuring improved use of feed or enhancing its nutritional value, by enhancing the host response toward disease or by improving the quality of its ambient environment" [6]. Beneficial microorganisms have been used successfully all around the world with wide purposes in different aquatic cultures such as inhibitory activity against primary pathogens [7], immunostimulatory effect and nonspecific immune response stimulation [8]; decrease of mortality and reduction of the infection of animals infected with different pathogens and increased rate of growth in turbot [9], atlantic salmon [10], rainbow trout [11] and Pacific oyster [12] and increase of production in channel catfish [13]. description of the autochthonous microbiota or assays directed to evaluate the responses of this species to probiotic or beneficial bacteria treatments. Thus, the knowledge of the indigenous microbiome of this economically important specie is required. Therefore, the aim of this work was to determine the cultivable microbiota of P. mesopotamicus and to select beneficial microorganisms to be further included in the design of a pharmaceutical preparation to be applied in the aquaculture of P. mesopotamicus. Samples and isolation of cultivable microorganisms Samples were taken from P. mesopotamicus specimens at different stages of their biological cycle in a fishery located in facilities of the Universidad Nacional del Nordeste. The first isolation was performed in autumn from 10 fishes of approximately 4 months old with mean values of 9.23 cm long and 4.25 g. Five of these animals belonged to a group affected with columnariasis disease (these isolates were not included into the potentially probiotic bacteria), while the other half were healthy specimens. For the second (summer) and third isolation (winter), 14 and 10 specimens, respectively, were sampled (Tables 1 and 2). The sampling process depended on the specimens weigh. Animals bigger than 5 g were sampled by scrapping dorsal and ventral skin (1 cm 2 ), gills, anus and mouth. In these specimens, the anterior, middle and posterior intestine sections were homogenized by using a Teflon pestle at 750 rpm in aseptic conditions. Small animals (weight ≤ 5 g) specimens were homogenized in different quantities of sterile distilled water (2 ml for those under 1 g, 5 ml for those between 1 and 3.5 g and 15 ml for those between 3.5 and 5 g) by using a Teflon pestle at 750 rpm in aseptic conditions. Samples were collected in two culture media: LAPTg (1% yeast extract, 1.5% peptone, 1% tryptone, 1% glucose, 0.1% Tween 80; pH 7.2) and Nutrient Broth (0.5% plurypeptone, 0.3% meet extract; pH 6.9) and stored at 4°C until processing. In order to isolate the spore-forming microorganisms, samples collected in Nutrient broth were heated at 80°C for 15 min and incubated for 3 hr at 37°C and later 24 to 72 hr at 37°C in Nutrient Agar (Nutrient broth added with 1.5% w/v agar). For the isolation of LAB, samples from LAPTg broth were inoculated in LAPTg agar (1.5% w/v) for 24 to 72 hr at 37°C. All plates were observed every 24 hr during 3 days and selected colonies were subcultured and stored at -20°C in their appropriate medium supplemented with 20% glycerol. Selection of beneficial microbial genera The isolates previously obtained were phenotypically characterized according to the following tests: Gram-Nicole staining, Schaeffer Fulton staining for spore forming bacteria, catalase reaction, and growth in anaerobic conditions, KNO 3 reduction and indole production [17]. Bacterial strains and culture conditions Except when cited, indigenous LAB were grown in LAPTg broth for 18 hr at 37°C, while Bacillus were grown under the same conditions in nutrient broth. Specific fish pathogens: Streptococcus agalactiae ATCC 27956, Streptococcus dysagalactiae subsp. dysagalactiae ATCC 27957, Lactococcus garvieae CRL 1828 (isolated from Lithobates catesbeianus) and L. garvieae (isolated from cows affected with mastitis) were grown in the conditions stated for LAB; other fish pathogens: Aeromonas salmonicida ATCC 33658, Yersinia ruckeri ATCC 29473, Klebsiella sp. (isolated from non-healthy fishes), Citrobacter freundii CA and CB (isolated from L. catesbeianus specimens) and meat spoilage bacteria: Listeria monocytogenes Scott A (isolated from meat), Pseudomonas aeruginosa 07 and 47 (isolated from L. catesbeianus specimens), Staphylococcus aureus (isolated from a clinical human sample), Escherichia coli M20DN3 and Salmonella enteritidis (isolated from meat) were grown in brain heart infusion (BHI) broth (Britannia ® ) and/ or nutrient broth for 8 hr at 37°C. Screening of beneficial properties Hydrogen peroxide production: The hydrogen peroxide (H 2 O 2 ) production was qualitatively determined by the plate method, employing horseradish peroxidase incorporated in 3,3',5,5'-tetramethyl-benzidine (TMB) agar medium [18]. Isolates grown as cited before were centrifuged at 3,000×g for 5 min, cells were washed twice with sterile saline solution and streaked in MRS plates containing 1 mM TMB and 2 U/mL peroxidase. After incubation for 48 hr, the plates were exposed to air. According to the color intensity acquired by the colonies, the isolates were classified as strong (blue), medium (brown), scarcely (light brown) or negative (white colonies) producers. Production of antagonistic substances: The production of antimicrobial metabolites of the isolates was tested using the agar-well diffusion assay. Selected isolates were grown for 18 hr at 37°C in LAPTg broth and centrifuged at 3,000×g for 10 min at 4°C. Supernatants were fractioned in two aliquots, one remained untreated and the other was neutralized, both were stored at -20°C until use. The indicator strains were grown in BHI or nutrient broth up to half of the exponential growth phase and inoculated at a concentration of 1 × 10 6 colony forming units/ml (CFU/ml) in 20 ml of nutrient or BHI soft agar (0.7% w/v) at 45°C and poured into Petri dishes. After solidification, 10 mm diameter wells were made into the agar, filled with 100 µl of pure (untreated) or neutralized (treated) supernatants, allowed to diffuse into the agar for 1 hr at room temperature and incubated at 37°C for 24 hr. The presence of antagonistic metabolites in the supernatants was revealed as an inhibitory area around the well. The inhibition was expressed, in millimeters, as the diameter of the inhibition halo. Biosurfactants production: The detection of surfactants was carried out by the oil spreading technique. Thus, a Petri dish was filled with a mix of distilled water and vegetable oil in a proportion 1:1 avoiding the mixture of both phases. Twenty milliliters of cell-free supernatants of selected isolates were dropped into the liquid surface. The loss of the drop form and the mixture with the vegetal oil phase was considered as a positive result. Degree of emulsification: Emulsifier activity was determined by adding 6 ml kerosene or hexadecane to 4 ml of cell free supernatant and vortexing at high speed for 2 min. The emulsion index was considered as the height of the emulsion layer, divided by the total height of the liquid sample, multiplied by 100. Quantifications were performed at 0, 4, 8, 24, 96 and 360 hr [19]. Bacterial surface properties Hydrophobicity index: Hydrophobic characteristic of the bacterial surface was determined by the microbial adhesion to hydrocarbon method (MATH) [20] applying their capability of partition in water and a polar solvent (hexadecane). Bacterial cultures of 18 hr at 37°C were centrifuged at 3,000×g for 5 min, washed twice with sterile neutralized distilled water and resuspended to an optical density The auto aggregation index score applied was: high (60-100%), medium (30-59%), low (0-29%) [18]. Compatibility within selected isolates: The compatibility between selected isolates was performed by using the agar-well diffusion assay detailed above, using selected isolates as indicator and producer isolates strains. The appearance of inhibition halos indicated the incompatibility between the isolates. Genotypic identification and BLAST comparison of selected strains: Microorganisms selected by their beneficial properties (Table 3) were submitted to genotypic identification. For this purpose the genomic DNA of each microorganism was extracted, purified and evaluated by PCR with specific primers in order to amplify de DNA region codifying for the region VI of the 16S RNA gene. The obtained fragment (of approximately 500 pb) was purified and sequenced. The results were analyzed online using the tools: http://rdp.cme.msu.edu/ html/analyses.html and http://www.ncbi.nlm.nih.gov/BLAST. All nucleotide sequences were submitted to the DDBJ ⁄ EMBL ⁄ GenBank databases. Fragments were also analyzed using the MEGABLAST software of the National Center for Biotechnology Information (NCBI) in order to search for the most

similar sequences and thoroughly examine whether they were already used as probiotics in aquaculture or not. Samples and isolation of cultivable microorganisms During the first isolation stage (autumn) 295 isolates were obtained, 41 of them were phenotypically identified as yeast, 64 as cocci and 190 as rods (Figure 1a). The second sampling (summer) resulted in 87 isolates, 47 of them were yeasts, while 34 and 6 were identified as cocci and rods, respectively ( Figure 1b). In the third sampling (winter), 140 isolates were obtained, 3 of them were yeasts while 67 were cocci and 70 rods (Figure 1c). At the end of the isolation process, a total of 522 isolates were available (91 yeasts, 165 cocci and 266 rods) for further evaluation. Selection of beneficial microbial genera In this step, the 431 bacterial isolates were identified by phenotypic tests in order to preselect those of interest in this study. Out of the total, only 30 isolates were partially identified as LAB and 8 as Bacillus ( Table 3). All of them were preselected to evaluate their beneficial properties. Screening of beneficial properties Thirty LAB and 8 Bacillus isolates were submitted to physiological studies to determine their beneficial characteristics by different assays related to surface properties and production of inhibitory metabolites. Production of antagonistic substances The evaluation of the antagonistic activity in the culture supernatant showed that 10 out of the 30 LAB strains were unable to inhibit any of the food-borne bacteria and specific fish pathogens used in this study ( Table 4). The other 20 isolates inhibited at least one of the indicator strains. Isolate 49 inhibited the growth of P. aeruginosa 47, while isolate A74B inhibited S. aureus. LAB isolates 16, 18, 66, 69 and 75 inhibited P. aeruginosa strains, S. aureus and E. coli (indicators group 1). Isolate 65 inhibited the indicators group 1 (IG1) and the fish pathogen Y. ruckeri. Isolate 66 inhibited the growth of IG1 and Klebsiella sp. Isolate 62 inhibited the growth of IG1 and C. freundii CA, while isolates 63, 64, 71 and 73 were able to inhibit the IG1 and C. freundii CB. Isolate 70 inhibited the IG1 and two important fish pathogens; A. salmonicida and Klebsiella sp., while isolate 74 showed a similar pattern by inhibiting the growth of IG1, Klebsiella sp. and Y. ruckeri. Two of the LAB isolates with the broader inhibitory spectrum were 61B and 72 inhibiting 9 and 7 of the indicator strains, respectively. The antagonistic activities described previously were due to the organic acids present in the culture supernatants. Isolate A34 inhibited the growth of A. salmonicida by organic acids and A. salmonicida and S. enteritidis by both, organic acids and non-acidic metabolites, observed in the inhibition caused by the neutralized supernatant. Isolate A35 showed a similar behavior, being able to inhibit the growth of A. salmonicida and Y. ruckeri only by organic acids and S. aureus by a combined effect of organic acids and non-acidic substances. On the other hand, only 3 of the Bacillus isolates were able to inhibit the growth of at least one of the indicator strains tested. Isolate A252 inhibits the growth ok Klebsiella sp. and S. enteritidis by a synergic effect of organic acids and nonacidic substances. Isolate A253 showed the same inhibitory spectrum as A252 inhibiting also the growth of S. aureus by organic acids only. Isolate A254 only inhibited the growth of S. aureus by a synergic effect of organic acids and non-acidic compounds (Table 4). It is important to point out that none of the isolates were able to inhibit the growth of S. agalactiae, S. dysagalactiae, L. monocytogenes or both strains of L. garvieae. Due to their inhibitory spectrum, the LAB isolates 61B, 70, 72, 74, A34 and A35 as well as the Bacillus isolates A252, A253 and A254 were selected for further studies. Biosurfactants production All the LAB isolates were negative to the drop collapsing method. However, only two Bacillus isolates (A254 and A256) were able to produce biosurfactants. Degree of emulsification Six Bacillus isolates (A252, A253, A254, A258, A259 and A260) were able to emulsify both organic solvents from time 0 h, while only two Bacillus isolates (A252 and A254), maintained this ability up to the end of the experiment. Isolate A256 emulsified only hexadecane, while isolate A257 showed to emulsify kerosene ( Figure 3). None of the LAB isolates demonstrated emulsification capacity. Cells surface's properties Hydrophobicity and auto aggregation: All the isolates showed hydrophobicity and auto aggregation values between 0 and 50% ( Figure 4). However, LAB showed high hydrophobicity values and low auto aggregation percentages, while an opposite behavior were observed for Bacillus isolates. The results evidenced that 6 (20%) LAB isolates (16,30,31,49, 63 and 67) were moderately hydrophobic, while the surface of the other strains had a hydrophilic nature; all the Bacillus strains were hydrophilic. Only one LAB isolate (49) showed a medium auto aggregation percentage, while all the others presented low percentages. Only one Bacillus isolate (A254) has a medium auto aggregation percentage, while all the other strains showed low auto aggregation. Thus, auto aggregating LAB (30 and 49) and Bacillus (A252) and the hydrophobic LAB (30, 31 and 63) were selected for further evaluations. Compatibility within selected isolates Compatibility tests were performed with all those isolates preselected after the screening of their surface and antagonistic characteristics and is summarized in Table 5. The results indicate that two isolates (61B and 74) showed to inhibit the growth of other preselected microorganisms. On the other hand, LAB isolates 31, 49, 63 and 72 growth scarcely in laboratory conditions; generating difficulties in the cells obtaining process for further assays. Then these six isolates were not included in the following studies. On the basis of their beneficial properties, a group of 8 isolates were deposited into the culture collection of CERELA and considered for the design of a potentially probiotic veterinary product to be used in aquaculture of P. mesopotamicus. Selected microorganisms were: LAB isolate 30 (renamed as CRL 1940) for having medium of both, auto aggregation and hydrophobicity indexes; LAB isolate 66 (renamed as CRL 1939), for being the highest H 2 O 2 producer; LAB isolate 70 (renamed as CRL 1941), for inhibiting two specific fish pathogens (A. salmonicida and Klebsiela sp.) and LAB isolates A34 and A35 (renamed as CRL 1937 and CRL 1938, respectively), for inhibiting two specific fish pathogens (A. salmonicida and Y. ruckerii). The selected Bacillus isolates were A252, A253 and A254 for being able to inhibit at least one of the specific fish pathogens, for their emulsifying abilities and isolate A254 for having, also, the highest auto aggregation index and biosurfactants production. These isolates were later identified by genotypic methodology. Genotypic identification and BLAST comparison of selected isolates The genotypic identification and the beneficial properties of the selected strains, as well as the accession numbers of the nucleotide sequences (available at the DDBJ ⁄ EMBL ⁄ GenBank databases) are shown in Table 6. Based on sequence alignments with these databases, out of the 100 results, a total of 93 and 5 sequences had an identity of 99% and 98%, respectively, with the sequence obtained from isolates CRL 1940 and CRL 1941. Out of these 98 microorganisms only seven were isolated from water environments and none of them were assayed as probiotics in this particular area. A total of 99 sequences had an identity of 95% with the sequence obtained from strain CRL 1939. Out of these microorganisms only three were isolated from water environments and none of them were assayed as probiotics. A total of 59 and 39 sequences had an identity of 100 and 99%, respectively with the sequences obtained from isolates CRL 1937 and CRL1938. Out of these microorganisms 22 were isolated from water environments, being only 12 isolated from fishes. Out of this twelve, one was isolated from farmed Tilapia (Oreochromis niloticus) and only described and characterized by molecular biology, two of them were isolated from farmed seabass (Dicentrarchus labrax) and evaluated only by their ability to inhibit pathogens and foodborne pathogens in invitro assays, one was isolated from Grass carp (Ctenopharyngodon idellus) and evaluated in vitro only in its cellulolytic activity. Finally one was isolated from a freshwater fish and seven from Mullet (Mugil cephalus), all of them were annotated in GenBank as potentially probiotic bacteria, however these results are shown as unpublished and we have not found any related publications. Results over the comparison of Bacillus strains showed that a total of 98 sequences had an identity of 99% with the sequence obtained from isolate A252. Out of these microorganisms 13 were isolated from water environments, being only 4 isolated from aquatic animals and only two evaluated in beneficial properties such as antitumoral in human gastric carcinoma cell lines (MC-4 and SGC-7901) and antimicrobial activity against plants pathogenic fungi. A total of 10 and 88 sequences had an identity of 100 and 99%, respectively, with the sequences obtained from isolates A253 and A254. Out of these microorganisms one was isolated from sea sediments, two from algae and one from the digestive tract of a deep sea eel; none was tested as probiotic microorganisms in in vitro or in vivo assays. Discussion The introduction of P. mesopotamicus to the fish culture system in the northern region of Argentina led to an increase in the production of aquaculture facilities. This rise is explained by the intrinsic resistance of this autochthonous fish to the environmental conditions. On the other hand, the lack of knowledge on some critical factors like the biological cycle, food requirements and immune system together with the stressing conditions of intensive production systems and the increasing use of antibiotics induces a decrease in the number of animals. These factors led to the need of novel techniques to increase surviving and mean weight in order to obtain a higher number of animals in each stage of the biological cycle. Probiotics emerge as a valid alternative to be widely and successfully applied all around the world in aquatic and terrestrial organisms. As demonstrated by Abraham et al. in ornamental fishes, the use of probiotic isolated from animals belonging to a different environment, makes uncertain the survival of probiotic microbes in the gastrointestinal tract of fishes [22]. These results support the need to isolate autochthonous microorganisms as putative probiotics to be tested in further "in vivo" assays. The results obtained in the first stage of this study indicate that there is a relative decrease in the number of yeasts during the cold seasons when compared with the summer period. The relative proportions between cocci and rods showed no tendencies due to the high variability between the cultivable bacteria obtained. This variability was described previously as one of the most important features of the gastrointestinal microbiota in fish, by the different environmental conditions that are very susceptible to a long list of external and internal factors [23]. These results support the idea of a seasonal variation of the indigenous microbiome that could be related to the animal behavior and food Inhibition is expressed as cm of the inhibition halo without the diameter of the well. SN: Supernatant N: Neutralized Supernatant intake during the cold seasons, period in which animals eat less and consequently show a slower growth than in wormer seasons. The indigenous microbiome of the gastrointestinal tract of fishes consists in a huge variety of microorganisms in which only some species can be considered as beneficial. Then, it is necessary to determine which groups or genera can be considered as potentially beneficial organisms. In the last years, specific entities have published different categories of "secure" microorganisms such as the GRAS (Generally regarded as Safe), FGM (Food Grade Microorganisms) and QPS (Qualified Presumptive as Safe) qualifications. Some LAB, and Bacillus species are included into these classifications. Based on these observations, the following aim of this research was to partially identify those isolates belonging to both groups. Only 30 LAB and 8 Bacillus isolates were obtained, representing only a 5.75% and a 1.53% of the total isolations, respectively. These results are in agreement with Sakata (1990), who establishes that Gram negative facultative anaerobes are the prevalent microorganisms in the digestive tract and symbiotic anaerobes could be [24]. The low incidence of LAB in P. mesopotamicus evidenced in this work, agrees with the

fact that, although they are not dominant in the indigenous intestinal microbiota of freshwater fish, some of them can colonize the gut [9,25]. The next step of the work was the evaluation of the beneficial properties of the LAB and Bacillus isolates. The hydrogen peroxide production is a desirable characteristic in potentially probiotics for its synergic action together with organic acids against pathogens and due to the reduction of potentially opportunistic pathogens in aquatic environments [6,26]. [17,27]. The production of antagonistic substances was also studied. Results indicate that 15 out of the 38 isolates were unable to inhibit the growth of assayed indicators. The preselection criterion for this stage was the ability to inhibit specific fish pathogens (A. salmonicida, Y. ruckeri or Klebsiella sp.). Thus, 12 LAB and 3 Bacillus isolates were preselected. The exogenous administration of surfactants or substances that promote its secretion is considered an adequate treatment for altered gastric mucosal barrier [28]. Besides, these substances are also promising compounds often showing antimicrobial and anti-adhesive properties, penetrating and removing mature biofilms from unanimated surfaces [29]. A wide range of microorganisms can produce these active compounds, mainly those belonging to the genus Bacillus. In this study the screening of biosurfactant-producing microorganisms was carried out using different methods. The production of biosurfactants and the capability of emulsify organic solvents was a selection criterion. 2 Bacillus isolates positive to the oil spreading technique, and 3 Bacillus isolates able to generate the most stable emulsions under the cited conditions were preselected. The adherence to the intestinal mucus layer is another important selection criterion for beneficial microorganisms and considered the first and key step in host colonization [30]. However, the difficulties in studying the bacterial adhesion through in vivo led to the use of in vitro models for the preliminary screening of potentially adherent isolates [31]. As a general rule, isolates adhering well to hydrocarbons are considered to be hydrophobic and those adhering poorly are considered hydrophilic. Then, hydrophobicity could indicate an advantage and important feature for bacterial maintenance in the gastrointestinal tract [32], being reported to be a qualitatively valid method to estimate the ability of an isolate to adhere to epithelial cells based on its hydrophobicity [33]. Based on the results obtained in this work, 4 LAB and one Bacillus isolates were preselected based on their highest hydrophobicity and auto aggregation indexes. The inclusion of more than one strain in a probiotic formula, demands to perform compatibility assays. From the 24 preselected isolates only LAB isolates 61B and 74 were incompatible and therefore they cannot be included in a mixed probiotic product. From the 38 isolates obtained in the first stage of the study, 8 were selected as potentially probiotic candidates for aquaculture of P. mesopotamicus taking into account general guidelines [34]. Selected microorganisms were: Pediococcus acidilactici CRL 1939 (isolate 66) for its high hydrogen peroxide production, Enterococcus faecium CRL 1941 (isolate 70) for its H 2 O 2 production and ability to inhibit two fish specific pathogens, E. faecium CRL 1940 (isolate 30) for its high hydrophobicity and auto aggregation indexes, E. faecium CRL 1937 (isolate A34) and CRL 1938 (A35) for their ability to inhibit two fish specific pathogens, Bacillus subtilis A252, A253 and A254 for being the only Bacillus isolates with inhibitory spectrum and emulsifying activity, A254 also showed a medium auto aggregation index and the production of biosurfactantes. The MEGABLAST comparison using de NCBI data base showed that there are no registrations of genetically similar microorganisms used as probiotics in aquaculture. Thus, selected microorganisms were included in the strain collection of the Laboratorio de Sanidad Animal of the Estación Experimental Agropecuaria Rafaela belonging to the Instituto Nacional de Tcnología Agropecuaria (INTA) under Budapest treaty [35]. The results obtained in this work represent the basis of the "in vivo" experiments planned to be performed in the next stage, as a way to determine the most suitable group of microorganisms, dose and biological cycle stage of the animals to be administered to fishes. Pragmatic considerations and approaches for measuring staff time as an implementation cost in health systems and clinics: key issues and applied examples Background As the field of implementation science wrestles with the need for system decision-makers to anticipate the budget impact of implementing new programs, there has been a push to report implementation costs more transparently. For this purpose, the method of time-driven activity-based costing (TDABC) has been heralded as a pragmatic advance. However, a recent TDABC review found that conventional methods for estimating staff time remain resource-intensive and called for simpler alternatives. Our objective was to conceptually compare conventional and emerging TDABC approaches to measuring staff time. Methods Our environmental scan of TDABC methods identified several categories of approaches for staff time estimation; across these categories, staff time was converted to cost as a pro-rated fraction of salary/benefits. Conventional approaches used a process map to identify each step of program delivery and estimated the staff time used at each step in one of 3 ways: (a) uniform estimates of time needed for commonly occurring tasks (self-report), (b) retrospective “time diary” (self-report), or (c) periodic direct observation. In contrast, novel semi-automated electronic health record (EHR) approaches “nudge” staff to self-report time for specific process map step(s)—serving as a contemporaneous time diary. Also, novel EHR-based automated approaches include timestamps to track specific steps in a process map. We compared the utility of these TDABC approach categories according to the 5 R’s model that measures domains of interest to system decision-makers: relevance, rapidity, rigor, resources, and replicability, and include two illustrative case examples. Results The 3 conventional TDABC staff time estimation methods are highly relevant to settings but have limited rapidity, variable rigor, are rather resource-intensive, and have varying replicability. In contrast to conventional TDABC methods, the semi-automated and automated EHR-based approaches have high rapidity, similar rigor, similar replicability, and are less resource-intensive, but have varying relevance to settings. Conclusions This synthesis and evaluation of conventional and emerging methods for staff time estimation by TDABC provides the field of implementation science with options beyond the current approaches. The field remains pressed to innovatively and pragmatically measure costs of program delivery that rate favorably across all of the 5 R’s domains. Page 2 of 10 Huebschmann et al. Implementation Science Communications (2022) 3:44 Contributions to the literature • The decision to adopt, implement, and sustain an evidence-based program is heavily influenced by cost. Often, staff time is a major driver of costs for a program. • Time-driven activity-based costing is widely used to measure implementation costs of a program, but current approaches to staff time estimation remain resource-intensive. • We reviewed conventional (uniform estimate, time diary, direct observation) and emerging electronic health record-based approaches to measure staff time used and compared these categories of approaches based on relevance to stakeholders, rapidity, rigor, resource requirements, and replicability. • The electronic health record provides the field of implementation science with new semi-automated and automated opportunities to assess time that can be rapid, rigorous, replicable, and low-burden in terms of resources required, but their relevance to a given setting/project may vary. Background The field of implementation science (IS) has made great progress in identifying critical approaches to translate evidence-based programs (EBP) into practice [1,2]. Despite this progress to guide the implementation of an EBP into a given health setting, persistent dissemination challenges include: (1) inconsistent "scaling up" to varied settings within a health system, (2) "scaling out" across different health systems remains rare, and (3) sustainment of these changes is difficult. When system-level decision makers lack information on the cost of implementing and sustaining EBPs, it deters dissemination and sustainment [3][4][5]. Some IS frameworks, including the Veterans Administration QUality Enhancement Research Initiative (VA QUERI) roadmap [6], seek to guide scaling up EBPs by considering different types of implementation costs within the following project phases: (1) pre-implementation, (2) implementation, and (3) sustainment [6]. During the pre-implementation and implementation phases, key cost considerations are as follows: (1) "capacity" for delivering the EBP, including the cost of staff time for both EBP delivery and the implementation strategy of staff training, and (2) comparing the costs of alternate implementation strategies. In the sustainment phase, the focus shifts to estimate the staff time needed to continue to deliver the EBP and implementation strategies, as well as other ongoing system costs such as program materials [5]. Recent reviews of cost assessment approaches for IS and improvement science have specified the need to track the staff time required for both EBP delivery and for implementation strategies used [5,[7][8][9]. Drilling down into the staff time costs for both of EBP delivery and implementation strategies is important because (1) staff time is a major source of costs for EBP delivery; (2) staff time is a costly element of certain implementation strategies, such as technical assistance and training; and (3) assessment of other types of costs, such as program materials, are more straightforward to track. The method of time-driven activity-based costing (TDABC) has been heralded as a relatively pragmatic approach to estimate the staff time required for these different tasks; accordingly, the use of TDABC in IS research has accelerated recently [3][4][5]. As developed by Kaplan et al. [3], TDABC methods specify costs across several steps of EBP implementation. A central aspect of TDABC is to create a process map that allocates the time for each staff actor to complete each process map step, inclusive of both EBP delivery and implementation strategies used [5]. However, a recent review of TDABC by Keel et al. [4] concluded that current approaches for staff time estimation in each step of a TDABC process map remain resource-intensive and called for the development of simpler and more rapid approaches with less resource burden [4]. Accordingly, the field would benefit from more pragmatic staff time estimation approaches, with balanced attention to rigorous and reliable data collection methods [10]. Thus, there is a need to contrast the conventional methods of staff time estimation with some novel and emerging electronic health record (EHR)-based methods that could address some of the current challenges. The purpose of this brief methodology report is to compare distinct categories of conventional and emerging TDABC approaches to staff time estimation according to the 5 R's model [11] for pragmatism that measures domains of interest to researchers and health system pressed to innovatively and pragmatically measure costs of program delivery that rate favorably across all of the 5 R's domains. Keywords: Costing, Costs and cost analysis, Implementation, Time-driven activity-based costing, Program delivery, Health workforce; Staff time estimation decision-makers: relevance, rapidity, rigor, resources, and replicability. In contrast to recent reviews and commentaries that only considered conventional approaches to staff time estimation by TDABC [4,5,[7][8][9][10], this paper also considers innovative automated and semi-automated EHR-based approaches, and compares these different approaches on each of the 5 R's domains. We also provide two illustrative case study examples that delineate why different staff time estimation approaches may be selected. This environmental scan of emerging pragmatic methods for staff time estimation provides the field of IS with options beyond the current standards of observation or asynchronous reporting, and presses the field to identify additional non-intrusive, real-time approaches to assessing implementation costs. Methods We conducted an environmental scan, including a literature search, for articles measuring the cost of staff time to implement healthcare-related EBPs. We searched Pub-Med using the following search terms: ("implementation cost" or "time-driven activity-based cost" or "microcost") and ("health*" or "clinic*"). The literature search was limited to articles in English over the past 5 years. Articles' references were hand-searched for additional articles. We also queried an online community of EHR users (Epic UserWeb) and colleagues with experience in EHR approaches to time capture: a clinical informatics nurse research scientist and two physician informaticists. Our intent was not to conduct a systematic review, but to use this environmental scan to identify existing categories of staff time estimation approaches, and to compare the relative pragmatism of these approaches using the 5 R's model perspective [11] (Table 1). While not exclusive to IS, the 5 R's was selected because it was developed to increase the pragmatism of health research and is an accepted model of pragmatic health research domains [11][12][13]. The 5 R's framework emphasis on relevance, rigor, and replicability are complementary with the approach that Cidav et al. took to track

TDABC according to the Proctor et al. framework [12] by specifying who/what/when/how often/for how long an individual delivers an implementation strategy, but also adds an explicit emphasis on rapid/low resource burden approaches [5]. Approaches to staff time estimation were evaluated from the perspective of system-level decision makers. Decision makers did not participate in the review process, but we considered their perspective on how a new EBP would impact their budget. Using a content analysis approach, two authors (KT and AH) reviewed the distinct approaches to staff time estimation in each article and placed them in categories named for common time capture terms [14]. We evaluated these categories of approaches from the 5 R's model perspective [11] (Table 1), providing more favorable ratings if they (1) rated high in relevance to stakeholders, rapidity, and recursiveness, rigor, and replicability and (2) required few resources. Results From our environmental scan, we identified several categories of approaches to estimate the staff time spent implementing EBPs as a part of TDABC [4,5,7,15]. These categories of staff time estimation approaches are applicable to EBP program delivery by managers, supervisors, and staff, as well as implementation strategies (e.g., training to deliver the EBP, and other time spent preparing for the program); time spent evaluating the program; and indirect time costs of the program on patients and care givers [5,7,15]. With the caveat that the approaches used to capture staff time were not always clearly described in our literature search, and a given study sometimes used more than one category of staff time estimation approach in concert [4], the most common conventional approaches reported were self-report using a time-reporting template or "time diary" [16][17][18] and uniform self-report estimates of time spent on certain activities [5,7,[19][20][21]. Some studies also reported a category of direct observation [22][23][24]. Using our pre-specified search terms, we found one study reporting use of an automated EHR-related approach [23]. Our broad environmental scan also identified other articles using semi-automated or automated EHR-based approaches for staff time estimation, including recommendations for their use and reporting [25,26]. Our summary of these TDABC categories of approaches are summarized in Table 2 from a 5 R's model perspective. Self-report/observation categories We identified numerous articles using conventional selfreport or observation approaches to estimate staff time [5,[7][8][9]28]. As described above, these began with a process map to identify each step of EBP delivery and then estimated the staff time required at each step of the process map using one of these approaches: (a) uniform estimate of time needed for a commonly occurring task, (b) retrospective self-report in "time diary", or (c) periodic direct observation. However, these approaches are somewhat resource-intensive, especially observation. Further, using these approaches, costs may not be feasible to capture during the sustainment phase when there are no grant funds to support observations and/or compilation of self-report data. Automated/semi-automated EHR-based approaches For programs in settings that have EHRs, recent approaches have emerged to automate the data collection partly or fully. Semi-automated approaches may include hard stops built into a specific EHR note type that "nudge" a user to input their time-this essentially embeds a contemporaneous time diary into the note. Incorporating a contemporaneous time diary into the clinic note allows staff to review their charting to guide their time estimate reported and may lessen recall bias by completing the time diary in "real-time. " In contrast, fully automated approaches require no action by the staff. Seven categories of fully automated EHR-based approaches to staff time estimation are possible. These 7 categories are not mutually exclusive and include time spent: (1) documenting care provided-including the time spent within specific types of encounters, such as time spent documenting an anticoagulation visit encounter in the second case example below, (2) time spent placing or refilling prescriptions, (3) managing the EHR inbox including patient messages, (4) managing orders as part of the team, (5) providing direct patient care, (6) work during scheduled work hours, and (7) work outside of scheduled work hours [25]. With automated approaches, data are collected in real time (e.g., by the EHR), which avoids recall bias and reduces the resource burden. However, depending on the EHR vendor or other software used to track time, there are limitations in what activities can be tracked, the accuracy of the tracking estimates, and the timeliness of retrieving the data. For example, when clinicians are multitasking and leave EHR windows open, time estimates may be inflated. As it relates to relevance, if the encounter type that is tracked is not specific to the EBP and it also captures other tasks, it may not be fully relevant and the rigor of measurement is decreased. Regarding resources, some automated EHR-based approaches require assistance from the EHR vendor and/or local analysts/informaticists at the outset to determine what data to collect and how to access them. In addition, although collected in real time, the data may not be accessible in real timedata access may also require help from the EHR vendor or a local analyst/informaticist. After the initial set up, there are some benefits to EHR approaches, notably that when programs reach a sustainment phase [3], an automated method that was set up in the EHR can continue to provide reports of the staff time needed for a certain type of clinical encounter. Case examples To further illustrate the tradeoffs of these different approaches, we provide an overview of the approaches employed in two real-world research case examples [26]. In the first case, the study used both self-report and semiautomated approaches for capturing time spent, whereas in the second case, the authors used both automated approaches and direct observation to develop a complete workflow process map and to validate the automated timing calculations. In Table 3, we provide the rationale for the approach selected in each case example, and at least one alternative approach that could have been used. The first case example is from a pilot type 2 hybrid implementation/effectiveness trial studying the delivery of an evidence-based physical activity coaching intervention in a primary care clinic [26]. Staff time costs included the following: (1) an implementation strategy of training existing staff to serve as coaches; (2) time spent delivering the 6 intervention telephone calls to each patient, and (3) time for the implementation strategy of coaches providing technical assistance to patients to share their physical activity data (FitBit©). Approaches to capture time varied across the different elements of the program (Table 3). For time spent training, a conventional selfreport time diary was used per the staff employer's preference, in order to allocate the time spent to the research grant for this one-time session. To optimally capture the time spent in each counseling session, a semi-automated EHR-based approach was used to avoid recall bias: a brief, required contemporaneous time diary was embedded within the behavioral coaching note template in the EHR (Epic Systems). This embedded time collection template can easily be replicated in Epic Systems and other commonly used EHRs by creating a "required field" for Benefits and Challenges mapped to Domains of 5 R's framework Self-report approaches Uniform estimates of time for tasks [5,7] -An average time is assessed for common tasks -Often used in concert with a process map to identify all tasks and the task frequency/actors -Can be self-tracked or an evaluation team member prompts individuals in each role to understand the time required for common tasks • Relevance is high to implementers-these tasks are part of the existing process • Rapidity is low-to-moderate-some delays to initially estimate data, but data collection is fairly rapid • Rigor is moderate-recall bias is minimized for common tasks, but this may be a bigger problem for rare tasks. Systems use this approach more for common tasks • Replicability is moderate-other systems likely use different process maps, so this may not be fully replicable • Resources required are moderate for research assistants or quality improvement teams-both to ensure process map is accurate and to send surveys and compile survey results. This poses a challenge to continue tracking in a sustainment period Retrospective time diary [16] -Time diary may include a template or time card that tracks time spent in different aspects of a project -Recorded retrospectively (e.g., time spent in last week or last month), but can be real time -Either self-tracked or may need to be interviewed or prompted to complete • Relevance is high-specific to the time spent on a relevant task for implementers • Rapidity is low-to-moderate-there are often delays to compile these self-report data, but templates are fairly quick to complete. Occasionally, these may be triggered automatically • Rigor is moderate, but recall bias is minimized for common tasks and by more frequent assessments with shorter recall periods (e.g., recall over 1 week is more rigorous than recall over a period of 1-3 months); as a negative, when using a time card type of approach, there is a more limited ability to capture variation of task type as compared to the uniform estimates of time for task approaches • Replicability is high-others may replicate this approach • Resources required are moderate-must be prompted to complete time diary (other than when it is triggered automatically) and compile survey results -thus at-risk to continue this in sustainment period Third party observer approaches Direct observation [22] -Specific observation template for 3rd-party to document time spent in different aspects of project and by different staff members • Relevance is high to implementers-these tasks are part of the existing process • Rapidity is low-this is a time-consuming process to observe and compile data, particularly for low-frequency events • Rigor is high due to direct observation-used as a "gold standard" comparison [27] • Replicability is high-others may replicate this approach • Resources required are high, and this is particularly challenging for low-frequency events due to large periods of "down time" for observers Table 3 Rationale for use of specific TDABC approaches in the case example pilot trials Rationale for this approach Alternative cost capture approaches (problems with this alternative) Case example 1: type 2 hybrid trial of physical activity coaching in primary care [26] Training (coach time) Self-report: retrospective time diary The coaches' employer received time diary data to track the time spent on this project, in order to allocate a proper portion of their salary to the research grant time that must be documented before closing the note. In contrast to an alert that fires and interrupts workflow, this approach only nudges staff if the template was left incomplete when signing the encounter. During the preimplementation phase, the coaches noted this approach fit their workflow and was minimally burdensome. The second example is a program evaluation of the staff costs of delivering care at an anticoagulation clinic for various phenotypes of patients-those who needed minimal adjustment to their treatment regimen and those who needed frequent adjustments [23]. As they sought to compare variable costs across patients in the existing anticoagulation clinic where baseline training had already occurred, the authors did not assess staff training costs. Instead, they used direct observation to detail a process map of each step in the workflow for a patient to engage with the anticoagulation clinic staff. This included multiple steps for in-person visits, from the time of check-in until the time of check-out, and the time spent by nurses and pharmacists between in-person clinic visits. Using a proprietary internal database, the authors captured automated data for the time spent by each member of the clinical team in each step of the process map workflow. They also used a subset of direct observation assessments to validate these automated measurements of staff time. Using TDABC, they calculated the costs of the staff time in each step of the workflow, and then differentiated the costs for patients who were well-controlled and not well-controlled. Although this internal database was proprietary to their system, other EHRs, including the Epic Systems EHR [25], also have this capacity to track staff time. Discussion This brief methodologic commentary compares several approaches to capturing the portion of implementation costs related to staff time-an important element of implementation according to the VA QUERI Roadmap [6] and other IS process models. In particular, approaches to capturing staff time are critical to transparently report to system decision-makers the time required to implement and sustain a program. Overall, the comparison of these approaches in

Table 2 may be considered as a balance of data optimization (i.e., rigorous/reliable) and efficiency in terms of a rapid return of relevant findings with low-resource requirements. In terms of the rigor/reliability, the observational approaches are most accurate, followed by the automated and semi-automated EHRbased approaches, and then the retrospective time diary approaches that are particularly prone to recall biases. In terms of efficiency, the semi-automated/automated EHR approaches stand out for their rapidity and for the limited resources needed after their initial set-up, followed by self-report. Observational approaches are the slowest and most time-consuming. It is interesting to further consider the relative merits of these approaches from the VA QUERI Roadmap perspective which dictates that estimates of staff time are most critical to assess in the Sustainment phase. Conventional self-report time diary and observational approaches are typically too burdensome for use in the Sustainment phase; however, the conventional self-report uniform estimate approaches could be pragmatic in this phase, as well as the semi-automated or automated EHR-based approaches. In contrast, during the pre-implementation planning phase of an EBP, estimation may be the only possible approach available if decision makers need data on the time required for alternate implementation strategies before these tasks have been pilot-tested. In sum, advances are needed in terms of highly rapid, rigorous, and low-resource time capture approaches, and the semiautomated and automated approaches described here provide innovative steps forward towards that goal. Strengths of this report include its summary of key emerging EHR-based semi-automatic and automatic approaches to capturing time and the concrete case study examples (Table 3). Further, the 5 R's model provided a systematic basis on which to evaluate the pragmatism of different approaches. In addition, reporting staff time as a cost is consistent with the recommendations from the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) [29] to "describe the methods for valuing each resource in terms of its unit cost. " However, depending on the EHR approach used, the automated approach may be challenged to separately report the distinct resource costs for the intervention and the implementation strategy, as others have recommended [5]. Although beyond the scope of this brief review, those applying these different approaches to staff time estimation should keep in mind the CHEERS recommendations to specify which staff are included (e.g., clinical staff, contracted coaches) and from what perspective (e.g., clinical health system staff, research staff ) [29]. Limitations include that our focused environmental scan on conventional self-report/observation approaches to staff time estimation and EHR-based semi-automatic and automatic approaches did not include all potential approaches relevant for IS, such as automated assessments by radiofrequency identification (RFID) tags or readers. In addition, our comparisons according to the 5 R's model are necessarily subjective. A future systematic review would expand and add rigor to this environmental scan. Automated EHR-based approaches have been used internally by health systems more often than in IS research; thus, there are some key limitations in terms of sparse prior reporting of details and validation of these approaches [25]. However, some of the described EHR approaches have been validated against direct observation and demonstrated that > 80% of the time the estimates are within 3 min of each other [27]. When used for research, it is reasonable to initially vet the accuracy of automated EHR-based approaches as compared to observation [25], as was done in case example 2-this is particularly important for complex processes that are prone to interruptions. Conclusions We summarized the strengths and limitations of different conventional and EHR-based semi-automated and automated approaches to measuring staff time as a cost for IS studies, with an emphasis on the 5 R's model as an index of factors that are important to stakeholders. This is critical to allow decision-makers to consider the feasibility of implementing and sustaining programs, based on the estimates of staff time required. Going forward, the field should continue to identify additional methods of estimating staff time (and other implementation costs) that are rigorous and replicable, and also relevant, rapid, and low-resource enough to be measured in a EBP sustainment phase. Needle tract seeding following percutaneous biopsy of renal cell carcinoma A 66-year-old man underwent computed tomography-guided needle biopsy of a suspicious renal mass. Two months later he underwent partial nephrectomy. Histology revealed a 30-mm clear cell renal cell carcinoma, up to Fuhrman grade 3. An area of the capsule was interrupted, which corresponded to a hemorrhagic area on the cortical surface. Under microscopy, this area showed a tongue of tumor tissue protruding through the renal capsule. A tumor deposit was found in the perinephric fat. These features suggest that tumor seeding may have occurred during the needle biopsy. INTRODUCTION Incidental detection of small renal masses is increasing. This has led to an increase in biopsy of small renal masses, a proportion of which are then proven to be benign. Needle biopsy of small renal masses is controversial owing to the risk of seeding malignant cells along the needle tract. Needle tract seeding is a rare event; the incidence is estimated to be less than 1 in 10,000 cases of all biopsies [1]. Eight other cases of needle tract seeding in a renal mass biopsy have been described in the medical literature, two as recently as 2013 (Table 1). Complications in these patients include distant metastasis and death. We report our experience of a man with renal cell carcinoma (RCC) seeding along a biopsy tract and compare the circumstances and biopsy techniques with reported cases in the literature. CASE REPORT A 66-year-old man was incidentally found to have a 32mm right lower pole renal mass on a computed tomography (CT) scan (Fig. 1). CT-guided needle biopsy of the renal mass ( Fig. 2) revealed Fuhrman grade 2 clear cell RCC. Two samples were obtained by use of a 16-gauge Temno core biopsy needle (CareFusion, San Diego, CA, USA) and a 22-gauge Francine needle. One pass was made for each needle sampling. No coaxial needle was used. Within 2 months, the patient underwent right partial nephrectomy. Perinephric fat was sent along with the excised renal lesion for histopathological analysis. No specific feature of tumor seeding was seen intraoperatively. Histopathology revealed a well-circumscribed 30-mm clear cell RCC, predominantly Fuhrman grade 2 with focal areas of grade 3. There was an area where the capsule was interrupted that corresponded to a hemorrhagic area on the cortical surface (Fig. 3). A tumor www.kjurology.org Needle tract seeding following percutaneous biopsy deposit was also noted in the perinephric fat. These features suggested that the tumor deposit in the fat was likely due to tumor seeding rather than a metastasis and that the tumor seeding could have resulted from the needle biopsy. His TNM staging was pT3a NX MX, at least stage 3 disease (American Joint Committee on Cancer, 7th edition, 2010) and his Leibovich score was 5 (intermediate risk). Six months after the operation, there was no radiological evidence of tumour recurrence on a CT scan. DISCUSSION Aside from the potential for false-negative results, a key risk of renal mass biopsy is seeding of the biopsy tract with malignant cells. Several factors in theory could affect the risk of biopsy tract seeding, such as needle size, the number of needle passes, and the use of a coaxial needle. Biopsy tract seeding has been reported in renal mass biopsies using needles as fine as 23-gauge and as large as 14-gauge [2][3][4]. Theoretically, a larger-bore needle would increase the risk of seeding owing to an increased area of defect on the surface of the tumor and an increased circumference or surface area of the needle. However, because of the scarcity of cases, it is difficult at this stage to accurately determine a relationship between needle size and the risk of seeding. It is also difficult because of underreporting to associate the risk of needle tract seeding with the number of needle passes through a tumor. Use of a coaxial needle allows multiple passes through the renal mass with only one pass through the surrounding normal tissue. This theoretically reduces the risk of needle tract seeding into normal tissue and potentially reduces patient discomfort as well [5]. Although it is interesting to note that a coaxial needle was not used in any of the currently reported cases of needle tract seeding af ter renal mass biopsy (Table 1), there are just too few cases to establish a firm relationship between the risk of biopsy tract seeding and the use of a coaxial needle. Visualization of larger coaxial needles on ultrasound or CT may be easier than with smaller biopsy needles [5], and this may improve accuracy. Histological evidence of biopsy tract seeding may not always be found after definitive surgery to remove the renal mass. Seeding into excised perinephric tissues can be found soon after surgery but seeding into surrounding muscle, fascia, and skin may only be apparent months, or even years, after surgery. As was seen with this case, the biopsy needle traversed skin, subcutaneous tissue, multiple muscle and fascia layers, and perinephric fat before reaching the renal lesion (Fig. 2). Thus, the tumor could theoretically seed into one or more of these tissues; seeding as superficial as the subcutaneous tissue has been reported (Table 1). This delayed presentation may increase the risk of adverse outcomes such as further metastasis and poorer prognosis. Time to presentation or diagnosis of tumor seeding after renal mass biopsy has ranged from 24 days to 84 months in previously reported cases where tumor seeding was not found on the initial histopathological analysis (Table 1). In conclusion, a common feature in all reported cases of needle tract seeding from a renal mass biopsy is that a coaxial needle was not used. However, because of the paucity of cases, there is currently no satisfactory association between the risk of needle tract seeding and needle size or the number of needle passes. It is important to consider that histopathological evidence of needle tract seeding may not be apparent in all cases, especially if seeding occurred beyond the excised tissues. CONFLICTS OF INTEREST The authors have nothing to disclose. Periconceptional Maternal Mediterranean Diet Is Associated With Favorable Offspring Behaviors and Altered CpG Methylation of Imprinted Genes Background: Maternal diet during pregnancy has been shown to influence the child neuro-developmental outcomes. Studies examining effects of dietary patterns on offspring behavior are sparse. Objective: Determine if maternal adherence to a Mediterranean diet is associated with child behavioral outcomes assessed early in life, and to evaluate the role of differentially methylated regions (DMRs) regulating genomically imprinted genes in these associations. Methods: Among 325 mother/infant pairs, we used regression models to evaluate the association between tertiles of maternal periconceptional Mediterranean diet adherence (MDA) scores derived from a Food Frequency Questionnaire, and social and emotional scores derived from the Infant Toddler Social and Emotional Assessment (ITSEA) questionnaire in the second year of life. Methylation of nine genomically imprinted genes was measured to determine if MDA was associated with CpG methylation. Results: Child depression was inversely associated with maternal MDA (Bonferroni-corrected p = 0.041). While controlling for false-discovery, compared to offspring of women with the lowest MDA tertile, those with MDA scores in middle and high MDA tertiles had decreased odds for atypical behaviors [OR (95% CI) = 0.40 (0.20, 0.78) for middle and 0.40 (0.17, 0.92) for highest tertile], for maladaptive behaviors [0.37 (0.18, 0.72) for middle tertile and 0.42 (0.18, 0.95) for highest tertile] and for an index of autism spectrum disorder behaviors [0.46 (0.23, 0.90) for middle and 0.35 (0.15, 0.80) for highest tertile]. Offspring of women with the highest MDA tertile were less likely to exhibit depressive [OR = 0.28 (0.12, 0.64)] and anxiety [0.42 (0.18, 0.97)] behaviors and increased odds of social relatedness [2.31 (1.04, 5.19)] behaviors when compared to low MDA mothers. Some associations varied by sex. Perinatal MDA score was associated with methylation differences for imprinted control regions of PEG10/SGCE [females: Beta (95% CI) = 1.66 (0.52, 2.80) – Bonferroni-corrected p = 0.048; males: -0.56 (-1.13, -0.00)], as well as both MEG3 and IGF2 in males [0.97 (0.00, 1.94)] and -0.92 (-1.65, -0.19) respectively. Conclusion: In this ethnically diverse cohort, maternal adherence to a Mediterranean diet in early pregnancy was associated with favorable neurobehavioral outcomes in early childhood and with sex-dependent methylation differences of MEG3, IGF2, and SGCE/PEG10 DMRs.

Background: Maternal diet during pregnancy has been shown to influence the child neuro-developmental outcomes. Studies examining effects of dietary patterns on offspring behavior are sparse. Objective: Determine if maternal adherence to a Mediterranean diet is associated with child behavioral outcomes assessed early in life, and to evaluate the role of differentially methylated regions (DMRs) regulating genomically imprinted genes in these associations. Methods: Among 325 mother/infant pairs, we used regression models to evaluate the association between tertiles of maternal periconceptional Mediterranean diet adherence (MDA) scores derived from a Food Frequency Questionnaire, and social and emotional scores derived from the Infant Toddler Social and Emotional Assessment (ITSEA) questionnaire in the second year of life. Methylation of nine genomically imprinted genes was measured to determine if MDA was associated with CpG methylation. Results: Child depression was inversely associated with maternal MDA (Bonferronicorrected p = 0.041). While controlling for false-discovery, compared to offspring of women with the lowest MDA tertile, those with MDA scores in middle and high MDA tertiles had decreased odds for atypical behaviors [OR (95% CI) = 0.40 (0.20, 0.78) for middle and 0.40 (0.17, 0.92) for highest tertile], for maladaptive behaviors [0.37 (0.18, 0.72) for middle tertile and 0.42 (0.18, 0.95) for highest tertile] and for an index of autism spectrum disorder behaviors [0.46 (0.23, 0.90) for middle and 0.35 (0.15, 0.80) for highest tertile]. Offspring of women with the highest MDA tertile were less likely to INTRODUCTION Numerous prenatal environmental exposures including smoking, stress, maternal obesity, and dietary factors have been associated with health and neurodevelopmental outcomes in children (Huizink et al., 2002;Wiebe et al., 2009;Steenweg-de Graaff et al., 2012;Gartstein and Skinner, 2017), with some exposures conferring life-long associations (House et al., 2016;Wyss et al., 2017). In a recent meta-analysis, pre-pregnancy obesity was estimated to be associated with a 51% increased risk of any neurodevelopmental impairment and a 62% increased risk of attention deficit hyperactivity disorder (ADHD) symptoms or ADHD related impairments . Prenatal nutrition may also have lasting effects on child neurodevelopment (Anjos et al., 2013;Lyall et al., 2014). While extensive research concerning maternal obesity, gestational weight gain, and child behavior has been conducted [for reviews see (Rivera et al., 2015;Contu and Hawkes, 2017;Sanchez et al., 2017)], less is known about particular types of maternal diets in humans and their effects on early childhood behavior. Far fewer studies have examined overall diet patterns (Borge et al., 2017). Among 23,020 children in the Norwegian Mother and Child Cohort Study, Jacka et al. (2013) reported that "unhealthy" maternal diet patterns during pregnancy were associated with increased externalizing child behaviors. A recent study in the United Kingdom suggested that maternal diets characterized by higher intakes of fruits and vegetables and lower intakes of meat and potatoes are associated with higher child IQ at age 8 years (Freitas-Vilela et al., 2017). In the Generation R study involving 3,104 children, investigators reported decreased levels of externalizing behaviors from children born to mothers who adhered to a Mediterranean diet during pregnancy, and increased child externalizing behaviors in children born to mothers who consumed a "Traditionally Dutch" diet consisting of high meat intake (processed and unprocessed), margarines and potatoes (Steenweg-de Graaff et al., 2014). Although mechanisms linking Mediterranean diet and outcomes are still unclear, epigenetic modifications are hypothesized as one mechanism to explain associations between environmental cues such as maternal diet, and child temperament/behavior (Harper, 2005;Lee, 2015;Arpon et al., 2016;Lorite Mingot et al., 2017); for review see (Lillycrop and Burdge, 2015). Animal studies have demonstrated changes in DNA methylation in the offspring of mice from dams fed different diets during pregnancy (Vucetic et al., 2010). In addition, dietary choline deficiency in dams has been shown to alter methylation patterns in mouse fetal brains (Niculescu et al., 2006). In humans, low maternal MDA has recently been associated with altered CpG methylation of the imprinted MEG3-IG control region in female offspring (Gonzalez-Nahm et al., 2017). In other studies, imprinted IGF2 was altered at birth in offspring with periconceptional exposure to famine during world war II (1944II ( -1945 (Tobi et al., 2009). Additional support for epigenetics as a mechanism for modulating behavior in humans is recent work identifying altered methylation in brain associated with schizophrenia and bipolar disorders (Xiao et al., 2014), altered cord-blood methylation patterns in DRD4 and 5-HTT associated with ADHD symptoms in children (van Mil et al., 2014), and work by Fuemmeler et al. (2016) relating methylation of the regulatory regions of imprinted genes with infant temperament. Sex-specific methylation differences are common in the genome (Singmann et al., 2015). As shown in Gonzalez-Nahm's work (Gonzalez-Nahm et al., 2017), differential methylation of imprinted genes from maternal dietary exposures in offspring can also be sex-specific. Additional evidence for sex-specific differences in the methylation of imprinted control regions (ICRs) in response to prenatal exposures comes from work showing that female infants, but not males, exhibited differential methylation of ICRs of IGF2 and H19 in response to maternal anxiety (Mansell et al., 2016). Likewise, lead exposures have been associated with sex-specific changes to the methylation patterns in multiple ICRs (Goodrich et al., 2015;Li Y. et al., 2016), and while not statistically significant, sex-specific differential methylation of the H19 and IFG2 DMRs were associated with prenatal exposure to cigarette smoke (Murphy et al., 2012a). In the current study, we sought to test the hypothesis that maternal adherence to a dietary pattern favorably associated with multiple health outcomes (Mediterranean) would be associated with offspring behavior and changes in methylation of ICRs, many of which have been associated with other behavioral offspring outcomes from prenatal exposures. NEST Cohort and Study Population Study participants included mother and infant dyads enrolled in the Newborn Epigenetics STudy (NEST), a prospective cohort for which accrual protocols have been previously described in detail (Liu et al., 2012). The target population was first trimester pregnant women visiting prenatal clinics of Duke Hospital and Duke affiliated clinics. To be included, women had to be aged 18 years and older and speak English or Spanish. From these, women with an established HIV infection or who intended to relinquish custody of offspring were excluded. To facilitate specimen collection at birth, women who receive obstetrics care at hospitals other than Duke Obstetrics or Durham Regional Hospitals were also excluded. Between 2009 and 2011, 1700 pregnant women were enrolled, and 396 were lost to the cohort due to fetal wastage (n = 113), refusal of further participation (n = 146), received obstetric care at an outside hospital (n = 114), or other reasons (n = 23), such that 1,304 (76.7%) women remained enrolled in the study up to the time of analysis. The recruitment protocol was approved by the Duke University Institutional Review Board. These analyses are limited to 325 mother child pairs (excluding multiple births) from which mothers completed both a periconceptional-food frequency questionnaire (FFQ) (University of Texas, MD Anderson Cancer Center Nutrition and Lifestyle Core Questionnaire 2008v.2) either upon enrollment or in their first trimester, and an Infant Toddler Social and Emotional Assessment (ITSEA) during the child's second year of life. Because we were interested in maternal diet and because of the potential epigenetic vulnerability due to the requirement to maintain methylation at imprinted DMRs during the reprogramming that occurs immediately post-fertilization, mothers were specifically asked to recall foods consumed at or near their last menstrual cycle. FFQ responses were converted to intakes of foods, food groups, energy, and nutrients by Nutrition Quest 1 . Child Behavioral Outcomes The ITSEA (Carter et al., 2003) was administered by a parent, caregiver or staff to children aged 12-24 months (mean = 13.9 months) during a NEST follow-up visit. The ITSEA tool has been validated and used extensively to examine social-emotional behavioral outcomes and temperament in early-childhood (Carter et al., 2003(Carter et al., , 2010. Scales of the ITSEA have demonstrated acceptable test-retest reliability and inter-rater reliability; Cronbach's alpha coefficients for the internalizing and externalizing domains were 0.80 and 0.86, respectively (Carter et al., 2010). The ITSEA consists of questions regarding Externalizing behaviors (with domains for aggressiondefiance, peer-aggression, and impulsive-activity), Internalizing behaviors (anxiety, separation-distress, depression-withdrawal, and inhibition-to-novelty), Dysregulation related behaviors (sleep, negative-emotionality, eating, and sensory-sensitivity), and Competency (attention, compliance, imitation/play, masterymotivation, and empathy). In addition, the ITSEA contains 13 questions to assess Maladaptive behavior, 10 questions to assess Social-Relatedness behavior and 8 questions to assess Atypical behavior. We also examined a composite autism spectrum disorder index calculated from the problem and competency portions of the ITSEA as described in Kruizinga et al. (2014). With the exception of the Social-Relatedness category and all Competency domains, increased scores indicate adverse behavioral outcomes. Infant Toddler Social and Emotional Assessment questions are scored from 0 to 2 (0 = Not True/Rarely, 1 = Somewhat, 2 = Very True/Often) and final domains are scaled by the number of questions in the domain. To examine associations of diet on ITSEA assessed child behaviors (Figure 1), and as many items were not normally distributed, even after log-transformation (Supplementary Figure 1), ITSEA behavior scores were divided FIGURE 1 | Maternal Mediterranean Diet Adherence (MDA) and Child Behavior Outcomes. For a given tertile of maternal MDA compared to tertile 1 (referent), the odds ratio (95% confidence interval) represents the risk of being in a higher tertile of behavioral outcome. Unadjusted (gray) and adjusted (blue) odds ratios (95% confidence intervals) are plotted. Estimates were adjusted for breastfeeding at least 3 months, age of child at behavioral assessment, maternal fiber intake, total calories, folate, education, diabetes, obesity, smoking, and age, as well as paternal age and child parity, premature birth, weight, race, and child sex. Q-values are shown and represent the false discovery rate (FDR) for each finding, which is automatically corrected for multiple testing. * Bonferroni-corrected p-trend = 0.041. into tertiles and associations with dietary exposures were assessed using ordinal logistic regression. We did not assess the peer aggression subscale of externalizing behavior due to lack of nonzero scoring (>2/3 of data were zero; data not shown). Mediterranean Diet Scoring As described in detail by Gonzalez-Nahm et al. (2017), pregnant women's MDA was calculated from the food frequency questionnaire (FFQ) upon enrollment soon after conception, concerning dietary consumption at or near their last menstrual period. In brief, intakes of food categories were adjusted by total energy to grams/1000 kcals and a modified version of Trichopoulou's Mediterranean diet scoring was used (Trichopoulou et al., 2003). Participants were scored a 0 for below-median consumption of meats (including red meat, pork, poultry, game, but excluding processed meats) and a 1 for abovemedian consumption of the following: fruit (including fresh, dried and frozen, but excluding juice), vegetables (excluding vegetable juice and white potatoes), fish, dairy (including full-fat dairy but excluding dairy desserts), whole grains, nuts and seeds (including nut-butters), beans and legumes (including soy), and the ratio of mono-unsaturated fat to saturated fat. Alcohol consumption was not scored as it is not recommended during pregnancy and was extremely low in this cohort. The final Mediterranean diet score ranged from 0 to 9 with 9 as most adherent (Supplementary Figure 2). This composite MDA score was divided into tertiles (referent = first tertile) and subsequently used to assess associations of MDA with child behaviors. The normal distribution of these scores (Supplementary Figure 2) was maintained by offspring sex and by the sub-cohort used to assess associations of maternal MDA on CpG methylation. Statistical Analysis For analysis of child behavioral outcomes, the following maternal covariates were included: age at delivery, education level (any college vs. none) pre-pregnancy obesity (BMI > 30 vs. BMI < 30), race (white, black, Hispanic, other), smoking during pregnancy, any of gestational, type I or type II diabetes, daily intake of folate (diet plus supplements: <400 µg, 400-800 µg, >800 µg), daily fiber intake, daily energy intake, and any breastfeeding of at least 3 months. We also adjusted for paternal age as well as child sex, child birthweight, full term status (≥37 weeks gestation or not), parity (nulliparous vs. not), and age of ITSEA assessment ( Table 1). As neither post-birth-breastfeeding-status nor child age at ITSEA would be expected to impact cordblood methylation patterns, these covariates were removed for assessment of associations of diet with differentially methylated imprinted gene DMR CpGs in cord-blood. All analyses were done using R version 3.4.2 (R Development Core Team. R Foundation for Statistical Computing, Vienna, Austria). For

associations of maternal diet (tertiles) with child behavior (tertiles), ordinal logistic regression was conducted using polr() from the MASS package (Venables and Ripley, 2002). Trend tests were conducted with ordinal logistic regression on tertiles of behavioral outcome while treating tertiles of maternal MDA exposure as a linear predictor (1 = least adherent, 3 = most adherent). In results, unless otherwise specified, trend p-values are not adjusted. For assessing the relationship between maternal dietary adherence and the percent methylation of CpGs (continuous) as well as the relationship between percent methylation of CpGs and child behavior (continuous), linear regression was conducted using lm() from the base R stats package. Correction for multiple comparisons was performed by (i) control of the family-wise error rate (FWER) at α= 0.05, and (ii) control of the false-discovery rate (FDR) at level 0.15. The use of FWER control is intended to highlight findings that pass the most rigorous false-positive standards, while the FDR procedure allows a proportion of false positives while maintaining a high proportion (0.85) of true discoveries. FWER control was achieved by applying the Bonferroni correction, augmented by permutation analysis for each set of tests. The permutation analyses used covariate-residualized of responses, permuted relative to the covariate-residualized predictors of interest. An alternate analysis used 10,000 permutations of DMR β values vs. maternal MDA, covariate-residualized and using the Child assessment age (m) 13.9 2.37 partial correlations as the test statistic. Control of the FWER over all DMRs was performed by using the minimum p-values over DMR as a statistic. In all instances, the permutation-adjusted p-values were only slightly more significant than the Bonferronicorrected values, and did not result in additional significant findings, and so the Bonferroni-corrected values are reported for simplicity. FDR q-values were computed using the Bioconductor qvalue package (Storey et al., 2015), version 2.10.0 using default settings and are, by construction, already multiple comparison corrected. Cohort Demographics Outcome Assessment Approximately half of the children assessed were male (53.5%) ( Table 1). The majority of mothers identified as white (42.5%), with 29.2% as black, 22.2% as Hispanic, and the remainder (6.2%) classified as other. Nearly all births were full-term (91.9%) and 43.7% of children were born to nulliparous women. Amongst mothers, 62.5% had some college, 11.3% smoked while pregnant, and 24.5% were classified as obese (BMI ≥ 30) at or near conception. The prevalence of any of gestational, type I or type II diabetes was 8.3% (Table 1). Nearly 70% of mothers reported breastfeeding for at least 3 months. Children were between the age of 12 and 24 months when assessed for behavior (average age, 13.9 months). The distributions and medians for assessed behavioral outcomes are summarized in Supplementary Figure 1. With the exception of social relatedness index and the Competency subscales (compliance, attention, mastery, imitation, empathy, and prosocial), lower scores indicate favorable behavior outcomes. Maternal Mediterranean Diet and Child Behavior Outcomes We examined associations between maternal MDA at or near conception with ITSEA assessed childhood behaviors in offspring 12 and 24 months of age. The distribution of maternal MDA is shown in Supplementary Figure 2. Maternal MDA scores were divided into tertiles with tertile 1 (referent) as the least adherent. Unadjusted and adjusted odds ratios for associations of maternal MDA and offspring behaviors assessed in the second year of age are displayed in Figure 1. Amongst Internalizing behaviors, when comparing offspring born to women with the lowest tertile of MDA, offspring of mothers in the highest tertile of MDA were less likely to score in a higher tertile of depression [OR (95% CI) = 0.28 (0.12, 0.64); p-trend = 0.002, Bonferroni-corrected p-trend = 0.041] and anxiety [0.42 (0.18, 0.97); p-trend = 0.041]. While not statistically significant, these associations persisted when comparing offspring of mothers in the middle tertile of MDA compared to those with mothers in the lowest level of MDA [depression: 0.59 (0.30, 1.16); anxiety: 0.66 (0.33, 1.31)]. The highest tertile of maternal MDA exposure was also associated with improved social relatedness behaviors [2.31 (1.04, 51.9); p-trend = 0.041] in offspring. Maternal MDA was not associated with Externalizing, Dysregulation or Competency behaviors (Figure 1). Child behavior can vary by sex; we examined if associations between maternal MDA and offspring behaviors did so (Supplementary Figure 3). In males, offspring born to mothers with the highest MDA scores were more likely to score in the next highest tertile of social relatedness behaviors [OR (95% CI) 3.20 (1.04, 10.25; p-trend = 0.032)]. Inverse associations of maternal MDA on depression, maladaptive, and ASD index behaviors were pronounced in females. Compared with daughters born to mothers with the lowest MDA scores, daughters born to mothers with MDA scores in the middle and high tertiles were much less likely to exhibit depression behaviors [middle tertile vs. low Maternal MDA and CpG Methylation in Offspring To test the hypothesis that associations of maternal MDA on offspring behavior are mediated through epigenetic mechanisms and taking into consideration work linking differential methylation control regions of imprinted genes with both offspring behavioral outcomes (Fuemmeler et al., 2016), and maternal diet (Rijlaarsdam et al., 2017), we examined the relationship of diet on CpG methylation of DMRs regulating nine imprinted genes. Building on prior research in our group on imprinted genes, we assessed the methylation status of 48 CpGs in the regulatory DMRs of nine imprinted genes for a subset of the cohort (n = 142 mother/child pairs). These include the DMRs of PLAGL1, H19, SGCE/PEG10, PEG3, NNAT, MEST, and two DMRs each for the DLK1/MEG3 and IGF2 domains. As mentioned in the introduction, since methylation of ICRs often vary by sex, we stratified on sex for these analyses and assessed whether maternal MDA was associated with the mean methylation percentage of CpGs at these loci with covariate-adjusted multiple linear regression (Figure 2). In females, maternal MDA was associated with an increase in mean methylation of the ICR of SGCD/PEG10 [β (95% CI) = 1.65 (0.52, 2.80); Bonferroni-corrected p-value = 0.048; Figure 2]. This translates to a 1.65% increase in mean methylation of the control region of IGF2 in offspring cord-blood for each unit increase in maternal MDA from zero to nine (Supplementary Figure 2). In males, after adjusting for a false discovery rate ( Figure 4). Further, while not statistically significant in females, increased maternal MDA was consistently associated across the interrogated DMRs with increased methylation of CpGs in the MEG3 DMR, the intergenic MEG3-IG DMR and the PLAGL1 DMR (Supplementary Table 1 and Supplementary Figure 4). CpG Methylation and Child Behavior For the significant associations of maternal MDA on offspring behavior (Figure 1) and maternal MDA on methylation of imprinted gene control regions (Figure 2), we examined associations of mean methylation on child behaviors (Supplementary Figure 5). In males, increased methylation of IGF2 was associated with increased ASD Composite Index behaviors and increased Atypical behavior, increased methylation of MEG3 was associated with increased Social Relatedness and decreased Maladaptive behaviors, increased methylation of SGCE/PEG10 was associated with increased Atypical behavior, while increased methylation of the PLAGL1 DMR was associated with increased Atypical and ASD Composite Index behaviors (Supplementary Figure 5). In females, increased methylation of the control region of SGCE/PEG10 was associated with decreased risk of depression, anxiety, and atypical behaviors. These estimates were unstable as evidenced by large confidence bounds and we were underpowered to conduct mediation analyses. In spite of this, we were able to identify multiple consistent associations of maternal MDA on child behavior, of maternal MDA on differential methylation of imprinted gene control regions, and of these DMRs on child behavior (Figure 3). DISCUSSION While much research has been conducted on the role of individual nutrients or foods on child outcomes, few studies have examined the potential role of the overall quality of maternal diet during pregnancy on child behavioral and other neurodevelopmental outcomes (Borge et al., 2017); the results so far are limited. In this study of a subset of the ethnically diverse NEST cohort, we observed a reduced risk of adverse child behaviors in relation to maternal diet FIGURE 2 | Maternal MDA and CpG Methylation. Females (light gray) and males (dark gray) were evaluated separately with linear regression for associations of maternal MDA on the average methylation status of the control region of 9 imprinted gene loci. Effect estimates for mean CpG methylation were adjusted for breastfeeding at least 3 months, age of child at behavioral assessment, maternal fiber intake, total calories, folate, education, diabetes, obesity, smoking, and age, as well as paternal age and child parity, premature birth, weight, race ( * FDR q < 0.15; n ranged from 51 to 75). FIGURE 3 | Sex Specific Maternal MDA and CpG Associations on Child Behavior. In females (orange), maternal MDA is associated with increased methylation in the control region of SGCE/PEG10 locus. In turn, this is associated with decreased odds of depression, anxiety, atypical, and Autism Spectrum Disorder (ASD) composite index behaviors, as is maternal MDA. In males (green), maternal MDA is associated with hypo-methylation of the control region of IGF2 and SGCE/PEG10 loci and hyper-methylation of the MEG3 control region which are, in turn, are consistently associated with maternal MDA associations on atypical and social relatedness behaviors. patterns around the time of conception. Maternal MDA was associated with child social relatedness behaviors, and inversely associated with the internalization subscales of depression and anxiety. Further, maternal MDA was inversely associated with atypical, maladaptive, and ASD behavioral patterns. To minimize maternal depression as a potential confounder of the association between maternal MDA and offspring depression (Figure 1), we included maternal post-natal depression in our adjusted model and estimates were materially unchanged (data not shown). In addition, although covariate proportions differed (except sex and child birthweight), between our cohort and the parent cohort (data not shown), estimates for our significant findings were materially unchanged between unadjusted and adjusted models (Figure 1). Our results suggest that during pregnancy, maternal adherence to the Mediterranean Diet, rich in legumes, vegetables, fish, and healthy fatty acids, and lower in red meat, may promote behavioral and emotional well-being in children. These findings are consistent with earlier studies reporting that maternal MDA, or to a dietary principal components pattern characterized as "healthy, " is beneficially associated with child behavioral and emotional problems (Jacka et al., 2013;Steenweg-de Graaff et al., 2014). In adults, adherence to this diet pattern is associated with better cognitive outcomes and decreased depression, suggesting a role of diet on cognition and emotional functioning (Crichton et al., 2013;Rienks et al., 2013;Trovato et al., 2014;Hardman et al., 2016;Aridi et al., 2017;Loughrey et al., 2017;Molendijk et al., 2018). Maternal adherence to this diet pattern during pregnancy has also been associated with protection from allergic disease in human offspring (Netting et al., 2014;Castro-Rodriguez and Garcia-Marcos, 2017). Given the rapid increase in the incidence of neurodevelopmental disorders in children (CDC, 2010;Visser et al., 2013Visser et al., , 2014, if replicated, our results relating maternal MDA during gestation with child neurodevelopmental outcomes may present new avenues for prevention. Consistent with our findings, laboratory studies suggest that maternal diet during pregnancy may affect offspring behavioral health. Numerous studies in rodents have found associations between maternal diet and offspring neurodevelopment and behavior, such as recent work associating high maternal caloric intake with increased anxiety in male rats (Balsevich et al., 2016). Offspring from female rats fed a high fat diet pregestational through lactation suffered cognitive deficits that were ameliorated by dams who also had access to exercise (Moser et al., 2017). Anxiety, cognition, and compulsive behaviors were altered in male mice weaned from mothers who had received gut microbiota from high fat diet mice (Bruce-Keller et al., 2017). Rather than reductions in fat or increases in energy intake, the Mediterranean diet is thought to beneficially influence health outcomes as a result of increases in intakes of micronutrients, including MUFAs, antioxidant vitamins, and phytochemicals (Ros et al., 2014). Most rodent research to date has focused on the high-fat paradigm when examining outcomes. Further, nearly all studies examine high-fat in the context of excess energy, further reducing the ability to elucidate whether observed changes in offspring are due to excess body weight, excess energy or fat intake, or some combination. Imprinted genes have parental allele specific silencing via epigenetics. These patterns are laid down during gametogenesis in imprinting control regions (Bartolomei and Ferguson-Smith, 2011). Germline ICRs are typically more

epigenetically stable throughout life and aberrant epigenetic marks for imprinted gene ICRs are associated with effects on metabolism, neurodevelopment, and growth. Because of our findings that maternal MDA is associated with changes in methylation patterns of imprinted genes, changes which can vary by sex, we did re-examine associations of maternal MDA on child behavior stratified by sex despite being underpowered to gain real insight. For a subset of our cohort, we were able to conduct targeted assessment of the CpGs in the control regions of imprinted genes (Supplementary Figure 4), providing evidence for novel associations of maternal diet with changes in methylation in imprinted ICRs of offspring. Our data suggest that high maternal MDA is associated with altered methylation of the MEG3 ICR independent of sex, and of the SGCE/PEG10, PLAGL1 and IGF2 ICRs in a sex-dependent manner. These changes in methylation, are in turn, consistently associated with offspring behaviors that are also associated with maternal MDA (Figure 3). We report here that maternal MDA is associated with increased methylation of the maternally expressed 3 (MEG3) ICR in males. Several lines of evidence suggest that increased methylation of the MEG3 ICR, regulating a known tumor suppressor (Zhou et al., 2012), is associated with decreased expression of MEG3, albeit with increased expression of the reciprocally imprinted DLK1 gene (Murphy et al., 2006). In genome-wide methylation arrays (Markunas et al., 2014), MEG3 DMR methylation was associated with maternal smoking and in human adults, DLK1 dysregulation has been associated with schizophrenia (Gardiner et al., 2012). In both mice and humans, hypo-methylation of the MEG3 control region has been associated with adverse neurobehavioral phenotypes (Kagami et al., 2015;Fuemmeler et al., 2016;Drobna et al., 2018). Microarray data from the mouse forebrain have shown differential spatial expression of the imprinted gene Gtl2 (aka Meg3) between the ventral and dorsal telencephalon of the mouse at a critical time point in the generation and migration of cortical neuronal populations (McLaughlin et al., 2006). In an effort to boost power, and given the sex independent associations of maternal MDA on the two control regions associated with MEG3, we examined if methylation at these loci mediated our findings of associations of maternal MDA on behavior, but were underpowered to draw conclusions (data not shown). We also found sex-dependent associations of maternal MDA and the methylation of the epsilon sarcoglycan and paternally expressed gene 10 (SGCE/PEG10) ICR. PEG10 is essential for proper placental development, and over-expression of PEG10 protein has been associated with multiple tumor phenotypes (Okabe et al., 2003;Li X. et al., 2016;Peng et al., 2016), while differential methylation of the SGCE/PEG10 ICR has also been associated with cancer (Sepulveda et al., 2016). Hypomethylation of the SGCE/PEG10 ICR has been associated with higher expression of PEG10 as well as maternal stress and depression, although mechanisms are still unclear (Liu et al., 2012). In males at age 1 and 3, increased methylation at PEG10 has been associated with greater weight for length ratios (Gonzalez-Nahm et al., 2018), and paternal obesity has been associated with decreased methylation of PEG10 in sperm . Less is known about Pleiomorphic adenoma-like protein 1 (PLAGL1 or ZAC1). PLAGL1 is a zinc finger protein associated with cell growth suppression and with transient neonatal diabetes mellitus (Kamiya et al., 2000;Varrault et al., 2001). We found maternal MDA associated with hypomethylation of the ICR in males. Interestingly, in the EDEN cohort, maternal alcohol and dietary vitamin B2 were associated with ZAC1 methylation while increased ZAC1 methylation was associated with estimated fetal weight, weight at birth and at 1 year of age (Azzi et al., 2014). Lastly, we report a strong association of maternal MDA with hypomethylation of insulin-like growth factor 2 (IGF2) in male offspring as well as an association of IGF2 hypomethylation with decreased atypical behavior (Figure 3). Hypomethylation of IGF2 and increased IGF2 protein have been associated with paternal obesity, as well as increased offspring obesity risk (Lawlor et al., 2012;Tobi et al., 2014;Dunford and Sangster, 2017). Decreased methylation of the IGF2 DMR has been associated with increased IGF2 transcripts (Murphy et al., 2012b) and protein (Hoyo et al., 2012). Differential methylation of IGF2 has been associated with both periconceptual exposure to extreme caloric restriction during the Dutch Famine (Tobi et al., 2009), and lower circulating folate concentrations (Hoyo et al., 2014). In the ALSPAC cohort, prenatal dietary patterns comprising high fat and high sugar were associated with lower IGF2 methylation and an earlier onset of ADHD behaviors (Rijlaarsdam et al., 2017). Dunford and Sangster provide a more thorough review of parental nutrition and offspring epigenetics in the context of metabolic syndrome risk (Dunford and Sangster, 2017). Additional studies are needed to replicate these findings, but together, these data support mechanistic relationships between high adherence to Mediterranean diet and shifts in methylation of regulatory sequences of imprinted genes in offspring. We adjusted for a wide array of covariates, including factors such as maternal education, BMI, maternal energy intake, and breastfeeding. Nonetheless, an important limitation of this research is uncertainty of the extent to which associations between child neurodevelopment and maternal diet quality during pregnancy may reflect differences in caregiving which are difficult to measure, or be influenced by confounding by factors such as maternal IQ (Der et al., 2006;Horta et al., 2015). Our study was limited to child outcomes in infancy; the persistence of effects in later life, and whether lasting effects may depend on childhood diet, remains uncertain. Though results are promising, more research is needed to elucidate whether this relationship is causal, and to identify the pathways through which maternal adherence to Mediterranean Diet may affect child behavior. Our study also had some limitations including small sample sizes for stratified analyses. Despite this, all of the overall findings of associations of maternal MDA on offspring behavior and maternal MDA on offspring mean DMR methylation have an FDR < 0.15, suggesting these associations warrant further studies to replicate these findings. A lack of dietary data in infants to evaluate the extent to which neurodevelopmental outcomes may have been driven by postnatal dietary exposures other than breastfeeding is a limiting factor of this study, although dietary variability should be somewhat homogeneous at the age of behavioral assessment [12-24 months (mean = 13.9 months)]. Ideally, one would have whole-genome methylation data to identify the strongest associations of diet with CpG methylation. Our study has only focused on the effects at imprinted DMRs; there are likely effects at many other regions throughout the genome that more comprehensive epigenetic analyses would reveal. Although maternal MDA was associated with half of the imprinted gene regions evaluated, many of these imprinted gene control regions are related to major cell proliferation pathways such as TGF-β and TP53 which would impact vast downstream signaling pathways, so this may be expected. Further, we have not examined any potential effects on other epigenetic regulatory elements, including histone modifications or the expression and actions of non-coding RNAs. Additional epigenome-wide studies are needed to both replicate and clarify these findings and to characterize an epigenomic signature associated with maternal diet. Recall bias in dietary recall questionnaires is often an issue. To minimize recall bias, mothers in our study were asked to fill out a FFQ about the foods they ate near conception inside of a dozen weeks. CONCLUSION We report novel significant associations of maternal periconceptional MDA both with positive neurodevelopmental phenotypes in offspring as well as associations of maternal MDA on differential methylation of CpGs in the control regions of imprinted genes. If confirmed in other studies, these findings may pave the way for early identification of adverse behavior risk in offspring and for tailored interventions. DATA AVAILABILITY The datasets for this manuscript are not publicly available because: Human subjects were used with consent in this study, but public release of data is not approved under the IRB. Requests to access the datasets should be directed to CH for approval. ETHICS STATEMENT This study was carried out in accordance with the recommendations of the Duke University Health System Institutional Review Board. It conforms with special protections for pregnant women described in 45 CFR 46, Subpart B. The protocol was approved by the Duke University Health System Institutional Review Board. All subjects gave written informed consent in accordance with the Declaration of Helsinki. AUTHOR CONTRIBUTIONS JH conceived and conducted the analyses, generated the figures and wrote the manuscript. MM, BF, FW, and CH supported the research and assisted with subject matter expertise, analysis plans, and manuscript writing and editing. RM, SG-N, ZH, JD, and SM each assisted with manuscript preparation, analysis support, and editing. FUNDING The research was supported by National Institute of Environmental Health Sciences of the National Institutes of Health (P01ES022831, R21ES014947, R01ES016772, R01HD084487, and P30ES025128) and by the U.S. Environmental Protection Agency (RD-83543701). Additional support was provided by the National Center for Advancing Translational Sciences (UL1TR001117). ACKNOWLEDGMENTS We thank Carole Grenier and her excellent skills and work with receipt and processing of the cord-blood specimens and pyrosequencing. Multiple medication use in older patients in post-acute transitional care: a prospective cohort study Background Older adults with a range of comorbidities are often prescribed multiple medications, which may impact on their function and cognition and increase the potential for drug interactions and adverse events. Aims This study investigated the extent of polypharmacy and potentially inappropriate medications in patients receiving post-discharge transitional home care and explored the associations of polypharmacy with patient characteristics, functional outcomes, and frailty. Methods A prospective observational study was conducted of 351 patients discharged home from hospital with support from six Transition Care Program (TCP) sites in two states of Australia. A comprehensive geriatric assessment was conducted at TCP admission and discharge using the interRAI Home Care assessment tool, with frailty measured using an index of 57 accumulated deficits. Medications from hospital discharge summaries were coded using the World Health Organization Anatomical Therapeutic Chemical Classification System. Results Polypharmacy (5–9 drugs) was observed in 46.7% and hyperpolypharmacy (≥10 drugs) in 39.2% of patients. Increasing numbers of medications were associated with greater number of comorbid conditions, a higher prevalence of diabetes mellitus, coronary heart disease, chronic obstructive pulmonary disease, dizziness, and dyspnea and increased frailty. At discharge from the program, the non-polypharmacy group (<5 drugs) had improved outcomes in Activities of Daily Living, Instrumental Activities of Daily Living and fewer falls, which was mediated because of lower levels of frailty. The commonest drugs were analgesics (56.8%) and antiulcer drugs (52.7%). The commonest potentially inappropriate medications were tertiary tricyclic antidepressants. Conclusion Polypharmacy is common in older patients discharged from hospital. It is associated with frailty, falls, and poor functional outcomes. Efforts should be made to encourage regular medication reviews and rationalization of medications as part of discharge planning. Whether careful deprescribing improves outcomes in frail patients should be the focus of randomized trials. Introduction Background Older adults with a range of comorbidities are often prescribed multiple medications, some of which may impact on their function and cognition, and many have a potential for drug interactions. 1 Studies showing evidence of benefit from pharmacotherapy have mostly been conducted in younger patients, and it is unclear how this translates to frail older patients. These patients are often excluded from drug trials; yet they are the largest consumers of medications. 1 Several studies have found current use of five or more drugs in well over a quarter of older community dwelling adults, [2][3][4] with higher prevalence in frail older populations and in hospitalized patients. 5 The assessment of frailty using various methods, including the Frailty Index, is being incorporated in recent studies of older adults and provides an insight into their accumulated deficits and reduced reserve. 7,8 The increased number of comorbidities requiring medications makes these patients prone to polypharmacy, yet their frailty status, together with the pharmacokinetic and pharmacodynamic changes that occur with aging, places them at risk of adverse events. The risks of polypharmacy include non-adherence, adverse drug reactions, drug-drug interactions, falls, fractures, poor nutrition, and mortality, [9][10][11][12][13][14] as well as increased exposure to potentially inappropriate medications (PIMs). 5,15 However, few studies have reported on the association of polypharmacy with functional outcomes in older patients. Aims The aims of the study were to 1. Explore the extent of polypharmacy in a cohort of older patients discharged from hospital to a home care program; 2. Assess the

relationship between polypharmacy and patient characteristics, functional outcomes, and frailty; and 3. Describe the prevalence of the most common medications in this cohort, with particular emphasis on PIMs. study design, setting, and participants A prospective observational cohort study of older persons discharged from hospital to a community-based Transition Care Program was conducted at six sites in two Australian states, Queensland and South Australia. The Transition Care Program (TCP) is designed to facilitate transitions from hospital to home for older people (aged 70 years and over or 50 years and over for the indigenous population), offering those with high care needs additional support during convalescence. 16 The program is therapy focused, providing a package of services which includes home help and personal care, physiotherapy and occupational therapy, nursing care, and case management over a maximum period of 12 weeks (average 7 weeks) post-discharge from hospital. 16 The provision of primary medical care to a Transition Care recipient is undertaken by their general practitioner. 16 Consecutive patients entering the TCP during the period from November 2009 to September 2010, who gave informed consent to participate, were eligible for the study. Recruitment details for the study, originally designed to examine the functional recovery trajectories of patients with high care needs, have previously been published. 17 Ethics approval was given by the University of Queensland Human Research Ethics Committee (HREC) as well as HRECs responsible for governance at each of the TCP sites. Data collection A comprehensive geriatric assessment using the interRAI Home Care instrument was conducted at TCP admission and discharge. The interRAI instruments comprise a suite of assessment tools to support assessment and care planning of persons with chronic illness, frailty, and disability across care settings, 18 with substantial reliability on core items in common. 19 The interRAI Home Care assessment collects data on multiple domains including sociodemographics, medical conditions, medications, physical and mental function, nutrition, and symptoms and syndromes such as mood, behavior, and continence. A number of scales embedded in the interRAI instruments combine single items belonging to a domain, such as activities of daily living (ADL), instrumental ADL (IADL), and cognition, which can be used to describe the presence and extent of deficits in that domain. 17,20,21 Trained assessors gathered data from multiple sources including from the patient, carers, medical and allied health staff, and hospital records. Medications from hospital discharge summaries were coded by pharmacy students using the World Health Organization Anatomical Therapeutic Chemical (ATC) Classification System and reviewed by a pharmacist and a geriatrician. Measures Medication exposure There is no universally accepted definition of polypharmacy in the literature. Some studies define it as use of five or more medications. [2][3][4]11 This has been supported in a recent study investigating polypharmacy cutoff points and risks of adverse outcome. 22 Moreover, recent studies have defined the use of ten or more medications as excessive polypharmacy 10,23 or hyperpolypharmacy. 24 Inclusion of over-the-counter medications and medications not consumed on a regular basis is also variable. In our study, polypharmacy status was categorized into three groupsnon-polypharmacy (0-4 drugs), polypharmacy (5-9 drugs), and hyperpolypharmacy (10 drugs) -based on regular medications. Drugs, vitamins, and mineral supplements administered on a regular basis through any recognized drug-delivery method were included in the analysis. Supplements without ATC codes, such as cranberry juice and primrose oil, were excluded. The American Geriatrics Society 2012 Beers Criteria was used to identify PIMs with a recommendation to avoid, regardless of patients' comorbidities. We included as PIMs those medications where the recommendation to avoid was strong and the quality of evidence was classified as moderate or high, also taking into account exposure to drugs above recommended maximum daily dose. 25 Table S1 lists the PIMs meeting these criteria. Frailty The frailty index (FI) was calculated using a well described methodology, 26 based on accumulated health deficits such as symptoms, signs, disabilities, and diseases measured in the interRAI Home Care assessment. Disability in ADL and IADL, impairments in general cognition and mobility, number of comorbidities, incontinence, and depressed mood were included as deficits. For each patient, deficits were added and divided by the total counted, here 57, to calculate an individual index score. Polypharmacy was excluded from the deficit count. The higher the score, the greater the number of deficits, and the more likely the patient is to be frail. In community-dwelling older people, 0.25 has been proposed as the cutoff between "fit" and "frail," with scores of 0.40 associated with dependence on others for activities of daily living. 27 Analysis To describe characteristics across polypharmacy groups, comparison of means (analysis of variance) or medians (Kruskal-Wallis test) for continuous variables was used, depending on distribution of the data. For categorical variables, chi-square or Fisher's exact test (where cell numbers are less than five) was performed. An exploratory analysis using logistic regression models tested the association between polypharmacy, frailty status, and functional outcomes. For the purpose of interpreting odds ratios, FI was multiplied by 10 so that the per-unit change was 0.1. 8 Patients with missing data were excluded from the relevant analysis, and percentages were reported as proportions of patients with available data. Significance level was set at P-value of 0.05. The SPSS IBM version 22 was used for analysis. Results Of the 351 TCP clients enrolled in the study, four cases had missing medication data. The remaining 347 cases were included in the analysis. The mean age (standard deviation [SD]) was 78.9 (±8.8) years, and 65.7% were females. The majority of patients discharged to the TCP needed ongoing support after hospitalization for orthopedic conditions (50.7%), including fractures (37.5%), medical conditions resulting in deconditioning (23.6%), and stroke (14.6%). The median length of stay in the TCP was 54 days (interquartile range 37-73 days). The number of regular medications taken ranged from 0 to 24, with a mean (SD) of 8.5 (±3.6). For "as needed" pro re nata (PRN) medications, the mean (SD) was 0.8 (±1.1). Only 14.1% of patients took 5 regular medications (nonpolypharmacy). Polypharmacy (5-9 drugs) was observed in 46.7% and hyperpolypharmacy (10 drugs) in 39.2%. The majority in the hyperpolypharmacy group (n=131; 96.3%) were taking between 10 and 15 regular medications, with five taking more than 15 regular medications. Table 1 shows the characteristics of patients according to polypharmacy categories at admission to the TCP. Patients with polypharmacy and hyperpolypharmacy had more comorbidities than the non-polypharmacy group and were more likely to have diabetes mellitus, coronary heart disease, chronic obstructive pulmonary disease (COPD), or depression. They were also more likely to have symptoms of pain, dizziness, and dyspnea. There was no significant association between polypharmacy categories and stroke, congestive heart disease, Parkinson's disease, or cancer. Considering frailty status and geriatric syndromes (including history of falls in the previous 90 days, impaired cognition, dependence in basic and instrumental ADL, and bladder incontinence), only the FI had a significant association with polypharmacy. Table 2 shows outcomes at discharge from the TCP according to polypharmacy status. The majority of patients continued living in the community (85.6%), 12.4% were readmitted to hospital, 0.9% were discharged to residential aged care facilities (RACF), and 1.2% died. Patients in the polypharmacy and hyperpolypharmacy groups were more likely than the non-polypharmacy group to fail to improve in ADL and IADL and were more likely to fall over the duration of the TCP. Multivariate models of functional outcomes (failure to improve ADL or IADL or falls over the TCP), with FI and polypharmacy groups as covariates, show that frailty status mediates the effects of polypharmacy. The odds ratios of ADL and IADL functional decline and falls for a 0.1 increase in FI are shown in Table 3. Table 4 shows the most common drug categories by polypharmacy group. The most commonly used drugs were analgesics (56.8%). Non-opioid drugs were prescribed more Anti-ulcer drugs (52.7%), statins (44.1%), aspirin, and anti-aggregates (43.2%) followed. Cardiovascular drugs were also commonly used. Beta blockers and angiotensin converting enzyme inhibitors were each prescribed in about a third of patients, while diuretics, calcium channel blockers, and angiotensin receptor blockers were each prescribed in about a quarter of the patients. Vitamin D and analogues 1457 Polypharmacy in older post-acute care patients were prescribed in 27.1%, while anti-resorptives and calcium were taken by 22.5% and 24.5%, respectively. A high proportion was prescribed antidepressants (30.8%) and laxatives (28%). Only nine patients were on antipsychotics and four patients on anti-dementia drugs. In all but a few drug categories (anticoagulants, oral hypoglycemics, anti-emetics, anti-Parkinson, antipsychotics, and anti-dementia drugs) the prevalence of each drug class increased significantly across the polypharmacy groups, with the hyperpolypharmacy group having the highest prevalence. The number of TCP patients taking at least one PIM was 41 (11.8%), with two persons taking two PIMs. Of Fisher's exact test). The commonest PIMs prescribed were tertiary tricyclic antidepressants (9.5%), particularly amitriptyline. Digoxin at a dose of 125 µg was prescribed in less than 2% of patients. Dipyridamol, promethazine, glibenclamide, and oral estrogens were each prescribed in only one or two patients. None of the patients was prescribed potent non-steroidal anti-inflammatory drugs which increase the risk of gastrointestinal bleeding and peptic ulceration. There were no patients on barbiturates or the antiparkinsonian agent, benztropine. Discussion The findings of this study showed that polypharmacy was significantly associated with frailty and poor functional outcomes. However, multivariate models of functional outcomes (failure to improve ADL or IADL or falls over the TCP), with FI and polypharmacy groups as covariates, show that frailty status mediates the effects of polypharmacy. This accords with previous findings which indicate that older adults who are frail are more likely to be exposed to multiple medications associated with increases in number of comorbidities. Conversely, multiple medications may exacerbate frailty. 24 While the association of polypharmacy with frailty and adverse outcomes has been shown in studies of community-dwelling older adults, 22,24,28 there have been few studies which have shown this relationship in the post-acute care setting. The majority of patients in our study (86%) were prescribed five or more medications per day. The mean number of drugs of 8.5 is higher than that reported in other studies of nursing home patients, community-dwellers, and day hospital patients, which report values of 3.7-7.9. 4,10,[29][30][31] The definition of polypharmacy and inclusion of vitamins, minerals, and over-the-counter medications was variable in these other studies, making comparison difficult. Similar to previous reports, 23,32 prevalence of diabetes mellitus, coronary heart disease, COPD, and depression were lower in the non-polypharmacy group, as were symptoms of dizziness and dyspnea. In contrast to a previous study, 23 measures of ADL, IADL, and cognition were not associated with polypharmacy at admission to the TCP. However, better functional outcomes in ADL and IADL were achieved with TCP rehabilitation for those on fewer medications. This was most likely because of their lower levels of frailty, which is a predictor of functional gain in rehabilitation patients. 8 Those with fewer medications were less likely to fall over the duration of the TCP, which is consistent with studies showing a relationship between polypharmacy and risk of falls. 33 A strong association between cognitive impairment and reduced rates of excessive polypharmacy has recently been described in nursing-home residents. 23 In contrast, our study did not find such an association, most likely due to the small number of patients with severe cognitive impairment. Analgesics were the most commonly prescribed class of medications, which may reflect the fact that the majority of patients had been hospitalized with fractures or for orthopedic procedures. While fractures were the commonest reason for hospitalization in patients admitted to the TCP, this was not mirrored by the use of anti-resorptives and vitamin D and analogues, which was lower than expected, given the importance of these medications in the prevention of osteoporotic fractures. 34,35 Though analgesic use was high, no patients were prescribed potent non-steroidal anti-inflamatory drugs, which are listed as PIMs under the Beers Criteria, 25 due to greater propensity for gastrointestinal side effects. The majority of PIMs that met Beers criteria were not prescribed for any of the patients in our study. The difficulties comparing our study with other published polypharmacy studies, due to different patient selection and polypharmacy definitions, are acknowledged. Our study has prospectively collected data on functional outcomes in a cohort of patients often excluded from clinical studies -frail elderly patients residing in the community but meeting criteria

for residential aged care. This is an important group of patients in which interventions can delay or avoid institutionalization. 17 Considerations should be given to enabling regular medication reviews and rationalization in patients enrolled in community rehabilitation and Transition Care Programs, by encouraging regular pharmacist and medical input. These interventions have been shown to improve appropriate prescribing and reduce drug-related adverse events, though results on number of medications prescribed have been variable. 6,[36][37][38][39] The strengths of our study are that the cohort is characteristic of older people eligible for post-discharge home-based care and representative of TCP recipients in particular, having been recruited across multiple sites in both rural and metropolitan communities. Few studies have explored associations of polypharmacy with functional outcomes after a period of longitudinal follow up. A study limitation is that the medication lists were documented by the interRAI assessors who transcribed or photocopied the patients' drug charts from hospital discharge summaries. It is acknowledged that this method of collecting medication data is not the current gold standard. To achieve complete medication reconciliation, multiple sources of information (including patient interview, general practitioner's letter, and dispensing history from the pharmacy) should be accessed. A further limitation is that the indications for each medication prescribed could not be determined. Conclusion Polypharmacy is common in older patients discharged from hospital to home-based care. It is associated with frailty, falls, and poor functional outcomes. Efforts should be made to encourage regular medication reviews and rationalization of medications by pharmacists and geriatricians in these frail patients with reduced physiological reserves. Use of medications associated with functional decline such as benzodiazepines and anticholinergics as well as other PIMs should be minimized. Comments regarding "A case of perforating injury of eyeball and traumatic cataract caused by acupuncture" Cite this article as: Kim ES, Yu SY, Han SB, Kim M. Decompression retinopathy after intravitreal bevacizumab and anterior chamber paracentesis in a patient with neovascular glaucoma. Indian J Ophthalmol 2016;64:861-3. This is an open access article distributed under the terms of the Creative Commons Attribution‐NonCommercial‐ShareAlike 3.0 License, which allows others to remix, tweak, and build upon the work non‐commercially, as long as the author is credited and the new creations are licensed under the identical terms. prelaminar, laminar, and retrolaminar optic nerve blood flow in monkeys. Invest Ophthalmol Vis Sci 1979;18:1030‐42. Comments regarding "A case of perforating injury of eyeball and traumatic cataract caused by acupuncture" Sir, We read with interest the article written by Shuang and Yichun (2016) entitled, A Case of Perforating Injury of Eyeball and Traumatic Cataract by Acupuncture. The article does trigger an alarm in the standardization of clinical acupuncture, but we are still confused about the case. Aiming to avoid similar accidents in the clinical practice, we decide to seriously analyze this case. Since the authors are ophthalmologists instead of acupuncturists, they do not give a detailed account of the acupuncture therapy in this article. Judging from the text and pictures, it is discovered that the perforated cornea and iris are the major cause of traumatic cataract and subsequently patient's vision loss. [1] However, there are several uncertainties about the case: First, we are curious about the selection of acupoints and its selecting principle, as well as the acupuncture manipulations adopted for cerebral infarction; second, as shown in the picture, what we find is only two holes with a diameter of 3 mm, which does not conform to the use of acupuncture needles with a diameter of 0.25 mm in clinical practice; furthermore, based on the position of the holes, it is puzzling that how we could reach the spot in acupuncture therapy. Why did the performer take no notice of the acupuncture dangers or patient's pain or bleeding at that time? If the performer did apply eye acupuncture therapy, [1] the extraocular horizontal insertion, [2] instead of needling into eyes, would be recommended. Due to the risk in needling acupoints around eyes, we are especially careful, and few people would choose them to treat cerebral infarction. We would like to clarify why the performer chooses that therapy. Due to this article's essence on the accident of acupuncture, we come to believe that a large number of acupuncturists would be interested in the article. Therefore, it is suggested that the authors should give more detailed description of the patient's history of present illness, which would objectively influence the acupuncture. were not clear. Second, I think you might misunderstand the size of the lesion. We cannot find any hole with the diameter of 3 mm, but only a 3 mm long self-sealed corneal injury could be observed, which we guess might be injured during pulling out the needle. Moreover, the diameter of the iris hole was approximately 0.25 mm, which we can find clearly from the figure. Third, during the acupuncture therapy, the patient did not feel seriously painful. He mistakenly reckoned the pain come from the acupuncture therapy (the stimulation of acupoints) rather than the corneal injury. Hence, he endured the slightly pain than usual without bleeding until the end. All the histories of the accident were obtained from his relatives. In China, acupuncture therapy has been widely used in ophthalmic clinic. The figure shows several acupoints around eyes [ Fig. 1]. Since closing to the eyeball, the periocular acupuncture may hurt the eyes in any careless situation. This article aims to attract the importance for the acupuncture therapy around eye acupoints. Moreover, I think acupuncturists know more about the selection of acupoints and its selecting principle for eye-related acupoints diseases. Sorry for the incomplete on this issue. 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. Authors' reply Sir, Thank you for your attention and comments on my case report. [1] Restricted to the expertise and knowledge, my article was mainly focusing on the injury of his eyeball. Thus, the details of descriptions for acupuncture treatment seem to be neglected. And I hope the following steps could deal with parts of your doubts. First, for the medical history of his acupuncture treatment, I have talked with his relatives. The cerebral infarction led to his hemiplegia and hemianopsia. They have chosen the acupoints around the eye for his hemianopsia treatment for 1 year and have got slightly improvement. Unfortunately, we cannot contact with the acupuncture therapist directly, and the exact acupoints and protocol for this medical accident New Insights Into the Persistent Effects of Acute Exposure to AFB1 on Rat Liver Aflatoxin B1 (AFB1) has mutagenesis, carcinogenesis and teratogenesis effects and mainly found in food crops and their processed foods. AFB1 exposure can cause acute or chronic liver poisoning, but there were few studies on the persistent effects of acute AFB1 exposure on the liver. In this study, rat liver injury models were established 2 and 7 days after single exposure to high and low doses of AFB1. The persistent effects of AFB1 single acute exposure (ASAE) on rat liver were analyzed from the phenotypic and genetic levels. The results showed that compared with the control group, liver function indexes, MDA content in liver and the number of apoptotic hepatocytes in model groups increased to the highest on the 2nd day after ASAE (p < 0.001). However, the changes of liver coefficient were most significant on the 7th day after ASAE (p < 0.01). The results of liver pathology showed that the liver injury was not alleviated and the activities of antioxidant enzymes GSH-Px and SOD were the lowest on the 7th day (p < 0.001). RNA-Seq results indicated that there were 236, 33, 679, and 78 significantly differentially expressed genes (DEGs) in the model groups (LA-2d, LA-7d, HA-2d, HA-7d) compared with the control group. Among them, the Gtse1 gene related to the proliferation, differentiation and metastasis of liver cancer cells, the Lama5 and Fabp4 gene related to the inflammatory response were significantly DEGs in the four model groups, and the differential expression of the immune system-related Bcl6 gene increased with the prolonged observation time after ASAE. In conclusion, ASAE can cause persistent liver damage in rats. The persistently affected genes Lama5, Gtse1, Fabp4, and Bcl6 possess the potential to be therapeutic targets for liver disease induced by AFB1. INTRODUCTION Aflatoxins (AFs) is a general term for a class of highly toxic secondary metabolites produced by Aspergillus parasiticus and Aspergillus flavus. Crops such as grain and oilseed are easily contaminated by AFs during planting, harvesting, storage and processing. In addition, its physical and chemical properties are stable. Therefore, there is a great risk of harming animal and human health through the food chain (Rushing and Selim, 2019). At present, more than 20 kinds of AF have been found, among which AFB 1 is the most toxic and most common, and the target organ of its toxic effect is mainly the liver (Hussein and Brasel, 2001). A warm and humid environment is easy to cause a large amount of AFB 1 pollution (Abrar et al., 2013), so acute or chronic poisoning is often caused by dietary exposure to AFB 1 . Table 1 lists the effects of different doses of AFB 1 and modeling time on the liver in some studies. It can be seen from the table that AFB 1 can cause acute or chronic poisoning in the liver. But viewed from the model, there are few studies on the lasting effects of AFB 1 exposure on the liver. AFB 1 is metabolized by the P450 enzyme system in the liver into the ultimate carcinogen aflatoxin B 1 -8, 9-epoxide (AFBO), which has strong oxidation ability and can induce the body to produce a large amount of reactive oxygen species (ROS) (Guengerich et al., 1998). AFBO can covalently bind to DNA guanine N7 to form adduct AFB 1 -N7-guanine (Lin et al., 2014), and oxidize guanine to produce DNA damage marker 8-OHdeoxyguanine (Bailey et al., 1996;Wang and Groopman, 1999). In addition, excessive ROS will break the dynamic balance of the redox system and cause oxidative stress reaction. It also attacks cells, causing cells damage and inducing apoptosis. Therefore, this study explored the liver injury in model rats by analyzing the changes of phenotypic indicators related to oxidative stress response. RNA-seq is a widely used parallel sequencing method. Through this technology, the complete transcript of the test sample and its expression level can be found (Wang et al., 2009). From the obtained data, we can not only know individual genes but also understand the functional pathways of gene enrichment through the analysis of functional database. RNAseq data has been used to analyze the underlying mechanism of carotid atherosclerosis (CAS). Differential and functional enrichment analysis of sequencing data from CAS patients and control group showed that inflammation and immune response were the potential pathogenesis of CAS, and significant DEGs CCR5, NPY, and NPY5R were potential therapeutic targets of CAS (Li et al., 2021). The method is also often used to study cancer therapeutic targets. Bioinformatics analysis of DEGs in ovarian cancer and normal tissue showed that they were mainly concentrated in metabolic, cell cycle regulation and antibiotic biosynthesis pathways. Meanwhile, 10 central genes including CCNB2, TYMS and KIF11 were analyzed . Sorafenib is the only FDA approved oral multi-target tyrosine kinase inhibitor that promotes apoptosis, reduces angiogenesis and inhibits tumor cell proliferation in the treatment of liver cancer. However, drug resistance and side effects of varying degrees were found in the process of drug use (Trojan and Zeuzem, 2013;Rota Caremoli and Labianca, 2014). Therefore, more potential genes for targeted treatment of liver diseases can be searched through gene sequencing technology. This paper explores the sustained effects of AFB 1 on the rat liver by establishing a model of ASAE. Based on this model, the RNA-Seq data were analyzed to find potential target genes for the treatment of liver diseases caused by AFB 1 , providing a new research basis for the development of targeted therapy for liver diseases. Animals A total of

36 male Wistar rats were purchased from Shanghai SLAC Laboratory Animal Co., Ltd. Body weight (BW): 180-200 g. All rats were housed in an environmentally controlled room at 22-24 • C with a relative humidity of 50-55% under a cycle of 12 h each light/dark. Food and water were available ad libitum. All animal procedures were performed in accordance with the Guidelines for Care and Use of Laboratory Animals of Shanghai University of Traditional Chinese Medicine and approved by the Animal Ethics Committee of Animal Center of China and Shanghai University of Traditional Chinese Medicine (Ethics number: PZSHUTCM210115014). Experimental Design After one week of adaptive feeding, 36 rats were randomly divided into 3 groups (12 in each group): blank group (CK): normal breeding; AFB 1 high-dose (HA) group: AFB 1 was intraperitoneal injection with a dose of 2 mg/kg BW, only once; AFB 1 low-dose (LA) group: AFB 1 was intraperitoneal injection at a dose of 1 mg/kg BW, only once (AFB 1 soluble in 4% DMSO). The experimental doses of AFB 1 were determined according to references and results of pre-experiment (Monmeesil et al., 2019;Deng et al., 2020). The above dose groups were determined based on references and preliminary experiments. The concentration of DMSO did not affect the normal growth of rats (Chen et al., 2020). Blood samples were collected from ocular venous plexus at the 1st, 4th and 7th day after ASAE and serum were prepared. On the 2nd day after ASAE, 6 rats in each group were randomly selected to weigh their final body weight. The rats were anesthetized with 10% chloral hydrate (350 mg/kg). After anesthesia, the rats were dissected, the blood of abdominal aorta was taken, the liver was separated, washed with normal saline, sucked dry and weighed. Liver tissues of an appropriate size were taken from the same part of the liver of each rat and soaked in 4% paraformaldehyde solution, which was fully fixed for histopathological study, and the rest tissues were frozen in liquid nitrogen and stored in a cryogenic refrigerator at −80 • C for use. The remaining 6 FIGURE 1 | Changes of (A) body weight, (B) liver index, and (C-G) liver pathological morphology with time after ASAE in rats (HE, × 100). 0: weight before ASAE; 1 and 4: weight on 1st and 4thr days after ASAE; 2 and 7: weight and index on 2nd and 7th days after ASAE. LA-2d and HA-2d: rats were sacrificed on the 2nd day after ASAE. LA-7d and HA-7d: rats were sacrificed on the 7th day after ASAE. The data are expressed as mean ± standard deviation. Different letters representing significant difference (p < 0.05) (the same below). rats in each group were treated with the same method on the 7th day after ASAE. Hepatic Pathological Examination After being fixed in 4% paraformaldehyde for 24 h, the liver was removed and dehydrated in graded alcohol series, transparent in xylene. Tissues were embedded in paraffin, and paraffin blocks were then cut to 4 µm thickness using a microtome (RM2016, Shanghai Leica Instrument Co., Ltd., Germany). These sections were deparaffinized, rehydrated and stained using hematoxylineosin (HE) and then analyzed by optical microscopy (Nikon Eclipse E100, Japan) to evaluate histopathological changes. Observe the morphologies of the liver and spleen tissues were observed at 100 ×. Determination of Biochemical Indices in Serum and Liver The serum was separated and measured the levels of alanine aminotransferase (ALT), alkaline phosphatase (ALP), aspartate aminotransferase (AST) and total bilirubin (TBIL) by automated chemistry analyzer (ADVIA 2120i, Hitachi, Ltd., Japan). The activities of SOD and GSH-Px, as well as the level of MDA were examined using commercial assay kits. Determination of Hepatocyte Apoptosis Rate The apoptosis rate of rat hepatocytes in the model groups were detected by TUNEL method. Paraffin sections were taken in sequence for deparaffinize and rehydrate, antigen retrieval, permeabilization, equilibration at room temperature, TUNEL reaction, BSA blocking, addition of double antibodies, DAPI counterstain in nucleus and then sealed for microscopic examination. Stained nuclei were blue under UV excitation and positive apoptotic cells were green. The number of apoptotic hepatocytes in the total field of view was counted as well as the total number of hepatocytes. Calculation formula: Hepatocyte apoptosis rate = (Number of apoptotic hepatocytes/Total number of hepatocytes) × 100%. Illumina RNA Sequencing Total RNA in the rats' liver was extracted using the MagMAX TM mir Vana TM Total RNA Isolation Kit following the specification (Thermo Scientific TM KingFisher TM Flex TM , Finland). Total RNA of each sample was quantified and qualified by Agilent 2100/2200 Bioanalyzer (Agilent Technologies, Palo Alto, CA, United States), NanoDrop (Thermo Fisher Scientific Inc.). 1 µg total RNA was used for following library preparation. Illumina RNA sequencing experiments used three parallel samples. The sequencing library construction and Illumina sequencing were conducted at GENEWIZ. Real-Time Quantitative PCR Analysis To verify the reliability of the expression profiles observed in the RNA-Seq data, four genes were selected for real-time quantitative PCR (Q-PCR) analysis with the same experimental samples as in the RNA-Seq experiments. Total RNA was extracted using a King Fisher Flex automated nucleic acid extractor and accompanying kit, and reverse transcribed after RNA electrophoresis to determine RNA quality. β-Actin was used as an internal reference and primers designed by Primer 5 are shown in Table 2, and relative expression was calculated according to the 2 − ct relative quantification formula. Statistical Analysis Results were analyzed using GraphPad Prism 8.0 (GraphPad Software, La Jolla, California), SPSS 25 (IBM, NY, United States) and Origin 8.0 software (Origin Inc., Northampton, MA, United States). Analysis of variance (ANOVA) was performed followed by Tukey's test with a confidence interval of 95% (p ≤ 0.05). Body Weight and Liver Index Changes in Rats During the experiment, the BW of the rats was weighed before ASAE and on the 1st, 2nd, 4th and 7th after ASAE ( Figure 1A). The BW of rats in the CK group showed an upward trend in the whole experimental cycle, while that in the model groups decreased and then increased. The organ index is one of the main indicators of the biological properties of animals (Zou et al., 2010). Figure 1B shows that there was no significant difference in liver index between rats treated on the 2nd day after ASAE and the CK group. Compared on the 7th day after ASAE with the CK group, the liver coefficient of the model groups was significantly higher (p < 0.05), and the HA group was significantly different from the CK group (p < 0.001). Histopathological Changes of Liver Tissue in Rats Representative hepatic histopathological examination results of each group were shown in Figures 1C-G. The rats in the CK group had a normal liver lobular architecture and cell structure. The nucleolus was clear, homogenous cytoplasm ( Figure 1C). Livers illustrated different degrees of pathological changes in the model groups. In LA-2d group, the liver exhibited moderate spotty necrosis and degeneration ( Figure 1D). In LA-7d group, moderate spotty necrosis, vacuolar degeneration, and mild inflammatory cell infiltration were observed in the portal area ( Figure 1E). Hepatocyte lytic necrosis, called hepatocyte bridging necrosis, was widely found in the liver pathological sections of rats in HA-2d group. This phenomenon was common in liver poisoning or severe chronic hepatitis ( Figure 1F). In HA-7d group, in addition to severe bridging necrosis of hepatocytes, there were also vacuolar degeneration of hepatocytes and fibrous tissue hyperplasia ( Figure 1G). Effect of AFB 1 on Liver Function Transaminase ALT and AST are common indicators to determine whether the liver is damaged. ALT mainly exists in the plasma of hepatocytes, and the elevation of ALT in serum reflects the injury of hepatocytes. AST mainly exists in mitochondria of hepatocytes, and elevated AST indicates more serious liver injury. ALP and TBIL are the detection indexes reflecting hepatic bile metabolism. After ASAE, ALT, AST, ALP and TBIL in the model groups increased first and then decreased (Figure 2). It reached the peak on the 2nd day after ASAE, which was significantly different from that in the CK group (p < 0.001). Except ALP, other indexes returned to normal level on the 7th day after ASAE. Analysis of Apoptosis in Rat Hepatocytes Apoptosis refers to the process by which cells are affected by abnormal factors and controlled by genes to end their life autonomously. As shown in Figures 3A-E, the nuclear membranes of normal hepatocytes were intact and stained blue; the nuclear membranes of apoptotic hepatocytes were clustered or fragmented and stained green. As can be seen from Figure 3F, compared with the CK group, the apoptosis rate in the model groups was significantly increased in a dose-dependent manner. However, the apoptosis rate of hepatocytes on the 7th after ASAE was slightly lower than that on the 2nd after ASAE, and there was still significant difference (p < 0.05). Hepatic Antioxidative Status and Lipid Peroxidation The content of MDA can reflect the degree of lipid peroxidation and indirectly reflect the degree of cell damage. In this study, the MDA content in model groups were significantly higher than that in CK group (p < 0.05) (Figure 4A), indicating that AFB 1 caused hepatocyte injury. It was also found that MDA content in the liver of the rats after ASAE not only showed a dose-dependent manner, but also first increasing and then decreasing trend with prolonged observation. However, it was still significantly higher than the CK group and did not return to the normal levels on the 7th day (p < 0.05). Studies have shown that AFB 1 can lead to the imbalance of the body's oxidative system homeostasis and cause oxidative stress response. The antioxidant enzymes SOD and GSH-Px are the body's first line of defense against free radicals. In this study, compared with the CK group, the activities of SOD and GSH-Px in the model groups decreased significantly (p < 0.05) (Figures 4B,C). And with the prolonged observation time after ASAE, the activities of SOD and GSH-Px became lower and lower (p < 0.001). Differential Expression Genes and Functional Analysis In this study, the DEGs between the model groups and the CK group were obtained by RNA-Seq technology, and a volcano map was drawn (Figures 5A-D). The volcano plot showed that LA-2d, LA-7d, HA-2d and HA-7d groups had 1246, 356, 2354, and 855 DEGs, respectively (FDR ≤ 0.05 and | log 2 FC| ≥ 1). From this result, it can be found that the number of DEGs was related to the dose of ASAE and the observation time after ASAE: the higher the dose, the more DEGs; the number of DEGs decreased with the prolonged time of continuous impact. Table 3 shows the classification and number of entries of KEGG pathways enriched in each group of DEGs. Looking at the total number of KEGG pathways in each model group, there were fewer on the 7th day after ASAE than on the 2nd day. The number of entries for each classification in each group shows that the main difference was in the metabolic category. This is related to the fact that the liver is the largest metabolic organ of the body as well as being extremely self-repairing. In order to explore the genes that were continuously affected and their functions, the screening criteria were strengthened, and the Venn diagram as shown was drawn (FDR ≤ 0.01 and | log 2 FC| ≥ 2) ( Figure 5E). Compared with CK group under the above filter conditions, the LA-2d, LA-7d, HA-2d and HA-7d groups had 236, 33, 679, and 78 differential genes, respectively. There were 211 common genes in LA-2d and HA-2d, 23 common genes in LA-7d and HA-7d, 20 common genes in LA-2d and LA-7d, and 56 common genes in HA-2d and HA-7d. In order to analyze the accuracy of the results, the first 20 independent DEGs were selected from each group, and all the ones less than (Tables 4, 5). By analyzing the DEGs in Tables 4, 5, it was found that DEGs of LA-2d and HA-2d were mainly enriched in pathways related to carbohydrate, lipid and amino acid metabolism, while DEGs of LA-7d and HA-7d were mainly enriched in four pathways, which were related to cell cycle regulation, inflammatory response and oxidation reaction. Lama5, Gtse1 and Fabp4 gene in DEGs were all highly expressed in the model

groups. Bcl6 gene was highly expressed on the 7th day after ASAE, and the expression level was higher than that on the 2nd day after ASAE. Lama5 gene enrichment in focal adhesion (ko04510) and metabolic pathways (ko01100). Focal adhesion pathway plays an important role in cell motility, cell proliferation, and cell differentiation (Ye et al., 2020). The dysregulation of focal adhesion is considered to be an essential step in tumor invasion, it can promote tumor invasiveness and metastasis (Shen et al., 2018). Metabolic pathway plays an important role in the metabolism of sugar, fat and protein in the liver. Metabolic dysfunction not only affects the normal growth of the body, but also causes various diseases. Gtse1 gene enrichment in p53 signaling pathway (ko04115). The protein encoded by the p53 gene is a transcription factor that controls the cell cycle. The occurrence of oxidative stress and DNA damage in the body can cause p53 mutation, resulting in cell cycle dysregulation (Lacroix et al., 2020;Bailey et al., 2021). Fabp4 gene enrichment in regulation of lipolysis in adipocytes (ko04923) and PPAR signaling pathways (ko03320). Regulation of lipolysis in adipocytes pathway plays an important role in the dynamic equilibrium of lipid droplet formation and decomposition (Yang and Mottillo, 2020). Dysregulation of this pathway leads to lipid accumulation in the liver causes hepatocyte damage and portal inflammation . PPAR signaling pathway plays an essential role in the maintenance of metabolic homeostasis, it can adjust the balance between anabolism and oxidation of adipose tissue to prevent peroxidation (Corrales et al., 2018). Bcl6 gene enrichment in FoxO signaling pathway (ko04068). The pathway can inhibit cell metabolism, growth, differentiation, oxidative stress, and aging, it is suggested to play a pivotal functional role as a tumor suppressor in cancers (Farhan et al., 2017;Lee and Dong, 2017). Quantitative Validation Quantitative validation of the predicted differential genes based on transcriptome sequencing data was carried out for Lama5, Ccnd1, Cdkn1a and Gpx2, which were randomly selected. RNA integrity is shown in the plots from electrophoresis experiments using a fully automated nucleic acid electrophoresis analyzer as in Figure 6. The validation results are shown in Table 6. The transcriptome sequenced genes with Q-PCR showed the same trend in different model groups. DISCUSSION In this study, a rat model of acute injury was established by a single intraperitoneal injection of 1 and 2 mg/kg AFB 1 , and then the rats were fed normally for 2 and 7 days, respectively. The aim was to study the persistent effects of ASAE on rat liver. We found that other indexes except antioxidant enzyme and liver coefficient showed a trend of recovery with the continuous influence time of the AFB 1 on liver. Among them, ALT, AST and TBIL returned to normal, ALP, hepatocyte apoptosis rate and MDA content still had significant differences compared with CK group (p < 0.05). Antioxidant enzyme activity continued to decrease with the prolonged exposure time of AFB 1 . The liver coefficient is rising. Meanwhile, the model was used to explore the persistent effects of acute AFB 1 exposure on liver genes, and we found that the number of DEGs decreased significantly over time. The data in this article shows that although there is no significant difference in the weight of rats after ASAE compared with that before ASAE, it still presents a downward trend. It illustrates that AFB 1 leads to the reduction of food intake in rats, and the decline of the body's metabolic capacity, so that the nutrients cannot be fully absorbed (Jindal et al., 1994). This is consistent with the results of previous research (Rastogi et al., 2001;Knipstein et al., 2015). Compared with the CK group, SOD and GSH-Px activities decreased, MDA content increased, and cell apoptosis rate increased on the 2nd day after ASAE, indicating that AFB 1 can cause oxidative stress reaction and liver cell apoptosis, resulting in severe liver damage. This result is similar to other published papers on acute liver injury caused by AFB 1 (Wang et al., 1991;Kumagai et al., 1998;Benkerroum, 2020;Ruggeberg et al., 2020). On the 4th day after ASAE, ALT and AST values decreased, but they were still significantly higher than the CK group (p < 0.05), which was consistent with previous studies (Monmeesil et al., 2019). At the same time, we found that ALT, AST, and TBIL returned to normal on day 7th after ASAE, but other phenotypic indexes still showed significant differences compared with the CK group. However, compared with the 2nd day after ASAE, the phenotypic indexes except antioxidant enzymes and liver coefficient showed significant differences, but showed a recovery trend. These results showed that AFB 1 still caused rats liver damage to a certain extent after feeding for 7 days after ASAE. The recovery of liver function indexes ALT, AST and TBIL may be related to the strong self-repair ability of liver (Ko et al., 2020), and the change trend of apoptosis rate and MDA content also indicates that the compensatory function of liver plays a role (Hall et al., 2021). The ALP value is related to whether the intrahepatic and extrahepatic bile ducts are blocked. The ALP value on the 7th day after ASAE was still significantly higher than that of the CK group, this is related to the fact that cholestasis caused by bile duct blockage does not self-adjust and recover in a short time (Boyer, 2013;Taylor et al., 2020). The changes of pathological section and liver coefficient further indicated that liver morphological damage was difficult to recover spontaneously in a short time. The antioxidant enzymes GSH-Px and SOD activities continued to decrease with the prolongation of AFB 1 influence time. On the 2nd day after ASAE, the change trend of antioxidant enzyme activity was similar to that reported in the previous paper . The changes of antioxidant enzymes on the 7th day after ASAE may be related to the fact that the liver damage of rats is not alleviated and the intake of antioxidant nutrients is insufficient due to decreased appetite (Michiels et al., 1994;Peskin, 1997;Liochev and Fridovich, 2010). However, the reasons for the changes in lipid peroxide content and antioxidant enzyme activity in this model need to be further explored through related experiments. We obtained the gene data of each experimental group by RNA-Seq technique and found that the number of DEGs was affected by the dose and duration of AFB 1 exposure, among which Lama5, Gtse1, Fabp4 and Bcl6 genes were continuously affected and associated with liver injury. Their mechanism of action is shown in Figure 7. Lama5 gene plays a role in promoting angiogenesis during tumor development and is a key promoter of liver metastatic growth (Yousif et al., 2013;Gordon-Weeks et al., 2019). Tumor-infiltrating myeloid cells can drive tumor cells to express Lama5 through NF-κB gene transduction, thereby promoting angiogenesis (Bonnans et al., 2014). The Lama5 protein chain is an important element of the extracellular matrix (ECM). Thus, the Lama5 gene links two markers of cancer progression (Inflammation and Angiogenesis) to extracellular matrix (ECM) protein deposition. Related studies had shown that down-regulation of Lama5 can reduce the formation of vascular branches, and even tend to normalize (Carmeliet and Jain, 2011;Di Russo et al., 2017). Fabp4 has also been shown to be involved in inflammatory responses. Cytokine levels, including TNF-α and IL-1β, were significantly increased in macrophages with high Fabp4 expression (Makowski et al., 2001). And Fabp4 plays an important role in transporting fatty acids between fat cells and cancer cells (Besnard et al., 2002). These effects ultimately may lead to fibrosis of tissues (Nieman et al., 2011). The proto-oncogene Bcl6 is an important transcriptional regulator of the immune system, and its upregulation is a marker of B cell accumulation through the germinal center (GC) transport process. In tumors, its dysregulated expression may FIGURE 7 | Mechanism of liver injury induced by Lama5, Gtse1, Fabp4, and Bcl6 genes. RAF: Regulation of lipolysis in adipocytes; ECM: Extracellular matrix; " ": vascular endothelial growth factor (VEGF); " ": promote/up-regulate; " ": inhibition. promote lymphoma by increasing B cell resistance to apoptosis and inhibiting B cell differentiation in the GC (Murakami et al., 1999;Niu, 2002). Low-grade B-cell lymphomas can occur in the liver and manifest as a characteristic massive infiltration of lymphocytes within the confluent zone. Lama5, Fabp4 and Bcl6 gene related to inflammation and immune response were all upregulated, and the expression of Bcl6 gene was higher at 7 days than at 2 days after ASAE. Their changing trend explained that the pathological phenomena of liver was not relieved the 7th day after ASAE to some extent. And there were still different degrees of liver cell necrosis, inflammatory cell infiltration and fibrous tissue hyperplasia. Gtse1 has been found to be overexpressed in a variety of cancers (Tian et al., 2011). Previous studies have shown that it inhibits apoptosis induced by the tumor suppressor protein p53 . Cell level experiments showed that Gtse1 was down-regulated, the protein content of p53 in cells increased, and the dephosphorylation of protein kinase B and cyclin B1 decreased, thus inhibiting the proliferation of HCC cells and promoting apoptosis (Guo et al., 2016). Our data showed that, the expression level of Gtse1 on the 7th day after ASAE was lower than that on the 2nd day after ASAE, which may explain why the apoptosis rate of liver cells decreased on the 7th day after ASAE. However, the specific regulatory mechanism needs further study. CONCLUSION In summary, ASAE can cause persistent liver damage. The experiment period of this study was 7 days, and the damage of liver during the extended experiment period could be further studied. The persistently affected Lama5, Gtse1, Fabp4 and Bcl6 gene can be potential target genes for the treatment of AFB 1 induced liver diseases, but their effectiveness needs to be further explored in experiments. DATA AVAILABILITY STATEMENT The data presented in the study are deposited in the NCBI repository, accession number BioProject ID PRJNA823496. ETHICS STATEMENT The animal study was reviewed and approved by Shanghai University of Traditional Chinese Medicine. Dose-response relationship between lower serum magnesium level and higher prevalence of knee chondrocalcinosis Background The aim was to assess serum magnesium levels in relation to prevalence of knee chondrocalcinosis in two population-based Chinese studies. Methods Data included in this analysis consisted of two population-based cross-sectional studies, i.e., the Xiangya Hospital Health Management Center Study and the Xiangya Osteoarthritis (XO) Study I. A bilateral knee anteroposterior radiograph was obtained from each subject. Radiographic knee chondrocalcinosis was present if there was definite linear cartilage calcification. Serum magnesium concentration was measured using the chemiluminescence method. We examined the relation of serum magnesium levels to prevalence of knee chondrocalcinosis using generalized estimating equations. Results The prevalence of knee chondrocalcinosis was 1.4% in the Xiangya Hospital Health Management Center Study (n = 12,631). Compared with the lowest tertile, the age, sex and body mass index (BMI)-adjusted odds ratios (ORs) of chondrocalcinosis were 0.59 (95% CI 0.40–0.87) and 0.49 (95% CI 0.33–0.72) in the second and the third tertiles of serum magnesium, respectively (P for trend <0.001). The prevalence of knee chondrocalcinosis in the XO Study I (n = 1316) was 4.1%. The age, sex and BMI-adjusted ORs of chondrocalcinosis were 0.67 (95% CI 0.34–1.30) in the second and 0.45 (95% CI 0.21–0.94) in the third tertile of serum magnesium when compared with the lowest tertile (P for trend = 0.030). Similar results were observed in men and women in both studies. Adjusting for additional potential confounders did not change the results materially. Conclusions Subjects with lower levels of serum magnesium, even within the normal range, had higher prevalence of knee chondrocalcinosis in a dose-response relationship manner, suggesting that magnesium may have a preventive or therapeutic potential for knee chondrocalcinosis. Electronic supplementary material The online version of this article (doi:10.1186/s13075-017-1450-6) contains supplementary material, which is available to authorized users. Background Calcium pyrophosphate deposition (CPPD) in articular fibrocartilage and hyaline cartilage, termed chondrocalcinosis, has been considered to be associated with both aging and osteoarthritis [1]. CPPD, together with basic calcium phosphate (BCP), are the two main components of calcium-containing crystals in cartilage in osteoarthritis [2,3]. While most subjects with chondrocalcinosis have no clinical symptoms (asymptomatic CPPD) [4], CPPD may promote articular degeneration (osteoarthritis with CPPD) and traffic of the crystals from articular cartilage into the joint space; and thus can

stimulate acute episodic crystal arthritis and can also lead to chronic inflammatory arthritis. To date, only a few risk factors have been identified for the occurrence of chondrocalcinosis. Studies have shown that prevalence of chondrocalcinosis varied in different ethnic groups. For example, prevalence of chondrocalcinosis was much lower in Beijing Osteoarthritis Study participants (1.8% in men, 2.7% in women) than that in white subjects among the predominantly white participants in the Framingham Study (6.2% in men, 7.7% in women) [5]. Several studies also found that knee chondrocalcinosis is associated with low levels of serum magnesium [6][7][8][9][10][11]. For instance, patients with hypomagnesemia (e.g., poor parenteral nutrition, short bowel syndrome, Gitelman syndrome or familial heredity) had a much higher prevalence of chondrocalcinosis (up to 23.1%) [6][7][8][9][10][11]. However, these findings were often based on subjects with extremely low serum magnesium, and it remains unclear whether more modest variations of magnesium levels observed in the general population are associated with the prevalence of chondrocalcinosis. Such potential associations and their quantification would have an implication in public health and clinical practices. To fill in this knowledge gap, we used data collected from two large population-based studies (i.e., Xiangya Hospital Health Management Center Study and Xiangya Osteoarthritis (XO) Study I) and examined the relation of serum magnesium levels with the prevalence of knee chondrocalcinosis and to determine the shape of the dose-response relationship between serum magnesium levels and the prevalence of knee chondrocalcinosis. Study population Xiangya Hospital Health Management Center Study Subjects included in this study were residents living in Hunan, China, who were undergoing routine health examination at Xiangya Hospital, Central South University, China. The study design has been published elsewhere [12][13][14]. In brief, routine health checkup included anthropometric (e.g., height, weight, etc.), basic clinical examination (e.g., blood pressure, heart rate, etc.), and biochemical (e.g., blood routine examination, hepatic function, renal function, trace elements test, etc.) and imaging tests (e.g., chest radiography, weight-bearing bilateral anteroposterior knee radiography, etc.). Subjects included in the current analysis were those who: (1) had health checkup between October 2013 and December 2015; (2) were age ≥40 years; (3) underwent a serum magnesium test; and (4) had bilateral weight-bearing anteroposterior radiographs. All interviewers, clinical examiners and x-ray technicians were trained by the principal investigators (CZ and GL) before the study began. Xiangya Osteoarthritis Study I (XO Study I) Subjects included in this study were a randomly selected sample of residents, age ≥50 years, from eight rural mountainous communities of Longshan County [16], Hunan Province, for a study of osteoarthritis. All of the villages in the selected communities were listed in a random order. Beginning with the first village in the first community, all residents age ≥50 years were invited to participate in our study. The village-to-village recruitment continued until the number of subjects in that community met the predetermined quota according to the Sixth National Census Data of Longshan County (2010). Subjects were recruited at the site near their home and were transported to the hospital for interview and clinical examination between November 2015 and January 2016. Trained health professional interviewers administered a standardized questionnaire that focused on sociodemographic factors, lifestyle habits, joint symptoms, joint functions, quality of life, dietary intake, medication use and other potential risk factors for osteoarthritis. Clinical examination, laboratory testing and radiography were also conducted at the time of interview. All interviewers, clinical examiners and x-ray technicians were trained under the supervision of the study principal investigators (CZ and GL). Blood biochemical analysis All blood samples were drawn after a 12-hour overnight fast and were stored at 4°C until analysis. The serum magnesium concentration was measured using the chemiluminescence method by Beckman Coulter AU 5800 (Beckman Coulter Inc., Brea, CA, USA). The interassay and intra-assay coefficients of variation were tested by low concentrations (0.60 mmol/L of serum magnesium) and high concentrations (1.00 mmol/L of serum magnesium) of standard human samples. The intraassay coefficients of variation were 1.86% (0.60 mmol/L) and 1.65% (1.00 mmol/L) for serum magnesium, and the inter-assay coefficients of variation were 1.87% (0.60 mmol/ L) and 1.70% (1.00 mmol/L) for serum magnesium. Measuring methods and reliability data of potential confounders (e.g., serum parathyroid hormone, iron, ferritin, total iron binding capacity, calcium, copper, zinc, phosphorus and vitamin D) are shown in Additional file 1. Assessment of radiographic knee chondrocalcinosis All radiographs were assessed by two orthopedic surgeons (TY and YLW) who were blinded to subjects' clinical symptoms and biochemical test results. Radiographic chondrocalcinosis was defined as present when there was evidence of definite linear cartilage calcification in the knee [17]. Specifically, prior to starting the assessment, these two orthopedists re-read 200 radiographs from the Osteoarthritis Initiative (OAI) to calibrate their reading. Formal readings of batches of randomly selected radiographs did not start until the readers reached a high level of agreement with previous readings from OAI (the specific cutoff of simple kappa for inter-rater reliability was 0.70). During the formal reading, one batch of knee radiographs (100 radiographs) consisting of 10 previously read radiographs selected at random and 90 unread radiographs were mingled and read. For each batch, 90 unread radiographs were used to test inter-rater reliability and 10 previously read radiographs were used to test intra-rater reliability, respectively. Two readers both assessed all radiographs and inconsistencies were resolved through discussion. The kappa value for reliability readings of radiographs from Xiangya Hospital Health Management Center Study was 0.72 (95% CI 0.67 − 0.77) for inter-rater reliability and 0.76 (95% CI 0.67-0.85) for intra-rater reliability. The kappa value for reliability readings of radiographs from XO Study I was 0.75 (95% CI 0.68-0.83) for inter-rater reliability and 0.80 (95% CI 0.65-0.94) for intra-rater reliability. Statistical analysis Continuous data were expressed as the mean ± standard deviation, and categorical data were expressed as proportion (percentage). The serum magnesium concentration was classified into three categories based on the tertiles distribution in each study population (i.e., ≤ 0.86, 0.87-0.91 and ≥ 0.92 mmol/L in the Xiangya Hospital Health Management Center Study and ≤ 0.89, 0.90-0.95 and ≥ 0.96 mmol/L in the XO Study I). We calculated the sex-specific prevalence of chondrocalcinosis for each category of serum magnesium. We examined the association of serum magnesium categories and prevalence of chondrocalcinosis using generalized estimating equations (GEE) [18], adjusting for the potential confounders (knee specific analysis). Odds ratios (OR) and related 95% confidence intervals (95% CI) of chondrocalcinosis among different categories of serum magnesium were calculated using the PROC GENMOD procedure in SAS with binomial distribution and logit links, and the lowest tertile of serum magnesium was considered as the reference. We also calculated the OR and the related 95% CI by using the category ≤0.70 mmol/L (dividing the lowest category of serum magnesium into two categories: ≤0.70 mmol/L and 0.71-0.86 in the Xiangya Hospital Health Management Center Study, and the category ≤0.70 mmol/L and 0.71-0.89 in the XO Study I) as the reference group. Specifically, in the Xiangya Hospital Health Management Center Study we first adjusted for age (40-49, 50-59, 60-69, ≥70 years), body mass index (BMI) (<25, ≥25 kg/m 2 ) and sex (male, female). Then we added each of the following covariates separately: serum iron (tertiles), serum ferritin (tertiles), serum calcium, serum zinc (tertiles), serum copper (tertiles), serum phosphorus (tertiles), education (high school or above, lower than high school) and occupation (manual labor, non-manual labor) in the regression model adjusted for age, sex and BMI. In the XO Study I, we first adjusted for age (50-59, 60-69, ≥70 years), BMI (<25, ≥25 kg/m 2 ) and sex (male, female), and then added each of the following covariates separately in the regression models. These covariates included knee injury, serum parathyroid (tertiles), serum iron (tertiles), serum total iron binding Mg magnesium, BMI body mass index capacity (tertiles), serum unsaturated iron binding capacity (tertiles), serum calcium (tertiles), serum 25(OH)D (tertiles), serum zinc (tertiles), serum copper (tertiles), serum phosphorus (tertiles), education (educated, non-educated) and occupation (farmer, non-farmer). In addition, we performed sex-specific analysis to examine the levels of serum magnesium and prevalence of chondrocalcinosis. The dose-response relationship between levels of serum magnesium and the prevalence of knee chondrocalcinosis was evaluated by restricted cubic splines regression with two knots defined by the tertile distribution of serum magnesium [19,20]. We also used the value 0.7 mmol/L as an additional knot, to evaluate the dose-response association between serum magnesium and the prevalence of knee chondrocalcinosis. Xiangya Hospital Health Management Center Study Of the remaining 12631 participants from the initially included 14,715 participants, 43.0% (n = 5428) were Mg magnesium, N number, CI confidence interval women, the average age was 52.3 ± 8.0 years, and the mean BMI was 24.5 kg/m 2 (SD = 3.3 kg/m 2 ). The prevalence of knee chondrocalcinosis was 1.3% in men and 1.5% in women ( Table 1). As shown in Fig. 1, serum magnesium, even within the normal range, was inversely associated with the OR for knee chondrocalcinosis in a dose-response-relationship manner (test for trend P = 0.03 for total population; P = 0.16 for the male population; P = 0.05 for the female population). The prevalence of knee chondrocalcinosis (knee-specific analysis) decreased from 1.3% in the lowest tertile of serum magnesium to 0.8% in the second, and 0.6% in the highest tertile of serum magnesium. Figure 2 showed the linear association between the serum magnesium and predicted prevalence of chondrocalcinosis. After adjusting for age, sex and BMI, the ORs for knee chondrocalcinosis were 0.59 (95% CI 0.40-0.87) in the middle and 0.49 (95% CI 0.33-0.72) in the highest tertile of serum magnesium, respectively, compared with the lowest tertile (P value for trend <0.001). Adding each of the potential confounders (e.g., serum iron, ferritin, calcium, zinc, copper and phosphorus) into the age, sex and BMI-adjusted model, or sensitivity analysis with exclusion of participants with chronic renal failure, also did not change the results materially. The association was also consistent in men and women (Table 2). When 0.70 mmol/L was used as an additional knot for spline regression, and the category ≤0.70 was used as reference group for GEE analysis, the results did not change significantly (see Additional file 2). XO Study I Of 1469 participants in the XO Study I, 1316 were included in the current analysis. About half (51.2%, n = 674) were women, average age was 62.9 ± 8.7 years and mean BMI was 24.2 kg/m 2 (SD = 3.6). The prevalence of knee chondrocalcinosis was 3.1% in men and 5.0% in women (Table 3). Serum magnesium, even within the normal range, was inversely associated with the OR for prevalence of knee chondrocalcinosis in a dose-response-relationship manner (Fig. 3, test for trend P = 0.06 for the total population; P = 0.25 for the male population; P = 0.14 for the female population). As shown in Table 4, the prevalence of knee chondrocalcinosis (knee-specific analysis) decreased from 4.8% in the lowest tertile of serum magnesium to 3.2% in the second, and 2.0% in the highest tertile of serum magnesium. Figure 4 shows the linear association between the serum magnesium and predicted prevalence of chondrocalcinosis. After adjusting for age, sex and BMI, compared with the lowest tertile, ORs for knee chondrocalcinosis were 0.67 (95% CI 0.34-1.30) in the middle and 0.45 (95% CI 0.21-0.94) in the highest tertile of serum magnesium, respectively (P value for trend = 0.030). Adding each of the potential confounders (e.g., knee injury, serum parathyroid hormone, iron, total iron binding capacity, unsaturated iron binding capacity, calcium, 25(OH)D, zinc, copper and phosphorus) into the age, sex and BMI-adjusted model, or sensitivity analysis with exclusion of participants with chronic renal failure or who used diuretics, also did not change the results materially. The association was consistent in men and women (Table 4). When 0.70 mmol/ L was used as an additional knot for spline regression, and the category ≤0.70 was used as the reference group for GEE analysis, the results did no change significantly (see Additional file 2). Discussion In two population-based cross-sectional studies, we found that serum magnesium, even within the normal range, was inversely associated with the prevalence of knee chondrocalcinosis in a dose-response-relationship manner. In both studies, prevalence of knee chondrocalcinosis among subjects in the highest tertile of serum Mg magnesium, BMI body mass index, PTH parathyroid hormone, TIBC, total iron binding capacity, UIBC unsaturated iron-binding capacity magnesium was approximately 50% lower than those

in the lowest tertile of serum magnesium. Such an association was consistent in men and women. A few case-reports and case-series studies have shown that chondrocalcinosis often co-occurs with hypomagnesemia or with hypomagnesemia owing to bowel syndrome and Gitelmen syndrome [7][8][9][10][11]. For example, in 2007 Richette et al. compared serum magnesium and the prevalence of chondrocalcinosis in 72 patients with intestinal failure and 72 age-matched and sex-matched patients with back pain [6]. They found that serum magnesium was significantly lower (P < 0.001) among patients with intestinal failure (0.75 mmol/L) than those with back pain (0.81 mmol/L). On the other hand, the prevalence of chondrocalcinosis was sevenfold higher (P = 0.006) among patients with intestinal failure (16.6%) than among patients with back pain (2.7%). Furthermore, serum magnesium, globular magnesium and 24hour urinary magnesium were significantly lower in patients with intestinal failure than those with back pain. Several in vivo studies have demonstrated that decreasing magnesium intake induces calcification formation in different animal models [21][22][23][24]. However, few in vivo studies, if any, have examined the effect of magnesium supplementation or intake on the inhibition of cartilage calcification. In a double-blind randomized clinical trial assessing the effect of a magnesium carbonate supplement among patients with chronic pyrophosphate arthropathy, there was no significant difference in the radiographic appearance of chondrocalcinosis between the treatment and placebo groups; however, the treatment group had a uniform trend towards improvement in pain, stiffness, effusion, tenderness and overall subjective and objective assessment compared with the placebo group over a 6month period [25]. This finding could be explained by the physiologic N-methyl-D-aspartate (NMDA) receptor antagonist property of magnesium, and consequently, increasing serum magnesium level may reduce the incidence of symptomatic calcium-deposition-related diseases [26]. Furthermore, in a large population-based study, Zhang et al. reported that the prevalence of knee chondrocalcinosis was much lower among participants in the Beijing Osteoarthritis Study than in their counterparts in the Framingham Study (age-standardized prevalence ratios = 0.34 in men and 0.43 in women); however, no difference in the levels of magnesium in the tap drinking water was found between the two cities [5]. Several mechanisms linking magnesium and chondrocalcinosis have been postulated [27]. These include that high serum magnesium may inhibit the formation of calcium phosphate apatite and of calcium-acidic phospholipidphosphate complexes [28,29], may increase the expression of calcification inhibitors (e.g., matrix gla protein, osteopontin, and bone morphogenetic protein 7), may decrease the expression of calcification promoters (e.g., alkaline phosphatase, bone morphogenetic protein 2 and runtrelated transcription factor 2) and apoptosis [30][31][32][33], may activate calcium-sensing receptors [34][35][36], modulate vitamin D receptor, fibroblast growth factor-1 receptor and its co-receptor klotho [36], or may block various calcium channels to impair excessive calcium uptake [37,38]. Several characteristics of our study are noteworthy. First, compared with other studies, the sample size of the current two studies was relatively large, and the prevalence of knee chondrocalcinosis found in these two studies (1.4% in the Xiangya Hospital Health Management Center Study and 4.1% in the XO Study I) was similar to another population-based study (i.e., the Beijing Osteoarthritis Study) conducted among residents in Beijing (1.8% in men and 2.7% in women) [5]. These findings suggest that the prevalence of knee chondracalcinosis is lower than in the predominantly white participants in the Framingham Study [5]. In addition, we also showed that prevalence of hypomagnesemia appeared to be lower among the participants in the current two studies than that among subjects in the Rotterdam Study [39]. Second, an inverse association between serum magnesium levels and prevalence of knee chondrocalcinosis was found in our two study populations and across the sex categories, and the magnitude of association was also quite similar when several potential confounders (e.g., knee injury, serum parathyroid hormone, iron, ferritin, total iron binding capacity, unsaturated iron binding capacity, calcium, copper, zinc, phosphorus and 25(OH)D) was added into the age, sex and BMI adjusted model, respectively, indicating the robustness of our findings. Third, we also showed that even within the normal range of serum magnesium, the levels of magnesium were still inversely associated with the prevalence of knee chondrocalcinosis in a dose-responserelationship manner. This finding suggests that increased serum magnesium not only among subjects with hypomagnesium but also among subjects who were within the normal range of magnesium may help to reduce the risk of chondrocalcinosis. The prevalence of radiographic chondrocalcinosis in the XO Study I was two to three times higher than that in the Xiangya Hospital Health Management Center Study. Such a difference may be due to the age difference between the two study populations in that the average age in the XO Study I (62.9 ± 8.7 years) was almost 10 years older than that in the Xiangya Hospital Health Management Center Study (52.3 ± 8.0 years). While the age distribution is similar in the XO Study I, and the Beijing Osteoarthritis Study restricted participants to 60 years and older, the prevalence in the XO Study I was higher than in the Beijing Osteoarthritis study. This may because of the heavy physical labor (82.4% of men and 91.6% of women were farmers) and frequent climbing (mountain area) of participants in the XO Study I, which of course warrants further study for confirmation. Our study has some limitations. First, this is a crosssectional study, thus, we cannot be certain that the causal relationship between serum levels of magnesium and prevalence of knee chondrocalcinosis. Second, we did not assess the dietary intake of magnesium in relation to the prevalence of knee chondrocalcinosis in the present study. While previous studies have shown that dietary intake of magnesium is strongly associated with the serum levels of magnesium [40,41], studies of dietary intake of magnesium and risk of knee chondrocalcinosis are warranted. Third, we did not use other views of the knee (e.g., lateral or skyline) to ascertain radiographic choncrocalcinosis; thus, it may have resulted in underestimation of the prevalence of radiographic CPPD. However, we postulate that such misclassification of the prevalence of knee CPPD is likely to be nondifferential; thus the association between serum Mg magnesium, CI confidence interval magnesium levels and prevalence of knee radiographic CPPD may be even stronger than we have observed. In addition, as almost half of cases of wrist, hip, symphysis pubis, and metacarpophalangeal joint chondrocalcinosis occur without knee chondrocalcinosis [42], the examination of the association of magnesium levels with a predisposition to form CPPD crystals would ideally include radiographs from other joint sites. Fourth, a synovial fluid test taken from the joint should be more sensitive and specific to detect crystals than radiographs, and some of the articular CPPD deposition (alone or mixed with BCP [2,3]) may be too small to be detected on radiographs; thus, our estimate of the prevalence of knee chondrocalcinosis is likely to be underestimated. However, as we stated previously, that prevalence estimate of knee chondrocalcinosis was similar to another study [5] conducted in Chinese subjects that used the same protocol to assess the presence of knee chondrocalcinosis. The proportion of persons with chondrocalcinosis identified on radiographs who develop clinical symptoms related to this is small. Fifth, data on some other potential confounders (e.g., osteoporosis and use of bisphosphonates) were not collected in either study and could impact the results. We performed sensitivity analysis to assess to what extent the residual confounding could explain the association, using the E-value proposed by VanderWeele and Ding [43]. The observed OR of 0.49 for prevalence with the highest levels of magnesium in Xiangya Hospital Health Management Center Study could be explained if an OR of an unmeasured confounder with either high serum magnesium or with the prevalence of chondrocalcinosis is at least 3.50, which is above and beyond the measured confounders. There should be a similar magnitude of residual confounders to explain the findings in the XO Study I. To our best knowledge we are unaware that there is such a strong confounder between magnesium and the prevalence of chondrocalcinosis. Our study findings have potential clinical implications. The existing in vivo and in vitro studies showed that magnesium supplementation may have a protective effect on cartilage or chondrocytes [44]. Two large epidemiological studies conducted in twins and in the general population reported an inverse association between serum magnesium levels and radiographic knee osteoarthritis [12,45]. Furthermore, several studies have reported that chondrocalcinosis may be a potential risk factor for osteoarthritis [46][47][48][49][50][51][52]. Thus, it is not unreasonable to speculate that increasing serum magnesium may be beneficial in decreasing the risk of osteoarthritis through reducing the incidence of chondrocalcinosis. Future studies are warranted to test this hypothesis. Conclusions We found that low serum magnesium, even within the normal range, was associated with higher prevalence of knee chondrocalcinosis in a dose-response-relationship manner. Future studies of dietary intake, including magnesium supplementary intake, on the risk of chondrocalcinosis are warranted. Additional files Additional file 1: Measuring methods and reliability data of potential confounders. (DOCX 15 kb) Additional file 2: Table S1. Association between serum Mg and knee chondrocalcinosis in the Xiangya Hospital Health Management Center Study (n = 12,631). Table S2. Association between serum Mg and knee chondrocalcinosis in the XO Study I (n = 1316). Figure S1. Association between serum Mg and OR of CC analyzed by spline regression (3 knots, 0,7 as reference) in the Xiangya Hospital Health Management Center Study (P = 0.049). Mg, magnesium; CC, chondrocalcinosis; OR, odds ratio; CI, confidence interval. Figure S2. Association between serum Mg and OR of CC analyzed by spline regression (3 knots, 0,7 as reference) in the XO Study I (P = 0.060). Mg, magnesium; CC, chondrocalcinosis; OR, odds ratio; CI, confidence interval. (DOCX 1720 kb) Toll-like receptor 3 deficiency in autoimmune encephalitis post–herpes simplex encephalitis The Toll-like receptor 3 (TLR3) pathway is a key component of the innate immunity that prevents replication of viruses in the CNS. Inborn errors of this pathway (TLR3-pathway deficiency), which includes defects in the genes TLR3, UNC93B1, TRIF, TRAF3, TBK1 and IRF3, occur in 10% of patients with herpes simplex encephalitis (HSE),1,2 and about 66% of these patients develop relapses of HSE.1 A recent study showed that 27% of patients with HSE develop autoimmune encephalitis (AE) in the weeks or months ensuing the infection.3 It is unknown whether TLR3-pathway deficient patients can also develop AE post-HSE. Here we report a patient with TLR3-pathway deficiency who developed HSE and a relapse of the viral infection followed by AE post-HSE, highlighting the fact that TLR3-pathway deficient patients should be carefully followed for both HSE relapses and AE. The Toll-like receptor 3 (TLR3) pathway is a key component of the innate immunity that prevents replication of viruses in the CNS. Inborn errors of this pathway (TLR3-pathway deficiency), which includes defects in the genes TLR3, UNC93B1, TRIF, TRAF3, TBK1 and IRF3, occur in 10% of patients with herpes simplex encephalitis (HSE), 1,2 and about 66% of these patients develop relapses of HSE. 1 A recent study showed that 27% of patients with HSE develop autoimmune encephalitis (AE) in the weeks or months ensuing the infection. 3 It is unknown whether TLR3-pathway deficient patients can also develop AE post-HSE. Here we report a patient with TLR3-pathway deficiency who developed HSE and a relapse of the viral infection followed by AE post-HSE, highlighting the fact that TLR3-pathway deficient patients should be carefully followed for both HSE relapses and AE. Case A 6-year-old healthy Caucasian girl, with a family history of maternal grandfather with recurrent herpetic keratitis, developed acute-onset headache, fever, decreased level of consciousness, aphasia, and right hemiparesis. Brain MRI showed bilateral hemorrhagic temporal lesions leading to suspect HSE. Signs of cranial hypertension precluded CSF studies; however, a positive herpes simplex virus 1 (HSV-1) PCR was obtained in the blood samples. IV acyclovir resulted in symptom improvement, and 3 weeks later the patient was discharged home with residual aphasia. Five months later, she was readmitted with severe headache and decreased level of consciousness. Brain CT showed new hemorrhagic temporal lesions. She required urgent decompressive craniectomy, and a brain tissue sample was HSV-1 PCR positive. Genetic studies showed that the patient and the mother carried a heterozygous missense mutation p.Glu110Lys (c.328G>A) in exon 2 of the gene TLR3 (NM_003265.2) (figure, A and B). Such mutation shows a very low allelic frequency (0.00082%, 1/121,316). Functional studies on patient's fibroblasts (figure, C) and monocyte-derived dendritic cells showed a decrease in TLR3-mediated activation

(methods in appendix e-1, links.lww.com/NXI/A141). The patient received a new 21-day course of IV acyclovir followed by oral valganciclovir. One year later, 17 months after the initial episode of HSE, she developed behavioral changes characterized by aggressivity and paranoid thoughts. CSF studies showed pleocytosis (15 cells/μL) and elevated protein concentration (100 mg/dL) but were HSV-1 PCR negative. CSF immunochemistry studies on rat brain tissue (figure, D, upper panel) and cultured live neurons showed *Shared first authorship. strong reactivity revealing the presence of neuronal antibodies. Autoantibodies against N-methyl-D-aspartate, γ-aminobutyric acid A, and other known receptors and cell surface proteins were all negative (methods in appendix e-1, links.lww.com/NXI/A141). The indicated patient's antibodies were absent in samples of CSF obtained during HSE (data not shown). With the diagnosis of AE post-HSE, she was started on high-dose IV steroids and immunoglobulins without a clear improvement. Subsequently, rituximab (2 doses 500 mg/m 2 2 weeks apart) resulted in neurologic improvement. A CSF sample obtained 4 months later was no longer reactive with brain. Discussion We report a young girl with recurrent HSE and a new TLR3 mutation associated with absent interferon-β responses to TLR3 agonist who several months after the first HSE episode developed recurrent HSE followed by AE. TLR3-pathway deficient patients, especially TLR3-deficient, are prone to develop HSE relapses, putting them at risk of developing AE. Indeed, a recent series of patients with HSE showed that 27% subsequently developed AE; none of them were investigated for TLR3-pathway deficiency but the current case indicates that patients with this deficiency are also at risk of AE. The symptoms of our patient (predominant behavioral abnormalities) are typical of AE post-HSE in patients older than 4 years 3,4 ; yet the long interval between HSE and AE (17 months since the onset of HSE and 12 months since the relapse of HSE) is somewhat atypical (median 26 days in children aged 4 years or younger and 43 days in patients older than 4 years), 3,5 but similar prolonged intervals occurred in 7% of patients in a recently reported series. 3 The exact mechanisms underlying this severe immune-mediated complication are unknown. It has been postulated that the neuronal destruction caused by the virus leads to a release of antigens, which in the context of severe inflammation results in neuronal autoimmunity, such as that shown in our patient. 3 In patients with TLR3-pathway deficiency, the increased number of episodes of HSE theoretically increases the risk of this autoimmune complication. Our case suggests that patients with recurrent HSE or family history of HSE or HSV keratitis should be assessed for TLR3-pathway deficiency. 1,6,7 If the function of this pathway is impaired, patients should be carefully followed for potential relapses of HSE or AE post-HSE. Systematic literature review of trials assessing recommended systemic treatments in hepatocellular carcinoma Aim: To identify and evaluate the similarity of all trials assessing recommended treatments for advanced hepatocellular carcinoma. Materials & methods: Single arm and randomized trials from any phase and published any time up to February 2021 were systematically searched. Results: From 5677 records reviewed, 50 trials were included in the review, and 24 for assessed for similarity. In the first-line (1L) setting, several trials assessing sorafenib were noted for enrolling patients with more severe disease and/or performance status than other 1L trials; trials within the second-line (2L) setting were generally similar. Median survival was <2 years in all trial arms. Conclusions: Trials assessing recommended treatments are largely similar and appropriate for quantitative comparisons of several efficacy and safety outcomes. duplicate. Trials were excluded if the intervention combined a recommended treatment with a non-recommended treatment or combined two treatments currently recommended only as monotherapies. After all relevant publications were identified, two independent reviewers extracted data from the articles and reconciled any discrepancies. A third independent reviewer was consulted as necessary and adjudicated where consensus could not be reached. A full list of variables extracted is provided in Supplementary Table 3. Risk of bias Criteria from the Centre for Reviews and Dissemination were used for assessment of the risk of bias in RCTs; the Newcastle-Ottawa Scale with modifications for cohort studies was used for single-arm studies [20,21]. The risk of bias assessment was conducted at the study level by two blinded reviewers and adjudicated by a third independent reviewer if necessary. Similarity assessment A qualitative similarity assessment was performed on all RCTs that compared at least two interventions of interest or placebo and any single-arm trials. Studies were first characterized as 1L or 2L based on trial inclusion criteria, then as single arm or RCTs, and finally by the drug class of the treatment used. Drug classes include tyrosine kinase inhibitors (TKI; i.e., lenvatinib, sorafenib), vascular endothelial growth factor receptor inhibitors (VEGFRI; i.e., bevacizumab, cabozantinib, ramucirumab, regorafenib), immuno-oncology agents (IO; i.e., atezolizumab, ipilimumab, nivolumab, pembrolizumab) or none (i.e., placebo, BSC). Trial inclusion/exclusion criteria were compared for the following characteristics (if described): BCLC stage or other tumor stages, ECOG performance status, Child-Pugh class, hepatic encephalopathy, ascites, bleeding risk, time since loco-regional treatment, and any excluded viral infections. Patient and disease characteristics assessed for similarity included mean or median age, proportion male, the proportion from the Asia-Pacific region, vascular invasion, extrahepatic spread, hepatitis B, hepatitis C, AFP, ECOG performance status, BCLC stage and Child-Pugh class. Efficacy and safety outcomes of interest were also reviewed. All proportions were converted to percentages for presentation. Missing data were reported as 'not reported'. Figures were made for patient characteristics (bubble charts), hepatitis prevalencegeographic region and disease severity-survival relationships (scatter plots), within-trial survival comparisons (bar charts), and disease stage/severity, survival, tumor response, and safety (histograms). As this study was intended to provide a qualitative review, median values and ranges were presented for survival outcomes, tumor response and safety outcomes. Of these 50 trials, all single-arm trials and any RCTs that compared a recommended treatment against a second recommended treatment or placebo were further examined for similarity; therefore, the similarity assessment included 24 trials (13 1L trials and 11 2L trials; Table 1). The two trials with both single arm and randomized groups were included in the similarity assessment, despite the fact that the randomized groups in these trials do not include two recommended treatment arms. Conversely, the 26 RCTs comparing a recommended treatment against non-recommended treatments (sorafenib, 25 trials; FOLFOX, one trial; Supplementary Table 4) were not examined for similarity. Risk of bias for the studies included in the similarity assessment can be found in Supplementary Table 5-8. Design & inclusion criteria Trial design and inclusion criteria were reviewed first for similarity. In terms of study design (Supplementary Table 9), blinding varied among the RCTs; most of the 1L RCTs were open label, but all the 2L RCTs were double blind. The majority of trials (18 of 24) were Phase II or Phase III, and half the trials (12 of 24) were global (i.e., included patients from countries within and not within the Asia-Pacific region); however, three trials included patients from outside Asia-Pacific countries [30,36,76] and nine trials included patients exclusively from Asia-Pacific countries [31][32][33][34][35]37,38,71,72]. Nine of the 12 global trials (i.e., CELESTIAL, CheckMate 040, GO30140, IMbrave150, KEYNOTE-240, REFLECT, REACH, REACH-2, RESORCE) provided subgroup analyses for patients from the Asia-Pacific region in the original publication or in separate analyses [87][88][89][90][91][92][93][94]. Inclusion criteria for general disease characteristics, such as the ECOG performance status, BCLC stage, and Child-Pugh class, were largely similar within the trials (Supplementary Table 10); however, some studies allowed patients with more advanced disease stage to enroll. These included four trials that allowed patients with ECOG performance status of 2 to enroll (including early trials for sorafenib) [35][36][37][38] and six trials that allowed patients with Child-Pugh class B to enroll [30,31,35,38,44,76]. Notably, all 2L RCTs had highly similar general inclusion criteria regarding ECOG performance status of 0 or 1, BCLC stage B or C, and Child-Pugh class A; patients with more severe disease were excluded from 2L RCTs. The time since loco-regional treatment was at least 4 weeks for any study mentioning this criterion. Specific disease characteristics assessed included hepatic encephalopathy, ascites, and bleeding risk; these criteria were either not reported or not allowed within varying timeframes of randomization or treatment initiation. Excluded viral infections were also mostly similar, with a co-infection of hepatitis B and C being the only hepatitis condition not allowed (Supplementary Table 11). In the 1L setting, up to 90% of patients had hepatitis B and 79% of patients had hepatitis C; the maximum proportions were lower in the 2L setting (56 and 45%, respectively). Most of the trials assessing patients in the 2L setting required that the only prior treatment was sorafenib; several required that no prior IO treatments had been administered (Supplementary Table 12). Patient & disease characteristics Minimum and maximum values for each trial arm are used to describe the demographic and disease characteristics shown in Figure 1. Patient age and sex were fairly consistent between studies, with single-arm trials and RCTs in 1L and 2L settings enrolling patients aged 59-68 years (mean) or 51-73 years (median); 71-93% of patients were male. The proportion of patients from the Asia-Pacific region varied widely. Trials included a wide range of patients with vascular invasion (0-88%) and extrahepatic spread (21-86%). There appeared to be a relationship between the proportion of patients enrolled from Asia-Pacific versus the prevalence of hepatitis B, but not the prevalence of hepatitis C (Supplementary Figure 2), which was expected given the regional epidemiology of hepatitis B [95]. Performance status & disease stage The ECOG performance status, BCLC stage, and Child-Pugh class of patients from each trial arm were also examined for similarity. The ranges of enrolled patients who were ECOG 0 (i.e., fully active), ECOG 1 (i.e., able to perform light/sedentary activity) and ECOG 2 (i.e., unable to perform work activities but capable of self-care) were 25-100%, 17-75% and 0-8%, respectively, in the 1L setting ( Figure 2). Notably, two of the early RCTs assessing sorafenib [37,38] enrolled a much lower proportion of patients who were ECOG 0 compared with other 1L single arm trials and RCTs; these two trials also enrolled patients who were ECOG 2. Aside from these two trials [37,38], the ECOG performance status of patients in 1L trials was largely similar. In the 2L setting, the ranges of enrolled patients who were ECOG 0 and 1 were 52-91% and 9-48%, respectively, and appeared broadly similar. Seventeen of the 24 trials in the similarity assessment reported patients' BCLC stage at baseline. Patients staged BCLC C (i.e., advanced disease) ranged 43-96% in the 1L setting (Supplementary Figure 4). Two trials in the 1L single-arm setting had a lower proportion of BCLC C patients than most other 1L trials, while the Sorafenib AP trial enrolled BCLC C patients almost exclusively [33,35,37]. Aside from these three trials, the BCLC staging appeared broadly similar in the 1L setting. In the 2L setting, BCLC C patients composed 76%-94% of patients; in this regard, 2L trials appeared broadly similar. Patients with a Child-Pugh score of 5-6 (i.e., Child-Pugh A, considered well-compensated) varied throughout the trial arms, ranging 43-100% in the 1L setting (Supplementary Figure 5) enrolled >20% Child-Pugh B patients [30][31][32]35]. One 1L RCT [38] enrolled only Child-Pugh B or C patients; other 1L RCTs enrolled <10% Child-Pugh B patients and were therefore considered similar in Child-Pugh class. In the 2L setting, one trial enrolled 55% Child-Pugh B patients [71]; all other 2L trials enrolled >90% Child-Pugh A patients and were considered similar. In summary, some trials in the 1L setting included patients who had more advanced disease based on ECOG performance status [30,37,38], BCLC stage [37], and Child-Pugh class [30][31][32]35,38]. All these trials assessed sorafenib as single arm or placebo-controlled RCTs. Patients in 2L trials were more similar in terms of these characteristics, aside from one single-arm trial for lenvatinib [71]. Efficacy results Survival The median overall survival (mOS) ranged from 3.5 to 19.2 months in the 1L setting and 7.3 to 22.2 months in the 2L setting. In 1L RCTs, the median (range) of mOS values was 13.4 (4.0-19.2) months in active treatment arms and 4.2 (3.5-7.9) months in placebo/BSC arms ( Figure 3); in 2L RCTs, the median (range) of mOS values was 10.2 (8.5-13.9) months in active treatment arms and 7.8 (7.3-10.6) months in placebo arms (Supplementary Figure 6). The association between

disease advancement and survival in the 1L setting was explored by plotting the proportion of ECOG 0 patients, BCLC B patients and Child-Pugh A patients versus the mOS for the trials reporting both these outcomes (Supplementary Figure 9). Based on qualitative review, the differences in mOS observed in 1L trials may be partially correlated with the underlying ECOG performance status of patients enrolled. The 1L and 2L RCTs also reported the hazard ratio (HR) and 95% confidence interval (CI) of survival based on the Kaplan-Meier curves of PFS and OS. In the 1L setting, all trials assessing sorafenib versus placebo/BSC found a significant improvement in both PFS and OS in favor of sorafenib [36][37][38]. Similarly, the combination of atezolizumab + bevacizumab was associated with an improvement in PFS and OS versus sorafenib [42,43]. The trial assessing lenvatinib and sorafenib demonstrated lenvatinib's non-inferiority [39], while the comparison of nivolumab versus sorafenib did not show a significant improvement in either outcome [40]. Last, the comparison of atezolizumab + bevacizumab versus atezolizumab alone found an improvement in PFS, but OS was not reported ( Figure 4) [44]. In the 2L setting, cabozantinib and regorafenib were associated with significant improvements in PFS and OS over placebo [77,79]; results for ramucirumab were mixed [78,80], and pembrolizumab did not significantly lower PFS and OS per the trial's specified criteria [81,82] (Supplementary Figure 7). Other time-to-event outcomes, such as the median duration of therapy (mDOT), median duration of response (mDOR) and median time to progression (mTTP) were less frequently reported than mOS and mPFS (Supplementary Table 13). NMA feasibility In addition to similar patient characteristics, the feasibility of performing an NMA relies on a network of treatments connected by common comparators assessed in RCTs and common outcomes reported within each network [16]. We examined the outcomes that could be indirectly compared through an NMA for all the treatments recommended by the NCCN Guidelines, based on the available RCT data and network structure (Supplementary Figure 12). Notably, FOLFOX could not be compared with other 1L treatments in an NMA because the in-trial comparator was doxorubicin [61]; since no other recommended treatments have been compared with doxorubicin in a trial setting, FOLFOX could not integrate into the treatment network. Similarly, nivolumab ± ipilimumab could not be compared with other 2L treatments in an NMA because its existing trials were single-arm studies. Based on the currently available RCT data, key efficacy outcomes (i.e., OS, PFS, ORR and DCR) could be indirectly compared via NMA for the majority of recommended treatments in both 1L and 2L RCT settings. Less commonly reported efficacy outcomes such as TTP and DOT could also be indirectly compared via NMA for most treatments. DOR could be indirectly compared in 2L RCTs only. Serious and discontinuation-related TEAEs could be indirectly compared via NMA in both 1L and 2L settings, but grade ≥3 TEAEs could be indirectly compared in the 1L setting only. Discussion The objective of this study was to determine the similarity of trials assessing recommended systemic treatments for advanced HCC. In general, we found that trials were largely similar based on design and inclusion criteria within respective 1L and 2L categories, aside from some older single-arm trials and placebo-controlled RCTs assessing sorafenib [30][31][32]35,37,38]. The design and patients' disease stage in 2L RCTs were also considered broadly similar and suitable for quantitative comparison. Trials differed in some key patient characteristics. While age, sex, and prevalence of vascular invasion and extrahepatic spread were relatively similar across trials, the proportion of patients infected with hepatitis B and levels of AFP varied widely. Hepatitis B infection and AFP level may represent treatment effect modifiers and should be evaluated closely in any potential meta-analyses. For example, a review of over 11,000 patients with HCC in USA found that hepatitis B-related cases were associated with a significantly lower risk of mortality compared with cases related to alcohol, metabolic disorder and multiple etiologies [96]. Similarly, AFP is a biomarker often elevated in patients with HCC who are more likely to have viral etiology, cirrhosis, larger or multiple tumors and vascular invasion [97]. Ultimately, the best approach may be to perform quantitative studies using analyses of patient subgroups with underlying disease characteristics or comorbidities as reported by individual study authors to control for specific patient characteristics [13]. This method enables more nuanced information for healthcare decision makers to make targeted recommendations for patients based on the presence or absence of specific disease characteristics, such as bleeding risk, hepatitis B infection or AFP levels. We also found generally robust reporting of key clinical efficacy outcomes, including mPFS, mOS and tumor response. There is broad consensus that mOS is the primary efficacy outcome of interest in determining a preferred treatment. Other efficacy outcomes, such as mPFS, mTTP and ORR, could be analyzed using NMA with relatively robust networks. However, there is some disagreement within the field on the usefulness of surrogate end points such as mTTP, mPFS and ORR [98][99][100][101], particularly for IO agents, which demonstrate different relationships between tumor response and disease progression than TKI and VEGFRI drugs [102,103]. In other words, although analyses of several efficacy outcomes may be possible in an NMA, the interpretation of analyses on surrogate end points must be performed with a deep understanding of the treatments involved and clinical implications of the findings. Indirect comparisons of serious TEAEs or TEAEs leading to discontinuation are also possible via NMA; however, we did not assess the practicality of indirectly comparing individual adverse events, and these analyses are typically less reliable due to heterogenous reporting and relative rarity. Limitations The purpose of this study was to qualitatively review the literature; therefore, no quantitative comparisons can be drawn from this study. We noted the limited information or absence of reporting by trial authors on patient characteristics and outcomes of interest, particularly when the primary study involved a clinical trial registry entry or conference abstract with minimal information. For example, mRECIST criteria are considered to be more sensitive than RECIST criteria in assessing tumor response and predicting survival [104]; however, as noted in the Results, mRECIST criteria were used less frequently than RECIST and would therefore be more difficult to assess for similarity in a quantitative analysis. Similarly, the reporting of TEAEs and TRAEs differed for specific safety events. We also did not assess the comparability of patient-reported outcomes, such as quality of life. However, we found that very few trials reported quality of life; therefore a robust assessment of similarity would not be possible. There are several treatments being assessed as monotherapies or in novel combinations in patients with HCC [105,106]. This review is inherently limited by the rapidly evolving trial landscape in this field. However, our review demonstrates the need for treatments with improved efficacy and safety profiles, as we found that the typical overall survival for patients using recommended treatments in the RCT settings was approximately 12 months, and approximately one in seven patients discontinued trial participation (based on mOS median values of 13.4 months and 10.2 months in 1L and 2L settings, respectively, and median discontinuation due to TEAEs of 13%). The relationship between performance status and survival (as well as other baseline disease characteristics) has been noted in both clinical trial and real-world studies [107,108]. The disease stage of patients enrolled in the clinical trials reviewed in this study appeared be less severe (e.g., ECOG 0, Child-Pugh class A) than the majority of patients who typically present with advanced HCC [11]. Conversely, there is limited guidance available on the treatment of patients with patients with more advanced disease (e.g., Child-Pugh class A) [109], since few interventional and observational studies have included patients with more severe disease [110][111][112][113]. Furthermore, most trials excluded patients based on other characteristics, such as bleeding risk and chronic infections; patients excluded for these reasons may experience worse outcomes with these treatments in real-world settings. Finally, treatments may be contraindicated or show worse safety profiles in patients with underlying comorbidities (e.g., hypertension) that were not explored here. In summary, efficacy and safety outcomes from individual trials and subsequent indirect comparisons may not be completely representative of the experience of HCC patients in the real world; treatment efficacy and safety should be considered in patients with more severe disease. Conclusion Patients with advanced HCC who are eligible for systemic treatment exhibit a wide spectrum in physical function and disease severity, as well as other key patient characteristics. Our objective was to assess the similarity and outcomes reported by trials that measured the efficacy and safety of systemic HCC treatments recommended by the NCCN Guidelines. To our knowledge, this study is first similarity assessment of trials that include only treatments recommended by the NCCN Guidelines in both 1L and 2L settings. After examining trial inclusion criteria, patient characteristics, and reported outcomes, we determined that randomized trials assessing recommended treatments for HCC are largely similar and appropriate for indirect comparisons, and several efficacy and safety outcomes are possible to analyze in a quantitative manner. Special consideration should be applied in determining the appropriate outcomes for analyses, as well as potential subgroups based on demographic and clinical characteristics, that will be robust, clinically relevant, and enable decision-making in healthcare. Future perspective There are several treatments under investigation for the treatment of advanced HCC, as well as trials assessing novel treatment combinations and/or sequencing of currently recommended treatments. These treatments, in combination with more widespread use of precision medicine approaches, have the potential to slow disease advancement and improve patient survival in the real world. Supplementary data To view the supplementary data that accompany this paper please visit the journal website at: www.futuremedicine.com/doi/suppl/10.2217/hep-2021-0003 Author contributions A Aly, F Benavente and J-D Rueda contributed to study conception, design and revisions to the manuscript; S Ronnebaum and D Patel contributed to study design, data analysis, drafting and revision of the manuscript. relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. Financial & competing interests disclosure No writing assistance was utilized in the production of this manuscript. Open access This • We performed an systematic literature review to identify all single arm and randomized trials assessing recommended treatments. • After reviewing 5677 records, a total of 57 references (representing 50 trials) were included in the systematic literature review. Of these, 31 records (representing 24 trials) were further assessed for similarity within 1L and 2L groups. • We found that trials in the 1L setting were largely similar, except that some of the early trials for sorafenib included patients with more severe disease, and that trials in the 2L setting were also generally similar. • The median (range) of mOS values from RCTs in 1L patients was 13.4 (4.0-19.2) months (active treatments) and 4.2 (3.5-7.9) months (placebo), and the median (range) of mOS values from RCTs in 2L patients was 10.2 (8.5-13.9) months (active treatments) and 7.8 (7.3-10.6) months (placebo). • Across all trials, the median (range) of patients experiencing grade ≥3, serious, and discontinuation-related treatment-emergent adverse events was 57 (range: 19-79%), 43 (range: 10-56%) and 13% (range: 0-56%), respectively. • Our review demonstrates the feasibility of quantitative comparisons of recommended treatments, as well as the survival, tumor response, and safety experienced by patients in 1L and 2L trial settings. Outcomes experienced by real-world patients may reflect more advanced disease stage, comorbidities and other factors. Combined Treatment of Heterocyclic Analogues and Benznidazole upon Trypanosoma cruzi In Vivo Chagas disease caused by Trypanosoma cruzi is an important cause of mortality and morbidity in Latin America but no vaccines or safe chemotherapeutic agents are available. Combined therapy is envisioned as an ideal approach since it may enhance efficacy by acting upon different cellular targets, may reduce toxicity and minimize the risk of drug resistance. Therefore, we investigated the activity of benznidazole (Bz) in combination with the diamidine prodrug DB289 and in combination with the arylimidamide DB766 upon T. cruzi infection in vivo. The oral treatment of T.cruzi-infected mice with DB289 and Benznidazole (Bz) alone reduced the number of circulating parasites compared with untreated mice by about 70% and 90%, respectively. However, the combination of these two compounds decreased the parasitemia by 99% and protected against animal mortality by 100%, but without

providing a parasitological cure. When Bz (p.o) was combined with DB766 (via ip route), at least a 99.5% decrease in parasitemia levels was observed. DB766+Bz also provided 100% protection against mice mortality while Bz alone provided about 87% protection. This combined therapy also reduced the tissular lesions induced by T. cruzi infection: Bz alone reduced GPT and CK plasma levels by about 12% and 78% compared to untreated mice group, the combination of Bz with DB766 resulted in a reduction of GPT and CK plasma levels of 56% and 91%. Cure assessment through hemocultive and PCR approaches showed that Bz did not provide a parasitological cure, however, DB766 alone or associated with Bz cured ≥13% of surviving animals. Introduction Discovered by the Brazilian physician Carlos Chagas one century ago, Chagas disease (CD) is a zoonosis caused by kinetoplastid flagellated Trypanosoma cruzi [1]. It is well known that CD is an endemic illness in poor areas of 15 developing countries of Latin America, affecting about 12 to 14 million people. Less well known is that CD is becoming a health problem in non-endemic areas such as Europe and United States largely due to the migration of infected people to these regions [2][3][4]. Due to the lack of an efficient therapy, mainly for chronic chagasic patients, and since it has been considered by many pharmaceutical industries to have limited economical potential, CD has been designated a neglected tropical disease [5]. The main route of transmission of T. cruzi infection to humans is through the feces of blood-sucking triatomine insects but other routes also exist including blood transfusion, transplacentally, organ transplantation, laboratory accidents and oral ingestion of contaminated food [6][7][8][9][10][11][12]. CD is the leading cause of infectious myocarditis worldwide, which is one of its most serious and frequent clinical manifestations observed during the chronic phase of the disease that appears in about 20-40% of infected individuals years or decades after the acute infection [13][14][15]. The available treatment is based on two nitroheterocycles: nifurtimox (4-[(5nitrofurfurylidene)-amino]-3-methylthio morpholine-1,1-dioxide), a nitrofuran produced only in El Salvador by WHO-Bayer, and benznidazole (N-benzyl-2-nitro-1-imidazoleacetamide), a nitroimidazole produced by LAFEPE-Brazil. Both are indicated for the treatment of all acute and early chronic cases, exhibiting about 60-80% efficacy [16]. Nevertheless, neither compound is highly effective against the late chronic phase (about 20% cures), both require long-term therapy in addition to displaying side-effects that can lead to interruption of the treatment. These deficiencies justify the search for new chemotherapeutic options [5,17,18,19]. Many studies have demonstrated the excellent activity of aromatic diamidines (AD), pentamidine and related compounds, against many pathogens, such as bacteria, fungi and protozoa [20]. Although the exact mechanism of action is still not precisely known, it has been proposed that the binding of these cationic molecules in the DNA minor groove, mainly at AT-rich regions, contributes, at least in part, to their effect upon trypanosomatids [16,[21][22]. We have previously reported the in vitro and in vivo activity of AD and analogues such as arylimidamides (AIA) upon T.cruzi [23][24][25][26]. A recent study demonstrated that the AIA, DB766 shows superior efficacy than Bz upon different parasite strains, including those naturally resistant to Bz [27]. Combination therapy represents a promising approach for the enhancement of drug efficacy since it (i) allows the use of at least two compounds that may act upon different cellular elements and metabolic pathways (ii) may reduce drug concentrations and number of doses thus contributing to the lowering of toxic effects, and (iii) may minimize the risk of drug resistance [28]. For the reasons stated above, the combination of different trypanocidal compounds merits exploration [16,[28][29]. Our present goal is to evaluate in vivo the combined effect of Bz with the prodrug DB289 and with the arylimidamide DB766, to determine if a scheme of therapy with these drugs could reduce toxicity and improve efficacy in an animal model for Trypanosoma cruzi-infection. Compounds The aromatic diamidine DB289 (pafuramidine maleate; 2,5bis [4-(N-methoxyamidino)phenyl]furan monomaleate) and the arylimidamide DB766 ( Fig. 1) were synthesized according to methodology previously reported by us [30][31][32]. A DB289 stock solution was made in a solvent consisting of sterile distilled water (99.4%), Tween 80 (0.1%), and ethanol, which was freshly prepared immediately before use each day. The route of administration used was oral gavage. DB289 and DB766 were dissolved in DMSO and then freshly diluted with sterile distilled water before use by intraperitoneal (ip.) or p.o. routes. The stock solution of Benznidazole (N-benzyl-2-nitroimidazol acetamide, Rochagan, Roche) was prepared in sterile distilled water with 3%Tween 80, and before use was diluted in sterile distilled water for p.o. administration. Parasites Bloodstream trypomastigotes (BT) of the Y strain were used throughout and were harvested by heart puncture from T. cruziinfected Swiss mice on the day of peak parasitemia, as described [33]. Mice infection Male Swiss mice (20-24 g) were obtained from the animal facilities of the Oswaldo Cruz Foundation (CECAL, Rio de Janeiro, Brazil). Mice were housed at a maximum of 8 per cage and kept in a specific pathogen free (SPF) room at 20-24uC under a 12/12 h light/dark cycle and provided with sterilized water and chow ad libitum. The animals were allowed to acclimate for 7 days before starting the experiments. Infection was performed by ip injection of 10 4 BT. Age-matched non-infected mice were maintained under identical conditions. Treatment schemes The animals were divided into the following groups (Table 1): uninfected (non-infected and non-treated); untreated (infected but non-treated); and treated (infected and treated with the different compounds combined or not with Bz) [24][25]. Drug therapy was performed by 20 daily consecutive doses (ip. and p.o., Table 1), starting at the onset of the parasitemia (5 dpi). In all assays, only mice with positive parasitemia were used in the infected groups. Parasitemia, mortality rates and ponderal curve analysis The level of parasitemia was checked by the Pizzi-Brener method. Mice were individually checked by direct microscopic counting of parasites in 5 mL of blood [34]. The mortality rates were checked daily until 60 dpi and expressed as cumulative mortality (% CM). Body weight was evaluated from 0 up to 60 dpi, and expressed as percentage of weight variation [24][25]. Histopathological analysis At 14 dpi (peak of cardiac parasitism and inflammation in this experimental model as described in de Souza et al., 2006 [24], the heart tissues were removed, cut longitudinally, rinsed in ice-cold phosphate buffered saline (PBS), and fixed in Millonig-Rosman solution (10% formaldehyde in PBS). The tissues were dehydrated and embedded in paraffin. Sections (3 mm) were then stained with hematoxylin-eosin and analyzed by light microscopy. The number of amastigote nests was determined in at least 60 fields (total magnification, 406) for each slide. The mean number of amastigotes' nests per field was obtained from at least three mice per group, with three sections from each mouse. Biochemical analysis At 14 day post infection (dpi), mice blood was collected and immediately submitted to analysis for biochemical determination of plasma tissular markers including glutamate pyruvate transaminase (GPT) and total creatine kinase (CK) using the Reflotron System (Roche Diagnostics, F. Hoffmann-La Roche Ltd, Basel, Switzerland) [35]. Cure assessment Cure criteria were based on two parasitological methods: polymerase chain reaction (PCR) and hemoculture assays which detect kDNA minicircle specific sequences or the parasite itself, respectively. Animals presenting negative results by both tests were considered cured. Briefly, after 40 days of drug treatment, about 700 mL of blood were collected from the heart of anesthetized mice and then 500 mL and 200 mL were used for PCR and hemoculture analysis, respectively. For PCR, the blood was diluted in 1:3 volume of guanidine solution (6 M guanidine-HCl/0.2 M EDTA), and heated for 90 s in boiling water in order to cleave the parasite kDNA network [36] and the PCR performed using the primers: (59AAATAATGTACGGG(T/G)GAGATGCATGA39) and (59GGTTCGATTGGGGTTGGTGTAATATA39), which amplify a 330 bp sequence from the kinetoplast DNA (aprox 120 000 copies per parasite), as previously described by Wincker et al. (1994) [37]. The PCR was carried out using a GeneAmpH PCR System 9700 (Applied Biosystems) as follows: one step at 94uC for 3 min (to activate the Taq DNA polymerase), 2 cycles at 98uC for 1 min and 64uC for 2 min, 38 cycles at 94uC for 1 min and 64uC for 1 min, followed by a final extension at 72uC for 10 min. The amplification products were detected on a 1.5% agarose gel electrophoresis following staining with ethidium bromide (5 mg/mL). For hemoculture, 200 mL of blood was added to 5 mL LIT medium and incubated at 28uC for 60 days, being weekly examined by light microscopy to detect epimastigote forms [38]. All procedures were carried out in accordance with the guidelines established by the FIOCRUZ Committee of Ethics for the Use of Animals (protocol approved -CEUA 0099/01). Results Since (i) in a previous study we found that a phenyl-substituted analogue of furamidine gave a trypanocidal effect upon a T. cruzi infection in vivo [23], and (ii) oral administration of only two doses at onset (5 dpi) and at parasitemia peak (8 dpi) of 25 and 50 mg/ kg/day of the furamidine prodrug (DB289) resulted in about a 60% decrease of parasitemia ( Fig. 2A), we evaluated the combination of DB289 with Bz to determine if an enhancement of efficacy against the parasite was observed. Our results show that although treatment with 25 mg/kg/day of DB289 or 50 mg/kg/ day of Bz, both alone, lowered the parasitemia peak levels by about 70% and 90%, the combined treatment reduced the number of circulating parasites at 8 dpi by more than 99% ( Fig. 2B; Table 2). Mice survival rates of about 85% and 40% were found for DB289 treated and untreated mice groups, respectively. The combination of the prodrug DB289 with Bz resulted in 100% animal survival (Fig. 2C). At three weeks post infection when the highest body weight lose is observed in this T. cruzi experimental model [27], both DB289 alone and Bz+DB289 showed considerable lose of mice body weight similar to or even as high as that for the untreated animals ( Fig. 2D). At 60 dpi, the group treated with only DB289 still showed high rates of weight loss, even more than the untreated mice group (Fig. 2D). Cure assessment evaluated by both hemoculture and PCR did not reveal a parasitological cure in any mice groups (Table 1). Since our recent study demonstrated the high efficacy of DB766 upon T. cruzi infection in vivo and in vitro [27], its combination with Bz was evaluated. Using the same therapy scheme as above, we found that the treatment of infected mice with 50 mg/kg/day of Bz or DB766 (ip.) or with DB766 (ip.) plus Bz (50 mg/kg/day each) resulted in decreases in mice parasitemia by about 90%, .96% and 100%, respectively, as compared to the untreated mice group (Fig. 3A, Table 2). The analysis of cumulative mortality revealed that while DB766 treated, Bz treated and untreated groups resulted in 50, 12.5 and 100% of death, the combined therapy of Bz plus DB766 resulted in 100% of protection, avoiding animal death (Fig. 3B). We found at 20 dpi, that the Bz treated and the combined therapy group showed partial restoration of the mice body weight compared to uninfected mice, however, the group treated with DB766 alone displayed even higher body weight loss than that of the untreated mice group (Fig. 3C). In these experimental groups, at the 14 dpi, we found that the T. cruzi infection led to an increase in biochemical markers, resulting in a rise of 2-and 29-fold for GPT and CK levels, respectively (Fig. 3 D-E). Bz treatment alone did not reverse the hepatic damage induced by parasite infection [35], however, DB766 alone or in combination with Bz produced a decrease of about 69% and 57% in GPT levels when compared to untreated control (Fig. 3E). Muscle lesions, as evaluated by CK plasma levels were decreased with Bz treatment by about 70%. Treatment with DB766 alone or combined with Bz resulted in reductions greater than 91% when compared to untreated control (Fig. 3D). Histopathological assays confirmed the high efficacy of DB766 combined with Bz which resulted in 100% reduction in cardiac parasitism as compared to untreated animals (data not shown). Cure assessment also revealed that the administration of 50 mg/ kg/day DB766 (ip.) plus 50 mg/kg/day Bz (p.o.) resulted in a 13%

parasitological cure, as evaluated by both hemocultive and PCR analysis (Table 1). Since DB766 shows high trypanocidal efficacy against T. cruzi in vivo on oral administration at 100 mg/kg dose [27], we evaluated another combination dosing regimen using sub-optimum oral doses of both DB766 (50 mg/kg/day by p.o.) and Bz (50 mg/kg/ day by p.o.). DB766 (p.o.) alone reduced parasitemia by only about 20% (Fig. 4A) and the mortality rates were reduced by about 25% (data not shown). The administration of Bz plus DB766 (both p.o) decreased the parasitemia by 54% (Fig. 4A, Table 2), and gave a 100% mice survival rate similar to that of Bz alone (data not shown). At 20 dpi, both Bz alone and the combined therapy provide a partial recovery of mice body weight, however DB766 alone displayed similar body weight loss to that of untreated animals (Fig. 4B). In the oral treatment studies, the only parasitological cure noted by the hemoculture and PCR methods was for the group treated with 50 mg/kg/day DB766 (p.o.) ( Table 1). Discussion Since the introduction of Nifurtimox and Bz (1960)(1961)(1962)(1963)(1964)(1965)(1966)(1967)(1968)(1969)(1970), despite the urgent need for new CD therapies, only allopurinol and a few azoles have moved to clinical trials, possibly due to (i) limited investment in this research field, (ii) an earlier mistaken concept that during the later phase of CD parasitism was absent and thus was not relevant to disease outcome and pathogenesis, and (iii) the absence of universal standardized protocols for drug screening [16]. The current profile for development of new drug candidates for CD includes 1) efficacy against different stocks, 2) efficacy against parasite forms relevant to the infection of mammalian hosts for both the acute and chronic phases, 3) oral administration in a minimum number of doses, 4) reduced toxicity, 5) yield high levels of tissue accumulation and long terminal half lives and 6) low costs [16,[39][40]. In this study using a murine model of T. cruzi acute infection, the trypanocidal efficacy of pafuramidine (DB289) and the arylimidamide DB766 either alone or in combination with benznidazole was evaluated over a relatively short period of treatment (20 days) employing both intraperitoneal and oral administration. Diamidine-containing compounds such as pentamidine and furamidine are DNA minor groove binders with broad-spectrum activity upon different species of human and veterinary pathogens [20]. DB289 is the orally active prodrug of DB75 (furamidine) that exerts microbicidal effects upon different pathogens including Trypanosoma brucei, Pneumocystis jiroveci and Plasmodium falciparum, [41][42][43][44][45]. Interestingly, DB289 showed good oral efficacy in murine models of human African trypanosomes (HAT) suggesting that sufficient quantities are absorbed from the mouse gastrointestinal tract, delivering this dicationic molecule across the gut mucosa [43,46]. Similar to the therapy of HAT [43], oral efficacy is a desirable feature for treatment of CD. However, in contrast to the studies performed with T. brucei [43], DB289 alone was not very effective against T. cruzi. This difference may be due to the fact that in contrast to the African trypanosomes, T. cruzi has intracellular stages living inside the cytoplasm of host cells, which represents an additional obstacle for drug access and delivery. We have found that the combination of Bz with DB289 improved the efficacy of the diamidine by reducing parasitemia and resulted in protection against mortality. In addition, this combined therapy provided a 9-fold enhancement of activity compared to that of Bz alone. Despite showing efficacy in Phase III clinical trials against HAT and initial indications of low toxicity in African, Asian, Caucasian and Hispanic populations, further studies with DB289 revealed considerable side effects leading to its withdrawal from advanced human trials [16,43]. In fact, when higher doses of DB289 ($100 mg/kg/day) were evaluated against T. cruzi infection in vivo (unpublished data of DGJB), higher numbers of circulating parasites and mortality rates were noticed as compared to untreated mice, perhaps a consequence of compound toxicity. A previous study by our group demonstrated the beneficial effect of DB766 upon T.cruzi in vivo: a ten-day regimen of treatment reduced both blood and cardiac tissue parasitism, resulting in 90-100% protection against death even with an infection with naturally resistant T. cruzi strain (Colombian) to benznidazole [27]. Also, this AIA ameliorated electric heart alterations and reduced hepatic and heart lesions induced by the parasite infection [27]. Despite the promising trypanocidal effects of this AIA via ip (up to 50 mg/kg/day) and by p.o (100 mg/kg/ day) routes which showed efficacy similar to Bz (100 mg/kg/day), DB766 (as well as Bz) failed to provide a parasitological cure [27]. This result may be a consequence of the highly stringent protocol employed (maximum of 10 days of drug administration). In fact, previous studies performed in T. cruzi-infected murine models with Bz and azoles reported high rates of parasitological cure only when dosing was continued for longer periods ($40 days) [47][48][49], which supports using longer periods of therapy for our combination studies. Our studies show that AIA are more active against T.cruzi than diamidine compounds [26,50]. The greater activity may be related to differences in their physical properties since AD are highly basic molecules with pK a values near 11 while AIA pK a values are near 7. At physiological pH, AD are protonated and thus cationic molecules while AIA are essentially neutral molecules enabling their passive diffusion through the plasma membranes of both parasites and host cells. This large difference in properties likely affects absorption and distribution and may play an important role in the different activity of these two classes of compounds. Our data showed that while DB766 alone reduced parasitemia giving a superior result to that of Bz, the combination of Bz and DB766 leads to undetectable parasitism, thus improving the efficacy of both compounds, especially Bz, whose potency was increased at least 20-fold. The improved activity of the combined therapy may reflect different targets (still incompletely understood for both) and/or effects upon different parasite forms. As reported, intracellular parasites must be considered the main parasite stage for drug targeting in CD since T. cruzi is an obligatory intracellular parasite [28]. As previously reported, DB766 displayed oral efficacy against an experimental T. cruzi infection when high but non-toxic doses (100 mg/kg/day) are employed [27]. However, when we evaluated the p.o. treatment with Bz and DB766 using sub-effective doses of both compounds (50 mg/kg/day each), the combined therapy only enhanced the activity of the AIA by 1.8 fold. The combined therapy showed a lower effect on parasitemia (but not on mortality rates) as compared to Bz treatment alone, suggesting an antagonistic effect that deserves to be further explored. One out of three surviving mice treated with DB766 by p.o. was cured as assessed by both hemocultive and PCR analysis. Although we did not find a considerable reduction in the mean parasitemia in this mice group, the cured animal was the one that displayed the lowest level of circulating parasitism, reaching undetectable parasitism (by light microscopy counting) after 23 dpi. Although no visible adverse effects were noticed for DB289 and DB766, when they were used alone, both increased the cachexia induced by the parasite infection. This effect raises the possibility of drug toxicity and/or up-regulation of inflammatory mediator levels such as TNF-alpha that is implicated in loss of mice weight during T. cruzi acute infection [51]. Although no detectable acute toxicity was observed in mice treated up to a cumulative dose of 200 mg/kg/day of DB766 [27] and 100 mg/kg/day DB289 [43], and our data showed reduced hepatic and muscle lesions during the treatment of infected mice with DB766, a detailed biochemical and histopathological analysis is needed to clarify this matter. The measurement of pro and anti-inflammatory cytokines in the plasma of infected and treated mice would contribute to the understanding of the possible role of these mediators upon drug toxicity and efficacy. Although no data is available for AIA, some studies suggest a regulatory effect by pentamidine upon proinflammatory cytokines [52][53]. Additionally, Bz down-regulates the synthesis of TNF-alpha by murine stimulated macrophages [54], ameliorates LPS-induced inflammatory response in mice by decreasing peak levels of this serum cytokine [55] and markedly reduces the production of pro-inflammatory cytokines and NOderived metabolites in experimentally T. cruzi-infected rats [56]. These data may explain the weight recovery found in Bz-infected treated mice as compared to untreated mice since this proinflammatory mediator is strongly expressed in T. cruzi-infected mice [56][57], and is thought to be related to mice weight loss [51]. In our studies, we found a correlation between mice cachexia and mortality rates, including in the DB766 groups (ip. and p.o.) that may explain the lower protection against animal mortality in the animal groups that only received the AIA. In conclusion, this study has shown that DB766 is much more potent in this mouse experimental model of T. cruzi infection than DB289 and that the trypanocidal activity is improved by combination therapy of both AD and AIA with Bz. Our data support additional studies with other diamidines and AIAs alone or in combination with other drugs with the goal of identification of new candidate therapies for the treatment of Chagas disease. Adjuvant radiotherapy for the treatment of stage IV rectal cancer after curative resection Supplemental Digital Content is available in the text Introduction The incidence of colorectal cancer has increased over the past decade, and it is currently the third-most common malignancy worldwide. [1] Approximately one-third of colorectal cancer patients present with lesions in the rectum. [1] Despite popular screening tests for early detection and treatment, approximately 20% of patients have metastatic disease at the time of diagnosis, which is associated with a poor outcome (5-year survival rate of 11.9%). [2] Treatment outcomes of patients with stage IV rectal cancer have improved with recent advances in protocols and refinement of surgical techniques. Additionally, the 5-year survival rate increased by 25% to 35% when curative resection of both primary and metastatic lesions was achieved. [3][4][5] As the survival of patients with stage IV resectable rectal cancer has improved, the issue of loco-regional control has gained increasing attention. Although most recurrences of colorectal cancer occur in distant organs, including the liver and lung, within the first 2 years following resection, the rate of pelvic failure has been reported to be approximately 30% to 35%. [5,6] Preoperative or postoperative radiotherapy (RT) has been reported to be effective for loco-regional control of advanced rectal cancer. It was also demonstrated that adjuvant chemoradiotherapy (CRT) significantly reduced the rate of local recurrence compared with adjuvant chemotherapy alone. The National Comprehensive Cancer Network guidelines recommended CRT as the standard therapy for patients with completely resected stage II/III rectal cancer. [7] However, despite the proven benefits of RT for the treatment of stage II/III rectal cancer, there is a lack of data regarding its potential role in the treatment of stage IV rectal cancer. [8,9] Several studies have reported the oncologic outcomes of patients who received pelvic RT for the treatment of stage IV rectal cancer; however, the results were inconsistent and the study population was both heterogeneous and subject to selection bias. [10,11] In the present study, we evaluated the potential impact of postoperative pelvic RT on the oncologic outcomes of patients with stage IV rectal cancer who underwent curative resection in 1 of 3 centers using propensity score matching analysis. A systematic review with meta-analysis was also performed, which may provide a more comprehensive overview of the impact of this treatment. Patients The medical records of 176 stage IV rectal cancer patients who underwent curative resection at the National Cancer Center, Seoul National University Hospital, or Seoul National University Bundang Hospital between August 2001 and December 2011 were retrospectively reviewed. All patients were >18 years of age, had histologically proven rectal adenocarcinoma located within 10 cm of the anal verge and synchronous metastasis, and had no history of other malignancies. Only stage IV rectal cancer patients who underwent curative resection for primary and metastatic tumors following adjuvant radiotherapy were included in the study. Primary rectal carcinomas underwent surgery according to the principle of total mesorectal excision. Patients who underwent palliative surgery had a pathological diagnosis other than adenocarcinoma, had preoperative radiotherapy, or had recurrent disease were excluded. Although there were no definite indications for pelvic RT in stage IV rectal cancer patients, 51 patients received pelvic RT at the discretion of the physician, which was reviewed by a multidisciplinary committee.

Pelvic RT was delivered to the entire pelvis at a dose of 45 Gy in 25 fractions, and was followed by a boost to the primary tumor of 5.4 Gy in 3 fractions over 5.5 weeks. All patients underwent computed tomography simulation for three-dimensional conformal RT. The superior border was placed at L5-S1 and the inferior border >3 cm caudal to the tumor level. The boost planning target volume included the gross tumor volume and mesorectum with a >2 cm margin in all directions. [12] Systemic chemotherapy was administered to 169 (96.0%) of the patients. With respect to the specific regimens, 5fluorouracil (5-FU) plus leucovorin (LV) or capecitabine alone were mainly administered to patients in the RT group, and 5-FU/ LV, capecitabine alone, 5-FU/LV/oxaliplatin, 5-FU/LV/irinotecan, capecitabine/oxaliplatin, and capecitabine/irinotecan were administered to patients in the non-RT group. The medical history of each patient was evaluated, as well as the results of a physical examination, routine blood tests, chest radiography, and other relevant studies including colonoscopy, computed tomography of the chest, abdomen, and pelvis, magnetic resonance imaging of the rectum, positron emission tomography, and preoperative carcinoembryonic antigen levels. Tumors were staged according to the Seventh Edition of the American Joint Committee on Cancer Staging System. Recurrence was defined as the first site of recurrent disease, and local recurrence was defined as the presence of adenocarcinoma within the rectal wall or mesorectum. The study protocol was approved by the Institutional Review Board of each center. Statistical analysis The baseline characteristics were compared between the RT and non-RT groups using the Student t test or Wilcoxon rank-sum test for continuous variables and x 2 or Fisher exact tests for categorical variables. To balance the differences in baseline characteristics between the 2 groups, propensity scores were generated using logistic regression modeling with all baseline variables for the likelihood of a patient receiving RT. Based on the score, each patient in the RT group was matched with 1 patient in the non-RT group. The rates of overall recurrence (including loco-regional and systemic) were compared between the matched groups using x 2 or Fisher exact tests. To assess the effects of RT on overall survival (OS), disease-free survival (DFS), and local recurrence-free survival (LRFS), we analyzed differences in the survival functions between the groups using the Kaplan-Meier method, and compared the survival curves between the groups using log-rank tests. A P value <0.05 was considered statistically significant. All statistical analyses were performed using SAS 9.3 (SAS Institute Inc, Cary, NC). Systematic review and meta-analysis A comprehensive search of the literature was performed in May 2015 using PubMed, Ovid, and Google Scholar to identify all publications in which oncologic outcomes among patients with stage IV rectal cancer who underwent curative resection and received adjuvant pelvic RT were evaluated. The key words were "radiotherapy" AND "rectal cancer" AND ("metastasectomy" OR "resection" OR "metastasis" OR "stage IV"). All 2-arm studies that compared outcomes between RT and non-RT groups were considered candidates. Studies regarding neo-adjuvant pelvic RT were not included due to the possibility of heterogeneity. Case reports or series that had <10 cases in each group, reviews, discussions, letters, and single-arm studies were also excluded. A sequential review of the title, abstract, and full-text of each report was conducted to select articles based on the selection criteria for our study. The outcomes of interest were the rates of loco-regional and systemic recurrence. The pooled risks for locoregional and systemic recurrence were computed and compared between the RT and non-RT groups. The relative risk (RR) and corresponding 95% confidence interval (CI) for each outcome was estimated using Mantel-Haenszel tests and RevMan 5.3 software (Cochrane, London, UK). Heterogeneity across the included studies was evaluated using Cochrane Q statistics and I 2 tests. When the value of I 2 was <30%, the included studies were considered to be homogenous, and a fixed effect model was employed. If the value was >30%, a random effect model was employed. The results were presented as forest plots. Publication bias was evaluated using funnel plots. Results A total of 176 patients were analyzed in this study. The baseline characteristics of the study populations are listed in Table 1. There were 51 patients who received RT and 125 who did not. The patients who received postoperative pelvic RT more frequently underwent abdominoperineal resection had a higher proportion of positive circumferential resection margins (CRM), Table 1 Characteristics of the study population before and after propensity score matching. Covariates Before matching After matching and more frequently received fluoropyrimidine-based chemotherapy than those who did not. Synchronous metastasectomy was performed in 49 (96.1%) patients of the RT group and in 118 (94.4%) patients of the non-RT group. Propensity score matching was conducted based on the 6 variables that were determined to be the most important for the final matching. These variables included T stage, CRM, number of metastases, operation type, angiolymphatic invasion, and histologic grade. As a result, 39 patients who received adjuvant RT were successfully matched with the same number of patients who did not. The mean follow-up period was 42.8 months (range, 1.7-143.3). The 2-year OS rate and median OS time were 81.8% and 45.2 months for the non-RT group, and 77.8% and 56.8 months for the RT group, respectively. No significant difference in OS was observed between the 2 groups (P = 0.388) (Fig. 1A). The 2-year DFS rates of the non-RT and RT groups were 22.6% and 35.8%, respectively (P = 0.709) (Fig. 1B). The median DFS times of the non-RT and RT groups were 14.0 and 13.3 months, respectively. During the follow-up period, pelvic local recurrence was observed in 8 (10.3%) of 78 patients, 7 (17.9%) of the patients who did not receive pelvic RT, and 1 (2.6%) patient who received pelvic RT. The 2-year LRFS rates of the non-RT and RT groups were 83.6% and 100%, respectively (P = 0.038) (Fig. 1C). The median LRFS was not reached in either group. The rate of systemic recurrence was 72.2%, and there was no significant difference between the RT and non-RT groups (69.2% vs 79.5%, respectively, P = 0.3). Systematic review and meta-analysis A total of 97 studies were identified in our initial search. After completing the selection process using the predefined inclusion and exclusion criteria, 4 studies were considered eligible for the meta-analysis (supplementary Figure 1, http://links.lww.com/ MD/B381). [10,11,13,14] There were no randomized controlled studies, and all studies were retrospective. Three hundred seventy-seven patients (129 who received adjuvant RT and 248 who did not) with stage IV rectal cancer who received postoperative RT after curative resection were included in the study. The median OS varied across the 4 studies, ranging from 27 to 61.8 months. The median DFS ranged from 11 to 18.2 months. None of the studies revealed significant differences in OS and DFS between the RT and non-RT groups. With the addition of the results from our study, the pooled risks for loco-regional and systemic recurrence were estimated and compared between the 2 groups. Overall, 455 patients (168 in the RT group and 287 in the non-RT group) were analyzed. The RT group had a significantly reduced risk for loco-regional recurrence compared with the non-RT group (RR; 0.48, 95% CI; 0.29-0.79, P = 0.004) (Fig. 2). There was no significant difference in the risk of systemic recurrence between the 2 groups (RR; 1.10, 95% CI; 0.96-1.25, P = 0.17) (Fig. 3). Relatively symmetricshaped funnel plots were generated in both analyses (Fig. 4), indicating a low possibility of publication bias. Discussion The role of adjuvant RT in the treatment of stage IV rectal cancer is not yet clear. Four studies have investigated the potential impact of adjuvant RT on oncologic outcomes among stage IV rectal cancer patients who underwent curative resection; however, the results were inconsistent. Kim et al [11] demonstrated that adjuvant pelvic RT significantly reduced the rate of local recurrence in patients with stage IV rectal cancer and synchronous hepatic metastasis. In contrast, An et al and Chang et al reported that the addition of pelvic RT to stage IV rectal cancer treatment after curative resection of the primary and metastatic lesions did not result in statistically significant beneficial effects in terms of the LRFS, DFS, and OS. [10,13] Lee et al [14] also concluded that pelvic RT did not improve loco-regional control and OS. Interestingly, pelvic RT did have oncologic benefits with respect to the LRFS in patients with pT4 disease. Considering the lack of strict criteria for adjuvant RT and the nonrandomized case assignments in all 4 studies, the inconsistent results might be attributable to heterogeneity in the characteristics of patients in the RT and non-RT groups, which could have resulted in selection bias. The present study employed propensity score matching analysis to match patients in the RT group with those in the non-RT group. This allowed us to control for the probability of a patient receiving 1 treatment over another, and reduced the risk of selection bias associated with nonrandomized case assignments. Another common weakness of previous studies was small sample size. Although a trend toward decreasing incidence of loco-regional recurrence in the RT group was observed compared to the control in all previous studies, the difference between the groups was not statistically significant, possibly due to an insufficient sample size. Indeed, Chang et al [13] suggested that at least 60 cases could be required in each group to determine whether there were statistically significant differences in the oncologic outcomes between the 2 groups; however, none of the studies included a sufficient number of cases. We analyzed a total of 455 patients (168 who received RT and 287 who did not) using meta-analytic methods to overcome the limitations of previous single-center studies involving small sample sizes. In the present study, adjuvant RT reduced the rate of locoregional recurrence, and the RT group had a significantly higher LRFS than did the non-RT group, after adjusting for the impact of other factors. The oncologic benefits were further supported by the results of the meta-analysis, which demonstrated that the RT group had a significantly reduced risk of local recurrence compared with the non-RT group. It is likely that adjuvant RT can have a protective effect against loco-regional recurrence after surgical resection in stage IV rectal cancer patients as well as in completely resected stage II/III patients. Although statistically significant, the actual difference in LRFS was not very large between the 2 groups in our study, compared with the difference observed with the meta-analysis. The hazard ratio of LRFS in the meta-analysis was 0.48, which means that the non-RT group had about a 2-fold higher chance of local recurrence than did the RT group. However, the hazard ratio is a relative measure of treatment effect, not the absolute risk. The total number of local recurrences in our study was only 8 (10.3%), and the actual difference of LRFS was not that large between the 2 groups. Loco-regional control of rectal cancer is clearly one of the most important issues for improving treatment outcomes and patient quality of life. Several studies have demonstrated that pelvic failure significantly affects patient quality of life in terms of fatigue, nausea/vomiting, and pain. [15][16][17] According to a study by Wong et al [17] , >80% of patients with locally recurrent rectal cancer experienced severe pain, and approximately 25% of them experienced bleeding, discharge, and urinary problems, which could significantly impact quality of life. Our results indicated that adjuvant RT did not increase OS or DFS times of patients with stage IV rectal cancer; however, it did have potential oncologic benefits in terms of reducing the risk of local recurrence, which could be important for improving patient quality of life. However, because RT itself can increase the number of morbidities, including bowel dysfunction, [18,19] additional studies are required to validate these conclusions. The rate of systemic recurrence of the RT group was comparable with that of the non-RT group in the present study. The meta-analysis demonstrated that the RT group had a 10% higher risk of systemic recurrence compared with the non-RT group, although the difference was not significant (P= 0.17). The higher rate of systemic recurrence in the RT group might have led to similar or slightly lower OS and DFS times in this group compared with the non-RT group, despite the potential benefits of RT for loco-regional

control. This could be explained by differences in the chemotherapy regimens between the 2 groups. Studies have shown that irinotecan-or oxaliplatin-based regimens resulted in superior outcomes to the conventional regimen of 5-FU plus leucovorin. [20,21] However, many clinicians tend to choose 5-FU-based regimens during RT because of the possibility of high toxicity with the other regimens. Consistent trends were observed across all studies, in that patients in the RT group were more frequently treated with 5-FU-based regimens, while patients in the non-RT group were more frequently treated with irinotecan-or oxaliplatin-based regimens. The heterogeneous chemotherapy regimens could have affected oncologic outcomes, including systemic control in the RT group. Several limitations of the present study should be taken into consideration when interpreting our results. First, the study had a retrospective design and was therefore subject to inherent biases, although we tried to minimize bias by using propensity score matching analysis. Second, although the meta-analysis was performed based on 5 studies and there were relatively consistent trends among the studies, all of the included studies were retrospective and there was no randomized controlled study. Therefore, the results were insufficient for drawing solid conclusions. Finally, our study was limited in that there was a lack of specific data on patient quality-of-life and the adverse effects of RT. Our results suggest that RT could have oncologic benefits for locoregional control in patients with stage IV rectal cancer who undergo curative resection, which is consistent with the findings of the metaanalysis. Pelvic RT could be considered in these patients to reduce the rate of loco-regional recurrence. Additional large-scale, wellcontrolled studies are required to confirm these conclusions. A window into the brain: An in vivo study of the retina in schizophrenia using optical coherence tomography Retinal nerve fibre layer (RNFL) thickness and macular volume (MV) can be measured in vivo using optical coherence tomography (OCT) providing a “window into the brain”. RNFL and MV are promising biomarkers in neurological diseases. This study explores the potential of RNFL and MV to detect axonal abnormalities in vivo in schizophrenia and their correlations with clinical features. OCT was performed in 49 patients (38 schizophrenia, 11 schizoaffective disorder) and 40 healthy controls matched for age and gender. Group comparisons were made between whole retina and quadrant RNFL thickness and MV using multi-level analyses. In patients, associations were sought between RNFL and MV with symptom severity (positive/negative). Patients and controls had similar whole retina RNFL thickness (p = 0.86) and MV (p = 0.64), but RNFL in the right nasal quadrant of the schizoaffective group was thinner than in the schizophrenia group (p = 0.02). In patients, positive symptom severity was associated with smaller MV (right β = − 0.54, p = 0.02; left β = − 0.49, p = 0.04). Normal MV and RNFL thickness suggests unmyelinated axons in patients with schizophrenia/schizoaffective disorder remain unaffected. Longitudinal studies using higher resolution OCT will clarify whether progressive RNFL and MV changes occur and whether they can be used as state or trait markers in schizophrenia. Introduction Schizophrenia lacks a clearly defined or diagnostic neuropathological signature. The best validated neuropathological abnormalities include possible reductions of neuronal density in the thalamus, but not in the cerebral cortex or hippocampus, although in these structures pyramidal neurons may have smaller bodies with reduced dendritic arborisation and spines (Harrison and Weinberger, 2005). Reductions in the number of some interneurons and in the number and function of oligodendrocytes, relevant to myelination and neuronal and synaptic integrity, complete the picture (Harrison and Weinberger, 2005). In keeping with these findings, neuroimaging studies have reported grey matter volume deficits in those with chronic schizophrenia (Ellison-Wright and Bullmore, 2009), those in their first episode of psychosis (Bangalore et al., 2009), in a prodromal phase of illness (Pantelis et al., 2003), or at high genetic risk of schizophrenia (Witthaus et al., 2009). Childhood onset schizophrenia also results in progressive grey matter volume loss in the early years of illness (Sporn et al., 2003). Compromised white matter integrity rather than volume loss has been reported in first episode (Price et al., 2007) and chronic schizophrenia (Kanaan et al., 2009). Longitudinal imaging studies have also suggested that at least in some patients there may be an accelerated loss of grey matter in comparison with healthy controls immediately after the first episode of schizophrenia (Sun et al., 2009), with similar results documented in chronic patients (Brans et al., 2008) and in those at ultra high risk, even before the first episode of psychosis (Takahashi et al., 2010). However, the precise neuropathological changes that underlie these neuroimaging findings remain to be determined. In vivo visualisation of the retinal nerve fibre layer (RNFL) can be achieved by optical coherence tomography (OCT), a non-invasive, fast imaging technique used to monitor retinal changes in glaucoma (Huang et al., 1991). There are no contra-indications to OCT, which opens a "window into the brain" by allowing measurement of RNFL thickness and macular volume (MV). The RNFL is composed of unmyelinated axons that traverse the retina, with the highest axonal density contained in the macula. OCT has recently been applied to the study of neurological conditions with diffuse and progressive brain pathology. RNFL thinning has been described in patients with mild cognitive impairment without dementia (Paquet et al., 2007). It correlates with the severity of cognitive impairment in patients with Alzheimer's disease (Iseri et al., 2006), and Berisha et al. (2007) have reported it in the early stages of Alzheimer's disease. RFNL thinning is also present in patients with Parkinson's disease, in whom loss of dopaminergic neurons is known to occur, not only in the cortex and basal ganglia but also in the retinal ganglion cells (Altintas et al., 2007). OCT reveals a variable degree of RNFL thinning (mean~20%) following an episode of optic neuritis (Trip et al., 2005). Reduced RNFL thickness is also described in multiple sclerosis (MS), even in patients without a prior history of optic neuritis and this correlates with magnetic resonance imaging (MRI) measures of brain atrophy (Siger et al., 2008). In healthy subjects, RNFL thickness has been reported to be associated with cognitive performance, particularly in those below 40 years of age (van Koolwijk et al., 2009). In glaucoma, around 20% RNFL thinning was reported at 2-month follow-up after onset of the disease (Fang et al., 2007), while rate of RNFL thinning in a longitudinal study of chronic glaucoma was less than 1% a year (Leung et al., 2011). In schizophrenia, widespread; albeit subtle, neuropathological abnormalities, together with electroretinogram (ERG) changes indicative of reduction in rod photoreceptors or changes in inter-neuronal retinal architecture (Balogh et al., 2007), have also been described in those at high genetic risk of schizophrenia (Hébert et al., 2009) and in autism, (Ritvo et al., 1988). Hence there is good reason for using OCT to search for disease biomarkers. We present here the first study, to our knowledge, to use OCT to study the retina in patients with schizophrenia and schizoaffective disorder. We hypothesised that retinal changes indicative of schizophrenia-related neuropathology would be present and quantifiable in our patients. We also looked for associations between OCT findings and disease symptoms. Subjects Patients recruited to this study were selected from a larger cohort (the West London First Episode Psychosis study) and had a firmly established diagnosis of schizophrenia or schizoaffective disorder using the DIP-DM Diagnostic Interview for Psychosis (Jablensky, 2000), which includes items from the Operational Criteria Checklist for Psychosis (OPCRIT) (McGuffin et al., 1991) and the WHO Schedules for Clinical Assessment in Neuropsychiatry (SCAN) (Wing et al., 1990). Substance abuse and dependence were established using the Alcohol and Drug Use Scales, (Drake et al., 1990). Two research nurses conducted the interviews, and patients were clinically reviewed by two experienced consultant psychiatrists (EJ and TB) to confirm the diagnosis 1 year after inclusion in the study. Forty-nine patients (36 males, 13 females), mean age 29.9 years (S.D. ± 8.74) with diagnoses of schizophrenia (38 patients) or schizoaffective disorder (11 patients) and a mean illness duration of 4.4 years (S.D. ± 3.6) at the time of OCT scanning were recruited. Forty patients were prescribed antipsychotics (39 atypical and 1 typical), 10 patients were prescribed antidepressants and four mood-stabilising medication concurrent with an antipsychotic, two were prescribed mood stabilisers only, one was prescribed night sedation only, and five were unmedicated. Scores on the Schedules for the Assessment of Positive and Negative Symptoms (SAPS and SANS) (Andreasen, 1981;Andreasen, 1983) were available for a subset of 32 patients. Severity of negative symptoms ranged from 0 to 16, mean 3.7, and positive symptoms ranged from 0 to 10, mean 2. All subjects included in the study had a normal ophthalmic examination including Humphrey 30-2 threshold visual fields test and logMAR visual acuity test. There were no differences between groups in were not recorded, but there was no evidence of optic disc cupping in any of our subjects. Exclusion criteria Exclusion criteria for all subjects were as follows: a) concurrent or previous systemic disease (e.g. diabetes or autoimmune disease) that could involve the eyes, b) a history of neurological or ophthalmological disease known to affect the visual pathway (e.g. glaucoma), c) high myopia (b−6D) as this may cause artefactual reduction in RNFL thickness and difficulty in fixation, d) previous head injury with loss of consciousness, and e) drug or alcohol dependence. Ethical approval was granted by the Joint Ethics committee of the UCL Institute of Neurology and University College Hospital NHS Foundation Trust, and this study was carried out in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki). Optical coherence tomography Using OCT, a cross-sectional image of the retina is produced by measuring the echo time delay of back-scattered infra-red light after it has passed into the eye and is bounced back using a low coherence light source and interferometer. OCT was performed by a neurologist (MK) trained and experienced in using the Stratus OCT3 device and software (Carl Zeiss Meditec Inc., California, USA) at UCL Institute of Neurology. RNFL images were acquired for each eye by taking a circumpapillary scan of 3.4-mm diameter to effectively intercept all nerve fibres converging toward the optic disc while avoiding inaccurate measurements resulting from peripapillary atrophy (Schuman et al., 1996). Retinal nerve fibre layer and macular volume measures The thickness of the RNFL quadrants (temporal, superior, nasal and inferior), were calculated by the OCT device software and represented by a line graph indicating RNFL thickness at all sections of the scanning circle (Fig. 1). With the Fast RNFL protocol, the mean of three circular 3.4-mm diameter scans, centred on the optic disc, was used to express RNFL thickness. MV was measured in each eye by taking six consecutive radial linear scans centred on the fovea, to provide six sets of equally spaced and intersecting scans, and to obtain a single average value (using the fast macular thickness map scanning protocol). The Stratus OCT3 device assigns a signal strength (10 being the maximum) to OCT images. These were rejected if signal strength was less than 7 or when the difference in signal strength between images for the two eyes was greater than 2 to ensure consistency in the quality of scans obtained. Statistical analysis Independent samples t-tests and chi-squared tests were used to compare age and gender between all patients and controls. A generalised linear mixed model (GLMM) approach was used to compare whole retina RNFL thickness, quadrant RNFL thickness and MV between three groups (controls, schizophrenia patients and schizoaffective disorder patients). Multilevel models were used to account for the fact that observations within each patient (i.e. right and left eyes) were not independent, while also controlling for possible confounding effects of age, gender and disease duration (Laird and Ware, 1982). Where association was shown between eye and disease subtype, linear regression analysis was used to investigate the effect of disease subtype on each eye independently. In the patient group, multiple linear regression analysis was used to investigate the relationship between severity of positive and negative symptoms with measures of RNFL thickness and MV in both eyes, while controlling for age and gender. Statistical significance was reported at p b 0.05. Power calculation Power calculation was based on a study reporting subtle decreases in RFNL thickness in the unaffected eyes of patients with MS due to subclinical axonal

damage (Fisher et al., 2006). We estimated that in order to detect a mean between-group difference in RNFL thickness of 9 μm with a standard deviation of ±14 μm and a standardised difference of 0.64, a sample size of 38 patients and 38 controls would be needed to achieve an 80% probability of detecting differences between the two groups at a 5% significance level. Based on Henderson et al. (2008), with a mean difference in MV of 0.27 mm³ between groups, our sample size would give a 70% probability of detecting a group difference at the 5% level of significance. Results Patients and controls were matched for age (t = 0.27, d.f. = 85, p = 0.79) and gender (χ² = 0.27, p = 0.36). Mean and 95% confidence intervals for whole retina RNFL, quadrant RNFL and MV measures in controls, schizophrenia and schizoaffective disorder groups are reported in Table 1. Group comparisons Using multi-level analyses, patients with schizophrenia or schizoaffective disorder did not differ from controls in whole retina RNFL or MV measurements ( Table 2). The left nasal quadrant RNFL appeared thinner than the right across all groups and also showed evidence of an interaction between eye and disease subtype (p = 0.002), with significant difference in right but not left nasal quadrant RNFL thickness between the patient groups. After adjustment, schizoaffective disorder patients had a thinner right nasal quadrant (mean − 17.9 μm, 95% CI = −31.0 to −4.9) than those with schizophrenia (p = 0.02). Left temporal quadrant RNFL appeared thinner than the right across all groups, and there was an interaction between side and disease subtype (p = 0.009), but there was no difference in right (p = 0.18) or left (p = 0.23) temporal quadrant RNFL thickness between groups. Symptom severity In patients, multiple linear regression showed a moderate association between lower MV and positive symptom severity (right β = −0.54, p =0.02; left β = −0.49, p = 0.04). When only those patients with schizophrenia were included in the analysis, a strong association was seen (right β = −0.85, p =0.04; left β = −0.85, p =0.05). MV was not associated with severity of negative symptoms, and RNFL was not associated with positive or negative symptom severity. Discussion This study did not detect any differences in whole retina RNFL thickness or MV between patients with schizophrenia or schizoaffective disorder and healthy controls. These negative findings suggest that in patients with schizophrenia or schizoaffective disorder there is no detectable loss of unmyelinated axons in the retina in the early years after disease onset. Based on average values from both eyes for mean and standard deviations of the RNFL thickness of our subjects, a retrospective sample size calculation showed that at least 300 subjects would have been required in each arm of the study to detect any significant differences between groups, had they been present, with 80% Fig. 1. Example of RNFL measurements taken from a healthy eye. Line of graph shows RNFL thickness of the scanning circle as seen around the optic nerve head in the photograph on the right. X axis = position on scanning circle and Y axis = RNFL thickness at different positions. TEMP 169 = temporal, SUP = superior, NAS = nasal, INF = inferior. The colours represent normal distribution percentiles. Green = 95 to 5%, yellow = 5 to 1%, red = 1 to 0%. power at a 5% level of significance. We therefore conclude that using this OCT methodology, RNFL variations are too subtle to be of value as a biological marker for schizophrenia. Differences in RNFL thickness were seen in the right nasal quadrant between patients with schizophrenia and schizoaffective disorder. Differences in temporal quadrant RNFL, however, were not associated with disease subtypes. These unexpected results cannot be explained by age or gender differences and should be interpreted with caution given the small number of patients with schizoaffective disorder in this study. Possible medication effects cannot be excluded, as half of the schizoaffective patients were also prescribed antidepressants or mood stabilisers as adjuncts to atypical antipsychotics. The association between severity of positive symptoms and MV in patients with schizophrenia needs replicating in a larger sample as it may represent state-dependent abnormalities similar to those reported by an ERG study (Balogh et al., 2007) that described transient retinal changes (i.e. decreased amplitude of the a-wave indicative of altered early visual information processing) that which during acute relapses correlated with the severity of positive symptoms and were attributed to state-dependent alterations in phospholipid metabolism and/or impaired dopaminergic transmission. Our results also suggest that axonal damage may be less important than myelin abnormalities in explaining the frequently reported reductions in brain volume in schizophrenia. There is growing evidence for disruption of white matter integrity in the early stages of schizophrenia (Price et al., 2006(Price et al., , 2007Pérez-Iglesias et al., 2010), which lends support to this possibility. The negative findings are also in keeping with our earlier studies in patient samples overlapping the one described here, which failed to demonstrate differences in cortical volume between patients and controls using surface-based morphometry (Gutiérrez-Galve et al., 2010) and only fronto-temporal white matter volume loss with no cortical changes observed when using magnetization transfer imaging or volumetric MRI (Price et al., 2006). The main limitation of this study is low OCT resolution, which may be inadequate to detect the subtle abnormalities that may be present in the early years of schizophrenia. Transverse resolution can reach a maximum of 10 μm in tissue and 15 μm in air using the Stratus OCT device (Fujimoto et al., 2004); hence, sensitivity is too low to visualise individual axons, which are typically 1 μm in diameter. This point has been made by Naismith et al. (2009), who detected RNFL thinning in optic neuritis in only 60% of cases, and detection rates were lower in those with good recovery. Newer technology using spectral domain OCT is faster and offers better axial resolution (5 μm compared to 10 μm) and more accurate macular thickness measurement than the Stratus OCT3 device used in our study. Most of our patients were receiving atypical antipsychotic medication at the time of data collection, and it is impossible to exclude the potential neuroprotective effects of these drugs (Navari and Dazzan, 2009), which may have obscured minor differences in RNFL between the groups. Despite mainly negative findings from this first exploratory study, there is a need for further OCT investigations to examine possible RNFL thinning and decreased MV in schizophrenia patients with longer illness duration. A longitudinal study would additionally be able to detect whether progressive retinal changes occur, at which stage of the disease this may happen and whether these retinal measures could be utilised as a trait or state marker of schizophrenia or schizoaffective disorder. Racism in oral healthcare settings: Implications for dental care‐related fear/anxiety and utilization among Black/African American women in Appalachia Abstract Objective To explore the association of racism in oral healthcare settings and dental care‐related fear/anxiety with dental utilization among Black/African American women in Appalachia. Methods We analyzed self‐report measures of racism in oral healthcare settings, dental care‐related anxiety and fear, recency of a dental visit, and demographic information from 268 pregnant women participating in the Center for Oral Health Research in Appalachia (COHRA) SMILE cohort. All participants self‐identified as African American or Black and resided in Appalachia (i.e., either West Virginia or Pittsburgh, PA). Results Over one‐third of the participants reported at least one instance of racism in oral healthcare settings, with “not being listened to” due to their race or color as the most frequent issue (24.4%). Clinically significant levels of dental care‐related anxiety and fear were reported by 14.3% of the sample. A mediational model demonstrated that the experience of racism in oral healthcare settings was a significant predictor of dental fear/anxiety, and that dental fear/anxiety was a significant predictor of dental utilization. There was a significant relationship between racism in oral healthcare settings and dental utilization only when mediated by the presence of dental care‐related fear and anxiety. Conclusions Together, experiences of racism in oral healthcare settings and dental care‐related fear/anxiety are predictive of decreased dental utilization for Black/African American women living in Appalachia. This study provides insight into racism in oral healthcare settings as a social determinant of dental anxiety/fear and inequities in dental utilization. INTRODUCTION Racism is a public health threat and has been linked to health and healthcare inequities [1][2][3][4]. It is the most often reported form of discrimination in US healthcare systems [3], and is linked to reduced healthcare utilization among other deleterious health outcomes [2][3][4]. Likewise, racial inequities exist in dental care utilization [1,2]. Racism in oral healthcare settings can be manifested through provider-and staff-patient interactions and clinical decision-making [5,6]. For example, African American/Black patients are disproportionately recommended for tooth extractions, provided with less pain management, issued fewer surgery referrals for oral cancer, given less time with providers, and are more likely to lack racial concordance with providers [7,8]. Yet, little is known about racism in oral healthcare settings [5] and how it might be related to dental fear/anxiety and ultimately racial inequities in utilization [8]. Krieger's Ecosocial Theory of Health Inequity highlights social forces, such as racism, that impact public health [9] Using this theory as a basis, this study focuses on the "dynamic pathway" [9] as one of the four ways public health inequities manifest. Within this pathway, we focus in this study on "everyday racism" [10] as ongoing, chronic social interactions with providers and staff in oral healthcare settings that produce inequity in oral health outcomes by race. These experiences may constitute one aspect of the pathway that reduces the probability of utilization of oral healthcare in Black/African American and other minoritized communities; they may be involved mechanistically with other factors, such as dental care-related anxiety and fear. Even with extensive documentation of overall barriers to dental care utilization, pathways involved in such utilization among Black/African American individuals have not been clarified [1,5,7,9]. Dental care-related fear/anxiety is a well-established predictor of dental care utilization in the general population [11]. Dental care-related fear/anxiety refers to the experience of negative anticipatory or procedural arousal related to professional dental treatment [11]. Individuals who report greater levels of dental fear/anxiety have a pronounced likelihood of delaying or avoiding dental treatment [11] . There is limited research, however, on the prevalence, genesis, or manifestations of dental care-related fear/anxiety among those identifying as Black or African American [12]. Further, there is a paucity of research on how racism and dental care-related anxiety/fear together influence dental care utilization for Black/African American patients. Yet these links are important as psychologically, anxiety/ fear and effects of racism have in common the anticipatory nature of danger, pain, or harm [13,14]. Research has consistently linked generalized anxiety and experiences of racism among Black/African American samples [15,14]. A study [15,16] involving Black/African American dental patients identified unique triggers of dental fear/anxiety that are different from those typically observed in majority cultural groups. Many of the Black/African American patients reported poor-quality care and/or mistreatment, stemming from prior experiences with oral healthcare providers, as triggers for their current dental care-related fear/anxiety. In other words, past experiences with racism in oral healthcare settings was noted as one of the triggers of current dental fear/anxiety, which contributed to Black/African American patients seeking dental care less often or to stop utilizing services altogether. Another study [17] found that participants who reported the emotional impact of experiencing racial discrimination in healthcare settings had a greater likelihood of underutilizing oral healthcare. Racism in oral healthcare settings therefore may reduce utilization of professional dental services [16] through a pathway that involves dental fear/anxiety. Along with dental fear/anxiety and potentially racism, there are other known factors that predict underutilization of dental services, which are considered in the current study. Those living in geographic regions like Appalachia have poorer oral health relative to many other regions nationally [18]. Additionally, pregnant women may underutilize dental care due to dental fear/anxieties about harm to the fetus and the reluctance of providers to provide treatment during pregnancy [19]. As such, Black/African American women residing in Appalachia are at the intersection of multiple social determinants of oral health outcomes, including gender, rurality, and race [20]. The intersectionality of various minoritized identities is

of critical importance in oral health and healthcare [21,22]. Appalachia spans 13 states, including all of West Virginia [22]. It includes many rural and remote areas, but a number of cities, too. Those who identify as Black or African American constitute $10% of the population of Appalachia and are the largest racial minoritized group therein [22]. The present study aimed to address several critical gaps in the literature regarding the prevalence of racism in oral healthcare settings and its association with dental anxiety/fear, and oral healthcare utilization among Black/African American women in Appalachia. Using Krieger's theory [9] as a basis, this study explores how everyday racism within oral healthcare settings may be associated with healthcare-seeking behavior (i.e., dental care utilization) via a pathway that involves dental carerelated fear/anxiety. We estimate the frequency of dental fear/anxiety and racism occurring in dental settings, specifically for those who embody underprivileged intersecting identities (i.e., race, gender and residing in Appalachia). Using mediation analysis, we explore the association of dental fear/anxiety in the relationship between experiences of racism in oral healthcare settings and dental care utilization. We hypothesized that racism experienced during oral healthcare would positively relate to dental fear/anxiety and, that dental care-related fear/anxiety would negatively relate to dental utilization in this sample of Black/African American women. We also hypothesized that dental fear/anxiety would mediate the relationship between racism in oral healthcare settings and dental care utilization. Participants The current study data (n = 268) is from the Center for Oral Health Research in Appalachia (COHRA) SMILE cohort. The COHRA longitudinal research project has been ongoing since 2002, focusing on the factors contributing to oral health inequities in Appalachia [23]. The COHRA SMILE cohort was initiated in 2018, and specifically recruited women who: (a) identify as Black or African American, (b) were in their 1st or 2nd trimester of pregnancy, (c) were 18 years and older, (d) reside in West Virginia or Pittsburgh, PA, (e) were healthy enough for an oral health screening, and (f) were willing to participate in follow-up visits. Sociodemographic variables Participants responded to questions about their educational level, household income, state of residence, and type of dental insurance, if any, selected from a larger sociodemographic questionnaire [23]. Oral healthcare utilization Utilization was measured using a single item, "About how long ago was your LAST visit to a dentist or dentist hygienist, including all types of dentists, such as orthodontists or oral health surgeons?" with the response options (1 = 6 months or less, 2 = 7 monthsÀ1 year, 3 = 1-2 years, 4 = 2-3 years, 5 = 3-5 years, 6 = more than 5 years and 9 = Do not know/Not Applicable). Participants also were asked if they ever had professional dental care of any type. These values were dichotomized into a variable indicating whether the participant had oral healthcare in the last year (coded as 1) or not (coded as 0), including those who reported they have never utilized professional dental care. Everyday Racial Discrimination in Dental Care Scale The 7-item scale is a modified version of the original 9-item, Everyday Discrimination Scale (EDS) [24]. The original measure was informed by Essed's Theory of Everyday Racism, which was developed with a sample of Black/African American women [10]. The ERDS-DC measures lifetime social interactions in oral healthcare settings that produce inequity in oral health outcomes by race. Example items include "Received poorer services than other people" and "Felt like a dentist, hygienist, or dental assistant was not listening to what you were saying." Participants are asked to indicate the frequency of these events with a 4-point scale (0 = Never; 1 = Once; 2 = Two or three times; 3 = Four or more times); the possible range of the total score is 0-21. Higher scores indicate more frequent experiences of everyday racism. Previous studies reported good internal consistency, ranging between α = 0.88-0.94 [24][25][26][27]. The scale has been previously validated [25,27,28]. Internal consistency for the ERDS-DC for the current sample was α = 0.88. Dental Fear Survey The Dental Fear Survey (DFS) is a 1-5, 20-item Likerttype self-report measure, with a possible range of 20-100, which assesses lifetime dental care-related fear/anxiety [29]. In addition to three subscales (i.e., avoidance of dental care, fear of specific dental-oriented stimuli, and physiological arousal), there is an omnibus item (DFS #20) that assesses overall dental fear/anxiety (i.e., All things considered, how fearful are you of having dental work done?"). Higher scores indicate greater dental carerelated fear/anxiety. The psychometric properties of the DFS and its relation to the DFS item #20 are well established [11,[29][30][31][32]. Internal consistency for the total DFS scale for the current sample was α = 0.95. For the purposes of this study, clinically significant fear/anxiety was that which was reported as "much" (i.e., 4) or "very much" (i.e., 5) fear and anxiety on the omnibus DFS item #20. Sample recruitment and data collection The study was approved by the Institutional Review Boards at West Virginia University and the University of Pittsburgh; participants signed printed consent forms. Using a community engaged approach; potential participants were recruited through three types of nonprobability sampling methods: (a) cluster sampling with relevant community organizations, (b) web-based sampling, and (c) snowball sampling. Participants were recruited between January 2018 and June 2021 for an inperson assessment visit, after a telephone or in-person screening interview to assess eligibility. The participants were involved in a comprehensive in-person biopsychosocial battery of assessments [22], followed by a telephone interview. Assessment of racism during dental care was accomplished privately during the in-person assessment, with the participants privately using a computer tablet. Participants were compensated for both the inperson assessment and the telephone interview. Statistical analysis Using the PROCESS macro for SPSS [33] we tested the hypothesis that racism in oral healthcare settings predicted dental utilization through dental fear/anxiety, using a logistic regression mediational model. Included as covariates were those with established influences on oral health behaviors and outcomes: (a) socioeconomic status, as reflected in educational level and household income; (b) dental insurance status (i.e., public insurance RESULTS Sociodemographic data are presented in Table 1. On average, the participants were 27.2 years old (SD = 5.5, range 18-43). Almost all (93.7%) had completed high school or further education, and 11.9% had a college degree or 16 Racism in Oral Healthcare Measure F I G U R E 1 Percent of participants endorsing each of the seven types of everyday racism experienced in oral healthcare settings on the everyday racial discrimination scale-dental settings [27]. The specific items are: "Treated with less courtesy than other people; Treated with less respect than other people; Received poorer services than other people; Had a dentist, hygienist, or dental assistant act as if he or she thought you were not smart; Had a dentist, hygienist, or dental assistant act as if he or she was afraid of you; Had a dentist, hygienist, or dental assistant act as if he or she was better than you; and Felt like a dentist, hygienist, or dental assistant was not listening to what you were saying." N = 268 postgraduate education. About three-fourths (n = 196; 73.1%) resided in an urban setting (i.e., Pittsburgh) and one-quarter (n = 72; 26.9%) in smaller cities or rural areas of West Virginia. Most participants had some form of dental insurance (n = 220, 82.1%). About half of the sample (n = 141; 52.6%) had public dental insurance (e.g., Medicaid). Of the 268 participants, 109 (40.7%) reported having had a dental visit in the past 6 months, while another 57 (21.3%) had a dental visit within the last year, and 34 (12.7%) reported never having had professional dental care. Over one-third (n = 100; 37.3%) of participants reported experiencing at least one instance of racism during dental care. Figure 1 displays the relative distribution across the seven forms of everyday racism assessed in the ERDS-DC. The average total score for the ERDS-DC scale was 1.7 (SD = 3.3). Figure 2 displays the distribution of levels of dental care-related fear/anxiety, based on the omnibus DFS item. The mean DFS total score was 34.1 (SD = 17.0). Overall, 44% of the women from the sample endorsed at least some dental care-related fear/anxiety; 14.5% of the women endorsed clinically significant levels of dental care-related fear/anxiety (i.e. those indicating much or very much dental fear/anxiety on DFS item #20). First, the mediational model was tested without covariate adjustments. The experience of racism was a significant predictor of dental fear/anxiety, B = 0.06, SE = 0.02, 95% confidence interval (95% CI) (0.013, 0.105), p = 0.0128, and dental fear/anxiety was a significant predictor of oral healthcare utilization, log odds estimate = À0.26, SE = 0.10, 95% CI (À0.458, À0.067), p = 0.0085. These results support the mediational hypothesis. More experiences of racism were associated with greater dental anxiety/fear; higher levels of dental fear were associated with decreased use of dental services. Racism alone was not a significant predictor of utilization, log odds estimate = 0.022, SE = 0.04, 95% CI (À0.0563, 0.1002), p = 0.5827. The indirect coefficient was significant, log odds estimate = À0.016, SE = 0.010, 95% CI (À0.0389, À0.0013), supporting a mediation between racism and dental fear/anxiety on the outcome of oral healthcare utilization. Converting the logit value to an odds ratio for the indirect effect of racism on utilization, as mediated by dental fear/anxiety, yielded an OR of 0.984 (0.9618, 0.9987), reflecting a small effect. This result suggests that racism indirectly decreases utilization. DFS 20 F I G U R E 2 Distribution of participant scores in percentages across the five levels of dental care-related fear/anxiety on the Dental Fear Survey's omnibus item #20; N = 268 between racism and dental fear/anxiety on the outcome of oral healthcare utilization. Converting the logit value to an odds ratio for the indirect effect of racism on utilization, as mediated by dental fear/anxiety, yielded an OR of 0.9807 (0.9527, 0.9970), suggesting a small effect. Of the covariates, only the public dental insurance variable was significant, log odds estimate = 0.78, SE = 0.28, 95% CI (0.2272, 1.3234), p = 0.0056. DISCUSSION This study is one response to the call for anti-racist research in dentistry. Results supported the hypothesis that racism in oral healthcare settings, together with dental care-related anxiety and fear, predicts oral healthcare utilization. There are few studies on racism as a social determinant of oral health inequities in the US [1,5]. Past research within the US has assessed lifetime racism or racism in general healthcare settings with single or two item measures [15][16][17]. In this study involving Black/African American pregnant women living in Appalachia, over one-third reported at least one prior incident of everyday racism in oral healthcare settings. Using a psychometrically-sound measure, this study highlights the nature of racism in these settings, with the most frequent occurrences of everyday racism being: (a) not listened to by dental providers, (b) being treated with less respect compared to others, and (c) receiving poorer service compared to others. These findings are consistent with the long-standing history of racism that plagues US healthcare systems. Additionally, though specific to racism in oral healthcare settings, results are like those found with racism reported by African American/ Black women generally in US healthcare settings [3]. Finally, these findings support the intersection of race and gender in experiences of racism with "not being listened to by dental providers" being the most reported form of racism and "being perceived a threat by dental providers" being the least form of racism in oral healthcare settings (which typically is associated with males). That is, in addition to common experiences of racism, there are some experiences unique to gender [10,20]. Additionally, 34 (12.7%) of the women in this sample reported never having professional dental care. This result may be related to previous findings [9,10,13] that racism is collective and impacts the lives of Black and African American people both directly and indirectly. As such, knowledge of another person's experience of racism in oral healthcare settings or personal experiences of racism in other healthcare settings may be a factor impacting one's dental utilization, which may contribute to avoidance of dental care services, even when one has not directly had that experience. In this analysis of Black/African American women living in Appalachia, high levels of dental care-related anxiety and

fear were comparable to the US as a whole in terms of clinically significant fears/anxieties (i.e., those indicating much or very much dental fear/anxiety on DFS item #20; 14.5% compared to $15%) [11]. Conversely, the overall dental care-related fear/anxiety score in this study was lower than that for White mothers in another COHRA sample [23] (i.e., M = 43.4, SD = 20.8; onesample test, t [267] = 8.93, p < 0.001). Even so, dental care-related fear/anxiety was predictive of dental care utilization in this sample as in the general population. Additionally, the findings also support existing studies that link experiences of racism with dental care-related fear/ anxiety. Of note, this is the first quantitative measure of dental care-related fear/anxiety among an exclusive sample of participants identifying as Black or African American. The pregnant women in this study are at the intersection [20] of race, gender, and residency in Appalachia, and so may be particularly vulnerable to adverse oral health outcomes. It is noteworthy, then, that their dental utilization is similar to that of the overall population in F I G U R E 3 Mediational model of the effect of racism experienced in oral healthcare settings on oral healthcare utilization, through dental carerelated anxiety and fear; N = 268 the USA, and that their dental fear/anxiety is comparable as well, if not lower than samples of White pregnant women. Dental utilization in the past year was similar to that found nationally (62.0% in the current sample compared to 63.0% across the US) [34]. Similarly, the small yet significant association of racism on dental care utilization could be attributed to the resilience of the participants in this sample [9,10]. Extensive documentation exists on how those from Black/African American communities have historically identified strategies to cope with and resist the deleterious impact of racism [12,15]. In this case, delaying or avoiding dental care may be an effect of racism in oral healthcare settings. Other possible coping responses include being assertive and changing providers [35], if possible. Findings from this study support the need for dentistry to pursue anti-racist care with Black/African American women, through education and policies that focus on racism as a social determinant of oral health [36], to ensure that Black and African American patients are listened to, provided good quality service, and treated with respect. Since negative effects of avoidance of oral healthcare are well established, the association of experienced racism in oral healthcare settings on dental care-related anxiety and fear is noteworthy and troubling. Being a dental patient can be uncomfortable and unsettling, particularly by those who are fearful and/or anxious about dental treatment (e.g., being closed in, laying in a supine position that feels out of control, tolerating the intrusion of fingers and instruments into the oral cavity, dealing with invasion into one's personal space/intimacy zone, fear of pain, among others) [11]. It is possible that for many people, any additional overlay of negativity may not only add but act synergistically with pre-existing dental anxiety and fear. LIMITATIONS All data in this study is self-report, with associated potential biases. The association of racism in oral healthcare settings, together with dental fear/anxiety on dental care utilization in the model was small, but statistically significant. Participants living in a region (i.e., Appalachia) with known disparities/inequities in oral health, and only pregnant women were included. This issue may limit the generalizability to other groups. Additionally, we had no information about the ethnic/racial identity of the participants' dental providers. Given that racial discordance between patient and provider may be associated with dental carerelated fear/anxiety, there is the possibility concordance or discordance affected these results [37]. Nevertheless, the current study has many strengths, notably including that a population group representing intersectionality [21] is the focus, with various identities that are minoritized and/or marginalized (i.e., women, being pregnant, identifying as Black or African American, and being from Appalachia). FUTURE DIRECTIONS Given considerable literature on the gendered nature of racism, studies exploring racism in oral healthcare settings among African American/Black men are essential. Additionally, research is needed with other minoritized population groups in other regions of the country. Qualitative interviews may be helpful to further investigate the vagaries of racism in oral healthcare settings. As already noted, future studies exploring racism in dental settings ought to consider asking about participant's dental providers' racial/ ethnic identity. CONCLUSIONS The psychological impact of racism in oral healthcare settings is a likely barrier to oral healthcare utilization. Findings from this study showed a significant relationship between racism in oral healthcare settings and oral healthcare utilization, when mediated by the presence of dental care-related fear and anxiety. This unique cohort of women provides insight on the nature of racism in dental settings, and into the intersectionality of race, gender, and residence in Appalachia. The number and percentage of women who have had experiences of racism in dentistry is concerning, and highlights the need to focus on racism as a social determinant of oral health, in order to promote equity so that all oral healthcare experiences involve respect and compassion. Pluripotent Stem Cell Studies Elucidate the Underlying Mechanisms of Early Embryonic Development Early embryonic development is a multi-step process that is intensively regulated by various signaling pathways. Because of the complexity of the embryo and the interactions between the germ layers, it is very difficult to fully understand how these signals regulate embryo patterning. Recently, pluripotent stem cell lines derived from different developmental stages have provided an in vitro system for investigating molecular mechanisms regulating cell fate decisions. In this review, we summarize the major functions of the BMP, FGF, Nodal and Wnt signaling pathways, which have well-established roles in vertebrate embryogenesis. Then, we highlight recent studies in pluripotent stem cells that have revealed the stage-specific roles of BMP,FGF and Nodal pathways during neural differentiation. These findings enhance our understanding of the stepwise regulation of embryo patterning by particular signaling pathways and provide new insight into the mechanisms underlying early embryonic development. Introduction Embryogenesis is a process by which the zygote develops into a complex and organized embryo. During early development in the mouse, the zygote first develops into a blastocyst containing the inner OPEN ACCESS cell mass (ICM) inside the trophectoderm. At embryonic day (E) 4.5, the blastocyst implants in the uterus and the ICM differentiates to form the primitive endoderm and the early epiblast [1]. At E5.5, the early epiblast develops into a columnar epithelial monolayer of pluripotent cells called the late epiblast [2]. At approximately E6.5, gastrulation commences with the formation of the primitive streak in the posterior of the epiblast. Epiblast cells that ingress through the primitive streak form the mesoderm and endoderm. The cells remaining in the anterior of the epiblast form the ectoderm [3]. As development proceeds, the ectoderm, mesoderm and endoderm generate most of the cell types and organs of the adult. For studies of early development, Xenopus, zebrafish and chick embryos are widely used, as they are easily raised and can be manipulated with various assays. Compared with these species, mice as mammals more closely resemble humans in developmental processes, but mouse embryos are difficult to manipulate, especially after implantation in the uterus. In past years, genetic, molecular and cell transplantation experiments in these animal models have established that the BMP, FGF, Nodal and Wnt signaling pathways play crucial roles in early embryonic patterning. However, there have been inconsistent results obtained from Xenopus, zebrafish and chick, probably caused by species-specific differences or by interference from complex germ-layer interactions. In the mouse embryo, knowledge of the molecular mechanisms governing early embryogenesis is limited due to the small size, complexity and inaccessibility of the early post-implantation embryo. Thus, mechanistic studies require an amenable system with embryonic properties but with the absence of the complexity that exists in vivo. Under these circumstances, pluripotent stem cells derived from different developmental stages have become an in vitro system for investigating the detailed molecular mechanisms through which signaling pathways regulate cell fate decisions. Both human and mouse embryonic stem cells (ESCs) are pluripotent cell lines derived from blastocyst-stage embryos [4][5][6][7]. Under appropriate culture conditions, ESCs differentiate into derivatives of all three germ layers, and the differentiation of specific cell types from ESCs is directed by a set of signals similar to that which regulates embryonic development in vivo [8][9][10][11]. Recently, another type of pluripotent stem cell, referred to as epiblast stem cells (EpiSCs), was derived from the late epiblast tissue of E5.5 mouse embryos. EpiSCs are molecularly and epigenetically distinct from mouse ESCs, but they share characteristics with human ESCs [12][13][14]. ESCs, corresponding to the ICM or the early epiblast state, combined with EpiSCs, which represent an in vitro equivalent of the late epiblast, form a novel system for studying the mechanisms of early embryonic development, especially for mechanistic studies at different developmental stages. In this review, we briefly summarize the key functions of the BMP, FGF, Nodal and Wnt signaling pathways in early embryogenesis, and then we discuss recent findings obtained from studies in ESCs and EpiSCs that reveal the stage-specific functions of BMP, FGF and Nodal signals. These findings begin to elucidate the mechanisms underlying different stages of early embryonic development. BMP Signaling The bone morphogenetic proteins (BMPs) are members of the transforming growth factor β (TGF-β) cytokine superfamily. BMP signaling has been shown to play a central role in ectodermal cell fate decisions. Using ectodermal explants (also called animal caps) from Xenopus blastula-stage embryos, researchers have shown that activation of the BMP pathway in ectoderm leads to the acquisition of an epidermal fate, whereas inhibition of BMP signaling by antagonists that are secreted by the Spemann organizer leads to a neural fate [15,16]. These results suggested that the ectoderm has a natural "default" tendency to differentiate into neural tissues unless it is instructed by BMP to become epidermis [17]. Since the "default model" was proposed, there has been debate concerning whether BMP inhibition is adequate for neural induction, because opposing results have been obtained with different assays in the chick. Initially, it was shown that grafts of cells expressing BMP4 or BMP7 failed to inhibit neural plate formation [18]. However, in epiblast explants from chick embryos, BMP4 showed a capacity to inhibit neural fate and promote an epidermal fate [19]. Furthermore, electroporation of BMP4 into the prospective neural plate inhibits the expression of the definitive neural markers (Sox2 and late Sox3), but it does not affect the pre-neural marker (early expression of Sox3). Therefore, BMP inhibition is probably required only as a late step during neural induction [20]. Other data relevant to the default model come from experiments showing that BMP antagonists are unable to induce neural character in the epidermal or extra-embryonic ectoderm of chick embryos [20][21][22]. Moreover, single or double BMP antagonist (chordin and noggin) mouse mutants show relatively little change in the initial size of the neural plate [23][24][25]. These results suggest that BMP inhibition is not sufficient to induce neural cells. The default model was not confirmed in mice until the generation of BMP receptor (Bmpr1a) knockout mice. Bmpr1a is the only type I BMP receptor expressed in the epiblast of implanted mouse embryos [26]. Knockout of Bmpr1a in mouse embryos, which completely inhibits BMP activity, was found to lead to premature neural differentiation of the epiblast accompanied by suppression of mesodermal fate [27]. Therefore, BMP inhibition is essential for neural differentiation in mice. However, the possibility that some other signals participate in neural induction cannot be excluded. That BMP inhibits neural differentiation continually in the epiblast rather than at a specific time point has been proposed [27]. The discrepancies of the data obtained from chick and mouse raise important questions, including whether there is a time point during which the BMP signal inhibits neural induction and what mechanisms are involved in this process. These questions are difficult to answer using only in vivo studies. Recently, some findings in pluripotent stem cells have shed light on these issues. In mouse ESCs, it was confirmed that BMP4 significantly inhibits neural differentiation, as it does in vivo [28]. Moreover, the addition of a BMP antagonist was found to result in an obvious increase in the number of neural cells [29]. Therefore, the ESC neural differentiation system is an amenable model in which to study the functions and mechanisms of BMP signaling.

BMPs have been implicated in maintaining the pluripotency of mouse ESCs through inducing the expression of inhibitor of differentiation (Id) genes to specifically block neural differentiation [30]. In addition to their roles in pluripotency maintenance, BMPs also induce mesodermal and epidermal differentiation in mouse ESCs [31][32][33][34][35]. The dual role of the BMP signal, to maintain ESC pluripotency and to induce non-neural differentiation, seems inconsistent. Recently, our studies in mouse ESCs and EpiSCs revealed that BMP signaling plays distinct roles during different stages of ESC neural differentiation [36]. We first found that cells at a specific period during mouse ESC neural differentiation mimic the late epiblast state and can be maintained as ESC-derived epiblast stem cells (ESD-EpiSCs). Thus, the ESC neural induction process can be divided into two stages: from ESCs to ESD-EpiSCs and from ESD-EpiSCs to neural progenitor cells. Using this system, it was revealed that BMP4 maintains ESC pluripotency by preventing cells from differentiating into late epiblast-like cells rather than by directly blocking ESC neural commitment. When the cells have been committed to becoming late epiblast cells, BMP4 cannot maintain their pluripotency, but it acts to inhibit neural induction by promoting mesodermal, epidermal and trophectodermal differentiation. Therefore, the late epiblast stage is the critical time point during which BMP4 switches its function from maintaining ESC pluripotency to promoting ESD-EpiSC non-neural differentiation. Based on this model, the molecular basis of the distinct functions of BMP4 was further investigated [36]. Ids, the direct downstream targets of BMP, were found to inhibit the conversion of ESCs into ESD-EpiSCs during the first stage, and to reduce ESD-EpiSC self-renewal, inhibit neural specification and promote mesodermal and trophectodermal differentiation during the second stage. FGF-Erk signaling was also found to be involved in the functions of BMP signaling during different stages. In ESCs, FGF-Erk activity was found to be reduced by short-term treatment with BMP4, whereas in ESD-EpiSCs, FGF-Erk activity increased during long-term treatment with BMP4. Therefore, BMP might perform its stage-specific roles by interfering with the FGF-Erk pathway. Due to the complexity of embryogenesis, the BMP signal may have divergent roles at other time points during the multi-step developmental process. For instance, BMP specifically induces an epidermal fate in ectoderm, which has been demonstrated in Xenopus. However, the ectoderm stage is usually overlooked during early embryonic development in chick and mouse because of its difficult accessibility and lack of markers. To confirm the default model of neural induction from ectoderm and the function of BMP during this process, an in vitro model of ectodermal cells needs to be established. We have recently identified ectoderm-like cells that form during mouse EpiSC neural and epidermal differentiation [37], and the signaling pathways involved in ectodermal cell commitment and neural differentiation are currently being investigated. FGF Signaling The studies reviewed above indicate a central role for BMP signaling in both the maintenance of mouse ESC pluripotency and neural inhibition. Data from different species show that the signaling of fibroblast growth factors (FGFs) also plays an important role during early embryogenesis through both BMP-dependent and -independent mechanisms. In Xenopus, it has been demonstrated that in addition to inhibition of BMP, pre-gastrula FGF signaling is also required in the ectoderm for the emergence of neural fates [38]. In chick embryos, the function of FGF in neural induction has been studied in detail. It was shown that inhibition of FGF signaling blocks neural induction [21]. Moreover, FGF initiates the onset of neural differentiation and suppresses BMP expression in the epiblast before gastrulation [19,21,39]. At later stages of embryogenesis, the major functions of FGF signaling are to induce mesoderm and to regulate movement during gastrulation [40,41]. FGFs control the specification and maintenance of mesoderm by regulating T box transcription factors in Xenopus [42,43], zebrafish [44,45] and mouse [46,47]. Thus, FGF signaling initiates the onset of neural differentiation in the epiblast before gastrulation, and it induces mesoderm formation and regulates gastrulation movements during later stages. However, an important role for FGF signaling in neural induction has not been conclusively demonstrated in mice. Mutation of Fgf4 or Fgfr2 results in peri-implantation lethality [48,49], which suggests that FGF signaling is required very early during embryogenesis. Because of this early lethality, it is difficult to determine the role of FGF signaling in neural induction. Fgf8 and Fgfr1 are also expressed in the blastocyst, but they appear to function later. Both Fgf8 mutants and Fgfr1 mutants die at late gastrulation with impaired axis formation and mesoderm specification [50,51]. In Fgf8 −/− embryos, patterning of the prospective neuroectoderm is greatly perturbed and the range of anterior neuroectoderm markers is widely expanded [46]. Another study showed that inhibition of FGF signaling in E5.5-E6.5 mouse embryos leads to a drastic increase in the proportion of embryos displaying ectopic expression of the neural marker Hesx1 [27]. Therefore, a positive effect of FGF in neural induction is not supported by the data from mouse embryos. Rather, FGF signaling negatively regulates the specification of neuroectoderm cell fate in post-implantation mouse embryos. Thus, the function of FGF signaling in the pre-implantation mouse embryo needs to be elaborated. The answer to this question may explain the neural induction effect of FGF observed in Xenopus and chick. Recently, the early roles of FGF signaling have become increasingly understood using mouse pluripotent stem cells. In 2007, Stavridis et al. found that inhibition of FGF/Erk during mouse ESC differentiation abolished neuronal induction with no reduction in levels of the pluripotency marker Nanog [52]. Kunath et al. further showed that mouse ESCs lacking Fgf4 or Erk2 and those treated with FGFR inhibitors not only resist neural induction, but also fail to undergo mesodermal differentiation even when BMP4 is added. Moreover, Erk2-null ESCs retain their expression of the pluripotency markers Oct4, Nanog and Rex1 when differentiated in adherent culture [53]. Therefore, when FGF signaling is inhibited, ESCs are deficient in the ability to commit to multiple lineages and stay in an undifferentiated state. These findings indicate that the FGF-Erk pathway primes ESCs for differentiation into a transitional stage that is analogous to the late epiblast state. It is possible that FGF signaling does not directly induce neural specification at the early stage of development but instead promotes ESCs to exit from self-renewal and enter a state in which they are more prepared for differentiation. Subsequently, Ying et al. found that FGF-Erk inhibitors, in combination with the inhibition of glycogen synthase kinase 3 (GSK3; which promotes cellular growth and viability), keep mouse ESCs in an undifferentiated state called the "ground state" [54]. Based on this finding, it was predicted that it should be possible to capture true ESCs from epiblasts of other species by blocking the FGF-Erk and GSK3 pathways [55]. Following this line of reasoning, germline-competent rat ESCs have been successfully generated by combining LIF with FGF-Erk and GSK3 inhibition [56,57]. Recently, human ESCs with biological and epigenetic characteristics similar to those of mouse ESCs were created by ectopic induction of Oct4, Klf4 and Klf2 combined with LIF and inhibition of the FGF-Erk and GSK3 pathways [58]. These data suggested that inhibition of FGF signaling blocks the transition of ESCs to a differentiation state. However, there was no direct evidence to suggest that this transition state is the late epiblast state. With the derivation of EpiSCs from mouse ESCs, this question could be resolved. In our recent report, we showed that either FGF4 or FGF2 can significantly increase the number of ESD-EpiSC colonies derived from differentiated ESC aggregates and that inhibition of FGF-Erk activity dramatically reduces the number of ESD-EpiSCs. Moreover, FGF2 partially counteracts the BMP4-induced reduction in ESD-EpiSC colony numbers [36]. These results strongly suggest an important role for FGF in the derivation of EpiSCs from ESCs. Recently, another group also reported the creation of an ESC-based system for isolating epiblast cells that are similar to EpiSCs, through which they confirmed that FGF signaling is crucial for priming ESCs to differentiate into the late epiblast state [59]. In addition to the early roles of FGF, the subsequent functions of FGF signaling starting from the late epiblast stage are being investigated in EpiSCs. First, as in human ESCs, FGF2 is required to maintain EpiSC pluripotency [12,13]. However, the mechanism through which FGF2 stabilizes the pluripotent state is different between these two cell types [60]. In human ESCs, FGF signaling in cooperation with SMAD2/3 signaling mediates NANOG expression, thereby actively promoting ESC self-renewal. In EpiSCs, however, FGF2 fails to regulate NANOG expression. Rather, it supports the epiblast state by inhibiting neuroectodermal induction, particularly by blocking Pax6 expression [60]. Blockage of FGF signaling in EpiSCs promotes rapid neural induction and subsequent neurogenesis [59]. These results are consistent with previous findings in mutant mice that FGF negatively regulates the specification of neuroectodermal cell fate in post-implantation mouse embryos [27,46,50,51,61,62]. In summary, FGF signaling plays distinct roles during different developmental stages. FGF signaling is crucial for priming ESCs to differentiate into the late epiblast state. It then acts to inhibit the subsequent transition to neuroectoderm. This conclusion is distinct from the notion that neural differentiation is promoted by FGF signaling. In the future, it will be interesting to determine the exact functions of divergent FGF ligands during commitment to the three germ layers from the epiblast and to elucidate the molecular mechanisms underlying these processes using EpiSCs. Nodal Signaling Nodal, a member of the TGF- family, plays central roles in early embryo patterning during the induction of mesoderm and endoderm and the specification of left-right asymmetry. In some cases, Nodal signaling is also referred to as "Activin/Nodal" signaling because another TGF-family member, Activin, binds to the same receptors as Nodal (with the exception of the coreceptor Cripto) and triggers similar intracellular events [63]. The roles of Nodal signaling are evolutionarily conserved and have been well established using molecular genetic studies in various animal models. In Xenopus, Activin and Xnrs (Xenopus homologues of Nodal) function as mesoderm inducers in whole embryos and explanted animal caps [64][65][66]. Inhibition of Nodal signaling was found to block the formation of mesoderm and endoderm [67][68][69]. Genetic studies in zebrafish and mouse have also provided strong evidence that Nodal signaling is essential for mesendoderm formation. Zebrafish that are double mutants for Cyclops and Squint, the orthologs of Nodal, lack head and trunk mesoderm and fail to form the germ-ring, an organizer analogous to the mouse primitive streak [70]. Similarly, mice lacking Nodal lose the primitive streak and most mesoderm [71]. In addition, Nodal signaling is essential for the formation and patterning of the anterior visceral endoderm (AVE) in post-implantation mouse embryos [72]. AVE produces the Nodal antagonists Lefty1 and Cerberus-1 (Cer1) to limit the extension of primitive streaks and to maintain the correct patterning of the epiblast [73]. Although additional signals such as BMP and FGF are also involved in gastrulation, genetic evidence suggests that Nodal signals are core players in the formation of the primitive streak and in epiblast patterning. Nodal signals may also act as anti-differentiation signals to maintain epiblast proliferation. Analysis of Nodal mutants and embryo explants suggests that Nodal signaling within the epiblast is essential for maintaining pluripotency determinants such as Oct4 and Nanog and for preventing precocious neuroectoderm differentiation [74,75]. In vitro studies have also confirmed that Nodal signaling is required for the maintenance of undifferentiated human ESCs and mouse EpiSCs [12,76]. Activin/Nodal signaling through SMAD2/3 activation is required to sustain the self-renewal of these pluripotent stem cells. It has been shown that SMAD2/3 binds to the NANOG promoter and thereby activates NANOG gene transcription in human ESCs [77,78] and mouse EpiSCs [60]. These results are consistent with findings in mouse embryos that Nodal is required to maintain epiblast pluripotency. Moreover, in vitro studies have revealed the corresponding molecular mechanism underlying this function. In mouse ESCs, EpiSCs and human ESCs, Activin/Nodal signaling is also known to drive differentiation toward mesendoderm [14,[79][80][81]. The functions of Activin/Nodal signaling in pluripotency maintenance and mesendoderm induction seem contradictory. Recently, Chng et al. provided an explanation for how Activin/Nodal signaling maintains pluripotency without inducing mesendoderm in human ESCs and mouse EpiSCs [82]. They found that Smad-interacting protein 1 (SIP1) has an essential role in the promotion of neuroectoderm differentiation and the suppression of mesendodermal genes induced by Activin/Nodal signaling. In turn, Activin/Nodal signaling cooperates with NANOG, OCT4 and SOX2 to control the expression of SIP1 in human ESCs, thereby

limiting the neuroectoderm promoting effects of SIP1. Therefore, SIP1 limits the mesendoderm inducing effects of Activin/Nodal signaling without inhibiting the pluripotency maintaining effects exerted by SMAD2/3. This conclusion was confirmed in mouse EpiSCs, implying that these mechanisms are conserved in different species and may operate in vivo during mammalian development [82]. Overall, there are still many unresolved questions about how Activin/Nodal signaling switches its function from pluripotency maintenance to mesendoderm induction. Future studies should focus on revealing the full molecular cascades by which Nodal signaling controls these cell-fate choices. Wnt Signaling Wnt signaling is mediated by blocking the activity of glycogen synthase kinase 3β (GSK3β), which promotes β-catenin degradation. Inhibition of Gsk3β activity thus stabilizes β-catenin, which activates the Wnt pathway. Wnt signaling plays multiple roles during early embryonic development. In both Xenopus and chick embryos, Wnt signaling represses neural fate and induces epidermal fate by attenuating the responsiveness of epiblast cells to FGF signaling [83][84][85][86]. In addition, Wnt signaling is required for primitive streak formation in chick and mouse embryos [87,88]. Conversely, a study in Xenopus found that Wnt signaling promotes neural development in ectoderm through inhibition of BMP4 expression [89]. Furthermore, Xiro1, a downstream target of the Wnt pathway, was shown to repress BMP4 expression [90]. These contradictory findings might be due to the different developmental stages during which the experiments were conducted. At early stages of development, Wnt signaling is activated on the dorsal side of the embryo, where it represses the activity of BMP signaling and specifies the dorsal character. Consequently, cells on the dorsal side give rise to neural cells [89]. However, at the blastula/early gastrula stages, Wnt signaling may suppress the generation of neural cells [86]. Therefore, it becomes necessary to inhibit Wnt signaling for normal neural cell development at later stages [91]. The functions and mechanisms of Wnt signaling have also been investigated in mouse ESCs. In agreement with observations in vivo, Wnt signaling is required for mesoderm differentiation in mouse ESCs. It has been shown that inhibition of endogenous Wnt signals in mouse ESCs prevents the expression of primitive streak-, endoderm-and mesoderm-associated genes and abrogates the functional development of mature mesodermal lineages [92]. However, β-catenin alone is not sufficient to promote primitive streak-associated gene expression, indicating that Wnt/β-catenin may cooperate with other signals to regulate germ layer induction. Another study in mouse ESCs showed that Wnt and Nodal signaling are required to act together for the formation of mesendodermal cells [10]. In summary, Wnt signaling cooperates with other signaling pathways to regulate mesoderm and endoderm differentiation in mouse ESCs. Recently, Sato et al. showed that Wnt activation stimulated by 6-bromoindirubin-3'-oxime (BIO), a specific pharmacological inhibitor of GSK3, sustains the expression of the pluripotent state-specific transcription factors Oct4, Rex1 and Nanog in mouse and human ESCs [93]. In addition, Doble et al. reported that overexpression of stabilized β-catenin in mouse ESCs inhibits neuronal differentiation and delays loss of pluripotency; moreover, β-catenin forms a complex with Oct4 and enhances Oct4 activity [94]. Therefore, Wnt/β-catenin signaling plays a role, in part through its interaction with Oct4, in the maintenance of pluripotency. However, Wnt signaling alone is not sufficient to maintain the ground state of mouse ESCs. Ying et al. showed that blockage of GSK3 in mouse ESCs enhances growth capacity and suppresses neural differentiation, but it also promotes non-neural differentiation. To block differentiation of ESCs into the cells that make up the three germ lineages, the combination of a GSK3 inhibitor and an FGF-Erk inhibitor is necessary [54]. At present, no published data fully elucidate the functions of Wnt signaling in EpiSC differentiation. Considering the multiple roles of Wnt signaling in early embryonic development, we speculate that Wnt signaling may play distinct roles at different stages of development, similarly to the BMP, FGF and Nodal pathways. Further investigation into the functions of Wnt and its crosstalk with other signaling pathways is needed, especially in EpiSCs. Outlook The BMP, FGF, Nodal and Wnt signaling pathways play important roles during embryogenesis. However, the mechanisms underlying cell fate decisions during vertebrate embryogenesis are complex. ESCs and EpiSCs provide in vitro systems for investigating the complex mechanisms through which signaling pathways play distinct roles during different developmental stages. Recently, published data [95] and our unpublished results suggest that it may be possible to derive ectodermal cells directly from mouse ESCs and EpiSCs. Ectodermal cells would work as a unique in vitro tool to study the mechanisms involved in ectoderm commitment and in neural and epithelial differentiation during embryonic development. Studies in pluripotent stem cells that correspond to different developmental stages would provide a foundation for efforts to guide the differentiation of pluripotent stem cells along selected developmental pathways for potential therapeutic use. Fast Detection of 2,4,6-Trinitrotoluene (TNT) at ppt Level by a Laser-Induced Immunofluorometric Biosensor The illegal use of explosives by terrorists and other criminals is an increasing issue in public spaces, such as airports, railway stations, highways, sports venues, theaters, and other large buildings. Security in these environments can be achieved by different means, including the installation of scanners and other analytical devices to detect ultra-small traces of explosives in a very short time-frame to be able to take action as early as possible to prevent the detonation of such devices. Unfortunately, an ideal explosive detection system still does not exist, which means that a compromise is needed in practice. Most detection devices lack the extreme analytical sensitivity, which is nevertheless necessary due to the low vapor pressure of nearly all explosives. In addition, the rate of false positives needs to be virtually zero, which is also very difficult to achieve. Here we present an immunosensor system based on kinetic competition, which is known to be very fast and may even overcome affinity limitation, which impairs the performance of many traditional competitive assays. This immunosensor consists of a monolithic glass column with a vast excess of immobilized hapten, which traps the fluorescently labeled antibody as long as no explosive is present. In the case of the explosive 2,4,6-trinitrotoluene (TNT), some binding sites of the antibody will be blocked, which leads to an immediate breakthrough of the labeled protein, detectable by highly sensitive laser-induced fluorescence with the help of a Peltier-cooled complementary metal-oxide-semiconductor (CMOS) camera. Liquid handling is performed with high-precision syringe pumps and chip-based mixing-devices and flow-cells. The system achieved limits of detection of 1 pM (1 ppt) of the fluorescent label and around 100 pM (20 ppt) of TNT. The total assay time is less than 8 min. A cross-reactivity test with 5000 pM solutions showed no signal by pentaerythritol tetranitrate (PETN), 1,3,5-trinitroperhydro-1,3,5-triazine (RDX), and octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (HMX). This immunosensor belongs to the most sensitive and fastest detectors for TNT with no significant cross-reactivity by non-related compounds. The consumption of the labeled antibody is surprisingly low: 1 mg of the reagent would be sufficient for more than one year of continuous biosensor operation. Introduction The fast and extremely sensitive detection of explosives [1][2][3] is one of the most relevant tasks to guarantee security in areas of public access. Many airplane passengers are confronted with some As antibody binding is a reversible process, antibodies with off-rates far longer than the column passage time are required to allow sensitive detection of the analyte. This means that high-affinity antibodies are needed for optimal assay performance. Furthermore, inactive antibodies would contribute to the background of this assay. Hence, highly purified antibody conjugates are preferable. Trinitrophenyl-BSA Conjugates and Indirect Competitive Enzyme-Linked Immunosorbent Assay (ELISA) Trinitrophenyl-(TNP)-BSA conjugate for the affinity column coating: 20 mg (0.3 µmol) of BSA was dissolved in 1 mL of 0.2 M NaHCO 3 , and 26.4 µL (4.5 µmol) of 5 (w/v)% aqueous trinitrobenzene sulfonic acid (TNBS) was added, vortexed and stored for one hour at RT and subsequently for 48 h at 4 • C. After incubation, 125 µL of 2 M NaH 2 PO 4 was added to adjust the solution to a neutral pH. A mean degree of labeling (DOL) of approximately 8 TNP per BSA was determined with MALDI-TOF MS (see Figure S2) TNP-BSA conjugate for indirect enzyme-linked immunosorbent assay (ELISA): In 5 mL of 0.2 M NaHCO 3 100 mg (1.5 µmol) of BSA were dissolved and 88 µL (15 µmol) of TNBS (5%) were added, vortexed and stored for 1 h at RT and subsequently for 48 h at 4 • C. After incubation, 625 µL of 2 M NaH 2 PO 4 was added to adjust the solution to a neutral pH. A DOL of 5 TNP per BSA was determined by MALDI-TOF MS (see Figure S14). ELISA procedure: Clear, high binding 96-well plates (MTP) with flat bottom were coated with 100 µL of a blend of 0.023 g/L TNP-BSA and 0.75 g/L BSA in phosphate-buffered saline (PBS) with 10 mM phosphate and 137 mM sodium chloride pH = 7.4 (100 µL per well). The plate was sealed with Parafilm, protected from light with aluminum foil and shaken at 750 rpm for 20 h at RT. The MTP was washed with PBS containing 0.05 vol% Tween 20 by an automated plate washer three times. Then 50 µL of diluted TNT in PBS ranging from 10 pM to 10 µM and 50 µL of 1:20,000 diluted A1.1.1-Dy-654 (approximately 17.5 µg/L) in PBS were added as quadruplicates and incubated for one hour at RT in the dark. After a washing step, 100 µL of 40 µg/L HRP-conjugated anti-mouse (H + L) IgG antibody in PBS with 1% BSA were incubated for one hour in the dark and the MTP was subsequently washed. Then 100 µL TMB substrate (Seramun Blau fast2) was incubated for 30 min and stopped with 100 µL of 0.25 M sulfuric acid. The absorbance was recorded with an Epoch2 Photometer at 450 and 620 nm. For the preparation of the TNP-BSA affinity column, 440 µL of the 7.8 eq. TNP-BSA solution was diluted with 2100 µL 0.1 M Na 2 HPO 4 pH 8.1 and incubated for one week at RT on the epoxy-functionalized raw column. of the labeling solution (3.8 nmol Dy-654-NHS) were added. The solutions were gently mixed by pipetting and incubated for 3 h at 14 • C in the dark and for 18 h in the dark at 4 • C. The solution was purified with a PBS conditioned PD Spintrap G-25. The conjugate was stabilized with 0.04 vol% ProClin300 and stored in the dark at 4 • C. Fluorescence Detector The custom detector based on an epifluorescence microscope setup was built from modular and affordable parts (see Table S2, Supplementary Materials). The detailed construction plans can be found in Figure S4 (SI). As a detector, the camera QHY174M-GPS from QHYCCD (Beijing, China) was used, featuring an IMX-174M (Sony) CMOS sensor. This sensor is thermoelectrically (Peltier) cooled, has 5.86 × 5.86 µm square pixels, and delivers a maximal dynamic range of 12 bit. Data are acquired via a USB 3.0 connection by a laptop with the Software SharpCap (Version 3.0.4074.0). In the excitation path ( Figure 2, orange color), a diode laser (70105582, Picotronic) with a measured center-wavelength of 638 nm with a full width at half maximum (FWHM) of approximately 3 nm (see Figure S7), and an optical output power below 1 mW was used. The laser was directed through an FL635-10 bandpass filter and guided at 45 • on a dichroic mirror (DMLP638R, THORLABS) with a cutoff wavelength of 638 nm; the excitation was focused by an infinity-corrected plan achromat 10×/0.25 Ph1 objective (415500-1605-001, ZEISS) on a microfluidic COC (cyclic olefin copolymer) chip. The microfluidic chip (10000091, Microfluidic ChipShop) features four 200 × 200 µm linear flow channels and is mounted on a custom holder for precise adjustment (see Figure S6). In the fluorescence path ( Figure 2, red color), the fluorescence of the label in the flow channel is collected by the objective, filtered by two stacked long-pass filters (FELH650, THORLABS), and focused with tube lens (focal length 165 mm, TTl165-A, THORLABS) on the sensor of the camera (QHY174M-GPS, QHY). The whole setup is mounted on an optical breadboard and protected from environmental light and dust with a box made of black cardboard. Fluorescence Detector The custom detector based on an epifluorescence microscope setup was built from modular and affordable parts (see Table S2, Supplementary Materials). The detailed construction plans can be

found in Figure S4 (SI). As a detector, the camera QHY174M-GPS from QHYCCD (Beijing, China) was used, featuring an IMX-174M (Sony) CMOS sensor. This sensor is thermoelectrically (Peltier) cooled, has 5.86 × 5.86 µm square pixels, and delivers a maximal dynamic range of 12 bit. Data are acquired via a USB 3.0 connection by a laptop with the Software SharpCap (Version 3.0.4074.0). In the excitation path ( Figure 2, orange color), a diode laser (70105582, Picotronic) with a measured center-wavelength of 638 nm with a full width at half maximum (FWHM) of approximately 3 nm (see Figure S7), and an optical output power below 1 mW was used. The laser was directed through an FL635-10 bandpass filter and guided at 45° on a dichroic mirror (DMLP638R, THORLABS) with a cutoff wavelength of 638 nm; the excitation was focused by an infinity-corrected plan achromat 10×/0.25 Ph1 objective (415500-1605-001, ZEISS) on a microfluidic COC (cyclic olefin copolymer) chip. The microfluidic chip (10000091, Microfluidic ChipShop) features four 200 × 200 µm linear flow channels and is mounted on a custom holder for precise adjustment (see Figure S6). In the fluorescence path ( Figure 2, red color), the fluorescence of the label in the flow channel is collected by the objective, filtered by two stacked long-pass filters (FELH650, THORLABS), and focused with tube lens (focal length 165 mm, TTl165-A, THORLABS) on the sensor of the camera (QHY174M-GPS, QHY). The whole setup is mounted on an optical breadboard and protected from environmental light and dust with a box made of black cardboard. Measurements All samples and buffers were injected with a Fusion 4000 dual syringe pump (Chemyx). The flow was directed to the column holder, which contained the monolithic column, and the eluate was subsequently passed through the flow channel of the microfluidic chip (see Figure 2). In order to ensure bioinert conditions, all connectors were manufactured from PEEK or PP, and the seals were made from silicone. The column holder was made by additive manufacturing and had custom PEEK connectors; the columns may conveniently be exchanged (see Figure S1). The typical backpressure of the monolithic column at a flow rate of 0.5 mL min −1 was approximately 2.7 bar (see Figure S3). Therefore, the actual overall system pressure is low enough to operate the whole fluidics with Measurements All samples and buffers were injected with a Fusion 4000 dual syringe pump (Chemyx). The flow was directed to the column holder, which contained the monolithic column, and the eluate was subsequently passed through the flow channel of the microfluidic chip (see Figure 2). In order to ensure bioinert conditions, all connectors were manufactured from PEEK or PP, and the seals were made from silicone. The column holder was made by additive manufacturing and had custom PEEK connectors; the columns may conveniently be exchanged (see Figure S1). The typical backpressure of the monolithic column at a flow rate of 0.5 mL min −1 was approximately 2.7 bar (see Figure S3). Therefore, the actual overall system pressure is low enough to operate the whole fluidics with standard plastic (PP) syringes and PTFE-silicone tubes. The total flow rates were limited to 0.5 mL min −1 to protect the microfluidic chip with the linear flow cell. Setup of the sensitivity test: Dy-654-COOH was diluted to from 5 to 25 pM in 5 pM steps and from 100 to 1000 pM in 100 pM steps and injected with a 12 mL syringe with PBS blanks in between the samples. The data were recorded with an exposure time of 5000 ms; the gain was set to four, and the sensor temperature to −5 • C. TNT detection: TNT was diluted from a stock solution in ethanol to 0.4 to 2 nM in 0.4 nM steps and from 4 to 20 nM in 4 nM steps in PBS with 0.1 (w/v)% BSA and mixed 1:1 with A1.1.1-Dy-654 (1:50,000, approximately 7 µg/L) in PBS with 0.1 (w/v)% BSA, incubated for five minutes and injected through a six-way-valve as shown in Figure 3. The data were recorded as described above. Biosensors 2020, 10, x FOR PEER REVIEW 6 of 18 standard plastic (PP) syringes and PTFE-silicone tubes. The total flow rates were limited to 0.5 mL min −1 to protect the microfluidic chip with the linear flow cell. Setup of the sensitivity test: Dy-654-COOH was diluted to from 5 to 25 pM in 5 pM steps and from 100 to 1000 pM in 100 pM steps and injected with a 12 mL syringe with PBS blanks in between the samples. The data were recorded with an exposure time of 5000 ms; the gain was set to four, and the sensor temperature to −5 °C. TNT detection: TNT was diluted from a stock solution in ethanol to 0.4 to 2 nM in 0.4 nM steps and from 4 to 20 nM in 4 nM steps in PBS with 0.1 (w/v)% BSA and mixed 1:1 with A1.1.1-Dy-654 (1:50,000, approximately 7 µg/L) in PBS with 0.1 (w/v)% BSA, incubated for five minutes and injected through a six-way-valve as shown in Figure 3. The data were recorded as described above. Figure 3. The data were recorded as described above. Fluorescence Detector The two stacked long-pass filters and the dichroic mirror efficiently prevent stray light from reaching the CMOS sensor. The excitation laser has a center wavelength of 638 nm and an FWHM of approximately 3 nm. In preliminary experiments, the use of an additional second long-pass filter has proven to be beneficial. The microfluidic chip is made from TOPAS ® (cyclic olefin copolymer, COC), which shows a very low autofluorescence at the used wavelength and has a smooth flat and transparent surface suitable for observation. The fluorescence is collected with the same 10X0.25 NA objective and guided through the dichroic mirror and passed through the long-pass-filters. Overall a high transmission for the fluorescence is to be expected as the filters show high transmission of about 90% at 654 nm, which is the peak wavelength of Dy-654 emission. The monochromatic sensor of the QHY is reported to exhibit a high quantum yield of approximately 50% at 650 nm and is, therefore, well-suited for the application. The position of the excitation laser spot on the microfluidic chip can be conveniently adjusted by the laser holder (see Figure S5) and is set in the center of the flow channel. Semi-Automated Data Evaluation with Python The images are recorded as a sequence of raw files (.fits) and have a native resolution of 1920 × 1080 pixels. The size of the laser-illuminated spot, which essentially defines the region of interest (ROI), is only a few pixels wide. Due to this small area, the exact position of the ROI in every frame (exposure time: 5000 ms) is of high importance for the correct and reliable data evaluation. It was Figure 3. The data were recorded as described above. Fluorescence Detector The two stacked long-pass filters and the dichroic mirror efficiently prevent stray light from reaching the CMOS sensor. The excitation laser has a center wavelength of 638 nm and an FWHM of approximately 3 nm. In preliminary experiments, the use of an additional second long-pass filter has proven to be beneficial. The microfluidic chip is made from TOPAS ® (cyclic olefin copolymer, COC), which shows a very low autofluorescence at the used wavelength and has a smooth flat and transparent surface suitable for observation. The fluorescence is collected with the same 10X0.25 NA objective and guided through the dichroic mirror and passed through the long-pass-filters. Overall a high transmission for the fluorescence is to be expected as the filters show high transmission of about 90% at 654 nm, which is the peak wavelength of Dy-654 emission. The monochromatic sensor of the QHY is reported to exhibit a high quantum yield of approximately 50% at 650 nm and is, therefore, well-suited for the application. The position of the excitation laser spot on the microfluidic chip can be conveniently adjusted by the laser holder (see Figure S5) and is set in the center of the flow channel. Semi-Automated Data Evaluation with Python The images are recorded as a sequence of raw files (.fits) and have a native resolution of 1920 × 1080 pixels. The size of the laser-illuminated spot, which essentially defines the region of interest (ROI), is only a few pixels wide. Due to this small area, the exact position of the ROI in every frame (exposure time: 5000 ms) is of high importance for the correct and reliable data evaluation. It was observed that within elongated measurement periods, the ROI position might be shifted slightly by a few pixels in x-and or y-direction. This behavior is most likely a result of the thermal expansion of the additive manufactured laser-holder ( Figure S5). In order to account for this shift, the correct position of the ROI must be determined automatically but precisely for every frame. By a semi-automated python-script, a Gaussian fit is utilized to determine the center pixel of the laser spot, based on which a pixel-square ROI is defined. Around the determined laser center, a square ROI region of 8 × 8 pixels is defined, and all 64 pixels inside are sorted by their intensity. The three most intense pixels are discarded to account for possible cosmic rays or hot pixels. Subsequently, the following five pixels are used to calculate the mean value, which is defined as intensity for the frame. To determine the signal intensity of an injected probe, for example, the 100 pM Dy-654-COOH solution, 16 frames are used to calculate the mean value and the standard deviation. To determine the starting point for the signal evaluation (f n ), the raw-signal is smoothed by a Savitzky-Golay filter with a polynomial of second-order and a window of five frames. Subsequently, the gradient of the smoothed data is calculated. The first sample point of the 1st derivative to fall below zero after the initial signal increase is picked (see Figure S8) and defined as f n , as it represents a stable signal as growth is completed. Based on this frame, f n and the next 15 frames are used to calculate the mean and the standard deviation of the signal. The employed antibody ultimately governs the quality of an immunoassay. Thus, the antibody used must be chosen carefully with sensitivity, selectivity, and stability in mind. Many manufacturers only give an order number of an antibody. The properties or even the true clone identity may remain unclear [69]. This creates a risk for reliable and reproducible assays, which must not be tolerated when this information is critical. Two commercially available monoclonal anti-TNT antibodies A1.1.1 (SDIX, USA) and EW75C (BBI Solutions, UK) from mouse IgG 1 subclass were evaluated for their affinity to TNT for validation purposes with indirect competitive ELISA as described in Section 2.2. Also, their specific cross-reactivities with compounds of interest and high explosives (see Table S2 and Figures S16 and S17) were determined. The clone A1.1.1 showed superior affinity to the analyte TNT by a factor of 30 ( Figure S18) compared to the clone EW75C and was, therefore, chosen for this project. Additionally, for both clones, a mass spectrometric antibody fingerprint, according to Tscheuschner et al. [69], was generated (see Figure S19), which may be useful to identify these clones in future work. The sensor system relies on the sensitive detection of fluorophore-labeled antibodies in the eluate of the affinity column. A proper choice of fluorophore is, therefore, of considerable importance. A suitable label should display desirable properties like high quantum yield, high photostability, excellent water solubility, reduced aggregation, and low non-specific binding. Additionally, no detectable cross-reactivity with the antibody is imperative. Based on the available laser excitation source of 638 nm and excellent performance on epoxy-functionalized glass substrates [70], the cyanine dye Dy-654 [ Figure 4] was chosen. The label features four sulfonic acids, which results in highly hydrophilic properties and a minimal tendency for aggregation. Furthermore, the dye showed no detectable cross-reactivity with the antibody A1.1.1 in preliminary experiments. The degree of labeling (DOL) for the A1.1.1-Dy-654 conjugate was determined with MALDI-TOF MS to be approximately 12 (see Figure 4) and the protein concentration of the A1.1.1-Dy-654 stock solution was determined to contain approximately 0.35 g/L antibody according to UV measurements. The test midpoint

(IC 50 ) was determined by indirect competitive ELISA to be 1.2 nM (see Figure 5), which is in excellent agreement with the literature stated value determined for the clone A1.1.1 of 1.3 nM [45]. The limit of detection (LOD) of the indirect ELISA was determined to be approximately 170 pM. 2,4,6-Trinitrophenyl-(TNP)-BSA Affinity Columns The degree of labeling of the BSA was determined by MALDI-TOF MS to be approximately 8 TNP molecules per BSA (see Figure S2, Supplementary Materials). In preliminary tests, the TNP-BSA and the BSA column showed no non-specific interaction for the label Dy-654-COOH or Dy-654-labeled human IgG (Avastin, Bevacizumab). The backpressure of a monolithic column was determined to be about 2.7 bar at a flow rate of 0.5 mL min −1 of PBS (see Figure S3). The columns were purged with a mixture of 80 vol% ethanol/water and stored immersed in this solution in a sealed vial at 4 • C for several months without noticeable degradation of column performance. 2,4,6-Trinitrophenyl-(TNP)-BSA Affinity Columns The degree of labeling of the BSA was determined by MALDI-TOF MS to be approximately 8 TNP molecules per BSA (see Figure S2, Supplementary Materials). In preliminary tests, the TNP-BSA and the BSA column showed no non-specific interaction for the label Dy-654-COOH or Dy-654-labeled human IgG (Avastin, Bevacizumab). The backpressure of a monolithic column was determined to be about 2.7 bar at a flow rate of 0.5 mL min −1 of PBS (see Figure S3). The columns were purged with a mixture of 80 vol% ethanol/water and stored immersed in this solution in a sealed vial at 4 °C for several months without noticeable degradation of column performance. Setup Optimization and Limit of Detection (LOD) of the Label The exposure time, the sensor gain, and the sensor temperature were varied to determine the optimal ratio of signal height and noise (S/N). To calculate the S/N, the signal difference between the signal intensity of 100 pM Dy-654-COOH dissolved in PBS and pure PBS was divided by the sum of the standard deviation of the Dy-654-COOH and the blank signal. The most substantial influence on the S/N was observed for the exposure time. Longer exposure times up to 5000 ms and even beyond, increased the S/N (see Figure S9). Increasing sensor gain reduced the S/N, especially at a gain >4 (see Figure S10). The temperature had no clear impact on S/N (see Figure S11), as the known hot pixels were already removed by the python script. An exposure time of 5000 ms and a gain of 4 was chosen to achieve a wide dynamic range and acceptable response times. The sensor temperature was set to -5 °C, which was the lowest temperature the camera was able to keep over longer times at ambient temperatures around 25 °C. These settings were applied in all measurements in this paper if not stated otherwise. Solutions of Dy-654-COOH were prepared in PBS from 5 to 25 pM to determine the LOD and limit of quantification (LOQ) for the label and dilutions from 100 to 1000 pM to assess the range of the linear response (see Figure 6, top). After a stable signal was achieved, 16 frames were used to calculate the mean and standard deviation for each dilution step, as described above. The LOD and Setup Optimization and Limit of Detection (LOD) of the Label The exposure time, the sensor gain, and the sensor temperature were varied to determine the optimal ratio of signal height and noise (S/N). To calculate the S/N, the signal difference between the signal intensity of 100 pM Dy-654-COOH dissolved in PBS and pure PBS was divided by the sum of the standard deviation of the Dy-654-COOH and the blank signal. The most substantial influence on the S/N was observed for the exposure time. Longer exposure times up to 5000 ms and even beyond, increased the S/N (see Figure S9). Increasing sensor gain reduced the S/N, especially at a gain >4 (see Figure S10). The temperature had no clear impact on S/N (see Figure S11), as the known hot pixels were already removed by the python script. An exposure time of 5000 ms and a gain of 4 was chosen to achieve a wide dynamic range and acceptable response times. The sensor temperature was set to -5 • C, which was the lowest temperature the camera was able to keep over longer times at ambient temperatures around 25 • C. These settings were applied in all measurements in this paper if not stated otherwise. Solutions of Dy-654-COOH were prepared in PBS from 5 to 25 pM to determine the LOD and limit of quantification (LOQ) for the label and dilutions from 100 to 1000 pM to assess the range of the linear response (see Figure 6, top). After a stable signal was achieved, 16 frames were used to calculate the mean and standard deviation for each dilution step, as described above. The LOD and the LOQ were calculated by the addition of 3 or 10 times the standard deviation of the blank sample. For this setup, a LOD of about 1 pM was achieved for Dy-654-COOH. The dynamic range was about a factor of 250, ranging from at least 4 to 1000 pM. A highly linear response (Figure 6, bottom) was achieved for Dy-654-COOH. the LOQ were calculated by the addition of 3 or 10 times the standard deviation of the blank sample. For this setup, a LOD of about 1 pM was achieved for Dy-654-COOH. The dynamic range was about a factor of 250, ranging from at least 4 to 1000 pM. A highly linear response (Figure 6, bottom) was achieved for Dy-654-COOH. Performance of Affinity Column To evaluate the performance of the TNP-BSA affinity column, a 1:100,000 dilution of the labeled antibody (≈3.5 µg/L) was injected onto the affinity column at varying flow rates, and the signal intensity of the eluate was monitored. At the lowest flow rate of 0.0625 mL min −1 the highest antibody retention of approximately 70% was observed. The antibody removal efficiency gradually declined until the highest flow rate of 0.5 mL min −1 was applied with approximately 54% retention (see Figure S12). When the dead volume of ca. 1 mL is considered (tubing, connectors, and the affinity column), a flow rate of 0.0625 mL min −1 would result in a dead time between injection and measurement of about 16 min. But this delay can be reduced to about 2 min if the highest flow rate of 0.5 mL min −1 would be applied. For all further measurements, a flow rate of 0.5 mL min −1 was used. TNT Measurements Samples from 2 to 10 nM of TNT in 3.5 µg/L of antibody conjugate were incubated for five minutes and injected at 0.5 mL min −1 to determine the dynamic range for the analyte TNT. The results showed that the linear range does not extend well beyond 2 nM for the chosen parameters (see Figure S13). The signal displays an asymptotic behavior. In order to determine the LOD (3s) and LOQ (10s) of the biosensor, dilutions containing 0 to 1.0 nM TNT in 3.5 µg/L labeled A1.1.1 were incubated for five minutes and injected. The raw data were evaluated as described above. From 0 to 1.0 nM TNT, a linear response was observed, and the LOD and LOQ were determined to be about 0.1 nM or 20 ppt TNT and 0.4 nM or 90 ppt TNT, respectively (see Figure 7). Biosensors 2020, 10, x FOR PEER REVIEW 11 of 18 To evaluate the performance of the TNP-BSA affinity column, a 1:100,000 dilution of the labeled antibody (≈3.5 µg/L) was injected onto the affinity column at varying flow rates, and the signal intensity of the eluate was monitored. At the lowest flow rate of 0.0625 mL min −1 the highest antibody retention of approximately 70% was observed. The antibody removal efficiency gradually declined until the highest flow rate of 0.5 mL min −1 was applied with approximately 54% retention (see. Figure S12). When the dead volume of ca. 1 mL is considered (tubing, connectors, and the affinity column), a flow rate of 0.0625 mL min −1 would result in a dead time between injection and measurement of about 16 min. But this delay can be reduced to about 2 min if the highest flow rate of 0.5 mL min −1 would be applied. For all further measurements, a flow rate of 0.5 mL min −1 was used. TNT Measurements Samples from 2 to 10 nM of TNT in 3.5 µg/L of antibody conjugate were incubated for five minutes and injected at 0.5 mL min −1 to determine the dynamic range for the analyte TNT. The results showed that the linear range does not extend well beyond 2 nM for the chosen parameters (see Figure S13). The signal displays an asymptotic behavior. In order to determine the LOD (3s) and LOQ (10s) of the biosensor, dilutions containing 0 to 1.0 nM TNT in 3.5 µg/L labeled A1.1.1 were incubated for five minutes and injected. The raw data were evaluated as described above. From 0 to 1.0 nM TNT, a linear response was observed, and the LOD and LOQ were determined to be about 0.1 nM or 20 ppt TNT and 0.4 nM or 90 ppt TNT, respectively (see Figure 7). Aqueous solutions of the conventional high explosives PETN, RDX, HMX, and TNT (5 nM) were incubated for five minutes with 3.5 µg/L of the labeled antibody and injected. No cross-reactivity could be observed at 5 nM, for all explosives, except TNT (see Figure 8). The results are in good agreement with the detailed characterization of the clone A1.1.1 (see Figure S17 and Table S2) and the literature. Aqueous solutions of the conventional high explosives PETN, RDX, HMX, and TNT (5 nM) were incubated for five minutes with 3.5 µg/L of the labeled antibody and injected. No cross-reactivity could be observed at 5 nM, for all explosives, except TNT (see Figure 8). The results are in good agreement with the detailed characterization of the clone A1.1.1 (see Figure S17 and Table S2) and the literature. Discussion In this paper, a sensitive and fast biosensor for the detection of 2,4,6-trinitrotoluene (TNT) at the ppt (ng/L or pM) level in water is presented. It is based on kinetic competition and hence displays some extraordinary properties. First of all, the biosensor is faster than most competitive immunoassays, which rely on the convergence to a solid-phase equilibrium. In our format, the analyte is incubated in a homogeneous solution with the respective antibody, which is a fast process. In addition, it is advantageous that this biosensor can be considered as quasi-continuous due to the very high capacity of the trapping column. This long-term measurement capability can even be extended by the regeneration of the trapping column, which, however, is not shown here. Another advantage is the calibration curve, which displays a positive slope in contrast to conventional competitive assays. There is a small delay of about two minutes between the analyte injection and the signal increase due to the dead volumes in the system. Shorter connections and higher flow rates could reduce this in the future. Although the general setup of the system consisting of a wide-field epifluorescence system is quite common, one of the aims of the project was to establish a highly sensitive laser-induced fluorescence detector (LIF), with low-cost equipment (see Figure S4). For many applications, the budget is limited, and hence, systems based on expensive high-end components may not find broad application. Fortunately, the prices of many semiconductor devices, such as cooled charge-coupled device (CCD) or CMOS cameras, have dropped considerably, without compromising their performance. Another decision is the choice of a suitable (fluorescence) label. During the last decades, many improved fluorescence dyes became commercially available, showing better quantum yield, ozone, bleaching and pH stability, less tendency for aggregation, water-solubility, ease of conjugation, and many more. Finally, we have chosen a highly hydrophilic dye, which can be excited with a wavelength of about 635 nm, for which small and cheap diode lasers are available. Perhaps it should be noted that other similar dyes also might be suitable for this purpose. Discussion In this paper, a sensitive and fast biosensor for the detection of 2,4,6-trinitrotoluene (TNT) at the ppt (ng/L or pM) level in

water is presented. It is based on kinetic competition and hence displays some extraordinary properties. First of all, the biosensor is faster than most competitive immunoassays, which rely on the convergence to a solid-phase equilibrium. In our format, the analyte is incubated in a homogeneous solution with the respective antibody, which is a fast process. In addition, it is advantageous that this biosensor can be considered as quasi-continuous due to the very high capacity of the trapping column. This long-term measurement capability can even be extended by the regeneration of the trapping column, which, however, is not shown here. Another advantage is the calibration curve, which displays a positive slope in contrast to conventional competitive assays. There is a small delay of about two minutes between the analyte injection and the signal increase due to the dead volumes in the system. Shorter connections and higher flow rates could reduce this in the future. Although the general setup of the system consisting of a wide-field epifluorescence system is quite common, one of the aims of the project was to establish a highly sensitive laser-induced fluorescence detector (LIF), with low-cost equipment (see Figure S4). For many applications, the budget is limited, and hence, systems based on expensive high-end components may not find broad application. Fortunately, the prices of many semiconductor devices, such as cooled charge-coupled device (CCD) or CMOS cameras, have dropped considerably, without compromising their performance. Another decision is the choice of a suitable (fluorescence) label. During the last decades, many improved fluorescence dyes became commercially available, showing better quantum yield, ozone, bleaching and pH stability, less tendency for aggregation, water-solubility, ease of conjugation, and many more. Finally, we have chosen a highly hydrophilic dye, which can be excited with a wavelength of about 635 nm, for which small and cheap diode lasers are available. Perhaps it should be noted that other similar dyes also might be suitable for this purpose. One of the most critical and often neglected issues is the selection of the antibody. Only very few (monoclonal) antibodies against explosives or TNT, respectively, are available. We were able to purchase two monoclonal anti-TNT-antibodies, the clones A1.1.1 and EW75C. We have characterized both antibodies in detail to determine their sensitivity and their cross-reactivity patterns. In most cases, the cross-reactivities (CR) were relatively similar (see Table S2). Both clones displayed a high cross-reactivity against 2,4,6-trinitroaniline and a relatively high CR against 2,4,6-trinitrobenzene and some dinitrobenzenes or dinitrotoluenes. A significant difference is the lack of CR against any nitro musk compounds of the clone EW75C, in contrast to A1.1.1 (see Table S2). The opposite is the case with some nitrophenyl alkyl acids. The most relevant difference, however, is the test midpoint in ELISA format, which is around 1.2 nM (270 ng/L) for A1.1.1 and 36 nM (8.2 µg/L) for EW75C (see Figure S17). Due to this significant sensitivity difference, we have chosen to proceed with A1.1.1 only. For identification purposes for both clones, antibody fingerprints have been prepared, which are shown in the Supplementary Materials ( Figure S19). Liquid handling is another issue in biosensor development. In our system, we rely on high-performance syringe pumps, which are robust, not too expensive, and deliver a nearly pulsation-free flow. The sample and the antibody are premixed and injected with a conventional 6-port injection valve. A real sampling system for air or wipe tests is lacking, however. The antibody trap is based on a monolithic affinity column developed in-house. It is manufactured from partially sintered borosilicate glass powder and coated via silane chemistry and conjugation with trinitrophenyl-derivatized albumin. The glass monolith is glued into a titanium shell and attached to standard high-performance liquid chromatography (HPLC) fittings. These glass monoliths have the advantage that they tolerate high flow rates while displaying low backpressures and show fast binding kinetics due to a lack of internal porosity. As a miniaturized flow-cell, a microfluidic COC chip was used. This polymer shows excellent optical properties, which are similar to glass, and is resistant to most chemical attacks, as from acids and bases. The optical system consists of a Zeiss objective for microscopes, a dichroic mirror, different filters, and a diode laser (635 nm, <1 mW). A cooled CMOS camera containing a Sony IMX-174M chip was used as the optical detector. In total, the cost of this setup was about <5000 EUR (including taxes, S4), which is quite moderate for this level of performance. In order to characterize the system, several tests have been performed. First of all, the sensitivity of the detector was examined with dilutions of the fluorescence dye (Dy-654-COOH). A highly linear relationship was obtained in the concentration range of 0-1000 pM. A limit of detection (LOD) of about 1 pM was found ( Figure 6). In another experiment, the trapping efficiency of the affinity column was examined ( Figure S12). An efficiency of 60-70% was achieved, which might indicate some impurities of the antibody. Any inactive antibody will lead to a higher background signal. Due to the shorter response time, the faster flow rate of 0.5 mL/min was chosen. In Figure 7, the detection of TNT was tested with the setup shown in Figures 2 and 3. A LOD of approximately 0.1 nM or 20 ppt TNT was achieved, which is significantly lower than the LOD of 60 ppt of a highly optimized competitive ELISA [45], and surpassing most biosensors based on the same antibody [4,26,30,31,38,[71][72][73][74][75][76][77][78][79][80][81][82][83][84]. Finally, some basic cross-reactivity tests with high explosives have been performed with the new biosensor format. It could be shown that only TNT leads to a significant signal at 5 nM. PETN, RDX, and HMX did not show any increased signal, even at much higher concentrations. Conclusions It could be shown that biosensors based on kinetic competition are very powerful and promising systems for the fast and highly sensitive detection of explosives, such as TNT. In contrast to many other biosensors presented for TNT and other nitroaromatic compounds, this approach does not depend on any specific physicochemical property of the target compound. Therefore, it can be easily transferred to all other explosives and substances of interest, for which suitable monoclonal antibodies or similar binders are available today or can be made. It has to be mentioned that monovalent binders, such as Fab fragments, are preferable to bivalent antibodies in this assay format to achieve maximum sensitivity. High speed and excellent sensitivity are also striking advantages. Although these systems are highly specific and hence primarily designed for mono-analyte detection, the setup can be easily parallelized and, therefore, transformed into a multiplex biosensor system. In particular, the use of a conventional CMOS camera offers the opportunity to detect many signals in parallel without the need for an additional detector or the use of highly expensive EMCCD cameras. Similarly, the beam of the laser diode might be split into several sub-beams to excite several flow channels on one chip at a time. The running costs of such a biosensor would be lower, as many would expect from an immunochemical device. With only one milligram of antibody, the biosensor could be operated continuously for more than one year. Finally, it has to be stressed that due to the use of extraordinarily selective antibodies as binders, these biosensors are not prone to false-positive signals, which is crucial for any real-world application in the security field. Confirmation of a unique species of Giardia, parasitic in the quenda (Isoodon obesulus) The ‘quenda genotype’ of Giardia was first identified in quenda (syn. southern brown bandicoots, Isoodon obesulus) in Western Australia in 2004. We aimed to formally describe this genotype as a species of Giardia, Giardia peramelis. Seventy five faecal samples positive for G. peramelis were obtained from quenda within the Statistical Division of Perth, Western Australia. These samples were used in morphological and molecular characterisation of G. peramelis. PCR amplification and sequencing was most successful at the 18S rRNA and ITS1-5.8s-ITS2 loci. Phylogenetic analyses placed G. peramelis external to the ‘Giardia duodenalis species complex’ and Giardia microti. This confirmed the uniqueness of G. peramelis, warranting classification as a separate species of Giardia. Study findings suggest quenda are a natural host for G. peramelis. Introduction A novel genotype of Giardia was isolated and described from a quenda (syn. southern brown bandicoot, Isoodon obesulus) in southwest Western Australia (Adams et al., 2004). This was believed to be a distinct species of Giardia, but the lack of samples from additional infected animals precluded formal description of this genotype as a separate species at that time. Since this initial description, the 'quenda genotype' of Giardia has been documented in quenda in several other locations in Western Australia (Thompson et al., 2010). It has also been identified using PCR from a calf in Western Australia, though it is unclear whether this reflected infection or cysts passing through the calf gut after ingestion of contaminated pasture (Ng et al., 2011a). The 'quenda genotype' of Giardia has not been isolated from other Australian marsupial species surveyed for Giardia spp. (McCarthy et al., 2008;Thompson et al., 2008Thompson et al., , 2010Ng et al., 2011b;Thomasz, 2014;Vermeulen et al., 2015). We undertook a parasitological survey of quenda in the greater Perth region, investigating the epidemiology of Giardia spp. infections in this species, and found infection with the 'quenda genotype' of Giardia to be common. We aimed to formally describe the 'quenda genotype' of Giardia as a separate species, Giardia peramelis, by describing the morphology of cysts and trophozoites and genetically characterising the parasite. It is recognised that for any parasite, once adequate data are available the names should be formalised at the species level (Brooks and Hoberg, 2000). This provides stability and is essential for effective communication. More specifically, we considered formal description of this parasite important to expand knowledge of the phylogenetic range of the genus Giardia; and in consideration of the public health significance of Giardia spp. in Australian marsupials, in differentiating zoonotic and non-zoonotic 'strains' of the parasite. Obtaining G. peramelis specimens Faecal samples were collected from quenda trapped across 51 locations in the Statistical Division of Perth, Western Australia. Faecal material was also collected from the large intestine of quenda carcasses, obtained opportunistically from the same area. All samples were obtained under Murdoch University Animal Ethics Permit (R2530/12), and Department of Parks Wildlife Regulation 17 (SF009640) and Regulation 4 (CE004287) permits. From each quenda, 2 mL faeces were thoroughly mixed in to 8 mL 10% buffered formalin, and 1 mL faeces were thoroughly mixed in to 8 mL 70% ethanol. Preserved faecal samples were stored at 4 C until analysis. The formalin-preserved faecal samples were screened for Giardia spp. cysts using immunofluorescence microscopy. Merifluor Cryptosporidium/Giardia kits (Meridian Bioscience, Inc. USA) were used according to manufacturer's directions for unconcentrated faecal samples. Slides were examined at 200X magnification. Samples positive for Giardia spp. were differentiated to a species level via PCR and sequencing (methodology in section 2.4). In addition, ten immunofluorescence negative samples from trapped quenda were randomly selected and subject to the same PCR and sequencing protocols as the immunofluorescence positive samples. G. peramelis morphological description-trophozoites Wet mounts were prepared from the small intestinal mucosa of two quenda carcasses, which were positive for G. peramelis on faecal testing (and were not positive for any other species of Giardia), and were considered sufficiently fresh, with minimal mucosal autolysis, for detection of trophozoites. The first third of the small intestine was gently scraped and the scrapings mounted on a microscope slide. The slides were examined for trophozoites using an Olympus BX50 microscope. A sample of mucosal scrapings from each quenda was also used to seed flasks containing Giardia media (section 2.2, below), and cultures were monitored regularly for the appearance of trophozoites. Excystation of G. peramelis cysts was attempted three times, using faecal samples from three quenda that were positive for G. peramelis by immunofluorescence microscopy and PCR (and not positive for any other species of Giardia). To purify G. peramelis cysts, 1 g of fresh faeces containing G. peramelis was homogenised in 1X PBS, passed through layers of gauze, and centrifuged at 0.6 G. Two wash steps were carried out, where the supernatant was removed, the pellet was resuspended in 1X PBS, and the sample was centrifuged at 0.6 G. Two sucrose samples were made-one to a specific gravity of 1.18, and another made

up to 0.8 M. The 0.8 M solution was layered on top of the SG 1.18 solution, and the purified sample was layered on top. The sample was centrifuged at 0.2 G, and cysts were collected at the water/sucrose interphase. 1X PBS was added to this isolation and it was centrifuged at 0.6 G, with the supernatant removed subsequently. The cysts were resuspended in a final volume of 1 ml 1X PBS, and examined under the microscope. G. peramelis morphological description-cysts To describe the morphology of G. peramelis cysts, faecal smears were prepared using formalin-preserved faecal samples from eight quenda that were positive for G. peramelis by PCR with sequencing (and not positive for any other species of Giardia). Smears were examined by bright field and Nomarski differential interference microscopy, using an Olympus BX50 microscope. G. peramelis cysts were photographed at 1000X magnification. ImageJ software (US National Institute of Health, Bethesda, Maryland), was used to measure cyst length and width. 2.4. G. peramelis molecular characterisation 2.4.1. DNA extraction and PCR amplification Amplification by PCR was attempted on all immunofluorescence microscopy positive faecal samples and the ten randomly selected immunofluorescence negative samples. Ethanol-preserved faecal samples were centrifuged to separate ethanol from faeces, with the ethanol supernatant discarded. Samples were then twice rehomogenised in distilled water, centrifuged and supernatant discarded. DNA extraction was then conducted using the Maxwell ® 16 Instrument (Promega, Madison, USA) as per manufacturer's instruction. Amplification by PCR was attempted at three loci: 18S rRNA, ITS1-5.8s-ITS2 and glutamate dehydrogenase (gdh). Initially, a semi-nested PCR protocol was employed to amplify a 130 bp product of the 18s rRNA, with primers RH11/RH4 and RH11/GiarR (Hopkins et al., 1997;Read et al., 2002). The PCR reaction was performed in 25 ml volumes, consisting of 1e2 ml of extracted DNA, 2.0 mM MgCl 2 , 1 Â reaction buffer (20 mM Tris-HCL, pH 8.5 at 25 C, 50 mM KCl), 400 mM of each dNTP, 0.4 mM of each primer, 0.5 units of Taq DNA polymerase (Fisher Biotec, Perth, Australia), and DMSO 5%. Amplification conditions were modified from Hopkins et al. (1997), and involved a denaturing step of 95 C for 6 min, then 40 cycles of 95 C for 30 s, 53 C for 30 s (56 C in the secondary round) and 72 C for 30 s, followed by a final extension of 72 C for 7 min. A nested PCR protocol was conducted to amplify a 330 bp product of the ITS1-5.8S-ITS2 region of the ribosomal gene, with primers developed by Caccio et al. (2010). The PCR reactions were the same as those used for 18s rRNA, but performed in 50 ml volumes. Conditions for amplifications were modified from Caccio et al. (2010), and involved an initial denaturing step of 95 C for 5 min, then 40 cycles of 95 C for 45 s, 59 C for 30 s and 72 C for 30 s, followed by a final extension of 72 C for 7 min. Finally, for gdh, a nested PCR protocol was used to amplify a 530 bp product, using the primer pairs Gdh1/Gdh2 and Gdh3/Gdh4 for the primary and secondary rounds respectively, as per Caccio and Ryan (2008). The PCR reaction was performed in 25 ml volumes, consisting of 2 ml of extracted DNA, 1.5 mM MgCl 2 , 1 Â reaction buffer, 200 mM of each dNTP, 0.4 mM of each primer, 1 unit of Taq DNA polymerase (Fisher Biotec, Perth, Australia) and DMSO 5%. Conditions for amplifications were the same for both rounds, and involved an initial denaturing step of 94 C for 2 min, then 35 cycles of 94 C for 30 s, 50 C for 30 s and 72 C for 60 s, followed by a final extension of 72 C for 7 min. Sequencing of amplified product PCR products were purified using an Agencourt AMPure XP system (Beckman coulter, Beverly, USA). Sequence reactions were performed using the Big Dye Terminator Version 3.1 cycle sequencing kit (Applied Biosystems), according to the manufacturer's instructions. Reactions were electrophoresed on an ABI 3730 48 capillary machine. Amplicons were sequenced in both directions, with resultant nucleotide sequences compared with published sequences on NCBI GenBank ® using the basic alignment search tool (BLAST). Further sequence analysis was conducted using the sequence alignment program Sequencher™ 4.8 (Gene Codes, Ann Arbour, MI, USA). Nucleotide sequence data reported in this paper are available in the NCBI GenBank ® database under accession numbers KU306911, KU306912, KU306913, KU306914 and KU306915. Results Faecal samples positive for Giardia spp. by immunofluorescence microscopy were obtained from 99 trapped quenda, and 11 quenda carcasses. Of the immunofluorescence positive samples, 63/99 trapped quenda and 11/11 quenda carcasses were confirmed to be infected with G. peramelis by PCR and sequencing at one or more loci. Thirty six immunofluorescence positive faecal samples from trapped animals either did not amplify by PCR, or amplified product failed to give a readable sequence. Of the ten randomly selected immunofluorescence microscopy negative samples tested by PCR and sequencing, one was positive for G. peramelis by PCR and sequencing at one locus. 3.1. G. peramelis morphology G. peramelis trophozoites were not detected in intestinal scrapings of the two tested quenda carcasses, or in the cultures seeded with mucosal scrapings from the two quenda, and all attempts at G. peramelis excystation failed. This precluded morphological description of G. peramelis trophozoites. Giardia peramelis molecular characterisation Of the 111 individual quenda faecal samples found positive for Giardia spp. by immunofluorescence microscopy or PCR, 75 (67.6%) were successfully amplified and sequenced at one or more loci. Sixty-four samples were sequenced at the 18s rRNA locus, 50 samples sequenced at ITS1-5.8s-ITS2 and two samples sequenced at the gdh locus ( Table 1). Two of the 75 samples sequenced at all three loci, 37 samples sequenced at both 18S and ITS1-5.8s-ITS2, and the remaining 36 samples sequenced at only one locus ( Table 2). Of the 116 sequences obtained, 114 were the novel G. peramelis; two sequences obtained at the 18S rRNA locus were suspected to be mixed infections of Giardia duodenalis/G. peramelis and Giardia canis/G. peramelis. These 'mixed infection' samples were also amplified at the ITS1-5.8s-ITS2 region locus-the resultant sequences were clearly identified as G. peramelis, with no ambiguous nucleotides evident. Utilizing the BLAST tool within NCBI, the 18S rRNA sequences were highly similar to the published sequence AY309064, previously reported as the 'quenda genotype'. Alignment in Sequencher™ revealed one single nucleotide polymorphism (SNP) between sequences obtained in this study (represented by QBY95 and QM22) and AY309064. In addition, one SNP was identified within a small subset of sequences obtained in this study, represented by sequence QBN13. A large region of extreme variability was identified between all sequences and another reported 'quenda genotype'. Of the 292 bp published for this sequence (accession number HQ398319), a region of approximately 150 bp showed extreme mismatch. Sequences obtained at ITS1-5.8S-ITS2 required the BLAST program selection of 'somewhat similar sequences' in order to achieve a result. The 'G. duodenalis species complex' was the primary match, but achieved low coverage and identities with all comparisons. Alignment of these sequences in Sequencher™ revealed one SNP, represented by samples QBY95/QM22 and QBN13. Similar BLAST results were obtained with the two gdh sequences. Phylogenetic analysis Very similar trees were obtained by neighbour-joining, Fig. 1. Cyst of Giardia peramelis-light microscopy. Fig. 2. Cysts of Giardia peramelis-immunofluorescence microscopy. maximum likelihood and maximum parsimony methods; only the neighbour-joining trees are presented here. Phylogenetic analysis of sequence data obtained at 18s rRNA confirmed that the Giardia genotype obtained from quenda in this study (G. peramelis) formed a separate clade with the 'quenda genotype' reference AY309064, and was distinct from all assemblages within the 'G. duodenalis species complex' and G. microti (Fig. 3). A similar topology was observed based on genetic data obtained at ITS1-5.8S-ITS2, with the exception that G. peramelis was also placed external to G. ardeae (Fig. 4). Phylogenetic analysis of gdh also placed the G. peramelis external to all assemblages within the 'G. duodenalis species complex' (results not shown). (Adams et al., 2004;Thompson et al., 2010). Site of Infection: unknown; presumably small intestine, based on the known Giardia spp. site of infection in other mammalian hosts (Monis et al., 2009). Prepatent and patent periods: unknown. Material deposited: A sample of formalin-preserved and ethanol-preserved G. peramelis cysts, and photomicrographs of G. peramelis cysts, were deposited at the Western Australian Museum (specimen registration no. WAM Z68785). Etymology: The specific epithet peramelis is derived from the subfamily Peramelinae/family Peramelidae/order Peramelemorphia-taxonomic classifications of the quenda, the only confirmed host of G. peramelis. This is in line with current taxonomic nomenclature for Giardia spp. (Monis et al., 2009). This study confirms that G. peramelis (previously the 'quenda genotype' of Giardia) is a unique species. Based on genotyping at 18s rRNA, all G. peramelis isolates in this study strongly aligned with the original 'quenda genotype' data, generated by Adams et al. (2004). We have further confirmed the genetic novelty of G. peramelis with genetic data obtained at ITS1-5.8S-ITS2, which identified G. peramelis as belonging to the genus Giardia but did not match any published sequences of Giardia spp. Phylogenetic analyses of both genes place G. peramelis external to the 'G. duodenalis species complex' and external to G. microti. This supports the proposals of both Adams et al. (2004) and Monis (2004, 2012) that G. peramelis is a novel lineage, distinct from described species. The extreme mismatch observed between the G. peramelis 18S rRNA sequences reported here and in Adams et al. (2004), and the previously reported 'quenda genotype' (HQ398319) obtained from a calf (Ng et al., 2011a), suggests that a partial 'quenda genotype' may have been sequenced from the calf, along with other genetic material. As G. peramelis cysts are morphologically indistinguishable from several other species of Giardia, molecular characterisation is required to differentiate G. peramelis from other Giardia spp. Our results suggest that the 18S rRNA and ITS1-5.8s-ITS2 loci are the most successful target regions for this purpose. The substantial number of quenda observed to be infected with G. peramelis, sampled across a large number of locations in this study, suggests that quenda are a natural host for G. peramelis. Conclusions This study confirms that G. peramelis (formerly the 'quenda genotype' of Giardia) is a unique species of Giardia. It expands on known genetic and morphological data of this species, and describes techniques used in the genetic characterisation of G. peramelis. Fig. 4. Phylogenetic relationships of Giardia peramelis isolates obtained in this study (quenda QBN13, QM22, QBY95) with published reference material available at the ITS1-5.8S-ITS2 locus. Evolutionary history inferred using the neighbour-joining method supported with bootstrap test of 1000 replicates (values > 50% shown). G. muris is used as the out group. Polymorphic Characterization, Pharmacokinetics, and Anti-Inflammatory Activity of Ginsenoside Compound K Polymorphs Polymorphism exhibits different physicochemical properties, which can impact the bioavailability and bioactivity of solid drugs. This study focused on identifying the polymorphs of ginsenoside compound K (CK) and studying their different behaviors in pharmacokinetics (PK) and pharmacodynamics (PD). Four CK polymorphs (form I, II, III, and IV) from organic solvents were characterized by scanning electron microscope (SEM), differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), Fourier transform infrared spectroscopy (FTIR), and powder X-ray diffraction (PXRD). A feasible LC-MS/MS method was exploited to determine the PK parameters. Form II displayed the most exposure, followed by form I, III, and IV. Notably, all forms showed sex dimorphism, and the bioavailability in the female group was about two-fold higher than in the male group. The PD properties were investigated in carrageenan-induced acute paw inflammation, and form II at 20 mg/kg showed significant inhibition of edema by 42.7%. This study clarified the polymorphic, PK, and PD characters of four crystal forms of CK, and the data suggested that form II had the best efficacy for drug development. Introduction Ginseng (Panax ginseng) is a well-known and prevalent traditional herb for its medicinal and healing properties in East Asian countries. Triterpene saponins, commonly known as ginsenosides, are the major pharmacologically active ginseng components [1,2]. 20-O-β-D-glucopyranosyl-20(S)-protopanaxadiol, or ginsenoside compound K, is one of the major intestinal metabolites of ginsenoside Rb1, Rb2, and Rc after oral administration [3]. We have isolated a fungus of Paecilomyces bainer sp.229, which could effectively produce compound K (CK) by the biotransformation pathway of Rb1

→ Rd → F2 → CK, and found a new approval of CK in treating rheumatoid arthritis (RA) [4][5][6]. It has been in phase I clinical trials in China as a novel oral candidate for RA therapy by Zhejiang Hisun Pharmaceutical Co., Ltd. (Zhejiang, China) [7]. In our previous studies, two single-crystal structures of CK solvates (dihydrate and methanol-water solvate) were reported [8,9]. Li et al. [10] further characterized another two solvates (nonstoichiometric hydrate and methanol solvate). Thus, the polymorphic phenomenon of CK is exhibited, but the influence of polymorphism on pharmacokinetics and pharmacodynamics is still unknown. Polymorphism is the ability of the same compound to be packaged in different ways in the solid state with different intermolecular interactions [11,12]. It exhibits different physicochemical properties [13,14] and can alter in vivo bioavailability, which finally modifies the drug efficacy [15,16]. Thus, appropriate analytical procedures to detect polymorphs and emphasize the importance of controlling the crystallization have great significance for the drugs administered in solid pharmaceutical dosages [17]. Therefore, it is crucial to pay attention to the solid-state forms at the beginning of CK development as an anti-RA candidate. The objective of this study was to prepare and characterize different CK polymorphs by SEM, differential scanning calorimetry (DSC), TGA, FTIR, and powder X-ray diffraction (PXRD). Furthermore, an HPLC-MS/MS method with a negative electrospray ionization source and a rat paw swelling model induced by carrageenan was set up to evaluate the bioavailability and anti-inflammatory activity of these polymorphs. Scanning Electron Microscopy Three techniques were used for the polymorphs preparation in the present study: (1) rapid solvent evaporation in a vacuum and high-temperature condition (form I); (2) gradual solvent evaporation in a small opening glass tube at room temperature (form II and IV), and (3) anti-solvent precipitation in water (form III) (see Section 4.3.1). During the solvent evaporation process, the solvents employed to affect the solute's molecular aggregates presented in a supersaturated solution are a determinant of the final crystal form [18]. Here, we obtained four CK polymorphs from different polar solvents: form I from ethanol, form II from acetone, form III from methanol and anti-solvent precipitation of water, and form IV from methanol. These four crystal forms of CK exhibited different morphologies by scanning with a high-power electron microscope ( Figure 1). Form I appeared amorphous solid, without a uniform and typical shape. Form II tended to clump together and presented an irregular granule structure. Form III tended to be scattered, and the crystal nucleus grew more evenly with thin flake form. In comparison, form IV showed a big block appearance and had a wavy surface under an optical microscope ( Figure S1). These results indicated forms I, II, III, and IV displayed the apparent differences due to the different molecular arrangements in different solvents, respectively. Thermal Analysis The DSC thermograms and TGA traces of the four polymorphs are shown in Figure 2. For all the forms, TGA (Figure 2A inset) showed slight weight losses below 140 • C (less than 2.0%), indicating the presence of any solvated or hydrated forms, and these multi-step weight losses were attributed to the different interactions between different solvents and CK molecules. There was a pronounced weight loss of about 95% in TGA ( Figure 2A) between 220 • C and 400 • C for every form, which corresponded to the CK molecule decomposition. However, there were noticeable shifts of the onsets in DSC curves ( Figure 2B), 267.5 • C for form I, 263.1 • C for form II, 255.9 • C for form III, and 250.2 • C for form IV, which indicated the different thermodynamical stabilities of these forms. Moreover, in DSC curves, form I exhibited only one endotherm peak of the decomposition. Form II had another two endothermal peaks at 115.4 • C and 186.6 • C, which were attributed to solvent desorption (acetone) and crystal melting, respectively. For form IV, the first solvent desorption (methanol) endothermal peak appeared at 120.1 • C, and the second melting endothermal peak followed at about 178.5 • C. As to form III, it had a broad peak from 50 to 100 • C, which was due to dehydration, and the weight loss was about 1% (Figure 2A). Meanwhile, it had no sharp endothermal melting peak as form II and form IV, so a phase transition might happen along with the solvate loss. Fourier Transform Infrared Spectroscopy (FTIR) The FTIR spectra ( Figure 2C) were performed to confirm the results obtained from the thermal analysis, and the characteristic FTIR peaks were summarized in Table 1. The O-H stretching vibrations of the four forms had obvious shifts, at 3405 cm −1 , 3362 cm −1 , 3423 cm −1 , and 3385 cm −1 for forms I, II, III, and IV, respectively ( Figure 2C, dashed line). In form IV, there was another small peak at 3677 cm −1 , which was the free O-H stretching Molecules 2021, 26,1983 3 of 13 vibration from the methanol molecules in this solvate, along with the bending vibration at 1308 cm −1 . Nonpolar C-H stretching vibration and CH 3 /CH 2 asymmetric deformation remained unchanged among the four forms. However, C=C stretching absorption peaks of form II and form III shifted to 1654 cm −1 and 1655 cm −1 , respectively, a little higher than those of form I (1637 cm −1 ) and form IV (1639 cm −1 ). The weak bands at 1710 cm −1 and 1250 cm −1 were assigned to C=O stretching and bending vibration, which indicated the acetone molecules in form II. The fingerprint regions (600 to 1100 cm −1 ) of the four forms were very different in peak position and shape ( Figure 2C, dashed line), which indicated the solvents interfered with the hydrogen bond formation between different glucosyl moieties of CK molecule in different ways. Thermal Analysis The DSC thermograms and TGA traces of the four polymorphs are shown in Figu 2. For all the forms, TGA (Figure 2A inset) showed slight weight losses below 140 °C (le than 2.0%), indicating the presence of any solvated or hydrated forms, and these mu of form II and form III shifted to 1654 cm −1 and 1655 cm −1 , respectively, a little higher than those of form I (1637 cm −1 ) and form IV (1639 cm −1 ). The weak bands at 1710 cm −1 and 1250 cm −1 were assigned to C=O stretching and bending vibration, which indicated the acetone molecules in form II. The fingerprint regions (600 to 1100 cm −1 ) of the four forms were very different in peak position and shape ( Figure 2C, dashed line), which indicated the solvents interfered with the hydrogen bond formation between different glucosyl moieties of CK molecule in different ways. . TGA showed a pronounced weight loss of about 95% between 220 and 400 • C for CK decomposition, and the inset showed a magnified view highlighting the weight loss blow 140 • C attributed to the solvent desorption. FTIR spectra (C) of the four forms of CK were from 400-4000 cm −1 , and the major changes observed in the spectra were marked with dashed lines and arrows. Powder X-ray diffraction (PXRD) patterns (D) of the four forms of CK were measured with Cu/K-α1 radiation (λ = 1.54056 Å) at 40 kV, 40 mA from 3 • to 40 • (2θ). Form I was in an amorphous state, and forms II, III, and IV were in crystalline structures. Figure 2D showed the PXRD data of the four forms, and the diffractograms indicated significant differences among them. There were only an extensive blunt band and no sharp peak on form I's diffraction curve, which was considered an amorphous state. Form II and form IV showed the strongest peaks at 2θ of 14.60 • and 14.89 • , respectively, while the most substantial peak of form III appeared at a much lower 2θ angle of 6 As shown in Table 2, the exposure to form IV was much lower than form I, and there was a 34% decrease in the bioavailability from form I to form IV. However, when the rats were treated with form II, a distinct increase in the area under the curve (AUC (0-∞) ) was observed compared with the rats treated with form I. Thus, form II had the maximum bioavailability among the four forms, with 4.07% in the male group and 7.87% in the female group, respectively. The pharmacokinetics (PK) values of form III were just between form I and form IV. Taken together, although the PK parameters in rats treated with different forms did not show statistically significant differences (Tables S3-S6), these results were of pronounced differences and indicated that the occurrence of CK polymorphs might alter its release step in the rat gastrointestinal fluids, reducing or increasing the extension of its absorption depending on the polymorphic form presented in the oral dosage. Note: AUC (0-t) : area under the plasma concentration-time curve from zero to the time of the last quantifiable concentration; AUC (0-∞) : area under plasma concentration-time curve from zero to infinity; t 1/2z : half-life; T max : time to maximum plasma concentration; C max : maximum plasma concentration; V z/F : apparent volume of distribution after administration; CL z/F : the total plasma clearance of the drug after administration. All values are presented as mean ± SD (n = 3). The significant effect of sex was investigated by one-way ANOVA for the pharmacokinetics (PK) parameters. ♂: male rats; ♀: female rats. *: p-value < 0.05 versus the male group. #: p-value < 0.05 versus form I. Sex-Related Impact on Pharmacokinetic Properties It was found that females had higher exposure levels of CK than males in healthy Chinese subjects [19], so we also compared the PK properties of these CK polymorphs in sex. The results showed that no significant sex difference was observed in dose-normalized exposure parameters in this single-dose i.v. administration (Tables S1 and S2). While in p.o. administration, a very sex-related impact on the pharmacokinetic properties of all the four tested CK polymorphs was found (Table 2 and Figure 3). For every CK form, the average C max and AUC in the female groups increased about two-fold compared to those in male groups. The AUC of form I and the C max of form III had statistically significant sex differences according to one-way ANOVA analysis. Although without statistical significance, the average V z/F and CL z/F were also much more different between sex. The results showed that no significant sex difference was observed in dose-normalized exposure parameters in this single-dose i.v. administration (Tables S1 and S2). While in p.o. administration, a very sex-related impact on the pharmacokinetic properties of all the four tested CK polymorphs was found (Table 2 and Figure 3). For every CK form, the average Cmax and AUC in the female groups increased about two-fold compared to those in male groups. The AUC of form I and the Cmax of form III had statistically significant sex differences according to one-way ANOVA analysis. Although without statistical significance, the average Vz/F and CLz/F were also much more different between sex-groups in p.o. than those in i.v. However, Tmax and t1/2z showed no apparent sex dimorphism on any polymorph in both i.v. and p.o. administration. Crystalline Polymorphic Impacts on Anti-Inflammatory Properties To further study the in vivo effects variation of different CK polymorphs, the pharmacodynamics was evaluated by measurements of time-dependent paw edema with carrageenan-induced acute inflammation (Figure 4). Compared with the normal group, edema reached the peak in the paw 3 h after carrageenan challenge with the maximum swell rate of 64% in the model group and then remitted gradually (Table S7). The paw edema in the drug-treated groups showed the most significant inhibition at 5 h post-injection compared to the model group. Remarkably, form II at 20 mg/kg signifi- Crystalline Polymorphic Impacts on Anti-Inflammatory Properties To further study the in vivo effects variation of different CK polymorphs, the pharmacodynamics was evaluated by measurements of time-dependent paw edema with carrageenan-induced acute inflammation (Figure 4). Compared with the normal group, edema reached the peak in the paw 3 h after carrageenan challenge with the maximum swell rate of 64% in the model group and then remitted gradually (Table

S7). The paw edema in the drug-treated groups showed the most significant inhibition at 5 h postinjection compared to the model group. Remarkably, form II at 20 mg/kg significantly reduced paw edema at 4, 5, and 6 h post-injection, with the inhibitions of 36.3%, 42.7%, and 40.0%, respectively. In contrast, form IV displayed no statistical difference from the model group at any time point post carrageenan injection. Further, form II showed a statistical difference of paw edema from form IV at 4 h post-injection (39.7 ± 6.4% vs. 58.6 ± 3.6%, p = 0.034) at the dosage of 20 mg/kg. These results strongly suggested that the treatment with CK polymorphs in rats had the capacity to diminish the acute inflammation subsequently induced by carrageenan in vivo, and the anti-inflammatory effects were definitely related to the pharmacokinetic properties of the polymorphs after oral administration. Fabs (%) 3.92% 6.84% 4.07% 7.87% 3.90% 5.85% 2.59% 4.07% Note: AUC(0-t): area under the plasma concentration-time curve from zero to the time of the last quantifiable concentration; AUC(0-∞): area under plasma concentration-time curve from zero to infinity; t1/2z: half-life; Tmax: time to maximum plasma concentration; Cmax: maximum plasma concentration; Vz/F: apparent volume of distribution after administration; CLz/F: the total plasma clearance of the drug after administration. All values are presented as mean ± SD (n = 3). The significant effect of sex was investigated by one-way ANOVA for the pharmacokinetics (PK) parameters. ♂: male rats; ♀: female rats. *: p-value < 0.05 versus the male group. #: p-value < 0.05 versus form I. Discussion Polymorphism is prevalent in the pharmaceutical field, which can impact the bioavailability of solid drugs [20,21]. CK was isolated and identified in 1972 [22], but data available on the polymorphism of CK were very limited [8][9][10]. Primarily, there is still no comparative study about the pharmacokinetic and pharmacodynamic properties of different CK polymorphs. In this study, four CK polymorphs were prepared and characterized, and significant differences were observed in crystal morphology. Form I was additionally proved as an amorphous state, with a halo pattern in PXRD and no melting peak in DSC [23]. The reason for the amorphous state of form I might be that it was produced from fast ethanol evaporation in diaphragm vacuum and high temperature (60 • C), thus the internal energy of the molecules was high, and the molecular movement was fast, which disturbed the formation the hydrogen-bonded aggregates of CK with the solvents for the growth of shaped crystals. Compared to form II and form III, PXRD peaks of form IV were more intense and sharp at high 2θ angles (20-30 • ), indicating that form IV had a better grain and a higher degree of crystallinity, which might occur due to the much slower crystallization rate. Form II peaks were less intense, indicating that form II probably had a more disordered structure and a discrete crystal size, as was observed in SEM ( Figure 1). As to form III, there was a broad peak from approximately 50 to 100 • C, but no sharp melting peak at about 180 • C, which was different from form II and form IV. According to our previous single-crystal X-ray diffraction analysis [8], form III was originally presented to be a dihydrate. However, in this study, the weight loss of water in this TG analysis was only one-fifth of the stoichiometric value calculated for the dihydrate dehydration ( Figure 2A). So, this form III was considered a novel anhydrous form of CK dihydrate by the water vaporization in the atmosphere during the storage or under dry nitrogen flow in DSC and TG analysis. Those having hydrogen bonding sites in the crystal can stoichiometrically incorporate water molecules into the crystal lattice of the drug and form a hydrate [24]. Conversely, the water molecules of a hydrate can be released from its crystal lattice under heating and drying conditions and transformed into anhydrous forms [25], even the amorphous form [26,27]. Thus, the detailed dehydration behaviors of form III deserved to be further investigated. Therefore, it was revealed that the four forms were different from each other and had different molecular arrangements in their crystal structures. By comprehensive comparison, form III was a novel anhydrous form of the dihydrate, while form I, form II, and form IV were almost but not the same as the reported forms [10,23,28]. The different crystal forms of an oral drug can affect its absorption and bioavailability and lead to differences in clinical efficacy [29,30]. However, to our knowledge, as an orally administrated drug, PK and pharmacodynamics (PD) studies of CK polymorphs have not been reported yet. Thus, in this manuscript, we conducted PK and PD properties of CK polymorphs for the first time. One of the important results, in oral administration, the sex dimorphism in the pharmacokinetics of these CK polymorphs was also observed in this study as in the previous report [7]. The maximum differences of the average AUC and C max between male and female groups almost twice folded. In contrast, the differences in intravenous injection were tiny, indicating the process of absorption in the gastrointestinal tract influenced the sex dimorphism more than the metabolism in circulation. Over the past few years, researchers have gradually begun recognizing the importance of sex in drug development and clinical applications. However, the reasons behind the sex-related differences are yet not to be fully clarified [31,32]. Indeed, the sex-related expression of metabolism enzymes [33], transporters in target organs [34,35], and hormones secretion [36] were found to play a role in the differences between sexes. Nevertheless, additional factors, such as intestinal motility [37], food [19], and P-glycoprotein (P-gp) expression [38], could simultaneously be involved in altering the pharmacokinetics of CK. Other than sex dimorphism, the variations of pharmacokinetics parameters among different CK polymorphs were also discovered as excepted. In our results, form II displayed the most potent bioavailable activity, followed by form I, III, and IV. The maximum differences of the average C max and AUC (0-∞) values compared to form II and IV occurred at 1.6-fold in male groups and 2.1-fold in female groups. Although there were remarkable differences, PK parameters did not present a statistically significant difference between genders [7,19], as well as among polymorphs. The reason was considered to the considerable individual variations of the CK pharmacokinetics; for example, C max and AUC displayed significant variations resulting in standard deviation values, which were even close to half the value of the corresponding mean values. However, in the multiple-dose trial in humans [7], females had a significantly higher dose-normalized C max and t 1/2 than those in males after administration of CK on day 1 and day 15, which might be relevant to accumulation or hepatoenteral circulation in vivo after repeated administrations. More investigation about the absorption, distribution, metabolism, and excretion of CK needs in the future to explain this significant individual difference adequately. Thirdly, in the classic carrageenan-induced rats' paw edema, irrespective of sex, the paw edema levels decreased significantly compared to the model group, indicating the promising anti-inflammatory activity of CK reported previously [39,40]. The reduction of paw edema levels could be considered to be in good agreement with the pharmacokinetic profiles of the four CK polymorphs. Form II displayed the most potent anti-inflammatory activity, followed by form I, III, and IV. As a relative initial study, all of these results greatly enriched the PK and PD knowledge of CK polymorphs, which could provide beneficial information for this promising drug candidate's crystal form selection. However, there were still some evident limitations: first of all, only four polymorphs were studied, and more crystal forms are likely to be discovered; secondly, more physicochemical properties, such as stability, dissolution rate, and solubility, will help explain the PK and PD differences; the thirdly, the novel form III presented the hydrate-anhydrous form transformation, and this transformation might occur under ambient conditions during drug processing, transportation, and storage. Evaluating the hydration and dehydration behaviors of form III is essential for developing a more stable formulation. Materials Raw CK was prepared in our lab [4] and higher than 98.0% in purity. Acetonitrile (HPLC grade) was purchased from Dikma Technologies Inc. (California, CA, USA). Ultrapure water used in the mobile phase was obtained by a Purist Pro water purification system (RephiLe Bioscience, Shanghai, China). All other reagents, solvents, chemicals, and solutions used were analytical reagent grades from Titan Scientific (Shanghai, China). Carrageenan (type IV) and indomethacin were purchased from Sigma-Aldrich (Missouri, MO, USA). Animals Sprague-Dawley rats (8 weeks old, 200 ± 30 g) were obtained from Shanghai SLAC Laboratory Animal Co., Ltd., (Shanghai, China). The animals were housed under a constant environment (temperature~25 • C; humidity~70%; and 12 h light-dark cycle) with free access to water and a rodent diet. Preparation of CK Polymorphs Form I: Raw CK was dissolved in ethanol (10%, w/v) in a pear-shaped flask and vaporized to dryness on the rotavapor with a diaphragm vacuum at 60 • C in a water bath to obtain form I (amorphous solid). Form II: Form I was dissolved in acetone (5%, w/v) in a glass tube (15 × 150 mm) at 60 • C in a water bath, and the solution was kept at room temperature overnight to obtain form II (granule-shaped). The tube was open for slow evaporation of acetone. Form III: Form I was dissolved in methanol (10%, w/v) in a glass tube (15 × 150 mm) at 60 • C in a water bath, and distilled water was added to dilute methanol to 65% (v/v). The solution was kept at room temperature overnight to obtain form III (flake-shaped), with the tube was closed [8]. Form IV: Form I was dissolved in methanol (20%, w/v) in a glass tube (15 × 150 mm) at 60 • C in a water bath, and the solution was kept at room temperature for 4 days to obtain form IV (block-shaped), with the tube open for slow evaporation of methanol [9]. The four CK forms obtained from the experiment were air-dried at room temperature and then evaluated by various crystal characterization methods. Scanning Electron Microscopy The photomicrographs of the crystallized samples were examined morphologically with a SU8010 scanning electron microscope operating at 2.0 kV (Hitachi, Tokyo, Japan). Thermal Analysis DSC measurements were performed on a DSC204 Phoenix calorimeter (NETZSCH, Selb, Germany) with a heating rate of 20 K/min from 20 to 400 • C. The dry nitrogen flow was about 50 mL/min. Data were analyzed using NETZSCH analysis software. TGA was obtained by a TGA8000 thermal analysis instrument (PerkinElmer, Waltham, MA, USA) at the same operating conditions as DSC, and data were analyzed using Universal Analysis software (TA Instruments, New Castle, DE, USA). Powder X-ray Diffraction PXRD data were collected at room temperature (25 • C) in the reflectance mode using a D2 PHASER diffractometer (Bruker AXS Inc., Madison, WI, USA) with Cu/K-α1 radiation (λ = 1.54056 Å) at 40 kV, 40 mA. Diffraction patterns (2θ) were collected from 3 • to 40 • at a step scan of 0.01 • with a scanning speed of 0.6 sec. Results were analyzed using MDI/JADE 6.0 software (Materials Data, Livermore, CA, USA). Fourier Transform Infrared Spectroscopy The FTIR spectra of all samples were measured with a Nicolet iS10 spectrophotometer (ThermoFisher, Waltham, WI, USA). The samples were ground and mixed with potassium bromide (KBr) at 1% dilution, and the pellets were prepared by compressing the powder. The spectra were recorded in the range 400-4000 cm −1 and analyzed by Origin 7.5 software (OriginLab, Northampton, MA, USA). Animals and Dosing Rats were divided randomly into 5 groups of six animals each (sex in half) and given 7 days to acclimatize to the facility before the experiments began. The four forms of CK were suspended in 0.5% sodium carboxymethyl cellulose (CMC-Na) solution and orally administered to rats at a dose of 20 mg/kg. A CK solution in a mixed solvent of dimethyl sulfoxide (DMSO)/castor oil/physiologic saline (3:5:92, v/v) was intravenously administered to the remaining group via the tail vein at a dose of 10 mg/kg. Blood samples were collected by retro-orbital puncture at the following times: 0.167, 0.333, 0.667, 1, 2, 3, 4, 6, 9, 12, and 24 h for the oral administration, and at 0.0083, 0.05, 0.133, 0.25, 0.5, 1, 2, 4, 6, and 9 h for

the intravenous administration. Plasma samples were obtained by centrifuging a heparinized blood tube at 4000 rpm for 5 min and then stored at −70 • C until analysis. Sample Preparation and Analysis The procedure for drug extraction from plasma samples was as follows: 0.2 mL acetonitrile, 0.2 mL plasma sample, and 10 µL IS (1 µg/mL) were added into a 1.5 mL Eppendorf-type tube. This mixture was vortexed for 1 min and then treated with 0.8 mL of methyl tert-butyl ether (MTBE) to precipitate the plasma proteins. After a mechanical vortex for 1 min and subsequent centrifugation (4000 rpm, 5 min), 0.8 mL of the supernatant was transferred and blow-dried with N 2 at 40 • C in a water bath. Finally, the residue was dissolved in 120 µL of mobile phase, and then 10 µL was subjected to HPLC for analysis. The assay was shown to be linear in the range 1.00-900 ng/mL with a linear regression equation of Y = 7.757 × 10 −3 X + 1.885 × 10 −3 (r 2 = 0.999). Intra-and inter-day precision and accuracy never exceeded 15% by analyzing six replicate samples. Statistical Analysis The CK pharmacokinetic parameters in rats, including the area under the curve (AUC (0-t) and AUC (0-∞) ), mean residence time (MRT (0-t) and MRT (0-∞) ), maximum concentration (C max ), time to maximum concentration (T max ), half-life (t 1/2 ), apparent distribution volume (V), and total plasma clearance (CL), were derived from plasma concentration-time curves by using non-compartmental analysis (NCA) with DAS 2.0 software (Drug And Statistics, Mathematical Pharmacology Professional Committee of China, Shanghai, China). The absolute bioavailability (Fabs) was calculated from the dose-adjusted ratio of AUC p.o. to AUC i.v. . One-way analysis of variance (ANOVA) by SPSS 11.5 (SPSS Inc., Chicago, IL, USA) was conducted to compare the primary PK parameters. Anti-Inflammatory Tests Rats were divided into 11 groups of six animals each (sex in half): the normal group, the model group, indomethacin group (1 mg/kg), and form I, II, III, IV groups (each form set 10 mg/kg and 20 mg/kg groups, respectively). Indomethacin and CK were ground and suspended in 0.5% CMC-Na solution. The rats were administered the drug orally for 3 days (qd), and those in the normal and model groups had the same volume of CMC-Na solution. One hour after the last administration, except for the normal group, 0.1 mL 0.1% carrageenan in sterile saline (0.9% NaCl) was injected into the plantar surface (i.pl.) of the right hind paw in order to induce inflammation. The same volume of saline was injected into the normal group. The volumes of the right hind paw of the rats were measured using a KW-7C plethysmometer (Nanjing KEW Bio-Technology, Nanjing, China) at different time intervals after injection (1, 2, 3, 4, 5, and 6 h). Paw edema and inhibitory rate were calculated as the following formula, and the differences among groups were compared with the t-test. Conclusions Here, we prepared four different polymorphs of CK and fully demonstrated the polymorphism characteristics. Remarkable differences in PK and PD parameters of the polymorphs were observed after orally administered to rats. Form II showed a maximum bioavailability and therapeutic efficacy and could be preliminarily inferred as an advantageous polymorph. The present study can serve as a basic step to further polymorphic studies of ginsenoside compound K. Table S1: Plasma concentrations at different time points after intravenous injection of CK in rats (10 mg/kg), Table S2: Summary pharmacokinetic parameters after intravenous administration of CK in rats (10 mg/kg), Table S3: Summary pharmacokinetic parameters after oral administration of form I in rats (20 mg/kg), Table S4: Summary pharmacokinetic parameters after oral administration of form II in rats (20 mg/kg), Table S5: Summary pharmacokinetic parameters after oral administration of form III in rats (20 mg/kg), Table S6: Summary pharmacokinetic parameters after oral administration of form IV in rats (20 mg/kg), Table S7: Effects of CK polymorphs on paw edema from rats induced by carrageenan. A hypothesis for a novel role of RIN1-the modulation of telomerase function by the MAPK signaling pathway Cancerous cells display abnormalities in the signal transduction pathways responsible for responding to extracellular growth factors, or mitogens. Mutations that alter proteins involved in these types of pathways can lead to inappropriate or unregulated cell growth, and therefore predispose the cell to become malignant. The critical role of the Ras/mitogen-activated protein kinase (MAPK) pathway in transducing growth signals to the interior of the cell and subsequently stimulating cell growth and proliferation is underscored by the fact that roughly one quarter of all human tumors contain mutant forms of Ras proteins. A particular focus on the signaling and membrane trafficking adaptor protein known as Ras interference 1 (RIN1) will reveal how this protein can potentially play a significant role in the development of the cancerous phenotype in certain cell types. Of equal interest is the possible connection between the Ras/MAPK pathway, and subsequent expression and enzymatic activity of telomerase–a key enzyme known to be overexpressed in roughly 85% of all cancers. RIN1 is a 783 amino acid (84 kDa) cytosolic protein that is involved in key steps of growth factor receptormediated endocytosis and can potentially moderate signaling through the MAPK pathways. RIN1, with its unique ability to compete directly with Raf for activation by Ras, could potentially influence signaling through more than one of the MAPK pathways. If so, RIN1 may then be able to exert a precise and selective effect on the downstream signal(s) of a MAPK target such as telomerase. Introduction Many cancerous cells display abnormalities, at least, in the two major signal transduction pathways (Ras/PI3-K/Akt and Ras/ MAPKs) responsible for responding to extracellular growth factors, or mitogens. Hanahan and Weinberg (2000), in a comprehensive review of the molecular hallmarks of cancer cells, list growth factor self-sufficiency as one of the six necessary physiological changes needed to convert a normal cell into a malignant cell. The authors also suggest that malignancy is acquired through a series of successive mutations that fall under two broad categories -dominant gain of function oncogenic mutations and recessive loss of function tumor suppressor mutations. Unlike normal cells that will only grow and divide in response to external growth signals, many cancer cells have acquired the ability to proliferate in the absence of extracellular mitogenic signals. The apparent autonomy is caused by three key factors: (1) Self-production and stimulation by growth factors in an autocrine fashion, (2) Mutant growth factor ce1l surface receptors, and (3) Mutant intracellular signaling proteins (Hanahan and Weinberg, 2000). Thus, it is this third category of deregulation that appears to be found ubiquitously in most tumors and to which mutants of the Ras signaling pathways fall under. Specifically, the critical role of the Ras/MAPK pathway in transducing growth signals to the interior of the cell and subsequently stimulating cell growth and proliferation is underscored by the fact that roughly one quarter of all human tumors contain mutant forms of Ras proteins (Hanahan and Weinberg, 2000). Therefore, a more detailed discussion of the significance of the MAPK pathway to cancer development is warranted in an effort to convey a more comprehensive understanding of the tumorigenic process. MAPK Signaling Pathways in Cancer At this point, it would be helpful to review the major steps of the Ras/MAPK pathway before proceeding to a more detailed consideration of the roles that downstream mediators of this widespread and critical cell signaling pathway play in cancer cell biology. The Ras/MAPK pathway is one of the principal means by which extracellular and/or mitogenic signals are transduced from the surface of the cell to the cell's interior (Fang and Richardson, 2005). The response often culminates in the nucleus with the transcription and expression of target genes that regulate cellular proliferation, differentiation, and development (Seger and Krebs, 1995;Wang et al., 2002). The interaction of a particular growth factor with its cell surface transmembrane receptor (also known as a receptor tyrosine kinase-RTK-) sets into motion a series of events that result in the activation of several different cytoplasmic protein kinases. RTK activation by binding of a growth factor ligand induces dimerization followed by autotransphosphorylation of specific tyrosine residues within the receptor. These phospho-tyrosine residues are then recognized and bound by an adapter protein such as Grb2 through its SH2 domains (Seger and Krebs, 1995). Activation of Ras through the exchange of GDP for GTP by a specific guanine nucleotide exchange factor (GEF) protein known as Sos, then initiates the activation of a cascade of cytoplasmic serine-threonine protein kinases which include RAF1, MEK, and ultimately ERK. Each of these kinases is activated in sequential order by the protein kinase immediately preceding it in the pathway. Once phosphorylated and activated, ERK then proceeds to activate a number of nuclear proteins involved in regulating cell growth and proliferation. The regulation often occurs at the level of transcription. One of the primary targets of activation by ERK are various types of transcription factors, including the Etwenty-six (Ets) family of transcription factors as well as c-Myc and c-Fos (Fang and Richardson, 2005). These transcription factors are potent stimulators of cellular proliferation (Seger and Krebs, 1995). Mut et al. (2012) provide evidence of the importance of the Ras/MAPK pathway in the activation of the E-twenty-six like transcription factor 1 (Elk-1) in U138 glioblastoma multiforme cells. The researchers demonstrate that these cells have a high basal proliferative rate that can be reduced in the presence of specific MEK or ERK inhibitors. The authors hypothesize that inhibition of the Ras/MAPK pathway with these types of enzymatic inhibitors prevents the ultimate phosphorylation and activation of Elk-1, which in turn prevents the transcription of specific early cellular proliferation genes such as c-Fos. Activated Elk 1 exerts its effect in the nucleus by binding to the promoters of genes containing a serum response element (SRE) motif. The expression of a number of important cell cycle stimulatory genes, such as Egr1 and c-Fos, can be stimulated by Elk-1. The results also indicate that stimulation with EGF results in a corresponding increase in the proliferative rate, which is most likely mediated by Elk-1 (Mut et al., 2012). For example, the knockdown of Elk-1through the use of siRNA does not result in an increase in proliferation even after EGF stimulation in U138 cells. The results also suggest that the PI3-K/Akt pathway plays an important role in the regulation of Elk-1activity. Inhibitors of this pathway do not prevent the phosphorylation of either ERK or Elk-1in the cytoplasm following EGF stimulation, but rather prevent the movement of these phosphorylated proteins from the cytoplasm into the nucleus (Mut et al., 2012). The step is obviously necessary to produce the full mitogenic response to EGF stimulation. Given the central role of the Ras/MAPK pathway in the proliferative response of cells to growth signals, it is logical that abnormalities in any one of its molecular components could lead to the uncontrolled cellular division characteristic of so many different cancers. For example, Fang and Richardson (2005) discuss the significance of the Ras/MAPK pathway in promoting growth, proliferation, and tumorigenesis in intestinal epithelial cells. Indeed, the importance of abnormalities in the MAPK pathway in promoting tumorigenesis is highlighted by the fact that this pathway is deregulated in about 30% of all cancers (Fang and Richardson, 2005). Abnormal MAPK signaling in colorectal cancer typically begins at the surface of the cell with overexpression and activation of EGF receptors. Additionally, protein kinase C (PKC), when activated, can also facilitate the binding of GTP to Ras, which ultimately leads to activation of the MAPK pathway. Like many other types of cancers, the development of colorectal cancer is a multi-step process involving mutations in specific cell cycle regulatory or signaling genes. Mutations in Ras, specifically the K-Ras isoform, are known to be an early step in colorectal carcinogenesis along with mutations in certain protein kinases such as BRAF (Fang and Richardson, 2005). Hyperactive MAPK signaling through EGF-receptor overexpression or mutant protein kinases in turn leads to the activation of various transcription factors, as mentioned earlier, which promotes cell growth and proliferation. Additionally, unusual MPAK signaling can also induce the expression of VEGF, which aids in tumor invasiveness and metastasis by promoting angiogenesis. Invasiveness of colorectal cancer might also be linked to increased synthesis of matrix metalloproteinases such as MMP7, which has been associated with abnormal

MAPK signaling (Fang and Richardson, 2005). Given the key oncogenic role of the Ras/MAPK pathway in a variety of cancers, it is no surprise that a number of MEK and ERK inhibitors are currently being tested as possible chemotherapeutic agents. RIN1 as a Key Effector in the MAPK Pathway Besides anomalous cell signaling through the Ras/MAPK pathway, there is a growing body of evidence that suggests that RIN1 might be critical in modulating the cell's response to mitogenic signals Smith and Ikura, 2014). The ability of RIN1 to have a moderating effect on signal transduction through the Ras/MAPK pathway lies in its ability to be activated by Ras. In fact, RIN1 has been shown to compete directly with RAF1 for activation by Ras , and this selective competition could have important implications for cancer biology research. For instance, could RIN1 be a useful target to help dampen or modify excessive signaling through the Ras/MAPK pathway in certain cancer cell lines? If so, could one then reduce the downstream proliferative response induced by the Ras/ MAPK pathway as a result of growth factor stimulation? We postulate that RIN1, due to its strong binding affinity for activated Ras, may be able to attenuate signaling through the MAPK pathway and thus potentially regulate the function of a key downstream MAPK effector such as telomerase. Recent experimental findings from our laboratory examining telomerase activity indicate that this as a distinct possibility. Before addressing these possibilities, it is necessary to first review the functional role of RIN1 at the cellular level. An interesting study by Milstein et al. (2007) indicates that RIN1 behaves as a tumor suppressor in breast cancer cells and that RIN1 expression is reduced in tumorigenic cells when compared to normal breast epithelial cells. The authors report that RIN1 expression is inhibited by overexpression of the repressive transcription factor SNAI1 in ZR75-1 breast cancer cells. Additionally, methylation of the RIN1 promoter in KPL-1 breast cancer cells is observed to silence RIN1 expression and contribute to the cancerous phenotype. Knockdown of SNAI1 or demethylation of the RIN1 promoter each restored RIN1 expression with the result of inhibiting key tumorigenic processes (Milstein et al., 2007). Han et al. (1997) and Wang et al. (2002) studied the biochemical properties of RIN1 and reported that RIN1 is an important downstream effector of activated Ras. RIN1 is able to bind to Ras through its Ras binding domain (RBD), which associates with an effector domain within Ras. Furthermore, biochemical analysis reveals that RIN1 has a high binding affinity for activated Ras and that it competes strongly with RAF1 for access to Ras. Molecular studies also indicate that RIN1 binds to 14-3-3 proteins in the cytoplasm and is able to interact with and be phosphorylated by the tyrosine kinase c-ABL. The ability of RIN1 to bind to Ras lies in its carboxylterminal domain, which contains a 433 amino acid sequence (between residues 294 to 727, Fig. 1) necessary for Ras binding . Different Ras effector proteins, such as RIN1 and RAF1, can vary significantly in terms of their overall primary structure but often display a high degree of similarity in the regions required for Ras binding (Ras binding domains). The Ras binding domains of effector proteins typically interact with a short amino acid effector sequence within Ras itself, and this interaction is heavily dependent on Ras being in its activated GTP-bound state. For RIN1, the carboxyl-terminal domain (or Ras Association-RA-Domain) mediates binding to both GTP-Ras and 14-3-3 proteins . The affinity of RIN1 for Ras can be seen in experiments utilizing antibodies against the GTP bound form of Ras. For example, RIN1 is co-immunoprecipitated with an overly active mutant allele of H-Ras in NIH 3T3 cells when treated with a particular Ras antibody. When the same antibody is pre-attached to Ras, RIN1 is not co-immunoprecipitated. The use of antibodies that bind to the switch II region of Ras also prevents the binding of RIN1. Additionally, the FIGURE 1. These Rin domains are human Rin (1,2,3) based on isoform 1 sequences. SH2 binds phospho-tyrosine residues. Vps9 acts as a guanine nucleotide exchange factor for Rab5. RA binds to activated Ras. Proline-rich (PR) binds to SH3 domains. effector binding domain within Ras itself is equally important in mediating the proper attachment of certain effector proteins to Ras . The point is illustrated by the fact that mutations in the Ras effector binding domain can selectively inhibit the binding of certain Ras effector proteins without affecting the binding of others. RIN1, for instance, can interact with a constitutively active mutant form of H-Ras (H-Ras V12 ). However, when this mutant form of H-Ras acquires additional mutations at amino acid positions 35 or 40, RIN1 binding is completely inhibited. The position 35 mutation, on the other hand, does not interfere with RAF1 binding, while a mutation at position 37 permits RIN1 association but prevents RAF1 binding . The ability of full-length RIN1 to interact with Ras is further highlighted by a variant form of RIN1 in which there is a 62 amino acid deletion. The naturally occurring truncated form of RIN1, known as RIN1 delta (D), is the result of alternative splicing of the RIN1 mRNA. This RIN1:D mutant exhibits a much weaker binding affinity for Ras when compared to fulllength RIN1 . Besides interaction with Ras, the carboxyl-terminal domain of RIN1 also mediates binding to 14-3-3 proteins Wang et al., 2002). These small acidic proteins exist as multiple isoforms in the cytoplasm (epsilon, beta, and zeta) and have been shown to be involved in mitogenesis and malignant transformation of cells through their interactions with signaling proteins. Both RAF1 and RIN1 share the ability to bind to 14-3-3 proteins via their Ras binding domains, and this binding typically results in the activation of RAF1 through a Ras-dependent mechanism that seems to enhance signal transduction functions . Since RIN1 competes directly with RAF1 for access to Ras, the binding of RIN1 to 14-3-3 proteins would reduce its potential to bind to Ras and thus allow for heightened signaling through the Ras/MPAK pathway. The binding of RIN1 to cytoplasmic 14-3-3 proteins would promote the oncogenic properties of cell growth and proliferation as a result of the increased access of RAF1 over RIN1 for activation by Ras. Interestingly, the deletion of the Ras binding domain within the carboxyl-terminal of RIN1 prevents RIN1 from binding to all isoforms of 14-3-3 proteins, as does the 62 amino acid deletion in the naturally occurring RIN1:D mutant . The interaction of RIN1 with 14-3-3 proteins appears to be largely controlled by a serine residue at position 351 within the Ras binding domain. The serine residue is phosphorylated predominantly by protein kinase D (PKD), and its phosphorylation is required for proper binding to 14-3-3 proteins . A mutation that substitutes alanine at this position blocks interaction with 14-3-3 proteins and results in an increased ability to suppress signaling through the Ras/MAPK pathway, presumably caused by an increased capacity to compete with RAF1 for access to Ras. The phosphorylation of serine 351 of RIN1 and subsequent attachment to 14-3-3 proteins may act as a suppression control mechanism in cells to unlink RIN1 from activated Ras by sequestering it in the cytoplasm . An important point is raised here; for RIN1 to effectively compete with RAF1 for access to Ras, it must be in the proper subcellular location. Wild type RIN1 in its nonphosphorylated state is weakly associated with the cell membrane and is in close proximity to interact with Ras, which is tightly associated with the plasma membrane. Also, the alanine substitution at position 351 in the mutant form of RIN1 allows for a significant shift to the plasma membrane, which may help to explain its suppressive effect on Ras signaling . However, when phosphorylated by PKD, wild type RIN1 is confined to the cytoplasm bound to 14-3-3 proteins. The phosphorylation of RIN1 by PKD therefore reduces its capacity to compete with RAF1 for binding to Ras. The amino-terminal of RIN1, like the carboxyl-terminal, plays an important role in mediating cell signaling through its ability to bind to the tyrosine kinase c-ABL (ABL1). ABL tyrosine kinases are known to be involved in various cellular functions including differentiation, division, migration, and adhesion (Hu et al., 2005). Additionally, the amino-terminal of RIN1 contains an SH2 domain capable of interacting with phospho-tyrosine residues on an activated receptor tyrosine kinase such as the EGF-receptor (Barbieri et al., 2004). RIN1 interacts with c-ABL most likely through a proline-rich sequence in its amino-terminal end and an SH3 domain in c-ABL . Upon binding to c-ABL in vitro, RIN1 becomes tyrosine phosphorylated and can then subsequently bind to an SH2 domain within c-ABL. The interaction does not seem to affect the catalytic activity of the enzyme however, and studies have shown that RIN1 has very little interaction with c-ABL in vivo. This reduced interaction is caused in part by the different cellular locations of the two proteins, with c-ABL being confined mostly to the nucleus and RIN1 to the cytoplasm . One interesting exception to this is the oncogenic BCR/ ABL fusion protein produced as a result of a translocation between chromosomes 9 and 22. The BCR/ABL is an unregulated tyrosine kinase that is localized primarily to the cytoplasm, where it stimulates cellular proliferation. The BCR/ABL is therefore in the correct location to interact with RIN1 and, indeed, RIN1 is able to bind to BCR/ABL. RIN1 appears to accentuate the tumorigenic, transforming properties of BCR/ABL Hu et al., 2005;Afar et al., 1997). Additionally, Hu et al. (2005) report that RIN1 is an activator of the ABL2 tyrosine kinase that is involved in the regulation of epithelial cell adhesion and migration. Specifically, RIN1 activation of ABL2 promotes phosphorylation of the adaptor proteins CRK and CRKL. This phosphorylation, in turn, produces conformational changes in CRK and CRKL, which influences cytoskeletal elements to inhibit cell motility. Cells deficient in RIN1 display reduced levels of phosphorylated CRKL and increased cell motility (Hu et al., 2005). Clearly, the biochemical profile of RIN1 suggests that it potentially plays a key role in modulating signaling through the Ras/MAPK pathway, given its ability to interact with multiple different effector proteins. The greatest capacity for RIN1 to moderate abnormal signaling, as detailed above, appears to lie in its ability to directly compete with RAF1 for access to Ras. The competition, in turn, could be a useful mechanism for dampening or attenuating signaling through the Ras/MAPK pathway by diverting the signal away from effector proteins downstream of Ras. The exact role that Rin1 plays in tumorigenesis remains unclear. For instance, Milstein et al. (2007) describe a tumor suppressor role for Rin1 in breast cancer cell lines. On the contrary, elevated levels of RIN1 may be associated with increased tumorigenesis and lower survival, as is reported by Wang et al. (2012) for nonsmall cell lung cancer. It should be noted that although RIN1 is thought to be expressed in most tissues, its expression is highest in brain tissues . The differential level of RIN1 expression could have important implications for research when investigating RIN1 function in various tissues and/or cell lines . RIN1 as an Early Modulator for Leading Telomerase Activity In addition to its ability to moderate mitogenic signaling through the Ras/MAPK pathway, RIN1 also acts as a guanine nucleotide exchange factor (GEF) for the small monomeric GTPase known as Rab5 (Tall et al., 2001). The Rab proteins are a diverse group of proteins that belong to the Ras superfamily of small GTPases and play critical roles in regulating the steps of endocytic vesicular transport. Specifically, Rab proteins regulate vesicular traffic from the plasma membrane by controlling cargo selection, vesicle formation, transport along the cytoskeleton, and fusion with intracellular target membranes (Stein et al., 2003;Hutagalung and Novick, 2011;Stenmark and Olkkonen, 2001). Certain Rab proteins also control the sorting of molecules for return to the plasma membrane or for degradation in lysosomes. The ability of Rab proteins to regulate the many complex steps of vesicular transport lies in their selective activation. There are approximately 70 or 90 different Rab proteins encoded in the human genome, each of which is selectively activated by binding GTP. Once activated, a particular Rab protein serves as a scaffold for the attachment of various effector proteins, which then subsequently direct the completion of

a specific step in the endocytic pathway. Rab5, which is activated by RIN1, has been shown to regulate vesicle budding and cargo selection from clathrincoated pits as well as early endosome fusion (Stein et al., 2003;Hutagalung and Novick, 2011;Stenmark and Olkkonen, 2001). Tall et al. (2001) report that RIN1 and Rab5 also play a crucial role in the receptor-mediated endocytosis of epidermal growth factor (EGF)-receptor, following stimulation by EGF. The endocytosis of EGFreceptor occurs through a Ras-mediated mechanism involving upstream activation of RIN1 by GTP-Ras and then subsequent activation of Rab5 by activated RIN1. The pivotal integrative role of Rab proteins in signal transduction is further illustrated in a study by Barbieri et al. (2004) examining the role of Rab5 in EGF-receptor mediated MAPK signal transduction. RIN1 interacts with Rab5 via its Vps9 domain, which also contains the GEF activation function for Rab5 (Galvis et al. 2009a). Galvis et al. (2009b) have previously reported the key function of the Vps9 domain of RIN1 in activating Rab5 through a series of mutational studies. Barbieri et al. (2004) demonstrate that a dominant-negative mutant form of Rab5 (Rab5:S34N) is capable of inhibiting activation of the MAPK pathway in mouse NR6 cells by preventing both the endocytosis and internalization of the EGF-receptor. The inhibition is particular to the MAPK pathway and does not interfere with other EGF induced kinase pathways. On the other hand, however, expression of wild type Rab5 or the RIN1 delta splice variant leads to increased MAPK activity and increased cyclin D1 expression after EGF stimulation-resulting in heightened cellular proliferation. The authors suggest that Rab5 activation is a key step in linking EGF stimulated endocytosis to signal transduction through the MAPK pathway (Barbieri et al., 2004). Although the exact biochemical link between Ras association and the Rab5 GEF activity of RIN1 is not entirely clear, the binding of Ras by RIN1 appears to strongly influence EGF-receptor endocytosis. In a manner similar to EGF-receptor endocytosis, the internalization of insulin receptor following the binding of insulin may involve other steps including Rab5 activation by RIN1 (Hunker et al., 2006). It is interesting to note that RIN1 sits at the intersection between cell signaling and receptor-mediated endocytosis for mitogens such as EGF and insulin. The increased rate of receptor-mediated endocytosis potentiated by Ras stimulation of the Rab5 GEF activity of RIN1 may be an important negative feedback mechanism by which Ras can divert signaling away from downstream effectors through RIN1. In this model, an increased rate of receptor-mediated endocytosis would favor the quick removal of receptors from the plasma membrane, followed by internalization and degradation (Hunker et al., 2006). The end result would be an attenuation of mitogenic signal transduction much in the same way as can be achieved by the direct competition of RIN1 with RAF1 for access to Ras. A number of Ras effector proteins have been shown to contain either a Ras association (RA) domain or a Ras binding domain (RBD). Although not identical in terms of sequence homology, these two domains do share a common ubiquitin fold structure that enables interaction with the surface of Ras (Erijman and Shifman, 2016;Takala and Ylanne, 2012). In a study of integrative Ras signaling, Smith and Ikura (2014) used NMR probes to determine a preferential Ras binding hierarchy among equal concentrations of various Ras effector proteins. The authors report that wild-type Ras binds preferentially to BRAF, followed closely by Rin1 (Smith and Ikura, 2014). Having examined the biochemical properties of RIN1 and how it fits into the larger signal transduction machinery of the Ras/MAPK pathway, it would be useful to elaborate on the possible connection between RIN1 and specific downstream effectors of the Ras/MAPK pathway known to induce cellular proliferation. These downstream targets, while varied, include the Ets family of transcription factors mentioned earlier as well as the telomerase reverse transcriptase enzyme. It is intriguing to wonder if RIN1, given its ability to compete directly with RAF1 for access to Ras, could have any influence on the cellular proliferation associated with classical MAPK targets such as the Ets transcription factors and/or telomerase. The signaling pathways that have been implicated in the stimulation of telomerase expression and activity usually involve the response to a mitogen such as EGF or IGF-1, and subsequent activation of a number of protein kinases belonging to the Ras/MAPK and PI3-K/Akt pathways (Inui et al., 2002;Seimiya et al., 1999;Zhou et al., 2013). It is feasible then that RIN1 could potentially have an effect on telomerase expression, given its role as a Ras effector molecule. If so, this raises the exciting idea of RIN1 as a potential therapeutic target in specific cancers. To appreciate the role of RIN1 in a therapeutic context, it is necessary to investigate what effect the regulation of the telomerase reverse transcriptase enzyme and the Ets family of transcription factors have on cellular proliferation. Telomere Structure and Function Telomeres are the regions of DNA that exist at the very tips or ends of linear chromosomes in eukaryotic cells. Recent evidence suggests that the telomere regions of chromosomes play vital roles in regulating normal cellular processes such as proliferation, aging, and senescence/apoptosis (Ramlee et al., 2016;Shay, 2018). The telomeric regions of chromosomes set a replicative limit on the number of cell divisions a normal cell can undergo before the induction of senescence and/or apoptosis occurs (Chung Low and Tergaonkar, 2013;Maciejowski and de Lange, 2017). The finite replicative capacity, which is approximately 50 to 70 rounds of cell division for most normal somatic cells, is largely due to the progressive loss of telomeric DNA with each round of cell division (Zvereva et al., 2010). The loss of telomeric DNA that accompanies each cycle of cell division is a consequence of incomplete DNA replication at the telomeres (Jafri et al., 2016). The telomere regions of vertebrate chromosomes are defined by a distinct structure that distinguishes telomeres from other areas of chromosomal DNA. In humans, a highly repetitive sequence of TTAGGG approximately 5 to 15 kb in length is tightly associated with a complex of six telomeric DNA binding proteins termed a shelterin complex (Stewart et al., 2012). The shelterin protein complex is critical for maintaining genomic stability at the telomeres and consists of the following six proteins: TRF1, TRF2, TIN2, TPP1, POT1, and RAP1 (Stewart et al., 2012). TRF1 and TRF2 (telomeric repeat factor-binding protein) each form separate homodimers that bind to the double-stranded DNA regions of telomeres. Both TRF1 and TRF2 act as negative regulators of telomere length and are linked to each other by the TIN2 linker protein (Walker and Zhu, 2012). The POT1 (protection of telomeres) binds to the 3' singlestranded guanine-rich overhang and mainly functions to inhibit the activation of DNA damage response pathways at the telomeres. The POT1 is connected to the other shelterin complex proteins through the linker protein TPP1 which binds to both POT1 and TIN2 (Heidenreich and Kumar, 2017). The RAP1 is a small protein that binds to TRF2 and aids TRF2 in preventing non-homologous end joining and chromosomal fusions (Stewart et al., 2012). The shelterin complex, which assists in the formation of T-loops of telomeric DNA, functions as a protective cap that prevents the activation of cellular DNA damage responses that would otherwise recognize the telomeres as double-stranded DNA breaks (de Lange, 2018;Stewart et al., 2012;Buseman et al., 2012). The major role of telomeres is to maintain genomic stability by acting as a buffer for the gradual erosion of DNA that accompanies each cycle of cell division. Each shelterin complex protein plays a clear role in telomere maintenance, and the loss of any one protein will result in reduced telomere protection (de Lange, 2018;Bandaria et al., 2016;Erdel et al., 2017). Telomerase is only expressed and active in a relatively small population of somatic stem cells where the depletion of telomeric DNA and the resulting onset of senescence or apoptosis would interfere with the normal functioning of specific tissues (Chung Low and Tergaonkar, 2013). On the other hand, the immortal replicative phenotype that is characteristic of the vast majority of cancer cells is attributed to high levels of telomerase activity. In fact, roughly 85% of all cancers display telomerase activity, and the activation of telomerase activity in these cells is a key step in the tumorigenic process (Zhu et al., 2010;Schmidt et al., 2016). Consequently, the selective inhibition of telomerase activity in cancer cells has emerged as an attractive therapeutic target (Holysz et al., 2013;Buseman et al., 2012). Telomerase Expression and Regulation Normal somatic cells lose approximately 50 nucleotides of telomeric DNA with each cycle of cell division as a result of incomplete DNA replication during the S phase. The replication problem is overcome in certain cells by extension of the 3' end of the template DNA strand by the telomerase enzyme. Telomerase is a ribonucleoprotein that functions as a reverse transcriptase by adding sequential repeats of the hexameric sequence TTAGGG to the 3' end of the template DNA strand; and once extended by telomerase, DNA polymerase α-primase (PαP) can then complete replication of the lagging daughter strand (Hockemeyer and Collins, 2015). The functional telomerase holoenzyme consists of two essential parts: an enzymatic protein component known as TERT that acts as a reverse transcriptase and an RNA template component known as TR (Heidenreich and Kumar, 2017). The TERT mRNA is first synthesized and processed in the nucleus and then exported to the cytoplasm for translation into protein. The TERT is then subsequently moved into the Cajal bodies where it assembles with TR to form the final functional telomerase holoenzyme. The active telomerase enzyme will then be recruited to the telomeres at the appropriate time for telomeric DNA synthesis (Hockemeyer and Collins, 2015;Podlevsky and Chen, 2012). Interestingly, the regulation of telomerase activity occurs mainly at the level of TERT transcription as TERT mRNA synthesis is highly regulated in most somatic cells (Leao et al., 2018). The synthesis of TERT mRNA appears to be the rate-limiting step in the regulation of telomerase activity as TERT gene transcription is highly repressed in most somatic cells (Zhu et al., 2010;Zhang et al., 2016). The telomerase RNA template (TR), however, is ubiquitously expressed in many cell types (Daniel et al., 2012). The TERT core promoter contains binding sites for several key transcription factors known to regulate TERT transcriptionthese include c-Myc, SP1, ER, Ets, AP1, and NF-κB (Zhu et al., 2010). The c-Myc binds to two E-box consensus sequences within the TERT promoter and is strongly linked to activation of TERT transcription. Similarly, SP1 plays a critical role in activating TERT transcription as the mutation of SP1 binding sites in the TERT promoter significantly reduces TERT expression (Cifuentes-Rojas and Shippen, 2012). Two estrogen response elements are located within the TERT promoter upstream of the transcription start site and enhance TERT expression when bound by ERα (Daniel et al., 2012). On the other hand, TERT expression is negatively regulated by the binding of tumor suppressor proteins such as Wilm's tumor-1 (WT1) protein. Additionally, p53 expression acts to inhibit TERT transcription. The p53 tumor suppressor protein has been shown to interact with both human telomerase-associated protein 1 (hTEP1) and SP1 to inhibit TERT promoter activity (Lu et al., 2013). When discussing transcriptional control of telomerase expression, it is necessary to examine the hTERT promoter. It is well established that the c-Myc transcription factor is strongly linked to cellular proliferation, and it comes as no surprise that telomerase expression is upregulated by c-Myc. c-Myc is able to interact directly with the hTERT promoter and stimulate hTERT expression. Transcription factors such as c-Myc stimulate gene expression by binding to specific sequences within the promoter region of a gene, and many cancer cells often accumulate mutations within the promoters of key cell cycle regulatory genes. In a large-scale study of 799 tumor samples, Huang et al. (2015) identified two specific mutations that were present in the hTERT promoter in a high percentage of the tumor samples. Specifically, these mutations C228T and C250T occur at 124 and 146 bp upstream of the hTERT translation start site and are prevalent in many different types of tumors. The authors conclude that each mutation creates a new binding site for the E-twenty-six (Ets) group of transcription factors which subsequently results in the upregulation of telomerase expression. Upregulation of telomerase activity has also been reported as a

consequence of the hypoxia commonly found in solid tumors, where the enzyme may aid in the stabilization of chromosomal damage induced by the low oxygen environment (Seimiya et al., 1999). Increased telomerase expression has also been observed in hepatocytes undergoing cell division in response to partial removal of liver tissue, and this expression is stimulated by both EGF and hepatocyte growth factor (HGF) (Inui et al., 2002). In each instance, the upregulation of telomerase expression was a direct response to signaling through the MAPK pathway. The central role that the MAPK signaling pathway plays in telomerase activation is illustrated in a study by Maida et al. (2002). The effect of EGF stimulation on telomerase expression was investigated by exposing A-431, ME180, MCF-7, and NIH3T3 cells to EGF for various periods of time, and then assessing hTERT mRNA expression. There was a significant increase in hTERT mRNA expression in all the cell types between 6 to 12 h after EGF exposure. The increase in hTERT mRNA expression in response to EGF stimulation was confined to these cell types which constitutively express telomerase and was not observed in telomerase negative cell types. The authors conclude the MAPK pathway is primarily responsible for EGF induced telomerase expression. A-431 cells stimulated with different concentrations of EGF for either 15 or 30 min showed a marked increase in ERK activity and hTERT expression. Exposure to the MEK inhibitor U0126 abolished this effect but exposure to PI3K or p38 inhibitors did not. Furthermore, the authors demonstrate that EGF exerts its influence on telomerase expression through the MAPK pathway via activation of the Ets family of transcription factors. This group of transcription factors is phosphorylated and activated by Erk, and the hTERT promoter contains two Ets binding motifs within it at the -23 and -18 positions. A-431 cells co-transfected with a wild-type Ets2 expression vector and an hTERT promoterreporter plasmid revealed high levels of EGF induced transactivation when compared to cells co-transfected with a truncated, dominant-negative Ets2 expression plasmid. A similar result was observed in cells transfected with either a normal hTERT promoter-reporter plasmid or a plasmid containing mutations in the Ets binding motifs (Maida et al., 2002). The MPAK pathway has also been implicated in the activation of telomerase activity and hTERT mRNA expression in estrogen receptor α (ERα) positive endometrial cells. Zhou et al. (2013) report that stimulation of Ishikawa ERα positive endometrial cells with estradiol (E2) results in increased phosphorylation p44/42 MAPK and increased hTERT mRNA expression. The hTERT promoter contains two binding sites for E2-ERα complexes, and the stimulatory effect of E2 on hTERT expression and telomerase activation was greatly reduced by exposure to the MEK inhibitor U0126 or ERK specific siRNA . The MAPK pathway therefore appears to play a prominent role in the estrogen-induced regulation of hTERT expression and telomerase activity. Telomerase activity can also be regulated to a lesser extent by post-translational modifications of the catalytic TERT subunit, such as phosphorylation and ubiquitination. Multiple protein kinases like c-Abl, protein kinase B (PKB) (also known as Akt), and protein kinase C (PKC) are able to influence telomerase activity (Wojtyla et al., 2011). Tyrosine phosphorylation of TERT by c-Abl tends to reduce TERT activity while phosphorylation of serine/threonine residues by Akt tends to stimulate TERT activity. Protein phosphatase 2A has been shown to reduce telomerase activity in specific cell types as TERT phosphorylation is necessary for its import into the nucleus. The TERT stability and half-life in the cytoplasm are controlled by E3 ubiquitin ligases that can target TERT for proteolytic degradation and thus prevent its entry into the nucleus (Cifuentes-Rojas and Shippen, 2012). Several splice variants of human TERT are known to exist each with differing levels of activity. Epigenetic regulation of the TERT promoter figures prominently for TERT expression as the TERT promoter is located within a region of highly condensed chromatin (Cifuentes-Rojas and Shippen, 2012; Lewis and Tollefsbol, 2016). Both hypoacetylated TERT promoter and CpG island methylation are also commonly associated with the silencing of TERT expression. Conversely, methylation of lysine 9 of histone 3 (H3K9) increases TERT expression, as does the hyperacetylation of other core histones through the inhibition of histone deacetylase complexes (HDACs) (Zhu et al., 2010). Several transcription factors are known to epigenetically regulate TERT expression through the recruitment of histone acetyltransferases (HATs) or HDACs to the TERT promoter (Lu et al., 2013). Telomerase expression and activity can be modulated by various signal transduction pathways. Wnt transcription factor to stimulate expression of both cyclin D and c-Myc. c-Myc is a well-documented activator of TERT expression (Wu et al., 2013). Inflammation is commonly associated with many different cancers, and NF-κB is a primary regulator of chronic inflammation linked with tumorigenesis and cancer progression. Recent evidence suggests that a reciprocal relationship may exist between NF-κB activity and TERT expression. NF-κB can bind upstream of the TERT transcription start site to stimulate TERT expression. The TERT, on the other hand, can reinforce NF-κB activity by binding to the p65 subunit of NF-κB to enhance transcription of inflammatory genes such as IL-6 and TNFα (Ghosh et al., 2012). Additionally, the RAP1 shelterin complex protein is a key regulator of NF-κB activity (Martínez and Blasco, 2011). Other signal transduction pathways such as the PI3K/ Akt and MAPK pathways also contribute significantly to the regulation of TERT expression and activity. The Aktmediated phosphorylation of specific effector proteins promotes the proteolytic degradation of p53 and the c-Myc competitor protein MAD1 (Peek and Tollefsbol, 2016). Additionally, the PI3K/Akt pathway has been tied to NF-κB activation as well as the inhibition of TGF-β signaling through Jab1 activation and SMAD4 degradation (Daniel et al., 2012). The TGF-β signaling pathway is an important inhibitory pathway that mediates cell growth, differentiation, and proliferation. TGF-β signaling can suppress c-Myc expression but is susceptible to inhibition by estrogen (Peek and Tollefsbol, 2016). Heeg et al. (2011) report that EGFreceptor overexpression in OKF6 oral-esophageal cells enhances TERT transcription via the Hif1-α transcription factor and directly stimulates telomerase activity through the Akt-dependent phosphorylation of TERT. Similarly, estradiol (E2) is reported to increase telomerase activity in endometrial cancer cells through MAPK induction of TERT transcription. Inhibition of either MEK or ERK resulted in decreased luciferase activity from a reporter plasmid containing the TERT promoter following treatment with E2 . In both normal and cancerous cells, Ras effector proteins such as RIN1 may be useful targets for modulation of signaling through the MAPK pathway. The detrimental effects of abnormalities in the Ras/MAPK pathway are well documented for a variety of different cancers. Cancer, in the simplest sense, is a disease of the cell cycle, and dissection of the complex molecular interactions which govern cancer cell growth and reproduction will ultimately shed light on instances where moderation of aberrant cell signaling may be possible. RIN1, a known Ras effector molecule, may provide an avenue for attenuation of signaling through the Ras/ MAPK pathway. If so, it is intriguing to contemplate what effect molecules such as RIN1 could have on downstream targets of the MAPK pathway known to influence cellular proliferation. Telomerase, for example, is widely expressed in many cancers and has been heavily implicated in tumorigenesis. As a target of the Ras/MAPK pathway, it is interesting to speculate on a possible connection between telomerase expression and the potential MAPK signal modifying ability of a Ras effector such as RIN1 (Fig. 2). In summary, abnormal telomerase expression and activity is a defining hallmark observed in many different cancers. The immortal replicative phenotype conferred by the enzyme is FIGURE 2. Diagram of the proposed model for the moderating influence of RIN1 on signaling through the Ras/MAPK pathway. The binding of a growth factor such as IGF-1 to its receptor on the plasma membrane activates Ras and subsequently the MAPK pathway. Activation of the MAPK pathway in turn activates various nuclear transcription factors involved in telomerase expression. RIN1 competes with Raf for binding to activated Ras. critical to the survival and proliferation of malignant cancer cells. The regulation of telomerase activity occurs primarily at the level of transcription and typically becomes abnormal and disordered during the tumorigenic process. The fact that telomerase is not expressed in most normal somatic cells has made the selective inhibition of telomerase activity an attractive chemotherapeutic target. Other strategies that seek to interfere with the signaling pathways responsible for the activation of telomerase expression could also be equally important from a therapeutic standpoint. Funding Statement: The authors received no financial support for the research, authorship, and/or publication of this article. Conflicts of Interest: The authors declare that they have no conflicts of interest to report regarding the present review article. The Influence of COVID-19 on Influenza and Respiratory Syncytial Virus Activities Background: Respiratory viral diseases have considerably declined since the COVID-19 outbreak, perhaps through influence by nonpharmaceutical interventions. We conducted a cross-sectional study using the CDC database to compare the pre- vs. post-pandemic flu activity (incidence) between the US states. Our secondary objectives were to estimate the association between flu activity and flu vaccination rates and compare the national trends of flu and RSV activities since the pandemic outbreak. Methods: We estimated the difference between pre-pandemic (April 2019–March 2020) and post-pandemic (April 2020–March 2021) flu activity between individual states using the Wilcoxon signed-rank test. The association between flu activity and immunization rates was also measured. Finally, parallel time trend graphs for flu and RSV activities were illustrated with a time series modeler. Results: The median (IQR) pre-pandemic flu activity was 4.10 (1.38), higher than the post-pandemic activity (1.38 (0.71)) (p-value < 0.001). There was no difference between pre-pandemic (45.50% (39.10%)) and post-pandemic (45.0% (19.84%)) flu vaccine acceptance (p-value > 0.05). Flu activity and vaccination rates were not associated (p-value > 0.05). Flu activity has declined since the COVID-19 outbreak, while RSV made a strong comeback in June 2021. Conclusion: Flu activity has significantly diminished throughout the pandemic while a sudden upsurge in RSV is a public health concern indicative of possible resurgence of other viruses. Flu vaccine acceptance neither changed during the pandemic nor influenced the diminished Flu activity. Introduction In March 2020, COVID-19 was declared a pandemic by the World Health Organization [1]. Since then, most countries have implemented preventive guidelines, including quarantine, social distancing, hand washing, and face mask usage in public places [2,3]. Although these measures were primarily implemented to prevent the spread of COVID-19, incidences of other respiratory viruses were also influenced by these nonpharmaceutical interventions [4,5]. Influenza (flu) and RSV are common human respiratory viruses, which traditionally cause significant mortality and morbidity [6]. An average of 28 million people were affected and 35,000 people died of flu annually in the USA in 2010-2020 [7]. In contrast, RSV has been the predominant cause of bronchiolitis and pneumonia in infants and also affects the elderly population aged 65 years and over [8]. While a flu vaccine is available, only a monoclonal antibody can be offered to prevent RSV [9]. Flu immunization reduces the risk of having flu by 40-60% [10]. A national survey from 2020-2021 suggested that the people vaccinated against flu were more likely to accept the COVID-19 vaccine, too [11]. However, to the best of our knowledge, the flu vaccine acceptance trend at the time of the pandemic has not been well-reported in the USA. The Centers for Disease Control and Prevention (CDC) publish weekly flu activity reports for every state. However, it is difficult to extract and comprehend the difference in flu activity between individual states from that vast database. Thus, a US political map with categorical highlights of individual states based on recent flu activity would perhaps be helpful to visualize the variation across the country. This variation in flu activity might be influenced by factors such as nonpharmaceutical interventions [12,13], while flu vaccine acceptance rates could also be another contributor [14]. RSV and flu cases declined drastically in 2020 [15,16]. However, RSV made a strong comeback in 2021 [17]. Although data on flu and RSV activities are available in the public domain, a comparative analysis between the trends and timelines of flu and RSV since the COVID-19 outbreak has not been well-described in the literature. From the public health perspective, understanding the variation in their trends would be of interest since the two viruses differ in the incubation period, mode of transmission, and

period of infectivity [18]. We conducted a cross-sectional study based on the data publicly available on the CDC website to compare the pre-vs. post-pandemic variation in flu activity between the fifty US states and present the findings on a US political map that could be readily visualized. The secondary objective was to estimate the association between the flu activity and the flu vaccine acceptance rate since the pandemic outbreak. Finally, we compared the national trends in flu and RSV activities over the last two years (2019-2021) and presented them as a parallel time trend graph. Data Collection We accessed the CDC database and collected weekly flu activity reports from individual states [19]. Information regarding flu vaccine acceptance was collected from the FluVaxView interactive website [20]. We further accessed the National Respiratory and Enteric Virus Surveillance System (NREVSS) [21] in October 2021 and collected data on RSV incidence. RSV activity reports from the last two years (from November 2019 to October 2021) were publicly available on the NREVSS website at the time of data collection. Data Analysis (a) Flu activity was reflected on the CDC website on a severity-based scale with 13 levels. In comparison, RSV activity was reported as the number of new cases. We normalized RSV cases on a "1-13 scale" to match with the flu activity and run comparative analysis. The flu vaccine acceptance rate represented the percentage of the immunized population (individuals aged six months or above). The flu immunization season lasts from July of the given year to May of the following year. The flu vaccine acceptance rate (percentage of the population) increased cumulatively in every passing month. (b) The difference between pre-(from April 2019 to March 2020) and post-pandemic (from April 2020 to March 2021) flu activity between individual states were compared using Wilcoxon signed-rank test (nonparametric data). Fifty US states were divided into five quartiles (cohorts) depending on the decline in flu activity. An online graphics tool was used to build a US political map [22] demonstrating five cohorts of US states based on the decline in flu activity. (c) Spearman's correlation analysis (nonparametric data) estimated the association between monthly flu activity and flu vaccination rates. (d) We compared both RSV and flu activities between the peak season in 2021 with the corresponding weeks of 2020. (e) Time trend analysis: since the incidence of respiratory viral infections has changed considerably since the COVID-19 outbreak, a time trend analysis was conducted for flu and RSV activities (between November 2019 and October 2021). We used the SPSS time series modeler [23] to generate a time trend graph for the parallel representation of both viral diseases. Results (a) The median (IQR) value of monthly flu activity during the pre-COVID months (from April 2019 to March 2020) was 4.10 (1.38) compared to the post-COVID months' (from April 2020 to March 2021) activity of 1.38 (0.71). The difference was statistically significant (Wilcoxon test: p-value < 0.001). Table 1 demonstrates pre-vs. post-COVID-19 flu activity in all US states, while Figure 1 categorizes the US states based on the rate of decline in flu activity. The variations between the states were random, and we did not find a trend according to the US regions. Discussion This cross-sectional study based on weekly flu activity (2019-2021) in all US states reemphasized the post-pandemic decline in reported flu cases. We also found that flu vaccine acceptance was primarily unchanged between the pre vs. post-pandemic periods, and the vaccine acceptance rate did not influence the flu activity. Finally, the time trend analysis illustrates a low flu profile throughout the pandemic (till October 2021), and peak RSV activity in the summer of 2021. Several studies conducted during the early months of the pandemic demonstrated an overall decline in non-COVID-19 respiratory viral infections [24,25]. Reports from other countries, including China, Australia, and France, also cited diminished flu cases since the pandemic was declared [26,27]. Stamm et al. described a parallel decline in flu and RSV incidence in Germany in 2020, while Tempia et al. reported a similar trend in South Africa [28,29]. Likewise, both flu and RSV cases declined in the USA in 2020. Flu vaccine acceptance behavior is known to be influenced by COVID-19. A UK-based national survey reported that the pandemic had motivated unvaccinated people to accept the flu vaccine [30], while another survey from Pennsylvania, USA, described net favorable changes in flu vaccine acceptance behaviors since the COVID-19 outbreak [14]. However, when the data from the CDC were analyzed, we found an unchanged flu vaccine acceptance rate since the pandemic outbreak. A sudden uptrend in RSV during COVID-19 was previously Discussion This cross-sectional study based on weekly flu activity (2019-2021) in all US states re-emphasized the post-pandemic decline in reported flu cases. We also found that flu vaccine acceptance was primarily unchanged between the pre vs. post-pandemic periods, and the vaccine acceptance rate did not influence the flu activity. Finally, the time trend analysis illustrates a low flu profile throughout the pandemic (till October 2021), and peak RSV activity in the summer of 2021. Several studies conducted during the early months of the pandemic demonstrated an overall decline in non-COVID-19 respiratory viral infections [24,25]. Reports from other countries, including China, Australia, and France, also cited diminished flu cases since the pandemic was declared [26,27]. Stamm et al. described a parallel decline in flu and RSV incidence in Germany in 2020, while Tempia et al. reported a similar trend in South Africa [28,29]. Likewise, both flu and RSV cases declined in the USA in 2020. Flu vaccine acceptance behavior is known to be influenced by COVID-19. A UK-based national survey reported that the pandemic had motivated unvaccinated people to accept the flu vaccine [30], while another survey from Pennsylvania, USA, described net favorable changes in flu vaccine acceptance behaviors since the COVID-19 outbreak [14]. However, when the data from the CDC were analyzed, we found an unchanged flu vaccine acceptance rate since the pandemic outbreak. A sudden uptrend in RSV during COVID-19 was previously reported, mostly from regional hospitals [17,31]. In contrast, our study comprehensively summarized statistics from a national database. Several factors influence the transmission of viral respiratory diseases. Susceptible host, virus, and the environment have a complex interplay which determines the successful spread of the disease [32]. While flu can be transmitted by contact, droplet, or aerosol, [33][34][35] RSV is predominantly transmitted via direct or indirect contact and often remains transmissible for a more extended period than flu [33,[36][37][38][39]. Understandably, nonpharmaceutical interventions such as hand washing, face masks, and social distancing helped to prevent flu, and we found that flu vaccination coverage certainly was not the reason behind remarkable flu containment. However, an untimely resurgence of RSV in 2021 was unique. In the past, the traditional winter peak of RSV helped to build immunity among the exposed children. During the first post-pandemic season, RSV was effectively prevented by nonpharmaceutical interventions [40]. Subsequently, the children became immunologically naïve against RSV, and infants did not receive passive immunity from their mothers either. Moreover, routine immunization is available against flu, but not against RSV. All these factors perhaps led to an early resurgence in RSV in countries like the USA, Australia, and South Africa [41][42][43]. Our study has several limitations. We compared the flu and RSV activities only, while other human respiratory viruses such as rhino-enterovirus, human metapneumovirus, and parainfluenza were not considered. The impact of nonpharmaceutical interventions on non-COVID-19 respiratory viral diseases has been well-accepted. However, mask policy and lockdown practices have changed several times in every state since the COVID-19 outbreak. Thus, direct cause-effect analyses of the preventive practices and flu or RSV activities were beyond the scope of this study. Furthermore, identifying the predictors of RSV resurgence was not the focus of our research. Nonetheless, this study should be of interest among the clinicians since the decline in flu activity was comprehensively summarized on the US political map. The parallel time trend graph of flu and RSV activities should also help visualize the pattern over the last two years. Finally, very few previous studies focused on the flu vaccination trend and its association with flu activity during the COVID-19 pandemic. Thus, our study results could be helpful to epidemiologists. Although nonpharmaceutical interventions helped to reduce non-COVID-19 respiratory viral diseases, we should not shift our focus from traditional respiratory viruses to COVID-19 only. The upsurge of RSV perhaps was not an isolated phenomenon and should be considered as an eye-opener as other respiratory viruses can have a similar comeback [25]. Some experts have even warned against the future resurgence of flu [44]. It is time for policymakers, stakeholders, and clinicians to be aware of the potential untimely resurgence of viral diseases and be more flexible accordingly. We found significant variations in the decline in flu activity between different states. Future researchers may focus on the potential factors that led to better containment of flu activity in some states and build informed preventive strategies. Finally, in-depth research on the predictors of RSV resurgence would be of importance, as a better understanding of the disease process from the epidemiological and virological perspectives would help to prevent similar circumstances in the future. Conclusions Flu activity had significantly declined since the pandemic outbreak and remained low until October 2021. RSV incidence also declined in 2020; however, it had an untimely peak in summer 2021. Flu vaccine acceptance neither changed during the pandemic nor was associated with declined flu activity. An untimely upsurge in RSV cases is a public health concern and indicates a possible future resurgence of other respiratory viral diseases. Author Contributions: Each author made substantial contributions to the conception or design of the work; the acquisition, analysis, and interpretation of data; each author also drafted the work or substantively revised it; and approved the submitted version and agrees to be personally accountable for the author's own contributions and for ensuring accuracy or integrity of any part of Institutional Review Board Statement: Not applicable since this study did not involve humans or animals. Geodemography, environment and societal characteristics drive the global diversity of emerging, zoonotic and human pathogens Abstract Understanding human disease, zoonoses and emergence is a global priority. A deep understanding of pathogen ecology and the complex inherent relationships at the agent–environment interface are essential to inform disease control and mitigation and to predict the next zoonotic pandemic. Here, we present the first analysis of social and environmental factors associated with human, zoonotic and emerging pathogen diversity at a global scale, controlling for research effort. Predictor–response associations were captured by generalized additive models. We used national level data to aid in policy development to inform control and mitigation. We show that human population density, land area, temperature and the human development index at the country level are associated with human, emerging and zoonotic pathogen diversity. Multiple models demonstrating society–agent–environment interactions demonstrate that social, environmental and geographical factors predict global pathogen diversity. The analyses demonstrate that weather variables (temperature and rainfall) have the potential to influence pathogen diversity. | INTRODUC TI ON Infectious diseases are a leading threat to both human health and the global economy (Johnson et al., 2015). In a database of 335 emerging infectious disease (EID) events (defined as those caused by newly evolved strains of pathogens such as multi-drug-resistant pathogens or pathogens that have recently entered human populations for the first time or pathogens which have recently increased in incidence) between 1940 and 2004, more than 60% had a zoonotic origin (note that zoonotic pathogens were defined as pathogens that originated in non-human animals) and events increased over time (Jones et al., 2008). Ovser time, we have seen the emergence of many novel pathogens such as severe acute respiratory syndrome coronavirus, H1N1 influenza, Ebola and Nipah virus (Morse et al., 2012;Murphy, 1998). Thirty new diseases including Legionnaires disease, acquired immune deficiency syndrome, hepatitis C, Nipah virus, Helicobacter pylori, severe acute respiratory syndrome, COVID-19, avian influenza, several viral haemorrhagic fevers and bovine spongiform encephalopathy/variant Creutzfeldt-Jakob disease have been reported in the last 50 years (Robin & Anthony, 2004). It is speculated that new infectious diseases could emerge anywhere across the globe at any time (Robin & Anthony, 2004), and the next human pandemic might have a zoonotic origin (Lancet, 2012) as evolution works on biodiversity to create novel pathogens. Anthropogenic disturbances may

play an important role in disease emergence and zoonoses. It has been argued that these disturbances primarily result from human incursions into pristine ecosystems and ecotones (Jones et al., 2013), and microorganisms often exploit such circumstances (Robin & Anthony, 2004); however, only limited information is available about the processes and actual facts of EIDs and zoonoses from wildlife globally. It is proposed that the interactions among ecological, biological and social processes in South East Asia enable microbes to exploit new ecological niches, leading to a high risk from emerging infectious diseases in this region (Coker et al., 2011). Rapid environment change and agricultural intensification have been described to play an important role in this phenomenon (Jones et al., 2013). On the contrary, it should also be noted that majority of the recurringly transmissible zoonotic pathogens (with an existing animal reservoir) come from domestic animals with rare occurrence from wildlife sources-except for a few specific diseases in which domiciliated animals such as rodents are the main reservoir hosts. Therefore, the demographic change in domestic animals might be an important factor associated with the emergence of zoonosis, and this needs to be investigated. Understanding the ecological interactions underlying emerging and zoonotic infectious diseases (Qi et al., 2020;Ward et al., 2020aWard et al., , 2020b is important for their mitigation and control (Johnson et al., 2015). Not much is known, and we need to look at all interactions and dynamics for EID and zoonosis to develop a more comprehensive contextual and specific understanding. Rapid increase in food trade and human travel is believed to be some of the main drivers of emergence of pathogens (Kilpatrick & Randolph, 2012). Modelling the dynamic factors underlying infectious diseases can assist the development of disease prevention and control strategies (Heesterbeek et al., 2015). Policymakers will benefit from the generated quantitative evidence for decision-making and public health policy formulation (Heesterbeek et al., 2015). To date, disease models to understand the factors responsible for pathogen diversity are limited. There are no mechanistic analyses that capture pathogen (human, zoonotic and emerging) diversity at the society-environment interface and their non-linear relationships. Predictive modelling using national data to inform country-specific control programmes is also lacking. Jones and colleagues (Jones et al., 2008), using a global database, analysed 335 EID events reported between 1940 and 2004. However, only the 'emerging infectious disease events' data were modelled; the existing non-linear relationships (between predictor and response variables) were not accounted for and a limited number of socioeconomic and environmental predictor variables were used for model building. Following this, the occurrence of 147 'emerging infectious disease events' for wildlife zoonoses (or just zoonotic origins) was modelled using demographic and environment variables (Allen et al., 2017). However, this study also had some disadvantages as the ongoing linkage of all the diseases with animals could not be established. Biogeographic grouping patterns of 187 human infectious diseases have also been described (Murray et al., 2015). Building on this previous research focussed on certain events/ diseases, we conducted a novel cross-sectoral assessment rigorously investigating the relative contributions of potential geodemographic, environmental and social factors on the diversity of human, zoonotic and emerging pathogens. Noteworthy among these factors are the driving forces fundamental for emergence of infectious diseases (Binder et al., 1999;Jones et al., 2008;Morens et al., 2010;Robin & Anthony, 2004). A deep understanding of pathogen ecology, as well as the complex inherent relationships at the agent-environment interface, is essential to inform disease control and mitigation and to predict the next pandemic zoonosis. Public health measures are important barriers to disease emergence and zoonoses (Lindgren et al., 2012;Mitchell, 2000). Therefore, the effects of healthcare facilities and expenditure on health, research and development on pathogen diversity were also explored. We would also like to stress and to make the reader aware of the various uncertainties and controversies surrounding the ontologies used in this and in previous studies conducted on this topic. As per the Joint WHO/FAO Expert Committee on Zoonoses, zoonoses are defined as 'those diseases and infections which are naturally transmitted between vertebrate animals and man (WHO/FAO, 1959)'. However, even the WHO uses different definitions of the term 'zoonosis'. For example, these are two other-and very different-definitions of zoonoses used by WHO: (a) 'A zoonosis is any disease or infection that is naturally transmissible from vertebrate animals to humans' and (b) that 'A zoonosis is an infectious disease that has jumped from a non-human animal to humans' (https://www.who.int/news-room/ fact-sheet s/detai l/zoonoses). This has created a difference in opinion regarding the infectious agents that should be considered as zoonotic pathogens. For example, some people believe that species-jumping or just zoonotic origin pathogens such as HIV or novel coronavirus (https://www.who.int/news-room/fact-sheet s/detai l/zoonoses) classified as zoonotic by the WHO should be considered non-zoonotic pathogens because these pathogens are currently not transmitted between non-human vertebrates and humans. In the current and other studies (Taylor et al., 2001;Woolhouse & Gowtage-Sequeria, 2005), the standard definition of the WHO (WHO/FAO, 1959) for infectious agent classification was used, whereas Jones et al. (2008) coined a different definition to categorize pathogens as zoonotic (i.e., pathogens that originated in non-human animals). Therefore, these modelling studies should be carefully interpreted and compared cautiously, due to differences in the ontologies used for zoonotic or emerging pathogens. Note that the results of our study might have been different if we had only considered the recurringly transmissible pathogens (requiring an animal reservoir for maintenance) between humans and non-human vertebrates as zoonotic pathogens. Lastly, the zoonotic or emerging pathogens in this study were classified as described in the methods. The isolation of these organisms from different hosts/countries only describes their zoonotic and emerging potential; some of these might not actually be emerging or zoonotic in these specific countries. | Microbe-country data set A species-location interactions data set developed by Wardeh et al. (2015) was used to determine viral, bacterial, parasitic and fungal microbes isolated from vertebrate hosts in different countries. The data set used in the analyses contained 13,892 unique microbe-country interactions (Appendix S1, p 130-408). Full details of microbe-country data set development are in the (Appendix S1, p 3). Note that there might be reporting bias due to factors such as difference in capacities in different countries and research focus oriented to prior decisions on risk. | Human, zoonotic and emerging agents Initially, we classified all microbes into human, zoonotic and emerging agents as reported by Taylor and colleagues (Taylor et al., 2001), and Woolhouse and Gowtage-Sequeria (Woolhouse & Gowtage-Sequeria, 2005). Zoonotic pathogens were classified as per the World Health Organization (WHO/FAO, 1959), as 'diseases and infections that are naturally transmitted between vertebrate animals and humans'. However, species such as HIV that are no longer transmitted between humans and animals were not considered zoonotic (Taylor et al., 2001). Emerging pathogens were defined as those 'that have appeared in a human population for the first time or have occurred previously but are increasing in incidence or expanding into areas where they have not previously been reported, usually over the last 20 years' (IOM, 1992). Human agents were defined as previously reported infectious human pathogens (Taylor et al., 2001) plus additional microbes reported to have a human host (Wardeh et al., 2015). The data set also included human/emerging/zoonotic agents that might not have been presented or reported in human hosts or have demonstrated their zoonotic/emerging potential in a given country but could infect humans and have been demonstrated to be zoonotic/emerging in some other parts of the world. This data set also had information about the microbe type (i.e., whether they were viruses, bacteria, parasites or fungi). | Country-specific parameters We examined published and official data and compiled information for 49 country-specific geodemographic, social, trade and environmental factors (Appendix S1, pp 4-5). | Pathogen-country interactions and countryspecific parameters The unique pathogen-country interactions data were merged with the country-specific variables to construct the analytical data set. The final data set had information about country-specific parameters and the numbers of human, emerging, zoonotic and total pathogens isolated from different countries (Appendix S1, pp 6-46). | Missing data Initially, we compiled information for 224 countries for 49 variables, but because of a lack of available information the following countries were excluded: Anguilla; Antarctica; Aruba; Bonaire, Sint Eustatius Islands. The final data set (for the initial modelling) contained information for 190 countries (Appendix S1, pp 6-46). | Predictors and outcome We used the country-specific parameters (Appendix S1, pp 4-5) as key predictors. Three host-pathogen-associated outcomes were explored: zoonotic emerging and human pathogen diversity at the country level, where the zoonotic/emerging/human pathogen diversity was defined as the total number of different zoonotic/emerging/human pathogens reported from any given country. | Statistical analysis Descriptive analyses conducted included the creation of histograms and scatter plots of predictors with outcomes. Data were tested for the assumptions of linearity and normality. A variable was log-transformed if the assumptions of normality were not met (Appendix S1, p 47). We also detected non-error, non-representative outliers (values >1.5-fold interquartile range) in some predictor variables in the data (original or log-transformed). These outliers were treated using winsorization, and all further analyses were conducted both using the original and the winsorized data set. | Generalized additive model (GAM) We built three separate models for the response variables of zoonotic, emerging and human pathogen diversity using both the original and winsorized data. Because of non-linear associations between some predictors and the outcomes, we compared additive models using non-linear splines with linear models and retained the model in which the data were a significantly better fit and had a lower Akaike information criterion value. A correlation matrix was then generated (separately for data with outliers and winsorized data) among the predictor variables to test for collinearity. Initially, only one predictor was retained for the multivariable models from among the predictors having a Pearson correlation coefficient >0.90 (Appendix S1, pp 48-49). From these variables, final multivariable models were constructed. We followed a forward stepwise approach and retained variables with p < .001, followed by retesting of all the variables with p < .25 and all nonsignificant variables. A final multivariable model was also re-tested by individually replacing a non-linear spline variable with a linear variable and retaining the model in which the data fitted significantly better and had a lower Akaike information criterion value. Statistical analyses were conducted in the R statistical program (R statistical package version 3·4·0, R Development Core Team [2015], http://www.r-proje ct.org). We used the parametric function 'linear' and non-parametric function 's spline (thin plate regression spline fit)' to fit the predictor variables. The function 'ti spline' was used to determine significant interactions among the combined parametric and non-parametric response variables. The GAM model fitting separated the linear trend of predictor variables from any other non-parametric association to determine significance of smoothing variables with non-linear patterns. The final models were examined for concurvity (a non-parametric analogue of multicollinearity), and model diagnostics were also examined. | Controlling for research effort To control for research effort, we included gross domestic expenditure on research and development (GERD in 000s US dollars) of countries into all the analytic models. All the non-significant variables were further re-tested in the models. | Prediction of the expected response The GAM models were used to predict the expected response variable using all the available data. In brief, we used data from all 224 countries (for predictions; Appendix S1, pp 50-71), wherever available for all the significant predictors reported in the GAM models. | Generalized additive modelling In total, all the predictor variables included in the GAM explained 74.8%-87.2% of the deviance in the different models. | Zoonotic pathogen diversity The logarithm of land area, human development index, logarithm of human population density, logarithm of mean annual temperature (average), logarithm of exports percentage of gross domestic product (GDP) and forest area percentage were associated with zoonotic pathogen diversity Table 1, Figure 1). There were signifi- Additional results, including the analysis of winsorized data, are presented in the (Appendix S1, pp 78-87). | Emerging pathogen diversity The GAM showed that emerging pathogen diversity is associated with the logarithm of human population density, agriculture land percentage, logarithm of land area, human development index value, national biodiversity index, longitude and the logarithm of sheep population (Table 2, Figure 3). After controlling for research effort, emerging pathogen diversity was associated with the logarithm

of human population density, logarithm of land area, human development index value and the logarithm of mean annual temperature average (Table 2, Figure 4). Further, significant interaction effects among the human development index value and the logarithm of land area, longitude, and the logarithm of human population density, human development index value, and the logarithm of mean annual temperature average, logarithm of mean annual temperature average and death rate per 1,000 people are reported (Figure 4). Additional results, including analysis of the winsorized data, are presented in the (Appendix S1, pp 88-106). | Human pathogen diversity The GAM showed a linear relationship of human pathogen diversity with the logarithm of human population density and logarithm of mean annual temperature average. Human pathogen diversity was also associated with the human development index value and national rainfall index (Table 3, Figure 5) Additional results, including analysis of the winsorized data, are presented in the (Appendix S1, pp 107-121). | Prediction of the expected response The predictive values of the response variable for all the GAM models are shown in the (Appendix S1, pp 108-112). A global richness map of pathogen diversity in different countries is shown in Figure 7. The analysis indicates high pathogen diversity in many countries of Latin and North America, Asia, Australia and Europe (Appendix S1, pp 122-126). | D ISCUSS I ON We conclude that geodemographic, environmental and social factors to be significant predictors of zoonotic/emerging/human pathogen diversity. We believe that this analysis will help our understanding of the diversity of human, zoonotic and emerging pathogens as well as the associated global risk factors. The data generated are valuable TA B L E 2 Generalized additive models of emerging pathogen diversity (outcome-log number of emerging pathogens) reported from different countries Significant terms and interactions are presented in Figure 6. TA B L E 3 Generalized additive models of human pathogen diversity (outcomelog number of human pathogens) reported from different countries for similar investigations in the future. Overall, we found a large range in the number of human, zoonotic and emerging pathogens reported from different countries. Lack of uniformity is likely due to differences in environmental, social and geographical factors at the country level, as well as potential differences in reporting. We used national level data in this study, so any country could use these models after collecting the desired data for their own policy development and to inform their control or mitigation actions. However, these models do not account for subnational variations; therefore, the results should be carefully interpreted for countries that have large subnational differences in geodemography, environment and societal characteristics. We also controlled for research effort to produce unbiased results; however, a certain level of uncertainty should be expected in the methods and outcomes. We believe gross domestic expenditure on research and development (GERD) per country to be a more appropriate proxy for controlling for research effort compared to a previously reported study (Jones et al., 2008) that accounted for these biases by quantifying reporting effort in an international journal (JID). The latter is biased; for example, scientists from non-English speaking countries are likely to have published their research in their native language journals. Many countries or institutes might have not published their research. GERD is a less biased representation of a country's research effort because laboratory support is more likely to translate into the number of pathogens reported from that country. GAM models indicated higher number of emerging pathogens (uncontrolled research effort) for countries in the longitude range of −50.0° to 0° and ≥150°. After controlling for research effort, countries in a longitude range of −100° to −30.0° (North and South American countries) and having high human population density were found to report a higher number of zoonotic pathogens. Although only marginally significant, higher number of emerging pathogens were noted in the lower latitudes. The potential effect of proximity to the equator and hemisphere needs to be further investigated. Certain variations were recorded in the analysis of winsorized data. For example, additional significant predictors-such as log mean annual temperature average, HDI value and longitude-were also associated with zoonotic pathogen diversity (after controlling for research effort); and the variable forest area percentage became non-significantly associated with zoonotic pathogen diversity (after controlling for research effort). After controlling for research effort, the predictors for emerging pathogen diversity (original and winsorized data) were similar except that the log health expenditure per cent GDP was additionally associated with emerging pathogen diversity (winsorized data). We accounted for non-linear relationships and used 49 socioeconomic and environmental variables during model development. By using pathogen biodiversity rather than EID events as our response variable, we included unknown or future disease emergence. We found human population density, land area, mean annual temperature (average) and human development index value to be associated with the overall diversity of human, emerging and zoonotic pathogens, and this association is not limited to EID events as reported previously (Jones et al., 2008). The national rainfall index was a significant predictor in the models of human and zoonotic pathogen diversity, but had no role for emerging pathogen diversity. Our analyses indicate that many socioeconomic and environmental factors are equally important for zoonotic and human pathogen diversity, as reported for disease emergence. Our analysis shows the national biodiversity index as a significant predictor in emerging pathogen diversity models, as for EID events reported by Jones and colleagues (Jones et al., 2008). We also report forest area percentage to be associated with zoonotic pathogen diversity. Forests are home to terrestrial animal biodiversity (Brockerhoff et al., 2017), and deforestation and forest incursions have been linked to zoonosis (Wolfe et al., 2005). We found that the diversity of emerging human pathogens was additionally correlated with longitude and death rate per 1,000 people. This needs to be further explored. However, we believe that availability of our data assist the countries where no predictor/response data are currently available to develop so that the desired information is available for analysis using our models. This will help overcome the limitations associated with incomplete data in the future. Overall, social and environmental factors and geography are significantly associated with global pathogen diversity. Finally, our analyses demonstrate that weather variables (temperature and rainfall) have the potential to influence pathogen diversity. Further research is required to assess the long-term impact of these variables. Similarly, the impact of climate change on pathogen diversity is a topic that needs to be researched. We believe future models based on simultaneous testing of host, agent and environment characteristics for prediction will shed more light on disease emergence and zoonoses. We conclude that weather variables, as well as forest and biodiversity conservation, have the potential to influence human, zoonotic and emerging pathogen diversity in the near future. ACK N OWLED G EM ENTS The authors thank the Department of Education and Training, Australian Government for providing the 2018 Endeavour Research Fellowship to the primary author to conduct this research. CO N FLI C T O F I NTE R E S T The authors declare no conflicts of interest. E TH I C A L A PPROVA L Informed consent for collection of epidemiological data was not required, as these data were already coded and available in the public domain. No identifiable personal information was used in this study. DATA AVA I L A B I L I T Y S TAT E M E N T The analysed data are available along with the manuscript. Sources of the raw data used in the analysis have been cited. Fermentation Enhanced Biotransformation of Compounds in the Kernel of Chrysophyllum albidum Chrysophyllum albidum Linn (African star apple) is a fruit with extensive nutritional and medicinal benefits. The fruit and kernel in the seed are both edible. Strains of lactic acid bacteria (LAB) were isolated from fermented seeds and assessed for probiotic characteristics. The extracts in both the unfermented and the fermented aqueous extracts from the kernels obtained from the seeds of C. albidum were subjected to analysis using the gas chromatography/mass spectrometry (GC-MS) method. This analysis identified the bioactive compounds present as possible substrate(s) for the associated organisms inducing the fermentation and the resultant biotransformed products formed. Three potential probiotic LAB strains identified as Lactococcus raffinolactis (ProbtA1), Lactococcus lactis (ProbtA2a), and Pediococcus pentosaceus (ProbtA2b) were isolated from the fermented C. albidum seeds. All strains were non hemolytic, which indicated their safety, Probt (A1, A2a, and A2b) grew in an acidic environment (pH 3.5) during the 48-h incubation time, and all three strains grew in 1% bile, and exhibited good hydrophobicity and auto-aggregation properties. Mucin binding proteins was not detected in any strain, and bile salt hydrolase was detected in all the strains. l-lactic acid (28.57%), norharman (5.07%), formyl 7E-hexadecenoate (1.73%), and indole (1.51%) were the four major constituents of the fermented kernel of the C. albidum, while 2,5-dimethylpyrazine (C1, 1.27%), 3,5-dihydroxy-6-methyl-2,3-dihydropyran-4-one (C2, 2.90%), indole (C3, 1.31%), norharman (C4, 3.01%), and methyl petroselinate (C5, 4.33%) were the five major constituents of the unfermented kernels. The isolated LAB are safe for consumption. The fermenting process metabolized C1, C2, and C5, which are possible starter cultures for the growth of probiotics. Fermentation is an essential tool for bioengineering molecules in foods into safe and health beneficial products. Introduction The seeds of plants are good sources of food for humans and animals because of their wealth of nutrients. However, not all seeds are edible. These non-edible seeds may have the potential to be transformed into useful purpose(s). Efforts are on the increase to reveal their potential and the benefits of some of these non-edible seeds since they may be transformed into useful purpose(s). Fermentation also transforms the physical properties of the seeds, leading to other products from the process, while at the same time transforming the natural phytochemicals. This is evident by the change in color, texture, and some other physical properties in products (liquid pap from cereals like sorghum, Additionally, natural fermentation provides an opportunity for a better overview of the biochemical and microbial activities during the domestic preparation of fermented foods. On the other hand, the limitation of this natural or spontaneous type of fermentation is the high risk of spoiling the microbial communities of foods, and the domination of microbial strains that could be dangerous for human health [20]. The microbes that confer beneficial health effects to their host when administered in adequate amounts are the probiotics [21]. Notable among the probiotics are the lactic acid bacteria that have been commonly utilized for several decades in the food industry [22]. Efforts are on the increase to isolate them from different sources and commercialize these lactic acid bacteria that are a dominant fermenting microorganism involved in the spontaneous fermentation of foods. The processing of the kernels of C. albidum with the aid of fermentation is yet to be reported. In this study, we employed gas chromatography-mass spectrometry (GC-MS) analysis to analyze the phytochemical compounds present in the raw and the fermented kernel of C. albidum. These efforts should further provide understanding of the nature of the biotransformation of such identified molecules during the fermentation, and the possible health benefits of the product of the fermentation. Identification of the Isolated Organisms Facilitating the Fermentation Three (3) lactic acid bacteria (LAB) isolates were obtained from the fermented C. albidum seeds. The isolates were named as follows: ProbtA1, ProbtA2a, and ProbtA2b. The three isolates were Gram-positive and catalase-negative. From the result of the biochemical identification, Probt (A1, A2a, and A2b) were identified as Lactococcus raffinolactis, Lactococcus lactis, and Pediococcus pentosaceus respectively ( Figure 1). Molecules 2020, 25 , x 3 of 14 other hand, the limitation of this natural or spontaneous type of fermentation is the high risk of spoiling the microbial communities of foods, and the domination of microbial strains that could be dangerous for human health [20]. The microbes that confer beneficial health effects to their host when administered in adequate amounts are the probiotics [21]. Notable among the probiotics are the lactic acid bacteria that have been commonly utilized for several decades in the food industry [22]. Efforts are on the increase to isolate them from different sources and commercialize these lactic acid bacteria that are a dominant fermenting microorganism involved in the spontaneous fermentation of foods. The processing of the kernels of C. albidum with the aid of

fermentation is yet to be reported. In this study, we employed gas chromatography-mass spectrometry (GC-MS) analysis to analyze the phytochemical compounds present in the raw and the fermented kernel of C. albidum. These efforts should further provide understanding of the nature of the biotransformation of such identified molecules during the fermentation, and the possible health benefits of the product of the fermentation. Identification of the Isolated Organisms Facilitating the Fermentation Three (3) lactic acid bacteria (LAB) isolates were obtained from the fermented C. albidum seeds. The isolates were named as follows: ProbtA1, ProbtA2a, and ProbtA2b. The three isolates were Grampositive and catalase-negative. From the result of the biochemical identification, Probt (A1, A2a, and A2b) were identified as Lactococcus raffinolactis, Lactococcus lactis, and Pediococcus pentosaceus respectively ( Figure 1). Identification of the Isolated Organisms Facilitating the Fermentation Gram-positive and catalase-negative reactions are preliminary ways to identify LAB. LAB has been reported to be Gram-positive and catalase-negative (Ismail, Yulvizar & Mazhitov, 2018). In the biochemical profiling of the isolates with the API 50 CHL system (for the identification of lactic acid bacteria) shown in Table 3, the positive results obtained for each of the forty-nine (49) substrates (sugars) demonstrated the ability of the LAB isolates to utilize the enzymatic activities, related to the fermentation of the carbohydrates. LAB is known to produce lactic acid as the end result of carbohydrate fermentation. Therefore, the lactic acid produced caused a drop in pH, resulting in the positive color change for any of the fermented sugar [23]. Identification of the Isolated Organisms Facilitating the Fermentation Gram-positive and catalase-negative reactions are preliminary ways to identify LAB. LAB has been reported to be Gram-positive and catalase-negative (Ismail, Yulvizar & Mazhitov, 2018). In the biochemical profiling of the isolates with the API 50 CHL system (for the identification of lactic acid bacteria) shown in Table 3, the positive results obtained for each of the forty-nine (49) substrates (sugars) demonstrated the ability of the LAB isolates to utilize the enzymatic activities, related to the fermentation of the carbohydrates. LAB is known to produce lactic acid as the end result of carbohydrate fermentation. Therefore, the lactic acid produced caused a drop in pH, resulting in the positive color change for any of the fermented sugar [23]. Table 3. The sugar metabolism pattern of the isolated lactic acid bacteria (LAB). Biotransformation of Molecules by the Fermenting Organisms This fermentation process did not change the indole and the norharman ( Figure 2). This finding may imply that the functional groups of these compounds are resistant to the change in acidic conditions that mostly led to phytochemical modifications during the fermentation process. The fermenting process metabolized C1, C2, and C5 ( Figure 2). This study suggests that the molecules were consumed by the supporting probiotics during the fermentation of kernels from seeds of C. albidum ( Figure 2). Hence, the three molecules are possible starter cultures for the growth of probiotics. The fermentation of the C. albidium also created two new molecules (Figure 2). These molecules are suspected to be the by-products of the metabolism due to the probiotics facilitating the fermentation [13]. This finding is in agreement with the reports of Goswam et al. [17] and Rezac et al. [18] stating that compounds of interest with more bioavailability can be obtained from existing foods depending on the kind of fermentation used for the processing. 2,5-Dimethylpyrazine (C1), 3,5-dihydroxy-6-methyl-2,3-dihydro-4H-pyran-4-one (C2), indole (C3), norharman (C4), and the methyl petroselinate (C5) were the predominant molecules in the unfermented C. albidum. Both identified metabolites (C1 and C2) in the unfermented C. albidum seeds are associated with the Maillard reaction and are of great benefit to the food industries due to their pleasant sensory qualities [24,25]. Both C1 and C2 are present in cakes, Tartary buckwheat tea, toasted oak wood, garlic oil, cane brown sugar, cocoa, and cocoa containing products [26][27][28][29][30]. Norharman (C4) is present in foods. It possesses a cytotoxic activity to cancer and Parkinson's diseases. It is useful in controlling xenobiotics and is not mutagenic. It has antimicrobial properties against Staphylococcus aureus, Bacillus subtilis, and the plant pathogen Agrobacterium tumefaciens [31,32]. This compound is in normal body constituents formed endogenously [32,33]. Norharman, which is endogenously derived from the pyrolysis of proteins and amino acids, is structurally related to the nonpolar heterocyclic aromatic amines. It consists basically of pyrrole, benzene, and pyridine rings [34]. The presence of ethanol, amines, amino acids, and acetaldehyde often influences the formation of norharman. Norharman is formed endogenously by cyclization of tryptamine with a carbon unit transferred from 5-methyltetrahydrofolate [35]. Indole (C3) is naturally derived from norharman depending on the level of acidity of the medium or food [36][37][38]. Ethanol (I-2) is a primary product of alcoholic fermentation [39]. The production of a proposed intermediate (I-4) for this fermentation was also previously implicated as a product of fermentation [40]. 3-Hydroxypropanal (I-5), otherwise known as lactaldehyde, is naturally produced from glycerol and is predominant in all living organisms including humans and bacteria [41]. The molecule (I-5) links with lactic acid production during fermentation [42]. The emergence of lactic acids as a product of this fermentation reflects the type of fermentation that occurred. Lactic acid is associated with anaerobic fermentation [43]. An intermediate (I-3) in this study was one of the several derivatives of butanoic acid. The butanoic acid and the associated derivative were intermediates for fermentation [44][45][46]. This particular derivative (I-3) of butanoic acid was linked to fermentation for the first time in this study. The formation of reduced compounds such as lactic acid, ethanol, and butanol, among others, is evidence of the occurrence of metabolic pathways shifting from acidgenesis to solventogenesis [43]. One intermediate (I-1), a primary α-hydroxy ketone with butan-2-one substituted by a hydroxy group at positions 1 and 3 or a secondary α-hydroxy ketone, has a close structural analogue of 3,4-dihydroxybutan-2-one, whose phosphorylated form (l-3,4-dihydroxybutan-2-one 4-phosphate) is a precursor for the synthesis of riboflavin [47]. Formyl 7E-hexadecenoate, an effective nitrogen removal stimulant found in aquatic duckweed, was reported in lactic acid bacteria enhanced fermented milk [48,49]. Methyl petroselinate (C5) was also previously detected as fermented products of Korean red pepper and the extract of miswak plant, Salvadora persica L [50,51]. Although the study most probably provides for the first time the transformation of methyl petroselinate (C5) to formyl 7E-hexadecenoate (P2) during fermentation ( Figure 2). This compound (P2), a component of the hexadecenoates derivatives identified in natural products, is a precursor for oleic acid biosynthesis [52,53]. This finding substantiates the report of Olawole, Okundigie, Rotimi, Okwumabua, and Afolabi [15], stating that fermented processed foods are chemically modified and nutritionally different from their raw sources. As stated earlier, fermentation is a metabolic reaction resulting from the activities of microorganisms that serve as agents for bio-transforming a particular compound to another. Fortunately, most of these biotransformations convert toxic compounds to safe and biologically active molecules [17,18,54]. This study gives credence to the potential of natural fermentation in improving the unique quality of agro-food crops, and the campaign targeting the detoxification of foods facilitated by probiotic lactic acid bacteria (LAB). Notably, lactic acid was the primary compound discovered in the fermented extract. The production of lactic acid by lactic acid bacteria during fermentation has been attributed to be one of the reasons for the increased preservation of fermented foods. The lactic acid produced along with other organic acids during fermentation has been confirmed to inhibit the growth of food spoilage microorganisms [14,55]. This ability to hinder the growth of microorganisms is a reason why lactic acid bacteria are majorly used as probiotics. The lactic acid, thus detected in our fermented extract ( Table 2), also suggests the presence of lactic acid bacteria during the fermentation, which could be isolated and further used as probiotics. The presence of lactic acid bacteria in most fermented retailed foods was previously reported, making them excellent sources of live lactic acid bacteria including probiotics that provide health benefits to humans [18]. The compounds that were eliminated in our fermented extract are likely substrates for the growth of these probiotic organisms. This finding further substantiates the fermentation as a bioengineering tool to remove undesired compounds from harmful food crops as well as to modify their sensory properties such as taste, appearance, texture, and aroma that is derived from consuming the food product for safe consumption [13,15,17]. These changes in the physicochemical properties of such fermented foods could have been due to the possible growth of the three isolated LAB (Lactococcus raffinolactis, Lactococcus lactis, and Pediococcus pentosaceus). The pyrazine compounds are receiving increasing attention because they possess unique organoleptic properties. This increasing attention has led to the discovery of several pyrazine compounds in foods. Several others have also been extracted or artificially synthesized, and promoted as organoleptic and flavor enhancing agents [17,58]. The 2,5-dimethylpyrazine was detected in both the fermented and the unfermented seed. The presence of these pyrazine derivatives could be due to Maillard reactions in the seed [59,60]. Compounds in these groups (pyrazines and pyridine) are known to give that characteristic pleasant smell of peanut. Plant Collection and Identification The wholesome and fresh C. albidum fruits were procured from a local market commonly called "Oja Ota" in Ota, Ogun State, Nigeria (6.6841 • N, 3.2153 • E). The plant (Voucher number: FHI 112796) was identified and authenticated by the Forest Research Institute of Nigeria (FRIN), Ibadan, Nigeria. Additionally, an ethical permit (CHREC/49/2020) was obtained from the Covenant Health Research Institute, Canaanland, Ota, Nigeria, for the research to proceed. Preparation of Flour from Kernel The seeds were dehulled manually to remove the kernels of the fruits. The kernels were dried with the aid of an oven at 40 • C for about 24 h until it was crunchy and crispy enough to be ground into a fine powder. The dried kernels obtained were afterward ground into a fine powder with the aid of an electronic homogenizer. A powder extract was collected from the kernels of both the fermented and the raw-unfermented seeds. The extract from the raw-unfermented seeds was to serve as the control facilitating the identification of the uniquely metabolized bioactives that may be due to the fermentation process. Production of Both the Unfermented and the Fermented Aqueous Extracts The ground powder from the kernels was steeped in distilled water (1:3 w/v), stirred together, and allowed to stand for 24 h, under ambient temperature (26-33 • C). After 24 h, the solution was filtered with the use of filter paper (Whatman no. 1). The filtrate obtained was divided into two portions (A and B). The first portion (A) was subjected to fermentation before analysis; and the second portion (B) was directly used for further investigation of the raw-unfermented sample. The first portion (A) of the filtrate obtained was transferred into an airtight container and was left to naturally ferment anaerobically for 72 h under ambient temperature (26-33 • C). The second portion (B) of the filtrate obtained was immediately transferred into the freezer (−20 • C) to facilitate the freezing of the sample and to prevent it from undergoing fermentation. Both filtrates (A and B) were then removed from the freezer and transferred into a Telstar Cryodos-50 freeze dryer for the sublimation of the aqueous solvent until the dried extract remained. The dried extract from filtrates A and B were then labelled as the fermented aqueous extract and the unfermented aqueous extract, respectively. Isolation of Lactic Acid Bacteria from the Fermented C. albidum Seeds An aliquot of 20 µL of the fermented C. albidum solution was transferred with the aid of a sterile nichome inoculating loop into sterile de Man Rogosa Sharpe (MRS) agar (Oxoid medium) plates previously prepared according to the manufacturer's instruction. This was thereafter streaked to observe the growth of distinct bacteria colonies. The plates were incubated microaerophilically using microaerophilic gas packs (Thermo scientific ltd., Basingstoke, UK) at 37 • C for 48 h. After incubation, single isolated colonies were picked with a sterile inoculating loop and cultured by repeated streaking on sterile MRS agar plates until pure cultures of the isolates were obtained. Gram Staining and Catalase Test Isolates were sub-cultured on sterile MRS agar plates and incubated under microaerophilic conditions at 37 • C for 24 h. The young cultures were smeared on slides and Gram-stained to observe for shape and

Gram-reaction. For the catalase test, the isolates were subjected to a reaction with hydrogen peroxide (H 2 O 2 ); each isolate was picked and placed on a slide, and a drop of hydrogen peroxide (H 2 O 2 ) solution (20 µL) was placed on each of them to observe for the formation of bubbles, where the bubble formation indicated a positive catalase reaction, while no bubble formation indicated a negative reaction. Biochemical Characterization of the Isolated Organisms To identify the organisms, an Analytical Profile Index (API) 50 CHL Kit (Biomereux Ltd., Marcy-l'Etoile, France) for the identification of lactic acid bacteria was used to obtain the biochemical profile of the isolates; this was done according to the manufacturer's instructions. The freshly sub-cultured 24 h LAB isolates were added into the API 50CHL basal medium and filled into wells on the API 50 CHL test strips. The strips were afterward incubated at 37 • C for 48 h and examined visually after 24 and 48 h of incubation. The indication of a positive or negative result was examined from the color change from a scale of 1 to 5, that is, from purple (1) to green (3) to yellow (5) at the 48 h-mark, and a value of >3 indicated a positive result. The results were afterward cross-referenced with the API ® databases with the use of the APIweb™ and based on the carbohydrate metabolism patterns obtained from the experiment, the names of the isolates were identified from the database. Sample Extract Concentration for Gas Chromatography-Mass Spectrometry (GC-MS) Analysis The remaining dried extracts (5 g) were re-dissolved in 25 mL of ethanol and was after that used for the GC-MS analysis. Procedure for GC-MS Analysis Shimadzu GC-MS (Model-QP2010SE, Osaka, Japan) equipped with a mass spectrometry detector was used for this experiment. The detector was set at an equilibrium time of 1.0 min, with an ion source and interface temperatures of 230 • C and 250 • C, respectively. The detector gain of 1.28 kV + 0.00 kV and a threshold of 2200 were attained during the experiment. The sample (1.0 µL) was injected at 250 • C using a splitless mode into the GC-MS equipped with a pre-conditioned Optima 5 MS capillary column (30 × 0.25 mm) with 0.25 µL film thickness at 60 • C, and the sampling time was also set for 2.00 min. The flow control mode of the sample was set at linear velocity, and a pressure of 144.4 kPa was used to derive a total flow of 22.6 mL/min. The flow within the column used was 3.22 mL/min at a linear velocity of 46.3 cm/s. The purge flow of 3.0 mL/min and a split ratio of 5.1 was used. The oven temperature was programmed for 60, 120, and 290 • C at a holding time of 2, 2, and 3 min, respectively, while the flow rate of 15 was attained during the last two ovens' programmed conditions. The NIST Mass Spectral Search Program for the NIST/EPA/NIH Mass Spectral Library (version 2.0 g, Gaithersburg, MD, USA) was used to perform the tentative identification. The common or scientific name of the identified compounds were obtained from the PubChem, Human metabolome, or yeast metabolome databases. The compounds detected by the tentative identification, with the percentage of coincidence (≥85%) were considered for the purpose of this study. Hence, all the compounds with either % area that was <1%, or with similarity index (≤85%) were manually excluded from the chromatographs (Tables 2 and 3) since the GC-MS was in scan mode during the analysis. The proposed chemical reaction mechanisms were drawn with ChemAxon MarvinSketch software (15.9.14, Budapest, Hungary). Conclusions This study created two new bioactive compounds with immense therapeutic uses, the potential to enhance aroma, and the sweet taste of foods with the aid of natural fermentation of the kernel from the seeds of the C. albidum. Three possible phytochemicals with possible potential to support the growth of probiotics were identified. Lactococcus raffinolactis, Lactococcus lactis, and Pediococcus pentosaceus are the isolated health beneficial lactic acid bacteria promoting the fermentation of the seeds of the C. albidum for their probiotic characteristics. These strains of probiotics identified during this fermentation could, in the future, serve as a starter culture for the growth of lactic acid bacteria in the food industry [19]. The need to better identify the specific role of the individual identified phytochemicals serving as a possible starter culture for the growth of lactic acid bacteria should be critically studied. It is also suggested to identify the specific fermentation organism that metabolized each of the consumed phytochemicals and to link such metabolism to a specific phytochemical product(s) identified in this study. This understanding should help to biosynthesize these compounds in the future. The application of LAB isolates is recommended for incorporation into functional food products. Supplementary Materials: The following are available online, Figure S1: Chromatogram generated during the GC-MS analysis of the unfermented aqueous extract of C. albidum, Figure S2: Chromatogram generated during the GC-MS analysis of the fermented aqueous extract of C. albidum, Table S1: Complete compounds present in the GC-MS analysis of the unfermented aqueous extract of C. albidum, Table S2: Complete compounds present in the GC-MS analysis of the aqueous extract of fermented C. albidum. Funding: We are highly appreciative of the funding received for the publication of this article by Covenant University, Ota, Nigeria. Nephrologists between power and vulnerability in times of technology ABSTRACT The doctor-patient relationship is often discussed from the perspective of patient vulnerability. Little attention is given to the vulnerability of nephrologists in their professional practice, a reality often affected by profound cultural transformation arising from technological development. Nephrology is based on research and procedure instrumentalization, both permeated with technology. In addition, the relationship between nephrologists and institutions is governed by market rules. Recent data showed a shortage of new nephrologists and the need to improve the technical training of new professionals, foster the establishment of interventional nephrology, and attract more graduating physicians to this medical specialty. Bioethics offers a different perspective on the issue, since it takes the subjective concerns of medical doctors and the social environments they participate in into consideration in order to enhance their ethical autonomy. These ideas may be discussed as part of undergraduate or specialization programs, thus reinforcing the acknowledgement of vulnerability as a condition and of the relevance of adopting a reflective attitude toward the events of everyday life that interact with the morality of nephrologists, so that risks are adequately faced having bioethical parameters as a reference. The doctor-patient relationship is often discussed from the perspective of patient vulnerability. Little attention is given to the vulnerability of nephrologists in their professional practice, a reality often affected by profound cultural transformation arising from technological development. Nephrology is based on research and procedure instrumentalization, both permeated with technology. In addition, the relationship between nephrologists and institutions is governed by market rules. Recent data showed a shortage of new nephrologists and the need to improve the technical training of new professionals, foster the establishment of interventional nephrology, and attract more graduating physicians to this medical specialty. Bioethics offers a different perspective on the issue, since it takes the subjective concerns of medical doctors and the social environments they participate in into consideration in order to enhance their ethical autonomy. These ideas may be discussed as part of undergraduate or specialization programs, thus reinforcing the acknowledgement of vulnerability as a condition and of the relevance of adopting a reflective attitude toward the events of everyday life that interact with the morality of nephrologists, so that risks are adequately faced having bioethical parameters as a reference. IntRoductIon Discussions about the professional lives of nephrologists generally focus on patient vulnerability. Attention is rarely devoted to debating the vulnerability nephrologists are faced with. This ethical-reflective study, therefore, aims to explore the latter type of vulnerability as it relates to the cultural transformation introduced by contemporary scientific development. After a few words on transformation, the circumstances interacting with the morality of nephrologists will be identified along with what it takes to tackle the resulting vulnerabilities. Medicine in a cOntext Of scientific transfOrMatiOn The changes occurred in the way physicians think are explained by the significant progress medical science has made in the study of the human body and the interplay between myriad factors. On the background are the great transformations engendered since the Renaissance and throughout the eighteenth century, with the formalization of the signs and symptoms of diseases combined with an anatomopathological view of ailments, whose implications are seen to this day 1 -particularly the idea of what constitutes disease and health, achievements such as immunization, and a fresh perspective on mental health, all of which helped give birth to the clinical rationale in effect today, based on the interpretation of signs exhibited by altered bodily structures to pinpoint the genesis of diseases. In addition to the ideas physicians had of health and disease and their corresponding bodily expressions, scientific knowledge has also revolutionized the tools used in medical practice. Two types of tools deserve special attention: the ones used to diagnose patients and the ones used to supplement clinical practice. In medical schools, teaching and practice were changed and molded around reductionism, an approach based on the premise that diseases linked to specific organs had to be diagnosed and treated. Based on this idea, the larger areas of medical practice (General Practice, Surgery, Pediatrics, Obstetrics and Gynecology) were subdivided into specialties. 2 The affirmation of subjective autonomy stemmed from a long process triggered by the revolution of scientific knowledge. The ideals proclaimed in the French Revolution (1789) planted the seed of citizen autonomy and laid the ground for patient autonomy in the context of healthcare -still a strong idea today. Interestingly, when physicians were empowered with new diagnostic tools and clinical resources, the power and influence they once had over their patients vanished, bringing medical paternalism to an end. In addition, their ability to diagnose patients becomes increasingly overshadowed by diagnostic equipment. These events explain why medical work changed profoundly in the twentieth century. Primarily after World War II, being a doctor became a less liberal job and a more technological one -a profession subordinated to the fast pace of scientific progress. Physicians once enjoyed a significant amount of autonomy, as they saw patients in their offices and had control over their schedules. They could choose to work in hospitals and discuss their fees directly with patients. But decision-making shifted predominantly to the hands of healthcare institutions -hospitals, clinics, outpatient units -and physicians had multiple employment ties in which employer profitability was a primary concern. In addition, the relationship between patients and physicians began to be mediated by other healthcare workers, health management organizations, hospital insurance companies, pharmaceutical companies, and the Government. In other words, medical autonomy was gradually diluted in multiple institutional partnerships and weakened by a brand of technological autonomy. Left with no choice, physicians are inevitably required to accept technology and seek new ways of feeling sufficiently autonomous to make decisions and attain acknowledgement. 3 The doctor-patient relationship currently unfolds predominantly in hospital settings -in wards, emergency and intensive care units -in which patients with acute manifestations seek care. Once their issues are resolved, the patients are discharged and transferred to outpatient follow-up, usually to be seen by other professionals and thus impeding the creation of a bond between patients and their doctors. This model contrasts with the experience of patients seen in their doctor's office, in which the two are tied for an undetermined length of time and treatment does not result in cure, as diseases are generally of a chronic nature. the iMpact Of transfOrMatiOn On nephrOlOgy Ever since it was established as a medical specialty in the early 1960s, nephrology has been strongly linked with research and renal replacement therapy (dialysis and kidney transplantation), both examples of the impact science and technology have had in medicine. Dialysis enabled the treatment of patients with acute and chronic kidney disease, a fact that attracted many medical doctors to nephrology for the possibility it offered of applying knowledge generated in basic research to clinical practice and for its uses in diagnosis and treatment. Besides, the related procedures were handled exclusively by nephrologists, namely hemodialysis or peritoneal dialysis;

creation of a vascular or peritoneal access for dialysis; percutaneous kidney biopsy for diagnostic purposes and transplant recipient follow-up. Pioneering nephrologists performed all such procedures, but as complexity and the number of patients grew -chiefly among the elderly and individuals with diabetes -some were reassigned to physicians from other specialties. Radiologists, vascular surgeons and others became essential elements in the work performed by nephrologists. However, nephrology procedures slowly lost their appeal over the years, as technical progress in other areas spurred greater scientific interest and superior income. Renal replacement therapy is deemed a high-cost medical procedure on account of the massive use of technologically sophisticated equipment and consumables. Primarily the Government assumed the payment for dialysis procedures, which resulted in a significant allocation of funds to dialysis clinics. Nephrologists have had an active role in the economic-administrative aspects of this environment, with varying degrees of involvement: some opted to remain distant from medical procedures to yet receive significant financial compensation, since the payment model for dialysis in effect to this day privileges the payment for materials and medication, including equipment and consumables, over physician fees. Dialysis units grew steadily to meet the care needs of a growing number of individuals with chronic kidney disease. Dialysis clinics -the majority of which privately held -concentrated in medium and large cities located mostly in Southern and Southeastern Brazil. 4,5 Newly graduated nephrologists saw the possibility of working in a medical specialty that combined profound clinical knowledge and the technical application of such knowledge. Besides, they also considered the prospect of becoming owners or partners in dialysis clinics as paths to attaining a certain level of financial stability. In this scenario, the interest in nephrology grew steadily for many years. However, the global economic crisis that shook the world in 2008 and the ensuing administrative adjustments made in healthcare changed things dramatically in nephrology. 6 The fees and payments for procedures in public and private institutions decreased significantly. These changes and the consequent drop in income have driven nephrologists to seek work in other areas, including intensive care, emergency and outpatient units. a new generatiOn Of nephrOlOgists The number of nephrologists entering the labor market has decreased in recent years, as seen in the lack of interest in training programs (internships or residencies) offered at teaching hospitals. 6 Surveys enrolling graduating medical students on the brink of choosing their specialties have probed into the issue. Patel and Balzer interviewed graduating medical students in the USA and Europe to conclude that students should build a "positive impression" of nephrology throughout their education by having access to more creative methodologies that utilize social media as an educational resource to foster closer integration between theory and practice and show the advantages of longitudinally following patients with chronic diseases. 7 The considerable decrease in the number of graduating nephrologists deserves attention. Stack et al. looked into the global disproportion between the growing number of patients with chronic kidney disease and the decreasing number of active nephrologists. The authors listed multiple factors contributing to students not opting for nephrology, among which are the lack of demonstration of the positive qualities of nephrology during the course of undergraduate education; procedures once performed solely by nephrologists have been taken over by other medical specialties; absence of a prospect of earning significant financial compensation; uninteresting working environment, repetitive work with little technological innovation and limited alternate procedures; inflexible working hours; little workforce replacement as senior physicians have opted to postpone retirement. 8 In a paper called "Is Nephrology specialty at risk??", Kalloo et al. borrowed the term "value-added" from the business world to state that a specialty is attractive when it strikes a good balance between technical procedures and medical consultation.9 Given the challenges young physicians choosing nephrology are likely to face, one might assume that the reasons to choose nephrology are connected to the desire to perform technology-based work, achieve financial stability, and improve quality of life, as indicated in the surveys cited above. The 1990s saw the birth of Interventional Nephrology, an attempt to enhance the technical skills of nephrologists -along with their autonomy -and meet the expectations of future specialists. Interventional nephrology has been strongly supported by scientific associations and the academia, and has been geared specifically toward "young nephrologists". 10,11 nephrOlOgist eMpOwerMent: challenges and prOspects The contrast between an early prosperous career for new grads and the less-than-exciting historical development of a medical specialty deemed by many as uninteresting begs questions about the future prospects of a career in nephrology. One ought to reflect and consider matters that extend beyond the job at hand. The challenge of picking a career is not an entirely technical decision, since training -and residency programs in particular -provides for the attainment of proper skill levels. The background question revolves around the attributes that add value to a health professional over and above technical knowledge. According to Fabri, "Professional practice builds moral personality, since assignments have an enormous power over how one builds oneself and one's relationships with others and the environment. It affects one's identity, health, and self-esteem". 12 This idea suggests that attractive and fulfilling careers rise above polarizing financial interests. Therefore, it is important to analyze the situations experienced by physicians in their daily practice that require perception, critical reflection, and decision-making skills. As described by Schraiber, until the 1960s the practice of medicine as a form of self-employment translated into proximity between doctors and patients, services rendered in the patient's residence or at the doctor's office, and a relationship in which trust was the main element; there, the physician took control of the relational duality of vocation and talent to practice medicine. However, it was not long before scientific progress coupled with capitalism profoundly changed medical practice, now supported by efficient technologies and new traits that significantly impacted the work and profile of medical professionals. These changes have altered individual and collective expectations around the job, along with the way technical interventions were carried out. Contrast with medicine as a form of self-employment is observed specifically in the technical, organizational, and mercantile arenas. Medicine is now practiced in institutional environments, and the relationship between doctors and patients is mediated through a formal contract; technical skill and accuracy are the sole qualities expected of physicians; patients establish trust with the institution at which they are given medical care; and other stakeholders are involved in providing care to patients. 3 This emerging environment flourishes in globalized society -known for its fascination with technological development -to directly affect the "means of production of goods, value, meaning, and relationships". 13 Technological development empowered physicians, but at the same time the fundamental values responsible for their identity and self-esteem were deeply affected, strongly influencing the way they act and feel. It is important to acknowledge the power of the transformations introduced by technological sciences since the last century, not to underrate development, but to wonder about how nephrologists might defend or even reinvent their identity and professional dignity within the new technological scene. The work of medical doctors is currently governed by the tenets of market economy, which force physicians to deal with countless morally challenging conflictive situations. It is imperative to realize the aggravation of moral conditions. The rules in effect today seek to maximize excellence, a pillar of the business world. The resulting empowerment allows physicians to occupy leading positions in society -a status reserved for the highly competitive, endlessly engaged in the search for up-to-date technical knowledge. This frantic search presupposes the achievement of excellence, a state where error, uncertainty, and doubt may only appear in insignificant levels. the challenge Of training nephrOlOgists The curriculum traditionally offered at medical schools reaffirmed the leading position physicians had in the performance of clinical procedures and found in paternalism its most significant representation. But the prioritization of patient and society demands has led to the development of new educational methods. The majority of medical schools have placed their chips on methods guided by a technological rationale, in which images and virtuality are employed, dismissing completely or relegating to a less important position the learning interpersonal relationships in favor of the acquisition of technical knowledge related to theoretical fundamentals and operational practice. Time is an essential element in teaching and learning activities. A few North-American universities have introduced digital self-learning in medical education, encouraging the development of thinking, memory, and relationship skills among peers and other healthcare workers, in a "team learning" model. 14,15 This model is also present in the medical practice of internship programs, in which diagnosis and treatment will no longer be the prerogative of physicians, but rather of the so-called artificial intelligence -robots and cognitive computing -as described in an article on innovation in healthcare published on newspaper O Estado de São Paulo: Recommending the best course of treatment for each patient and analyzing scans are some of the tasks that will cease to be performed exclusively by physicians -or humans. Progress in areas such as artificial intelligence, cognitive computing, and machine learning has enabled software to read and cross-reference information to levels deemed impossible for the most brilliant experts. However, according to Mariana Perroni, the medical coordinator of Healthcare Transformation at IBM, it is not about replacing healthcare workers, but empowering them. "A specialist would have to study 20 hours a day to keep up with everything that is published in his area. Hard work and good intentions are no longer enough to ensure effective, high-quality care", said Perroni, also an intensive care physician. 16 Empowerment through technology assumes that medical doctors will no longer undergo classical education by reading manuals or books. Information comes in quickly and easily, as selected by the individual requesting it, reflecting the choice for a shortsighted solution, often void of context. In an environment with nearly no criticism or reflection, alien interests may be represented in the information or knowledge acquired by physicians without going through the analysis and review that would otherwise enable autonomous choice. Society has also changed significantly in the way it relates to work. The work of physicians is somewhat paradoxical, since "their acts are manual-dependent, yet highly technological; and, at the same time, practice and technological knowledge have been socially appropriated by physicians only". 3 The social prestige inherent to being a specialist contrasts against the progressive loss of control physicians have experienced in different areas of their profession. However, one might argue that the strategies for medical education should provide students and teachers with a set of fundamental principles, in which "all action is preceded by reflection", 17 and prepare future physicians to acknowledge their limitations and consciously accept their vulnerability when faced with the ethical conflicts present in medical work. Attention to patient fragility should not exclude the consideration of physician vulnerability. Fragility is a condition intrinsic to human beings related to our finitude and insufficiency. Direct contact with the world and its contingencies may elicit our vulnerability (vulnus = injury) at any time. Fragility might be thought of as a condition needed in vulnerability. 13 Medical doctors do not see themselves as fragile or vulnerable; nonetheless, they are exposed daily to risks of all sorts -ethical ones included. It is important that physicians develop critical awareness to understand how their professional context works and the position they occupy before the complex set of factors affecting what they do, namely: -The legal implications of paid employment and discussions on compensation and fair value; the insufficiency of basic labor rights; working hours not always regulated by the Brazilian Consolidated Labor Laws (CLT); -The frequent demand from employers to engage in business-to-business instead of employment relationships with physicians, since the first ensues lesser labor risk and lower tax liabilities; -Gifts offered by pharmaceutical companies in exchange for less than moral or ethical favors; -The business of a nephrologist requires management skills, assuming that the work of nephrologists is a "business"; -Local conditions to perform the job properly in terms of space, time, and equipment. -The relationship with health management organizations, in which rules are imposed on physicians (contract-based relationships; acquisition of membership). Physicians provide care to individuals in situations of fragility for an indefinite period of time; therefore, patient needs might include psychical, social, and cultural demands that require attention and skill before a bond is formed

between the parties. Issues such as depression, death, decreased autonomy, and palliative care are often discussed in nephrology. In addition to the multiple dysfunctions presented by their patients, physicians must also deal with the situations arising from the interactions with other healthcare workers involved in treatment. The ethical and moral assumptions stemmed from the doctor-patient relationship come to fruition only when the parties involved listen to each other. However, this necessary freedom has been the target of external pressures coming from various sources as a result of the changes that Western civilization has undergone in recent centuries, more specifically since the Industrial Revolution. Technology -a symbol of the new era -is gradually taking over the space once dedicated to verbal interaction, a fundamental attribute for the establishment of productive actionoriented dialogue. 18 conclusIons The lack of interest young physicians have had in nephrology derives from the scarcity of technological procedures performed in this medical specialty and from the disregard with which matters related to the work of nephrologists and the social contexts they interact with are treated when medical students are introduced to the specialties they will eventually choose from. These issues call for an ex-ante reflection on vulnerability as a human condition in the field of ethics and require that decisions be made based on values acquired previously by medical professionals. Considering these matters is in itself a substantial exercise of citizenship. Vulnerability does not refer exclusively to the organic fragility inherent to our species. One has to understand that the formative process of personality, morality, and ethics -all indispensable items in one's ability to deal with vulnerability -is not completed in everyone; it relies on a number of variables such as the environment one has lived in since early childhood and the acknowledgement of the participants needed for the formation of a subject's character, considering all interests at play. 17 Western society is governed by the logic of capitalism. Willingly or not, physicians must be aware of the power that governs, standardizes, and determines the actions influencing medical acts. Taking it as a given, as inescapable fact, strengthens the position of those looking to unilaterally satisfy their interests. 19 Adopting a political stance in this highly privatized environment is a challenge that enables the implementation of efforts more in tune with the art of medicine, whose essence is based on the bioethical principles of beneficence and nonmaleficence. It is extremely important to understand how deeply the health-disease complex connects to the social relationships established between patients and healthcare workers under the influence of a hegemonic power. Writing about this topic, Foucault drew attention to "politically active concrete subjects" working on the background of social interactions, particularly in healthcare, submitting physicians to the interests of power holders and promoting exclusion in various degrees and levels. This exclusion affects patients and healthcare workers alike and manifests in various forms and shapes with different degrees of visibility. 20 Nephrology is a medical specialty known for having close ties with technology and development, in addition to institutional relationships governed by the principles of market economy. Nephrologists are faced with the issues troubling their patients, in situations that often involve risk of death or significant morbidity. Pressure from all sides, conditions imposed by disproportionately stronger stakeholders, and lessthan-exciting prospects of enjoying a fairer and more dignifying professional life are elements present in the daily life of nephrologists. By acknowledging vulnerability, nephrologists strengthen their autonomy and take a firmer stance in their interactions with other stakeholders. As subjects aware of their condition, newly graduated physicians may take over the status of apprentices of nephrology with a more thorough understanding of the experiences they are likely to go through in their careers. Augmented Reality Technology for the Positioning of the Auricle in the Treatment of Microtia Supplemental Digital Content is available in the text. INTRODUCTION In auricular reconstruction, placing of the reconstructed auricle in the proper position will greatly affect the result of the procedure. Although anthropometric analysis has revealed detailed data, ie, the shape, size, and angle of the human ear, 1,2 it is still challenging to use these data in surgery. Due to previous works of several surgeons, some parts of auricular reconstruction have become routinary. 3,4 However, deciding the auricular position is still performed manually, eg, transferring the nonaffected side image using transparent film. Recently, augmented reality (AR) technologies, which make it possible to overlay computer-generated images onto patients' bodies, have become an operative auxiliary tool. [5][6][7] In this report, we describe an application of the AR technology in auricular reconstruction. PATIENT AND METHODS The clinical application of the AR technology is approved by the Ethical Committee of Osaka Medical College (No. 2316). We used an AR device to position the auricle of a 10-year-old male with right congenital small concha type microtia. A day before the surgery, 3-dimensional (3D) photographs were taken, while the patient was in a supine position. Before taking the photographs, 3 dots were drawn around the affected right ear. Then, 3D photographs of the right and left face including the auricles were taken using VECTRAH1 (Canfield Background: The positioning of the auricle is a key factor in successful ear reconstruction. However, the position of the ear is usually determined by transferring the auricle image of the nonaffected side to the affected side using a transparent film. Augmented reality (AR) is becoming useful in the surgical field allowing computer-generated images to be superimposed on patients. In this report, we would like to introduce an application of AR technology in ear reconstruction. Methods: AR technology was used to determine the position of the reconstructed ear of a 10-year-old male with right microtia. Preoperative 3-dimensional photographs of the nonaffected side were taken using VECTRAH1. Then, the image was horizontally inverted and superimposed on the three-dimensional image of the affected side with reference to the anatomical landmarks of the patient's face. These images were projected onto the patient in the operation room using Microsoft's HoloLens. The design and positioning of the auricle was done with reference to the AR image. To confirm the accuracy of the AR technique, we compared it to the original transparent film technique. After the insertion of the cartilage framework into the skin pocket, the position and shape of the reconstructed ear was confirmed using the AR technology. Results: The positioning of the reconstructed ear was successfully performed. The deviation between the 2 designated positions using the AR and the transparent film was within 2 mm. Ideas and InnovatIons PRS Global Open • 2020 Scientific, Parsippany, N.J.). The 3D images were exported to Blender, a free-and open-source 3D software. Then, the 3D image of the nonaffected side of the face was inverted to the right and left. This image was superimposed to the image of the affected side with reference to the anatomical landmarks on the face, eg, ala of nose, lateral canthus, eyebrow, angle of mandibular, and facial contour. Finally, the image of the facial outline of the affected side with the 3 dots was created using the software. As a result, the following images were obtained: 1) auricular image of the affected side, 2) reversed left and right nonaffected auricular image aligned on the affected side, and 3) image of the outline of the face of the affected side with the 3 dots (see Figure, Supplemental Digital Content 1, which displays a 3D photograph of the face of the affected side with the 3 dots indicated by a, b, and c (left) and a 3D photograph of the nonaffected side which was inverted to right and left was superimposed with reference to the anatomical landmarks on the face, eg, ala of nose, lateral canthus, eyebrow, angle of mandibula, and facial contour (middle); image of the outline of the face of the affected side with the 3 dots was prepared (right); http://links.lww. com/PRSGO/B307.). These images were exposed to HoloLens (Microsoft Corp., Redmond, Wash.), a headmounted mixed reality device. The operation was performed under general anesthesia with the patient in a supine position. The prepared 3D images were projected onto the surgical field using HoloLens (Fig. 1). By aligning the 3 dots on the 3D image of the facial outline to the 3 dots on the patient's face, it can be superimposed on the surgical site 5 (Fig. 2). Then, the 3D image of the affected side was displayed and the position was confirmed and finely adjusted. Finally, the reversed 3D image of the right and left nonaffected sides was displayed and the outline of the opposite auricle was transferred (Fig. 3) (see Video, [online], which displays the surgeon projecting the image of nonaffected ear on the operative field and drawing the position of reconstructed ear via HoloLens). Positioning of the auricle was also done using a transparent film. The outline of the normal auricle and facial landmarks, eg, ala of nose, lateral canthus, eyebrow, and sideburns, were marked on the film then it was flipped and the position was transferred to the affected side. After the skin incision, the remnant cartilage above the concha was removed. The three-dimensional costal cartilage framework of the upper two-thirds of the auricle was placed into the skin pocket. Then, the frame was draped with skin flap and suction was applied to simulate the draping of the skin. The shape and position of the reconstructed auricle was compared to the normal auricle by using HoloLens. RESULTS The positioning of the reconstructed auricle was successfully performed. Real-time comparison of the reconstructed auricle with the normal was made possible using the AR device. The 2 designated positions using the AR device and the transparent film were almost identical in the longitudinal direction. On the other hand, the auricular design using the AR device was 1-2 mm shorter in the lateral direction. The time it took to align the 3D images and draw the design was less than 10 minutes. DISCUSSION In auricular reconstruction, the AR technology can effectively support the positioning and confirm the shape of the reconstructed auricle after the costal cartilage framework was transferred. To determine the position of the auricle, level, axis, and distance from the orbit should be considered. 1,2 In our technique, the image of the nonaffected side and the affected side can be superimposed finely using computer software with reference to not only the anatomical landmarks on the face but also the facial contour. Historically, transparent film has been used to transfer the outline of the nonaffected auricle to the affected side. 8,9 To execute the design correctly, it is necessary to set the film in the appropriate position. Patients often have low hairline, and available anatomical landmarks are limited to the lateral canthus, eyebrow, and ala of the nose. These landmarks are away from the auricle; therefore, slight shifts in alignment can greatly affect the position of the auricle. Furthermore, these right and left landmarks may not be exactly the same. In hemifacial microsomia patients, it is challenging to apply the transparent film technique. Therefore, an AR device can help project a detailed simulated image onto the surgical site with less error. In this case, the auricular design was 1-2 mm shorter than the design using transparent film in the anterior-posterior axis. It may be because the 2 mm error must be due to the image of the standing ear which did not perfectly reflect the image of the auricle after the first stage of operation. There are also reports that some error may occur in the alignment method. 7,10 At this early stage, these kinds of errors can be easily optimized with improvements in the system and the device. The AR technique is costly and it time-consuming to prepare the images. However, we hypothesized that an AR device has a potential to reduce the time taken for the positioning of the reconstructed auricle in the operation room and can help improve the accuracy. CONCLUSION An AR device can effectively overlay computer-generated images onto the surgical site. This technology can be a promising tool in auricular reconstruction. Skin microcirculatory reactivity assessed using a thermal challenge is decreased in patients with circulatory shock and associated with outcome Background Shock states are characterized by impaired tissue perfusion and microcirculatory alterations, which are directly related to outcome. Skin perfusion can be noninvasively evaluated using skin laser Doppler

(SLD), which, when coupled with a local thermal challenge, may provide a measure of microcirculatory reactivity. We hypothesized that this microvascular reactivity would be impaired in patients with circulatory shock and would be a marker of severity. Methods We first evaluated skin blood flow (SBF) using SLD on the forearm and on the palm in 18 healthy volunteers to select the site with maximal response. Measurements were taken at 37 °C (baseline) and repeated at 43 °C. The 43 °C/37 °C SBF ratio was calculated as a measure of microvascular reactivity. We then evaluated the SBF in 29 patients with circulatory shock admitted to a 35-bed department of intensive care and in a confirmatory cohort of 35 patients with circulatory shock. Results In the volunteers, baseline SBF was higher in the hand than in the forearm, but the SBF ratio was lower (11.2 [9.4–13.4] vs. 2.0 [1.7–2.6], p < 0.01) so we used the forearm for our patients. Baseline forearm SBF was similar in patients with shock and healthy volunteers, but the SBF ratio was markedly lower in the patients (2.6 [2.0–3.6] vs. 11.2 [9.4–13.4], p < 0.01). Shock survivors had a higher SBF ratio than non-survivors (3.2 [2.2–6.2] vs. 2.3 [1.7–2.8], p < 0.01). These results were confirmed in the second cohort of 35 patients. In multivariable analysis, the APACHE II score and the SBF ratio were independently associated with mortality. Conclusions Microcirculatory reactivity is decreased in patients with circulatory shock and has prognostic value. This simple, noninvasive test could help in monitoring the peripheral microcirculation in acutely ill patients. Electronic supplementary material The online version of this article (10.1186/s13613-018-0393-7) contains supplementary material, which is available to authorized users. Background Circulatory shock is a life-threatening condition affecting about one-third of patients admitted to the intensive care unit (ICU) [1,2]. Regardless of the underlying pathophysiological mechanisms, the hallmark of shock states is altered tissue perfusion, which if not rapidly corrected leads to organ dysfunction and death [3,4]. Recent data have highlighted the prognostic importance of microcirculatory abnormalities in patients with shock using noninvasive bedside techniques, such as sublingual video-microscopy [5][6][7][8]. The hallmark of these alterations is decreased capillary density, and microvascular blood flow with increased heterogeneity of perfusion [9,10]. Interestingly, these microcirculatory abnormalities are not explained by routinely measured global macrohemodynamic variables, making it attractive to assess them directly [11,12]. The microcirculation can also be studied by evaluating the response generated by a hypoxic stress event, such as during a transient vascular occlusion test (VOT) [7,8]. Near-infrared spectroscopy (NIRS) or laser Doppler techniques can be used to indirectly or directly evaluate a transient increase in flow after a VOT [13][14][15]. Studies using these techniques have shown that endothelial reactivity is impaired in sepsis and is associated with organ dysfunction and outcome [16,17]. However, VOTs are not easily standardized (duration of occlusion and/ or tissue oxygen saturation [StO 2 ] reached) [16,18]. In addition, VOTs may alter local metabolism [19]. Alternative methods to evaluate microvascular recruitment are, therefore, of interest. Local heating of the skin may represent an alternative means of evaluating microvascular reactivity. Skin laser Doppler (SLD) (also known as laser Doppler flowmetry) can be used to assess skin blood flow (SBF) during a thermal challenge [14]. This technique uses an optical fiber to direct light from a low-power laser source to the skin and to collect the back-scattered light. The shift in light wavelength is proportional to the red blood cell velocity in the studied skin area, providing a noninvasive measurement of SBF expressed as arbitrary perfusion units (PUs) [20]. New SLD flow probes can heat the explored tissue in a controlled way, making it possible to perform a dynamic test of capillary reactivity by increasing local temperature [14]. However, there are no published data evaluating this test in patients with circulatory shock. We hypothesized that reactivity of the skin microcirculation, evaluated as the skin blood flow ratio, during a thermal challenge would be impaired in patients with circulatory shock. We also assessed whether these alterations could be explained by other hemodynamic parameters and were correlated with patient outcome. Methods This prospective, observational study was conducted in our 35-bed Department of Intensive Care. Institutional Ethical Committee approval was obtained, and informed consent was obtained from each participant or the next of kin. Protocol To assess the most appropriate probe position for an SLD thermal challenge, we first studied 18 healthy volunteers. They were comfortably seated in a quiet, temperature-controlled room for at least 15 min before each experiment. Heart rate, respiratory rate and hemoglobin saturation were evaluated noninvasively by pulse oximetry using a Siemens SC 9000 monitor (Siemens, Erlangen, Germany). Noninvasive measurements of mean arterial pressure (MAP) were taken in the opposite arm to that used for the SLD blood flow measurements. We then studied, in 2013, a cohort of 29 critically ill adult patients admitted with a diagnosis of circulatory shock, defined as the need for norepinephrine infusion to maintain a MAP of at least 65 mmHg, associated with an altered mental status, acute oliguria defined as a urine output < 0.5 ml/Kg/h or an arterial lactate level > 2 mmol/L [4]. Screening of ICU admissions, collection of data and SLD measurements were performed by doctors not involved in patient management. All SLD measurements were taken as soon as possible after completion of initial resuscitation, i.e., when an adequate arterial pressure for that patient had been reached (determined by the treating physician), and norepinephrine doses had been stable for 1 h. A second SLD measurement was taken, when possible, 48 h after the first measurement. Having analyzed the results from our first patient cohort, we repeated the study in a second, confirmatory cohort of critically ill patients with circulatory shock (using the same definition) admitted in 2015. Patients were evaluated by an investigator who had not been involved in the initial study and at just one time point during the first day of hospitalization after initial resuscitation. We collected demographic and clinical data on admission and classified patients as having sepsis or not, based on standard criteria [21]. At the time of each SLD measurement, we collected all available hemodynamic and respiratory data from ongoing patient monitors and recorded the central body temperature. We also collected biochemical and laboratory data from clinical records, including the most recent blood gas analysis. The APACHE II score [22] was calculated using the worst data during the first 24 h in the ICU. The sequential organ failure assessment (SOFA) score [23] was calculated from the data present at the time of the SLD measurements. Patients were grouped according to ICU outcome (dead or alive) for further analysis. Skin laser Doppler measurements All SLD measurements were taken using the PeriFlux System 5000 monitor (Perimed, Jarfalla, Sweden), and data were continuously collected (PeriSoft software 2.5.5; Perimed) for further analysis. For the thermal challenge, a small angled thermostatic SLD probe 457 (Perimed) with 0.25 mm fiber separation was used. This probe also allows skin temperature measurement at site of application. The SLD machine emits a beam of laser light with a wavelength of 780 nm that allows skin evaluation at a depth of 0.5-1 mm. The initial skin temperature was measured with the thermostatic probe prior to each measurement. In the healthy volunteers, the probe was placed on the skin of the volar face of the proximal forearm and on the palm of the same arm. In the patients, the SLD probe was placed on the forearm without the arterial line. The probe was kept in position using the double-sided tape provided with the SLD monitor. All participants were asked to abstain from any activity during the study period to prevent any possible artifacts in the recorded signals [13]. To limit differences between subjects in the basal temperature, SBF was recorded at a local skin temperature of 37 °C allowing for at least three minutes of stabilization. Thereafter, a thermal challenge was performed by increasing the probe temperature abruptly (0.1 °C/s) from 37 to 43 °C. After 9 min at 43 °C, the SBF was recorded (Additional file 1: Figure S1). We chose to increase the temperature to 43 °C, because our preliminary experience indicated that using lower thresholds (39 °C or 41 °C) reduced the amplitude of the response, making it more difficult to detect potential differences. We calculated the SBF ratio (SBF obtained at 43 °C/SBF obtained at 37 °C) as a simple measure of microvascular reactivity in the explored area. Statistical analysis Statistical analyses were performed using SPSS 22.0 (IBM, New York, NY) software. Variables were assessed for normality of distribution using skewness and kurtosis tests and Q-Q plots. Continuous variables are presented as means ± standard deviations or median values with percentiles (25-75%) depending on the presence or absence of normality. Categorical data are presented as numbers of events and percentages. Repeated measurements were compared using a paired Student's t test or Wilcoxon signed rank test, as appropriate. Comparisons between different cohorts were made using an unpaired t test or Mann-Whitney U test as appropriate. Proportions were compared with a Chi-square test or Fisher's exact test as appropriate. We plotted the sensitivity and specificity using a receiving operating characteristics (ROC) graph, and the area under the curve (AUC) was calculated for the different variables as a measure of their ability to predict mortality. To assess possible explanatory variables correlated with the SBF ratio, we plotted individual data on graphs and calculated the Pearson or Spearman correlation coefficient (r) as appropriate. Univariate and multivariate analyses (binary logistic regressions) were performed to identify the ability of different variables to predict ICU mortality, calculating the odds ratios and their respective 95% confidence intervals. In a post hoc analysis considering that the APACHE II score was directly correlated and the SBF ratio inversely correlated with mortality, we calculated the ratio between the two factors (APACHE II score/SBF ratio) in an attempt to improve their prognostic value. A two-sided p value less than 0.05 was considered as significant for all analyses. Healthy volunteers The main characteristics of the volunteers are listed in Additional file 1: Table S1. Comparisons of measurements taken in the hand and the forearm are shown in Fig. 1 and Additional file 1: Table S2. Baseline values of SBF at 37 °C were lower in the forearm than in the hand, but the SBF ratio was higher in the forearm. Patients: initial cohort The main demographic and baseline characteristics of the 29 patients are listed in Table 1: 16 (55%) had septic and 13 (45%) cardiogenic shock. Eleven patients died in the ICU (44%). Times from ICU admission until the first measurement were similar in survivors and non-survivors: 6 (4-17) versus 7 (5-14) h (p = 0.59). The patients had a similar baseline SBF to the healthy volunteers, but a lower SBF at 43 °C, which resulted in a lower SBF ratio ( Table 2). SBF values at baseline and at 43 °C were similar in survivors and non-survivors, but survivors had significantly higher SBF ratios ( Table 2). A comparison of the SBF ratios in volunteers, survivors and non-survivors is shown in Fig. 2. SLD measurements were repeated at 48 h in 20 of the patients. At that time point, the survivors had a somewhat higher SBF ratio than the non-survivors although the differences were not statistically significant [3.1 (2.4-4.3) vs. 2.3 (1.5-3.1), p = 0.08] (Additional file 1: Figure S2). The difference in the SBF ratio between T0 and T48 was not significant in survivors or non-survivors (Additional file 1: Figure S3). There were no significant correlations between the SBF ratio and any hemodynamic or blood gas-derived variable, administration of vasoactive drugs (Additional file 1: Figure S4), presence of sepsis (Additional file 1: Table S3), degree/severity of organ dysfunction (SOFA score), initial forearm temperature or the difference between the central to forearm temperature (Additional file 1: Figure S5). The SBF ratio and the APACHE II score had similar ROC AUCs to predict survival (0.73 [0.55-0.91] vs. 0.74 [0.56-0.92], respectively) (Fig. 3). There was no significant correlation between the SBF ratio and the APACHE II score (r = 0.047, p > 0.05), so that we calculated their combined ratio, which resulted in a larger ROC AUC to predict survival than that for either measure alone (0.90 [0.79-1.00], Fig. 3). Considering

a cutoff ratio of 3 for the SBF ratio, the specificity and sensitivity were 91 and 56%, respectively, to predict survival. Considering a cutoff value of 2 for the SBF ratio, the specificity and sensitivity were 36 and 83%, respectively. In multivariable analysis, taking into consideration the SOFA score, the blood lactate concentration, the mean arterial pressure, the norepinephrine dose and the presence of sepsis, only the APACHE II score and the SBF ratio were independently correlated with mortality (Additional file 1: Table S4). Patients: confirmatory cohort The demographic and baseline characteristics of the 35 patients in the confirmatory cohort are listed in Table 1: 28 had septic (80%) and 7 cardiogenic (20%) shock. Eighteen patients died in the ICU (51%). SBF values at baseline were similar in survivors and non-survivors, but survivors had significantly higher SBF values at 43 °C and larger SBF ratios (Table 2). There were no significant correlations between the SBF ratio and any hemodynamic or blood gas-derived variable, dose of vasoactive drugs, presence of sepsis (Additional file 1: Table S3), degree/severity of organ dysfunction (SOFA score), initial forearm temperature or the difference between the central to forearm temperature (all p values > 0.05). The ROC AUCs to predict survival were 0.85 (0.73-0.98) and 0.77 (0.60-0.94) for the SBF ratio and the APACHE II score, respectively. The ratio between the APACHE II score and the SBF ratio resulted in an ROC AUC to predict survival of 0.91 (0.82-1.00) (Fig. 3). Considering a cutoff ratio of 3 for the SBF ratio, the specificity and sensitivity to predict survival were 83 and 71%, respectively. Considering a cutoff value of 2 for the SBF ratio, the specificity and sensitivity for prediction of survival were 50 and 88%, respectively. Discussion In this study, reactivity of the skin microcirculation during a thermal challenge was compromised in patients with circulatory shock, and this impairment was related to outcome. This phenomenon could not be explained by other hemodynamic variables and was independent of regularly used scoring systems, such as the SOFA and the APACHE II score. As far as we know, this is the first report to assess the skin microvascular reactivity (or vasodilatation) during a thermal challenge in shock patients. Several previous reports have used an ischemic challenge to assess microcirculatory reactivity [7] and shown that post-occlusive reactive hyperemia is impaired in patients with sepsis and septic shock, and this has been correlated with poor outcomes [15][16][17]24]. However, this condition is not pathognomonic of sepsis and is also present in hemorrhagic [25] and cardiogenic [19,26] shock. Our data also show that the degree of impairment in microvascular reactivity during a thermal challenge was similar in septic and cardiogenic shock. Nevertheless, studies in patients with just cardiogenic, septic or hemorrhagic shock may be of interest to further investigate whether the effects of a thermal [11,12]. Importantly, an ischemic challenge can significantly alter local metabolism. Hence, some NIRS-derived variables, such as the descending slope (the rate of decrease in tissue oxygen saturation during a VOT and correlated with tissue oxygen consumption), may be affected by some degree of ischemic preconditioning in the studied tissue, limiting the reproducibility of results over a short time period [19]. The thermal challenge that we propose allows the SBF to return to baseline within minutes, enabling the test to be repeated in the same area. If there are concerns about alterations in local metabolism, the challenged area is very small (compared to the complete extremity involved during a VOT), so that the test can be repeated in an adjacent area. The mechanisms underlying a thermal challenge are different to those of an ischemic challenge. As the skin temperature increases, the SBF increases until a maximum plateau value is reached [27][28][29]. More specifically, an initial short-lived phase of rapid vasodilation, which is neurally driven, is followed by a more gradual but protracted increase in SBF that is dependent on local nitric oxide (NO) production stimulated by endothelial NO synthase [27][28][29][30][31]. Thus, the use of this very simple and noninvasive test may provide an indirect assessment of the local NO pathway. NO plays a major role in the local control of the microcirculation and in its interaction with red blood cells [32,33]. It also plays an important role in the pathophysiology of sepsis and septic shock [34,35] and of cardiogenic shock [36,37]. Interestingly, during endotoxin infusion in healthy volunteers, recruitment of the skin microcirculation through a thermal challenge that involves NO-dependent pathways was impaired, whereas post-occlusive reactive vasodilation was not [38]. Thus, the worse outcome in patients with shock who had compromised skin microvascular reactivity after the thermal challenge may reflect altered NO pathways. Nevertheless, physiological or basic science studies need to be conducted to verify the role of the NO pathway in our observations. Two important technical aspects should be discussed further: the position of the probe on the forearm and the duration of the thermal challenge. Intuitively, the skin area used to perform SLD with a thermal challenge should have the highest potential recruitment even if the baseline SBF is relatively low. SLD studies have shown that the cutaneous microcirculation in healthy volunteers is very heterogeneous with lower blood flow values in the forearm than in the palm and lower values in the metacarpal spaces than in the fingertips [14,39]. Similar observations have been made for other areas of the body, such as the torso, feet and face, and this heterogeneity has been confirmed by other techniques evaluating skin perfusion [40][41][42][43][44]. When performing a thermal challenge in different cutaneous regions, Metzler-Wilson et al. [40] showed that Fig. 2 Comparison of the capillary reactivity, evaluated as the skin blood flow ratio, in the healthy volunteers and the patients with circulatory shock (initial and confirmatory cohorts). *p < 0.05 compared with healthy volunteers; $p < 0.05 compared with survivors Fig. 3 Receiver operating characteristic (ROC) curves for the predictive value of the APACHE II score, the skin blood flow (SBF) ratio and the APACHE II/SBF ratio (initial and confirmatory cohorts) recruitment was higher in the forearm than in the palm. We confirmed this finding in our healthy volunteers. Table 2 Comparison of skin laser Doppler (SLD) variables between volunteers and patients and between survivors and non-survivors Different temperature thresholds and heating times have been proposed when performing a thermal challenge [28,29,38,40,41]. However, 30 min may be needed to reach the maximal SBF during the test [28,29], a period that is too long in critically ill patients who often need rapid changes in therapy. During pilot studies (data not shown), we observed that the best temperature to demonstrate a response was 43 °C. We also observed that we were not able to identify the initial short-lived phase of rapid vasodilation in all patients (this does not mean that the neural stimulus was absent, but just difficult to track with this technology), but when present it had already disappeared at 9 min and its analysis was difficult to standardize (time to peak versus maximum peak versus slope). For these practical reasons, we arbitrarily chose to limit the duration of the thermal challenge to 9 min. Hence, we did not evaluate the maximal recruitment of the skin microcirculation, but a surrogate value (the slope or speed of the microcirculation recruitment) within a reasonable observation period. Other limitations should also be acknowledged. First, we included a relatively small number of patients, potentially limiting the statistical power of our analyses and possibly accounting for the lack of significant difference in microvascular reactivity between survivors and nonsurvivors, particularly at 48 h. Moreover, although we found similar results in our two cohorts, external validation of our data is still necessary. Second, although the forearm is a very accessible area in most critically ill patients and has the additional benefit of high potential recruitability with a thermal challenge, studies using different skin locations may also be informative. Further studies are also required to describe the evolution of these alterations over time and their changes with specific therapies (fluid bolus, transfusions, change in vasopressor dose, etc.). Although we focused on mortality as an outcome measure, future studies could also specifically assess relationships with other clinical outcomes, including changes in SOFA scores over time. Third, the presence of fever may represent a confounding factor that should be taken into account; however, the local temperature on the forearm was always less than 37 °C and there was no apparent correlation with the resulting SBF ratio. Fourth, we did not collect data from a control cohort of elderly patients (a factor that can influence the vascular response); however, our control group was designed to provide a picture of the normal physiological response and, more importantly, our main findings are related to the differences between survivors and non-survivors. Fifth, we did not compare the results of the thermal challenge to results from other concomitant measures of the microcirculation. However, our recent observations indicate that VOTs can induce ischemic preconditioning and alter local metabolism [19], which would need to be taken into account if a skin thermal challenge was performed in the same area. Finally, although it is impossible to extrapolate our skin measurements to other microvascular networks in different organs, the strong association we found with outcomes and the simplicity of this technique makes it attractive for future research. Conclusion SLD with a thermal challenge is a technique that allows simple, noninvasive evaluation of skin microcirculatory reactivity. Using this technique, we have demonstrated that reactivity of skin microcirculation during a thermal challenge is compromised in patients with circulatory shock, is related to outcome and combination with the APACHE II score can improve its prognostic value. Genetic trans-Complementation Establishes a New Model for Influenza Virus RNA Transcription and Replication The influenza A viruses genome comprises eight single-stranded RNA segments of negative polarity. Each one is included in a ribonucleoprotein particle (vRNP) containing the polymerase complex and a number of nucleoprotein (NP) monomers. Viral RNA replication proceeds by formation of a complementary RNP of positive polarity (cRNP) that serves as intermediate to generate many progeny vRNPs. Transcription initiation takes place by a cap-snatching mechanism whereby the polymerase steals a cellular capped oligonucleotide and uses it as primer to copy the vRNP template. Transcription termination occurs prematurely at the polyadenylation signal, which the polymerase copies repeatedly to generate a 3′-terminal polyA. Here we studied the mechanisms of the viral RNA replication and transcription. We used efficient systems for recombinant RNP transcription/replication in vivo and well-defined polymerase mutants deficient in either RNA replication or transcription to address the roles of the polymerase complex present in the template RNP and newly synthesised polymerase complexes during replication and transcription. The results of trans-complementation experiments showed that soluble polymerase complexes can synthesise progeny RNA in trans and become incorporated into progeny vRNPs, but only transcription in cis could be detected. These results are compatible with a new model for virus RNA replication, whereby a template RNP would be replicated in trans by a soluble polymerase complex and a polymerase complex distinct from the replicative enzyme would direct the encapsidation of progeny vRNA. In contrast, transcription of the vRNP would occur in cis and the resident polymerase complex would be responsible for mRNA synthesis and polyadenylation. Introduction The influenza A viruses are the causative agents of yearly epidemics of respiratory disease and occasionally more severe pandemics [1]. The latter are the consequence of transfers from the avian virus reservoir to humans by either genetic reassortment or direct adaptation [2]. Thus, current occasional infections of humans with highly pathogenic H5N1 avian strains have raised fears about a possible new pandemic of great severity. The influenza A viruses belong to the family Orthomyxoviridae and posses a single-stranded, negative-polarity RNA genome made up by 8 RNA segments, that form ribonucleoprotein (RNP) complexes by association to the polymerase and the nucleoprotein (NP). Such RNPs are independent molecular machines responsible for transcription and replication of each virus gene and contain an RNA-dependent RNA polymerase composed by the PB1, PB2 and PA subunits [3]. The polymerase complex recognises the RNA promoter comprising both 59-terminal and 39-terminal sequences of each segment, by preferentially binding the 59terminal end [4][5][6], and in this way stabilises a supercoiled conformation of the RNPs [7]. Upon infection of susceptible cells, the parental RNPs are first transcribed in the nucleus (primary transcription). Transcription

initiation takes place by a cap-snatching process whereby the viral polymerase recognises the cap structure of cellular pre-mRNAs in the nucleus, cleaves these some 15 nt downstream the cap and utilises such capped-oligonucleotides as primers to copy the virus template RNA [8]. Transcription finalises by reiterative copy of the virus polyadenylation signal, an oligo-U sequence located close to the 59-end of the template [9,10]. Synthesis of new virus proteins is required to proceed to RNP replication [11], that takes place first by the generation of complementary RNPs (cRNPs). These RNPs are structurally analogous to those present in the virions (vRNPs) but contain complete positive-polarity copies of the virus RNA segments, that are neither capped nor polyadenylated. The structural differences between the vRNP transcription and replication products (mRNAs and cRNPs) led to the proposal of a transcription-to-replication switch by which the parental RNPs would change from capped-RNA-dependent to de novo initiation, from polyadenylation to full copy of the template, and in addition would induce encapsidation of the RNA product into new RNPs (reviewed in [12]. Such notion has been challenged recently by a new model proposing that parental vRNPs can directly synthesise cRNA but require newly synthesised polymerase and NP to stabilise the product in the form of cRNPs [13]. The cRNPs accumulate to low levels but serve as efficient templates for the synthesis of large quantities of progeny vRNPs that can be transcribed (secondary transcription) and eventually be incorporated into progeny virions [3]. Much information has been obtained during recent years on structural aspects of the RNPs [14,15] (R. Coloma, unpublished results) and their components, like the NP [16,17], the polymerase complex [18,19] and specific domains of the polymerase subunits [20][21][22][23][24]. Likewise, a number of host cell factors have been identified that may play important roles in the transcription and replication processes [25][26][27][28][29][30][31][32][33][34]. However, much remains to be learned about the detailed mechanisms for RNP transcription and replication. For instance, it is not clear whether the polymerase complex present in the template RNP is able to synthesise the progeny vRNA or whether the replicative complex directs the encapsidation of progeny RNA it into a new vRNP. Likewise, it has been assumed that the polymerase complex present in the vRNP accounts for viral mRNA synthesis, but it is not clear whether other vRNPs or other soluble polymerase complexes perform this step in trans. In this report we used efficient in vivo recombinant replication and transcription systems and defined polymerase mutants specifically affected in either transcription or replication to answer these questions. Our results are consistent with a new model whereby polymerase complexes not associated to the template cRNP are responsible for the replicative synthesis of vRNPs in trans and polymerase complexes distinct from the replicative one specify the encapsidation of viral RNAs. On the contrary, no vRNP transcription could be detected by other RNP or a soluble polymerase complex in trans, suggesting that it takes place by the activity of the RNP-associated polymerase complex. The experimental approach To gather information on the mechanisms of influenza virus transcription and replication we have adopted a genetic transcomplementation approach. This is based on the ability to reconstitute in vivo an efficient transcription-replication system that mimic these steps of the infection cycle and is more amenable to experimental manipulations [15,35]. Furthermore, the vRNP products can be efficiently purified, their structural and biological properties can be easily analysed [14,18,20] and they can in turn be used as templates for further rounds of in vivo replication. Essential for these approaches is the availability of well-defined mutants to be used as genetic markers. We have earlier described point mutants in the PB2 subunit of the viral polymerase that are defective in viral RNA replication but fully efficient in virus transcription [36]. Likewise, we have recently reported polymerase PB2 mutants that are affected in the cap-binding activity and hence are defective in cap-snatching, but retain their capacity to replicate virus RNPs [20]. A polymerase complex distinct from the replicative enzyme becomes associated in trans to the newly synthesised vRNA Using the approaches indicated above we first addressed the question whether the replication deficiency of point mutants within the N-terminus of PB2 [36] could be rescued in trans by coexpression of PB2 point mutants defective in cap-binding [20]. Cultures of HEK293T cells were co-transfected with plasmids encoding PB1, PA, NP and a deleted NS virus replicon (clone 23, 248 nt in length; [14,15]). In addition, either PB2wt or PB2 mutants R142A or F130A (replication-defective) or mutant E361A (transcription-defective) were co-expressed. Alternatively, pair wise combinations of these PB2 mutants were co-expressed (R142A+E361A and F130A+E361A). Among the PB2 proteins expressed, either wt or the replication-defective mutants R142A or F130A were His-tagged at the C-terminus, a modification that does not alter their biological activity and allows the efficient purification of the in vivo RNP replication progeny [18]. The expression levels of all PB2 mutants were shown to be similar to that of PB2wt (Fig. S1) and the untagged PB2wt was used as a control for purification (see diagram of the experimental setting in Fig. 1A). After incubation, the cell extracts were used for Ni 2+ -NTA-agarose purification as described in Materials and Methods and the accumulation of progeny RNPs was determined by means of Western-blot assays using anti-NP sera. The purification of the complete RNPs was verified by Western-blot with antibodies specific for PB2 and PA (Fig. 1B). This strategy allows measuring the replication capacity of the RNPs formed in vivo, as omitting any RNP element or using a defective point mutant leads to undetectable RNP accumulation [15,36,37]. Amplification of virus RNPs was expected for wt and mutant polymerase containing transcription-defective PB2 (E361A), but not for those containing replication-defective PB2 (R142A and F130A). However, since no tag is present in the former mutant, only RNPs derived from cultures containing PB2His were expected in the Ni 2+ -NTAagarose purified material. This was indeed the case, as shown in Fig. 1B. If the transcription-defective polymerase were able to rescue in trans the defect in replication of polymerase mutants R142A or F130A, one would expect the accumulation and purification of RNPs containing these mutant PB2. The results obtained by the co-expression of pairs of replication-and transcription-defective polymerases indicate that such prediction is hold (Fig. 1B). The transcription-defective mutant could rescue both R142A and F130A alleles and similar rescue was obtained when other transcription-defective mutants, like H357A, K370A, F404A [20] were used (Fig. S2). The progeny RNPs contained the replication-defective PB2 allele, since (i) they could be purified by Ni 2+ -NTA-agarose chromatography and (ii) the mobility of the PB2 subunit in the Western-blot assay corresponded to the His-tagged subunit and not to the untagged one. It is important to mention that only His- Author Summary The influenza A viruses produce annual epidemics and occasional pandemics of respiratory disease. There is great concern about a potential new pandemic being caused by presently circulating avian influenza viruses, and hence increasing interest in understanding how the virus replicates its genome. This comprises eight molecules of RNA, each one bound to a polymerase complex and encapsidated by multiple copies of the nucleoprotein, in the form of ribonucleoprotein complexes (RNPs). These structures are responsible for virus RNA replication and transcription but the detailed mechanisms of these processes are not fully understood. We report here the results of genetic complementation experiments using proficient in vitro and in vivo recombinant systems for transcription and replication, and polymerase point mutants that are either transcription-defective or replication-defective. These results are compatible with a new model for virus replication whereby a polymerase distinct from that present in the parental RNP is responsible for RNA replication in trans and the progeny RNP is associated to a polymerase distinct from that performing replication. In contrast, transcription is carried out in cis by the polymerase resident in the RNP. tagged PB2 protein was detected in the purified RNPs and not the untagged counterpart, indicating that no transcription-defective polymerase was co-purified (Fig. 1B). Furthermore, the phenotype of the rescued RNPs was tested by determination of their in vitro transcription activity (Fig. 2). Since the transcription-defective mutants had alterations in their cap-binding pocket, they show low in vitro transcription activity when a mRNA is used as a capdonor, whereas cap-independent transcription is observed with a general primer as the dinucleotide ApG [20]. The transcription activity profile of rescued RNPs using ApG or b-globin mRNA as primers was identical to that of wt RNPs, as expected, and not to that of mutant E361A, that is unable to use b-globin as primer [20] (Fig. 2 and Fig. S3). In the experimental approach used, the reconstitution of a RNP from the viral proteins and genomic RNA has to take place first and its subsequent amplification would account for the bulk of RNPs that can be purified from the transfected cells. Since a RNP template with the replication-defective polymerase does not replicate [36], only transcription-deficient polymerase could perform RNP replication. The results obtained (Figs. 1, 2) demonstrate that a polymerase complex distinct from that responsible for RNP replication (replication-defective versus transcription-defective) is incorporated into the progeny vRNP and suggest that a replication-defective polymerase can direct the encapsidation of the progeny vRNA, i.e. can bind the 59-terminus of newly synthesised vRNA and direct the incorporation of NP monomers into the progeny vRNP. It could be argued that the incorporation of the replication-defective polymerase to the progeny RNP might occur by exchange with replicationcompetent during purification in vitro. Two lines of evidence argue against such possibility: (i) Our transcription experiments verify that the polymerase present in an RNP complex is stably bound (see below) and (ii) The data reported by Wreede et al. [38] suggest that the binding of a polymerase complex to the 59terminal sequence of viral RNA can not competed by a preexpressed polymerase. In fact, the average rescue efficiency obtained (55+/218%) (Figs. 1, 2) was very high, and is in line with the possibility that both types of soluble polymerase complexes, transcription-and replication-deficient, are incorporated in the progeny viral RNA, around half of which would not be detected because are not His-tagged. Non-resident polymerase complexes are responsible for the synthesis of vRNA in trans The rescue of viral RNPs containing the mutant R142A polymerase complex, as described above, enabled us to purify these RNPs and use them as templates for a second in vivo reconstitution experiment in which instead of a template RNA we introduced the rescued and purified R142A mutant RNPs in the system. This strategy ensured that only replication-defective RNPs are used as templates for in vivo replication and allowed us to ask whether the resident polymerase complex or a distinct, soluble polymerase is responsible for replication of RNPs in vivo. The concentration and biological activity of these purified RNPs was first controlled by Western-blot and in vitro transcription. The results are presented in Fig. 3B and show that higher yields were obtained for RNPs containing the E361A mutation in PB2 than those containing the R142A mutation. This was expected, as the latter could only be amplified by trans-complementation (see Fig. 1 above). The transcription phenotype of these purified RNPs was in agreement with the mutations present in PB2 (Fig. 3B, right panel). Therefore, cultures of HEK293T cells were co-transfected with purified RNPs containing either the R142A mutation or the E361A mutation in PB2, plasmids encoding PB1, PA, NP and a plasmid encoding either PB2-His R142A (replication-defective) or PB2-His E361A (transcription-defective) (see Fig. 3A for a diagram of the experimental setting). As controls, the RNPs were cotransfected with empty pCMV vector or the expression plasmids were transfected in the absence of template RNPs. The intracellular accumulation of progeny RNPs was determined by Ni 2+ -NTA-agarose purification, Western-blot and in vitro transcription as indicated above and the results are presented in Fig. 3C and To verify these results and to determine the polarity of the progeny RNA, similar experiments were carried out and the RNA present in the purified his-RNPs was analysed by hybridisation with positive-and negative-polarity RNA probes comprising the NS sequence. The results reinforced the data obtained by Western-blot and indicated that most of the progeny RNPs are vRNPs (Fig. 4), as previously reported [15]. The accumulation and phenotype of the progeny RNPs was also verified by in vitro transcription using either ApG or b-globin as primers (Fig. 5).

The accumulations observed paralleled those shown in Fig. 3 Fig. 1 were tested for in vitro transcription primed with either ApG (red) or b-globin mRNA (green). The data are presented as percent of maximal value and represent the averages and ranges of two independent complementation experiments. The transcription activities parallel the values of NP accumulation presented in Fig. 1 In some cases, purified RNPs containing replication-deficient or transcription-deficient polymerase were also transfected. The expected progeny RNPs are also depicted, depending on the indicate that a polymerase complex distinct from that present in the template RNP can perform the replicative synthesis of viral RNA. The high level of replication detected by trans-complementation suggests that virus RNA replication mostly occurs in trans. It could be argued that the mutation R142A in PB2 might destabilise the polymerase-promoter complex, allowing the efficient replacement by a polymerase complex containing the E361A mutation. However, RNPs containing the R142A mutation are as efficient in transcription as wt RNPs, suggesting that they are not affected in promoter recognition. Transcription of vRNPs takes place in cis and cannot be stimulated by non-resident polymerase complexes in trans It is well established that vRNPs can transcribe mRNAs in the absence of any newly synthesised viral proteins (primary transcription) [39,40] and highly purified recombinant RNPs can transcribe in vitro [18] (R. Coloma, unpublished results). However, it is not clear whether transcription takes place intramolecularly, i.e. in cis, or a RNP can transcribe another RNP. To test this possibility we reconstituted in vivo two genetically distinct RNPs, one containing a cat virus replicon (with the cat negative-polarity ORF flanked by the UTRs of the NS segment of influenza virus), the other one being the NS deletion mutant clone 23 [14,15]. Both RNPs contained a His-tagged PB2 subunit to allow purification by affinity chromatography as indicated above but two PB2 alleles were used, either wt or mutant E361A, which is defective in the recognition of the cap structure [20]. Purified RNPs were used either separately or in combination for in vitro transcription with ApG or b-globin as primers and the transcription products were analysed by denaturing polyacrylamide gel electrophoresis and autoradiography. The results are presented in Fig. 6. As expected, the purified wt RNPs were active, both when ApG or b-globin were used as primers (Fig. 6A). The RNPs containing the mutation PB2 E361A could transcribe mRNA with ApG as primer, but did so less efficiently when using b-globin mRNA as primer donor (Fig. 6A). These results allowed us to test whether a purified, wt clone 23 RNP could rescue the transcription activity of mutant E361A cat RNP in trans, since the mRNA products would be distinguishable by size (720 nt versus 248 nt). The wild-type cat RNPs could transcribe efficiently, both when incubated on their own and when mixed with clone 23 RNPs (Fig. 6B, middle panel). The cat RNPs containing PB2 E361A only produced background transcription levels and no increase in the amount of cat mRNA was observed when wt clone 23 RNP was co-transcribed (Fig. 6B, right panel). Quantisation of the relevant bands indicated that the increase in cat transcript in the co-transcription of clone 23 RNP+E361 cat RNP versus the transcription of E361 cat RNP was less than 3% of the cat transcript value obtained by co-transcription of clone 23 RNP+wt cat RNP. These results suggest that, at least in vitro, no transcription in trans among different RNPs takes place. However, the possibility still persists that a soluble polymerase complex is able to transcribe a vRNP template in trans. To analyse this alternative we generated in vivo recombinant RNPs containing the negative polarity cat virus replicon, purified them by affinity chromatography as indicated above and used them to transfect HEK293T cultures. Alternatively, the cultures were co transfected with the purified cat-containing RNPs and plasmids expressing the polymerase subunits (see Fig. 7A for a diagram of the experiment). As no plasmid expressing NP was used, no in vivo replication of the RNPs can take place [41,42]. Two RNP versions were used, either wt or transcription-defective (containing PB2 E361A mutant). Three alternative alleles were used to express in vivo PB2, generating wt polymerase, transcription-defective E361A or replication-defective R142A polymerase complexes, and various RNP-polymerase combinations were used in cotransfection experiments. In this way the experiment would mimic the situation of primary transcription (transfection of purified RNPs) or secondary transcription (co-transfection of RNPs with plasmids expressing the polymerase complex). At 24 hours posttransfection total cell extracts were prepared and the CAT protein accumulation was determined by ELISA. To ensure that the purified RNPs used for transfection were biologically active, two assays were carried out. First, their transcription activity was determined in vitro. As shown in Fig. 7B, there was a good correlation between the concentration of the RNPs, as determined by Western-blot with anti-NP and anti-PA antibodies, and the their capacity to synthesise RNA in vitro. Furthermore, the relative activity when using ApG or b-globin mRNA as primers verified that the purified mutant RNPs contained the E361A mutation (Fig. 7B, 361). In addition, the biological activity of the purified 361 RNPs was verified in vivo, by their co-transfection with plasmids expressing the polymerase subunits and NP. The results are presented in Fig. S4 and indicate that they can serve as templates for replication and transcription in vivo. Expression of the polymerase subunits did not yield any detectable CAT protein, as expected (Fig. 7C, Pol wt), but the transfection of wt purified RNPs lead to clearly measurable CAT accumulation (Fig. 7C, RNP wt) and co-expression of wt RNPs with wt polymerase did not lead to any increase of CAT accumulation (Fig. 7C, Pol wt-RNP wt). As control transfections with CAT-containing cellular extracts indicated that the carry-over of the protein was in the range of 10 23 to 10 24 (data not shown), the CAT protein generated by transfection of wt purified RNPs should represent primary transcription. In agreement with their transcriptiondefective phenotype, transfection of purified mutant 361 RNPs produced much less CAT accumulation (Fig. 7C, RNP 361). No significant increase in the level of CAT protein was observed by co-transfection of the RNPs containing the E361A mutation with polymerase-expressing plasmids, neither wt nor mutant polymerase and no correlation was observed between the accumulation of CAT and the phenotype of the polymerase co-expressed ( Discussion The processes of virus RNA replication and transcription usually require the action of one to several virus-specific proteins, notably the RNA-dependent RNA polymerase (RdRp), and various host cell factors (for a review see [43]. To unravel the complex procedures involved, genetic experimental approaches have been particularly useful. For example, genetic data have strongly supported the requirement of RdRp oligomerisation for RNA replication in several virus groups, like poliovirus [44,45], HCV [46,47] and Sendai virus [48,49]. Early studies on the dominance of RNA-synthesis negative ts mutants of VSV suggested that the oligomerisation of virus factors involved in RNA replication is an essential step in the process [50], a conclusion that could be also verified in the poliovirus system [51]. More generally, the multimeric nature of complex viral systems, as the virus particles, has profound consequences in the apparent phenotype observed [52,53]. In the case of influenza, early data on the intragenic complementation of mutants affecting the PB1 and PA proteins suggested the potential role of virus polymerase interactions in the infectious cycle [54][55][56] and the recent biochemical evidence for virus polymerase oligomerisation supported such contention [57]. Here we have taken advantage of the availability of well-established recombinant systems for RNP replication and transcription and wellcharacterised polymerase mutants to address specific questions on the mechanisms of these processes. Due to the segmented nature of the influenza virus genome it is essential to use mutant polymerases having phenotypically distinct mutations in the same subunit, thus avoiding the problems of reassortment. Hence, we have used point mutants of polymerase PB2 subunit that abolish RNA replication but transcribe normally (R142A or F130A) [36] and/or mutants that are defective in cap-recognition and transcribe poorly, but replicate virus RNA normally (E361A among others) [20]. With these experimental tools we have asked whether the polymerase complex present in an RNP actually perform the replicative or transcriptional synthesis of RNA and whether the polymerase complex present in the progeny RNP is identical to that performing replicative synthesis of RNA. Our results will be discussed on the basis of the model presented in Fig. 8, in which only the replication step cRNP-to-vRNP is presented. The results shown in Figs. 1 and 2 indicated that two such phenotypically distinct mutant polymerases can complement to perform viral RNP replication in vivo and demonstrated that a replication-defective polymerase can be incorporated into progeny RNPs. These results are consistent with the model presented in Fig. 8A, step 4, that suggest that a polymerase complex distinct from that performing replicative synthesis is involved in the recognition of the 59-end of the progeny vRNA. This model is also consistent with the results published earlier indicating that a pre-expressed polymerase can protect newly synthesised cRNA [13,38]. The identity of the replicative polymerase complex could be tested by directly transfecting mutant RNPs as templates for the replication reaction and asking whether coexpressed replication-defective or transcription-defective polymerase complexes could carry out the replication process in trans. The results shown in Figs. 3 and 4 demonstrated that a polymerase complex genetically distinguishable from that present in the parental RNP was able to perform replication and became incorporated into the progeny RNPs. These results are compatible with the model presented in Fig. 8A, steps 2-4, whereby a soluble polymerase complex would interact with that resident in the parental RNP and gain access to the 39-terminal sequence in the promoter. Such polymerase-polymerase interaction is supported by the genetic data presented here, by the intragenic complementation reported earlier [54,55] and by the oligomerisation of influenza polymerase in vivo [57]. Although not shown in Fig. 8A, we can not exclude that a host factor(s) participate in the polymerase-polymerase interaction and in fact several nuclear factors have been described previously that could play such a role [25][26][27][28][29][30]32]. The trans-replication model depicted in Fig. 8A, steps 2-4 relates to the cRNP-to-vRNP phase in replication. However, earlier data published on the protection of newly synthesised cRNA by pre-expressed polymerase would suggest that the vRNP-to-cRNA phase can occur in cis, since a pre-expressed, catalytically inactive polymerase allowed the accumulation of cRNA in cicloheximide-treated, virus-infected cells [13]. According to the model proposed here, the soluble polymerase complex would act as replicative enzyme by de novo initiation and elongation through the NP-RNA template (Fig. 8, steps 3-4). We propose that the 39-end of the parental RNA is used repetitively for further initiation rounds, thereby leading to several progeny vRNPs generated from a single cRNP template. For simplicity, the model presented in Fig. 8A does not show the interaction of the new incoming replicative complex with the parental polymerase bound to the 59-end of the template, but such interaction might be required. In view of our previous evidence on polymerase-polymerase interaction [57], an appealing possibility is that the replicative polymerase complexes would oligomerise to form a fixed replication platform along which the NP-RNA template would move 39-to-59 to generate many progeny vRNPs. Such strategy has a precedent in other positive-strand RNA viruses [44][45][46][47] and would be consistent with the localised synthesis of influenza virus RNA in the nucleus [58,59]. A critical point in the generation of progeny vRNP is the recognition and packaging of its newly synthesised 59-end. Our results are compatible with the proposal that a polymerase complex distinct from the replicative enzyme can protect the newly synthesised vRNA (Fig. 8A, step 4) and this event probably represents the sequencespecific step in the encapsidation of RNA into progeny RNP. The subsequent incorporation of successive NP monomers would be directed by polymerase-NP interactions [60], that have been shown as essential for RNP replication [61,62], as well as by the NP-NP oligomerisation [17] (R. Coloma, unpublished results). Another critical point in the replication process is the displacement of the parental polymerase complex bound to the 59-end of the template, a step necessary to avoid polyadenylation (see below) and to allow the complete copy of the RNA. Our results do not permit us to distinguish whether the elimination of such interaction is transient or permanent, but an attractive possibility would be that

the reiterative copy of the NP-RNA template on a fixed platform of replicative polymerases would force the displacement of the parental polymerase bound to the 59-terminal sequence. Such displacement could be transient, leading to the replacement of the parental polymerase by each successive replicative polymerase, or permanent, leading to a linearised NP-RNA complex (Fig. 8A, step n). In contrast to the positive complementation obtained for the replication process, no trans-complementation could be detected in the transcription assays using either in vitro (Fig. 6) or in vivo experiments (Fig. 7). In vitro transcription of a recombinant RNP containing a cap-binding defective polymerase could not be rescued by a wt RNP holding a template of different length (Fig. 6). Similar negative results were obtained in vivo, by transfection of a cap-binding defective RNP and co-expression of wt or replicationdefective but transcriptionally functional polymerase (Fig. 7). These results are not compatible with the possibility of transcription among viral RNPs in trans and do not support the possibility of a soluble polymerase transcribing an independent RNP. Furthermore, these results indicate a high stability of the polymerase binding to the RNP structure during the transcription process, as no polymerase exchange could be functionally detected. In view of the lack of detectable transcription in trans, we propose the model presented in Fig. 8B for the generation of viral mRNAs. The resident polymerase complex would be transcriptionally activated by recognition of the cap-containing cellular mRNA and proceed to cap-snatching and elongation of the virus transcript (Fig. 8B, step 2), but still keeping hold of the 59terminal sequence of the promoter [63]. Such process would lead to a running knot structure with a diminishing loop length (Fig. 8B, steps 2-4) until the transcribing polymerase reaches the oligo-U polyadenylation signal [10]. Due to steric hindrance, the polymerase would stutter around the oligo-U sequence and generate a 39-terminal polyA (Fig. 8B, step 4). For simplicity, the model presented in Fig. 8B shows the transcribing RNP in a linearised form, but the polymerase complex should recognise the 39-terminal side of the promoter at some time later in the process, in order to recycle and allow further rounds of transcription. This model for transcription of RNPs in cis is compatible with the fact that parental RNPs perform primary transcription as a first step in the infection and with the possibility to rescue virus by transfection of purified virion and/or recombinant RNPs [64,65]. It also would fit with the correlation of vRNA and mRNA levels of the various RNA segments along the infection cycle [66,67]. In summary, using genetically distinct RNA polymerase complexes, we have presented direct evidence for trans-comple-mentation during the influenza virus RNA replication process. These results are compatible with a new model for viral RNA replication whereby a template RNP would be replicated in trans by a soluble polymerase complex and the progeny RNP encapsidation would be specified in trans by a polymerase complex distinct from the replicative enzyme. In contrast, no transcription in trans could be detected in vitro or in vivo and hence we propose a model for cis-transcription of the RNPs whereby the resident polymerase complex would be responsible for mRNA synthesis and polyadenylation. Biological materials The HEK293T cell line [68] was used throughout. The origins of plasmids pCMVPB1, pCMVPB2, pCMVPB2His, pCMVPA and pCMVNP, as well as pHHclone 23, have been described [20,64]. Plasmid pHHCAT was kindly provided by A. Rodriguez. The antibodies specific for PB2 and PA have been described [69,70]. Antibodies specific for NP were prepared by immunisation with purified His-NP protein. Mutant PB2 plasmids including mutations in the N-terminal region [36] or the cap-binding site [20] have been reported earlier. The mutations F130A, R142A, E361A, H357A, K370A and F404A were transferred to pCMVPB2 by swapping the appropriate restriction fragments. The genotype of the mutant plasmids was verified by sequencing. Biochemical techniques Western-blotting was performed as described [30]. The replication of RNPs in vivo was determined as described [20]. In brief, cultures of HEK293T cells were transfected with plasmids pCMVPB1, pCMVPB2His (or mutants thereof) or pCMVPB2 (or mutants thereof), pCMVPA, pCMVNP and pHHclone 23. In some experiments pHHclone 23 plasmid was omitted and purified clone 23 RNPs were transfected instead, 24 hours after plasmid transfection. Total cell extracts were prepared at 24 hours posttransfection and used for purification by affinity chromatography on Ni 2+ -NTA-agarose as indicated above and the accumulation of progeny RNPs was determined by Western-blot with anti-NPspecific antibodies. The transcription of RNPs in vivo was assayed by transfection of purified NSCAT RNPs into HEK293T cells. The cultures were first transfected with plasmids pCMVPB1, pCMVPB2 (or mutants thereof) and pCMVPA [20] and 24 hours later were further transfected with purified His-tagged NSCAT RNPs under the same conditions. At 24 hours post RNP-transfection, total cell extracts were prepared and the CAT protein concentration was determined by ELISA (GE Healthcare). RNA analyses To determine the transcription activity of purified RNPs, samples were incubated in a buffer containing 50 mM Tris-HCl-5 mM MgCl2-100 mM KCl-1 mM DTT-10 mg/ml actinomycin D-1 u/ml RNAsin-1 mM ATP-1 mM CTP-1 mM UTP-10 mM a-P 32 -GTP (20 mCi/mmol) and either 100 mM ApG or 10 mg/ml b-globin mRNA, for 60 min at 30uC. The RNA synthesised was TCA precipitated, filtered through a nylon filter in a dot-blot apparatus and quantified in a phosphorimager. To analyse the transcription products, similar reactions were carried out but the specific activity of the labelled GTP was increased to 200 mCi/ mmol. The synthesised RNA was isolated by treatment with proteinase K (50 mg/ml) for 30 min at 37uC in TNE-1% SdS and phenol extraction. The RNA was ethanol precipitated, resuspended in formamide loading buffer and analysed by electrophoresis in a 4% polyacrylamide-urea denaturing gel. To analyse the progeny RNA, purified RNPs were incubated with proteinase K (200 mg/ml) in a buffer containing 100 mM NaCl-5 mM EDTA-0.5% SDS-50 mM Tris.HCl, pH 7.5 for 60 min at 37uC, phenol extracted with ethanol precipitated. Samples of the purified RNAs were denatured by boiling for 3 min in 7.5% formaldehyde-10SSC and were fixed onto nylon filters. Replicate filters were hybridised at 37uC with full-length NS riboprobes of either positive-or negative-polarity in a buffer containing 6SSC-40% formamide-0.5% SDS-5xDenhart's mixture-100 mg/ml single-stranded DNA. After washing at 60uC with 0.1SSC-0.1%SDS, hybridisation signals were quantitated in a phosphorimager. Figure S1 Expression of wild-type and mutant PB2 proteins. Cultures of HEK293T cells were transfected with plasmids encoding wt or mutant PB2 proteins, as indicated. Total cell extracts were prepared and analysed by Western-blot using anti-PB2 antibodies as described in Materials and Methods. The position of the PB2-specific signals of His-tagged (PB2His) or untagged PB2 is indicated to the right. To verify the biological activity of the PB2 E361A RNPs used in the experiments described in Fig. 7, cells were transfected with polymerase subunits and NP-expressing plasmids and further transfected with the purified RNPs. At 24 h post-transfection of the latter cell extracts were prepared and the CAT protein accumulation was determined by ELISA. As control, single transfection with polymerase subunit and NP-expressing plasmids was performed in parallel. The Impacts of Servant Leadership and Organizational Politics on Burnout: A Research among Mid-Level Managers Burnout, as a factor related to an individual’s success, is defined as being exposed to behaviors that negatively affect performance and mainly refers to chronic stress. This research focuses on the effects of servant leadership on burnout through organizational politics. Data for the study were collected by survey method with participation of 401 employees from 49 different organizations. Obtained data were subjected to factor analysis with SPSS in order to check the internal consistency of the sample. Thereafter, by using AMOS software, confirmatory factor analysis were conducted and with structural equation modeling method research hypothesis of the study were tested. The results revealed that burnout was not emerged by servant leadership through organizational politics. Even though servant leadership is related to a change by creating a vision, this change process does not affect employees negatively because it comprises long-term plans and occurs spontaneously. Introduction As in many areas, information age has led to changes in leadership styles. Today's leaders, who want their followers to perform at higher levels and collaborate to achieve company goals, serve to their followers to achieve the best. Robert K. Greenleaf defined this leadership style as "servant leadership" and highlighted that for those leaders serving to followers and organizations become a way of their life. Servant leaders encourage their followers to generate new ideas while solving organizational problems and they place importance to collective decision-making (Stone, Russell, & Patterson, 2004). Furthermore, servant leaders emphasize objectiveness for the benefit of employees and they put information-oriented working forward. Moreover, this leadership is related to environmental factors; for instance, in static and stable environments creating the culture of servant leadership is recommended (Smith, Montagno, & Kuzmenko, 2004). On the other hand, there are some similarities between transformational leadership and servant leadership such as teaching effort, respect, confidence, vision, integration, and effects on followers. The focus of the leader is the main difference between transformational leadership and servant leadership theories. Transformational leaders focus on the behaviors of employees and organizations, and they create leader-follower commitment to achieve organizational goals. On the other hand, servant leaders focus on well-being of their followers and they believe that organizational goals can only be achieved by doing this. While the focus of servant leadership is serving to followers, transformational leadership focuses on the organization that followers belong to and eventually achieving organizational goals. Servant leaders also care about individuals instead of abstract concepts and this shows that these leaders value the contents of the processes rather than the results (Stone et al., 2004). Servant Leadership Although studies regarding servant leadership were revealed first by Robert K. Greenleaf in the beginning of 19th century, the concept gained its importance in '90s and created a quiet revolution in workplaces all around the world. While servant leaders increase the quality of organizational outcomes, they try to develop individuals as well (Spears, 2004). In addition, another characteristic of servant leadership is conceptualization that means having great dreams and looking beyond the daily realities. Likewise, "empathy", "awareness", and "focus on human development" are also among the basic characteristics of servant leadership (Spears, 2004;Stone et al. (2004). Some prior studies questioned whom is served by servant leaders and which factors make leaders as servant leaders. In these studies, "being" and "doing" attributes were determined as significant features of servant leaders and these attributes comprises the leader's self-concept and primary intent. It was stated that primary intent (doing) comprehends "serving others first, not leading" and self-concept (being) comprehends "being servant and steward, not leader or owner" (Sendjaya & Sarros, 2002). It is possible to infer that major goals of servant leaders are to determine what their followers need and eliminate their deficiencies in order to create a powerful organization (Vinod & Sudhakar, 2011). Organizational Politics Although being an effective concept, organizational politics has taken relatively little place in the literature until 1970s. However, later studies indicate that organizational politics should be analyzed for a better understanding of organizations (Ullah, Jafri, Gondal, & Khyzer Bin Dost, 2011). Recent empirical studies support that employees' ideas regarding organizational politics are better understood by focusing on perceived politics instead of applied tactics (Vigoda-Gadot & Talmud, 2010). The concept of organizational politics can be defined from different perspectives. While it's related to the activities carried out to achieve specified targets in public institutions, in private organizations it comprises striving to achieve personal goals instead of organization's (Andrews & Kacmar, 2001). Political struggles in organizations mostly depend on the characteristics of those organizations rather than individual differences. However, many researches put forward that situational conditions and organizational culture evoke political behaviors. For instance political behaviors can be observed when the sources of organizations diminish in time. Some studies revealed that organizational politics prevent achieving organizational goals because it comprises functional difficulties that caused by external authorities. These external authorities benefit from conflict, power struggles, and tactical battles and they cause more complicated problems in decision-making (Ullah et al., 2011). It can be said that political behaviors derive from mature dilemmas (Jafariania, Mortazavib, Nazemic, & Bulld, 2012), for instance in an environment where individuals tend to achieve their own goals rather than achieving organizational goals or

especially in organizations where individuality is valued (Cropanzano, Howes, Grandey, & Toth, 1997). Therefore the structure of an organization should comply with legal and other regulations and cooperation within the organization should be ensured. In addition, organizations may become a structure formed from the impact of negative public and social sanctions in case of violation of certain regulations (Randall, Cropanzano, Bormann, & Birjulin, 1999). The increase in these sanctions may cause an increase in fear and anxiety and also exhaustion of individuals inside organizations . Furthermore, decisions made in large enterprises are often dictated by politics and participation in political decision-making is important as a matter of life and death (Ullah et al., 2011). Previous findings showed that employees' perceived trust and social support have an important effect on the relationship between organizational politics and various job-related outcomes (job satisfaction, organizational commitment, stress, burnout etc.). In other words, if trust and social support are dominant characteristics in the organization then the negative effects of organizational politics are expected to decrease in time (Vigoda-Gadot & Talmud, 2010). In this study Vigoda-Gadot's organizational politics scale was used and the sub-dimensions of the scale were determined as fair in pay & promotion, fair reward and fair opinion. For sure, the deterioration in the perceptions of fairness derives from the disparity between initially perceived justice and the results (Whisenant, 2005). The term "fair pay & promotion" means that all employees at the same level have equal pay and promotion opportunities. Similarly, "fair reward" refer to providing same rewards in terms of financial and non-financial to the individuals who perform equal. Finally, "fair opinion" states that employees are not evaluated and treated according to their ideas or opinions. The study is based on the assumption that individuals will not behave politically regarding these three conditions since those behaviors will also be restricted to external factors (Jafariani et al., 2012). Burnout The concept of burnout was first mentioned by Freudenberger (1974) and was conceptualized as an "occupational hazard". Freudenberger described burnout as "a loss of energy or power due to failure, degradation, and overload or exhaustion of individual's internal resources due to unmet demands" (as cited in Sağlam Arı & Çına Bal, 2008 Results As employees' perceptions regarding their leader's persuasion and encouragement characteristics increase, their perceptions regarding the leader's fairness in organizational politics also increase thus burnout feelings is less experienced. Therefore H1a, H1b, H1c, H1d, H1e, H1f, H3a, H3b, H3c and H3d were accepted and H3e and H3f were rejected. Despite that, relationship between leaders' visionary characteristics and fairness in organizational politics was found to be insignificant. Thus, H2a, H2c (partially), H2e, and H2f were accepted, and H2b and H2d were rejected. On the other hand, through the results it can be said that as the perception of servant leader's vision increases, burnout decreases. Another finding is that there is no significant relationship between communication capability of servant leaders and the perception of fair rewards. In addition, as servant leaders provide better communication with their followers, perceptions regarding fair pay & promotion and fair opinion were positively affected. The insignificant relationship between communication and fair rewards could be explained by emphasizing that employees generally focus on organizational practices and they care less about temporal issues such as rewards. Thus H4a, H4b, and H4c were accepted. PHLPP1 promotes neutral lipid accumulation through AMPK/ChREBP-dependent lipid uptake and fatty acid synthesis pathways Summary Infiltration of arterial intima by foamy macrophages is a hallmark of early atherosclerotic lesions. Here, we investigated the potential role of Ser/Thr phosphatase PHLPP1 in foam cell development. PHLPP1 levels were elevated in OxLDL-exposed macrophages and high-fat diet (HFD)-fed zebrafish larvae. Using overexpression and knockdown approaches, we show that PHLPP1 promotes the accumulation of neutral lipids, and augments cellular total cholesterol and free fatty acid (FFA) levels. RNA-Seq analysis uncovered PHLPP1 role in lipid metabolism pathways. PHLPP1 interacted with and modestly increased ChREBP recruitment to Fasn promoter. PHLPP1-mediated lipid accumulation was attenuated by AMPK activation. Pharmacological inhibition or CRISPR/Cas9-mediated disruption of PHLPP1 resulted in lower lipid accumulation in the intersegmental vessels of HFD-fed zebrafish larvae along with a reduction in total cholesterol and triglyceride levels. Deficiency of phlp-2, C. elegans PHLPP1/2 ortholog, abolished lipid accumulation in high cholesterol-fed worms. We conclude that PHLPP1 exerts a significant effect on lipid buildup. INTRODUCTION Foam cells, the hallmark of atherosclerosis, are lipid-laden macrophages that build up and form plaques leading to cardiovascular diseases (CVDs), which contribute to the highest global mortality (Aviram, 1999;Barquera et al., 2015). The formation of foam cells is initiated via macrophage engulfment of modified LDL (oxidized LDL, OxLDL) on its binding to either CD36, LDLR, and scavenger receptors class A (SR-A) promoting dramatic changes in cellular lipid metabolism (Chistiakov et al., 2017). The engulfed OxLDL enters lysosomes, is metabolized to free cholesterol and subsequently might form cytotoxic crystals (Tabas, 1997). The free cholesterol is effluxed out via ATP-binding cassette (ABC) transporters such as ABCG1 and ABCA1 or further gets converted into nontoxic cholesteryl ester by Sterol O-acyltransferase I (SOAT1) (Ouimet and Marcel, 2012). Neutral cholesteryl ester hydrolase (nCEH) aids in the reconversion of cholesteryl ester to cholesterol in blood and the effluxed cholesterol is tracked into reverse cholesterol transport (Cuchel and Rader, 2006;Jeong et al., 2017). The balance between uptake, cholesteryl ester formation, and efflux decides the cell 0 s fate (Yuan et al., 2012). PH domain and leucine-rich repeat protein phosphatase (PHLPP) exist as two isozymes, PHLPP1 and PHLPP2 with 50% global amino acid sequence identity. PHLPP1 further comprises two alternatively spliced variants PHLPP1a and PHLPP1b, differing in their amino terminus. Although the tumor suppressor role of PHLPP1 is actively pursued, its involvement in different facets of metabolic syndrome remains largely unexplored. Toward this, our laboratory has previously shown that PHLPP1 augments skeletal muscle insulin resistance and impairs glucose homeostasis by inducing ER stress, a key driver of metabolic syndrome (Behera et al., 2018). Furthermore, findings from different research groups including ours highlight the importance of PHLPP1 in controlling macrophage inflammatory responses (Alamuru-Yellapragada et al., 2017a;Alamuru et al., 2014;Katsenelson et al., 2019), prompting us to investigate the significance of PHLPP1 in the development of foam cells which display abnormal lipid accumulation and altered inflammatory program. In vitro foam cells show altered PHLPP1 levels We and others have previously reported that PHLPP1 restrains pro-inflammatory responses of macrophages which are commonly dysregulated in atherosclerosis (Alamuru-Yellapragada et al., 2017a;Katsenelson et al., 2019). Thus, we investigated the possible involvement of PHLPP1 in the development of foam cells, a hallmark of atherosclerosis. For this, we examined the expression of PHLPP1 in in vitro foam cells using OxLDL-exposed macrophages as the model system. In line with our earlier observation (Alamuru-Yellapragada et al., 2017a), stimulation of RAW 264.7 cells with bacterial endotoxin for 24 h depleted PHLPP1 protein levels ( Figure 1A). Furthermore, treatment of RAW 264.7 cells with OxLDL resulted in the dose-dependent elevation of PHLPP1 protein levels up to 24 h ( Figure 1A). PHLPP1 transcripts were also enhanced at 24 h as revealed by qPCR analysis (Figures 1B and 1C). Post-24 h of OxLDL exposure, PHLPP1 mRNA and protein levels were reduced (Figures 1C and 1E). Elevated circulating 7-ketocholesterol, an abundant oxysterol found in atherosclerotic plaques and processed cholesterol-rich foods, is associated with heightened cardiovascular risk and death (Hayden et al., 2002;Rao et al., 2014). Similar to OxLDL treatment, 7-ketocholesterol treatment also enhanced PHLPP1 levels up to 24 h ( Figure 1D), which declined 48 h posttreatment ( Figure 1E). The levels of fatty acid synthase (FASN) and the lipogenic transcription factor carbohydrate responsive element binding protein (ChREBP) were also augmented in 7-ketocholesterol exposed macrophages. Consistent with the role of 7-ketocholesterol in ER stress induction that supports foam cell formation (Pedruzzi et al., 2004), ER stress markers such as CHOP, ATF6a, and TRB3 were elevated up to 6-9 h post-7-ketocholesterol treatment ( Figures 1D and S1C). In accordance with our previous findings, PHLPP1 levels increased up to 3 h and then declined at both mRNA and protein levels on exposure to the ER stress inducer thapsigargin (Figures S1A and S1B) (Behera et al., 2018). Parallel experiments confirmed the induction of ER stress in thapsigargin-treated cells ( Figure S1B). Carbohydrate responsive element binding protein functions as a transcriptional activator of PHLPP1 ChREBP, also known as MLX-interacting protein-like (MLXIPL), is a key transcriptional activator regulating the glucose-dependent expression of triglyceride and fatty acid synthesis genes. Immunoblot analysis of 7-ketocholesterol-treated macrophages showed similar induction patterns of PHLPP1 and ChREBP proteins suggesting the possibility of ChREBP-mediated upregulation of PHLPP1 ( Figure 1D). Computational analysis of the PHLPP1 promoter revealed an evolutionarily conserved ChREBP binding site between À488 and À473 bp ( Figure S2). Therefore, we next determined the effect of ChREBP on PHLPP1 levels using Western blot, qPCR, and ChIP analysis. Overexpression of flag-tagged ChREBP in RAW 264.7 cells upregulated PHLPP1 protein and mRNA levels ( Figures 1F and 1G). Consistently, PHLPP1 and FASN protein levels were reduced in OxLDL and 7-ketocholesterol-treated cells upon ChREBP knockdown (Figures 1H and S1D). ChIP experiments confirmed ChREBP recruitment to Phlpp1 promoter ( Figure 1I). PHLPP1 escalates lipid accumulation in in vitro foam cells We next examined neutral lipid accumulation (cholesteryl ester and triglycerides) in RAW 264.7 cells upon PHLPP1 overexpression or knockdown using Oil Red O staining. Forced expression of PHLPP1a enhanced neutral lipid accumulation in OxLDL-treated cells as well (Figures 2A and 2B). Neutral lipid buildup in OxLDL-exposed PHLPP1-ablated cells was robustly reduced to levels observed in untreated control cells (Figures 2A and 2C). Overexpression of PHLPP1a or PHLPP1b but not PHLPP2 increased basal levels of neutral lipids (Figures 2A-2D). Additional analysis showed that ectopic PHLPP1 expression enhanced total cholesterol and FFA levels in untreated RAW 264.7 macrophages ( Figures 2E and 2F). RNA-seq analysis reveals the role of PHLPP1 in lipid and cholesterol metabolism RNA-seq analysis of differential gene expression under three different experimental conditions was performed ( Figure 3A): set 1: control cells and OxLDL-treated control cells (100 mg/mL) for 24 h; set 2: control cells and PHLPP1-knockdown cells; set 3: OxLDL-treated control cells and OxLDL-treated iScience 25, 103766, February 18, 2022 iScience Article PHLPP1-knockdown cells ((100 mg/mL) for 24 h). Volcano plots presented in Figure 3B show robust differential regulation of genes under the three conditions. With a 2-fold change and FDR < 0.05 as the cut-off, 121, 569, and 1,303 genes were significantly altered in sets 1, 2, and 3, respectively ( Figure 3C). Importantly, supporting the findings in Figure 1, fatty acid biosynthesis, fatty acid elongation, fluid shear stress & atherosclerosis, and biosynthesis of unsaturated fatty acid pathways were highly enriched among the Top 25 hits along with those associated with reported roles of PHLPP1 such as apoptosis and insulin signaling among others in set 3 ( Figure 3D). Several cholesterol biosynthesis genes were downregulated in OxLDL-treated PHLPP1-knockdown cells ( Figure 3E). HMG-CoA reductase, a rate-limiting enzyme in the cholesterol biosynthesis pathway and the target of statins and AMPK-mediated inhibitory phosphorylation, showed a prominent decrease. Furthermore, the expression of another rate-limiting enzyme, squalene epoxidase (SQE), and terminal with densitometric quantification (bottom) and mRNA (B and C) levels after OxLDL treatment in RAW 264.7 cells; Statistical analysis was performed using one-way ANOVA followed by Bonferroni multiple comparisons test for (C) (*p < 0.05 vs Control, 0 h; a p < 0.05 vs 24 h). (D) Protein levels of PHLPP1, ChREBP, and ER stress markers upon 7-ketocholesterol treatment under different time points. Densitometric analysis is shown in Figure S1C. (E) PHLPP1 protein levels upon 48 h OxLDL and 7-ketocholesterol treatment along with densitometric quantification (Bottom). (F and G) PHLPP1 protein (F) and mRNA (G) levels upon ChREBP overexpression; Statistical analysis was performed using two-tailed t-test for (G) (*p < 0.05 vs Vector). (H) PHLPP1 and FASN protein levels in ChREBP knockdown cells treated with 7-kC or OxLDL. Densitometric analysis is shown in Figure S1D. iScience Article enzymes in cholesterol biosynthesis (DHCR7 and DHCR24) were also reduced in PHLPP1-ablated cells. Fatty uptake and synthesis-associated genes such as CD36 (OxLDL uptake), Fasn (fatty acid synthesis), ELOVL6 (fatty acid elongation), and SCD1 (fatty acid desaturation) were reduced in PHLPP1-deficient cells ( Figures 3E and S3). Supporting the role of PHLPP1 in restraining macrophage pro-inflammatory responses, PHLPP1-knockdown cells showed significant upregulation of genes including IL6, IL1-b, CD40, and CD86 on OxLDL

exposure ( Figure 3E). Expectedly, pathway analysis for set one also listed atherosclerosis condition as one of the top 20 hits; Furthermore, we observed that synthesis of unsaturated fatty acids and AMPK signaling, both pathways associated with lipid metabolism, are among the top 20 hits in untreated PHLPP1 knockdown cells ( Figures S3A and S3B). (D) Neutral lipid staining in cells overexpressing PHLPP isoforms; Statistical analysis was performed using one-way ANOVA followed by Dunnett 0 s multiple comparisons test (*p < 0.05 vs Vector). (E and F) Total cellular cholesterol (E) and FFA levels (F) in RAW 264.7 cells upon PHLPP1 overexpression; Statistical analysis was performed using two-tailed t-test (*p < 0.05, **p < 0.01 vs Vector). Data are a representative of at least three independent experiments. Numerical data are expressed as mean G SEM. iScience Article knockdown substantially blocked OxLDL uptake by RAW 264.7 cells ( Figure 4B). Protein levels of CD36, a selective receptor for cholesteryl ester uptake, and other lipid biosynthesis genes SOAT1 and FASN were elevated in macrophages upon PHLPP1 overexpression. Consistently, depletion of PHLPP1 by shRNA robustly reduced expression of CD36, SOAT1, and FASN ( Figure 4E). Very high PHLPP1 expression caused a reduction in protein levels of ABCG1, a lipid efflux transporter and as expected, PHLPP1 knockdown increased ABCG1 protein levels ( Figure 4E). However, levels of another cholesterol efflux transporter ABCA1 were unchanged upon PHLPP1 ectopic expression. Surprisingly, neither PHLPP1 overexpression nor PHLPP1 knockdown showed any major impact on fluorescent cholesterol efflux ( Figures 4C and 4D). Prolipogenic effects of PHLPP1 are mediated through carbohydrate responsive element binding protein and AMPK Based on the transcriptomics analysis, we reasoned that PHLPP1 may target a key lipid-metabolism-associated transcription factor. We analyzed the enrichment for transcription factors based on differentially (1) ChREBP is a key lipogenic transcriptional factor and was earlier reported to be critical for the regulation of lipid metabolism in liver, b-cells, and so forth. (Iizuka et al., 2020;da Silva Xavier et al., 2006). (2) PHLPP1 promoted the expression of ChREBP targets Fasn, ELOVL6, and so forth. (Figures 4E and S3C). Overexpression of ChREBP resulted in the robust accumulation of neutral lipids which was stronger than the accumulation observed in PHLPP1 overexpressing macrophages ( Figure 5D). Ectopic expression of ChREBP resulted in a subtle increase of cholesterol but a more prominent increase in FFAs (Figures 5E and 5F) along with the induction of FASN (da Silva Xavier et al., 2006) ( Figure 5G). However, CD36, SOAT1, and ABCG1 protein levels remain unchanged ( Figures 5G and S5A). Immunoprecipitation iScience Article experiments revealed an interaction between ectopically expressed PHLPP1 and ChREBP in HEK 293T cells ( Figure 5H). Pull-down of endogenously expressed ChREBP in RAW 264.7 cells also coprecipitated PHLPP1 validating the interaction between them ( Figure 5I). Incubation of PHLPP1 and ChREBP proteins enriched from HEK 239T cells resulted in higher phosphate release, suggesting that ChREBP may be a direct PHLPP1 substrate ( Figure 5J). Previous studies showed that the nuclear translocation and DNA binding activity of ChREBP is controlled by PKA and AMPK-mediated phosphorylation of ChREBP residues Ser 196 and Ser 568 (B and C) Plot and table indicating the -log 10 p value (B) and terms odds ratio, and combined score (C) used by the Enrichr software to represent the output (scatterplot). p values are computed based on alterations in the number of genes that are known primary or secondary targets of transcription factors; Odds ratio indicates a measure of association between input and outcome; combined score: log(p) *z, p is p value, and z = deviation from the expected rank. (D) Neutral lipid staining by Oil Red O in ChREBP overexpressing cells along with quantification (top); Images are taken at |3203| magnification. Each image is a representative image of four different experiments with a minimum of eight different fields in each experiment. Statistical analysis was performed using two-tailed t-test (*p < 0.05, vs Vector). (E and F) Total cellular FFA (E) and cholesterol levels (F) upon ChREBP overexpression. Statistical analysis was performed using two-tailed t-test (*p < 0.05, vs Vector). (G) Protein levels of foam cell markers in ChREBP overexpressing cells. Densitometric analysis is shown in Figure S5A. (H and I) Pull-down assay: HA-PHLPP1 and Flag-ChREBP interaction in HEK293T cells (H) and endogenous interaction (I) between PHLPP1 and ChREBP in RAW 264.7 cells; (J) Malachite green assay: phosphate release upon the incubation of Flag-ChREBP with SFB-PHLPP1 in vitro. Statistical analysis was performed using twotailed t-test (*p < 0.05, vs Vector). (K) AMPK Thr 172 phosphorylation in ChREBP or PHLPP1 overexpressing RAW 264.7 cells. Densitometric analysis is shown in Figure S5B. iScience Article residues, respectively (Sakiyama et al., 2008). Thus, to further examine the effect of PHLPP1 on ChREBP, ChIP assay was performed to monitor ChREBP recruitment to Fasn promoter. ChIP-qPCR analysis revealed an increase in ChREBP recruitment to Fasn promoter under basal conditions ( Figure 5L). The enrichment of ChREBP at Fasn promoter under basal conditions was modestly enhanced in PHLPP1-overexpressing cells. Overexpression of PHLPP1 or ChREBP reduced the phosphorylation of AMPK pThr 172 phosphorylation ( Figures 5K and S5B). AICAR-mediated AMPK activation lowered CD36 and FASN protein levels in PHLPP1 overexpressing cells ( Figures 5M and S5C). Treatment of cells with activator-3 (Act-3), which we recently described (Bung et al., 2018), to directly activate AMPK, also resulted in the reduction of CD36 and FASN protein levels despite PHLPP1 overexpression ( Figures 5M and S5D). Consequently, AICAR and Act-3 led to a significant reduction in lipid accumulation in OxLDL-treated PHLPP1 overexpressing cells ( Figure 5N). To our surprise, we noticed that ChREBP-dependent neutral lipid buildup in macrophages was not influenced by AMPK activation ( Figure 5N). Pharmacological inhibition of PHLPP1 reduced neutral lipid accumulation in high-fat diet-fed zebrafish larvae BLAST analysis showed high global sequence identity (Similarity 73.7; Identity 62.8%) between zebrafish and human PHLPP1, indicating high conservation and the utility of zebrafish as an in vivo model for investigating PHLPP1 function. Neutral lipid staining of 13-dpf HFD-fed larvae ( Figure 6A) using Nile Red revealed robust lipid accumulation in the head, trunk, tail, and intersegmental vessels (ISV) regions ( Figures 6B-6D). HFD feeding of zebrafish larvae resulted in an increase in total cholesterol and FFAs in larval homogenates ( Figures 6E and 6F). In addition, HFD-fed zebrafish larvae showed higher PHLPP1 protein levels ( Figure 6G). We assessed the importance of PHLPP1 for lipid accumulation in HFD-fed zebrafish larvae. Exposure of HFD-fed larvae to 10 mM NSC117079, a commonly used small molecule PHLPP inhibitor (Jackson et al., 2013), resulted in a substantial reduction in neutral lipid accumulation in the head, trunk, and tail regions of zebrafish larvae ( Figure 6H). Additionally, a marked reduction in the neutral lipid buildup of the ISV region of HFD-fed PHLPP1 inhibitor-treated larvae was evident ( Figure 6I). Although NSC117079 treatment resulted in a modest decrease in total cholesterol levels, the triglyceride levels were strikingly reduced in inhibitor-treated larvae ( Figure 6J). PHLPP1 deficiency attenuates lipid accumulation in high-fat diet-fed zebrafish larvae and high cholesterol diet fed C. elegans To probe the role of PHLPP1 in the zebrafish model of lipid accumulation using a genetic approach, a CRISPR-Cas9 strategy was employed to knockout PHLPP1 function in zebrafish ( Figures S6A-S6E). Neutral lipid accumulation in head, trunk, and tail regions of HFD-fed PHLPP1-knockout larvae was lower than that in their control counterparts ( Figure 7A). Furthermore, quantification of the ISV Nile red staining showed diminished lipid accumulation in knockout larvae ( Figure 7B). Total cholesterol levels were also reduced in HFD-fed PHLPP1-knockout zebrafish larvae ( Figure 7C). Phenotypic measurements revealed that body length was reduced in HFD-fed PHLPP1-knockout larvae but this difference was not statistically significant ( Figures 7D and 7E). On the other hand, linear body width of the PHLPP1-knockout larvae was significantly lowered ( Figures 7D and 7F). To test the evolutionary significance of PHLPP1 in lipid metabolism, we studied its impact on lipid accumulation in HCD-fed C. elegans. phlp-2 is the only conserved nematode gene in the PHLPP family with predicted serine/threonine phosphatase activity and is orthologous to human PHLPP1 (48% similarity with PHLPP1a) and PHLPP2 (44% similarity) (Shaye and Greenwald, 2011). Consistent with in vitro and zebrafish larval findings, inactivation of phlp-2 by RNAi ( Figure S6F) resulted in a robust reduction in HCD-fed neutral lipid buildup ( Figures 7G and 7H). Consistently, elevated triglyceride levels in control HCD-fed worms were normalized upon phlp-2 ablation ( Figure 7I). DISCUSSION Atherosclerosis is a progressive inflammatory disease of large arteries. It is clinically manifested as CVDs such as myocardial infarction, unstable angina, stroke, and sudden cardiac death which collectively constitute the leading cause of death worldwide and low to middle-income groups of countries account for more than 75% of CVD deaths. iScience Article harboring lipid deposits and dead cells is formed. Vascular smooth muscle cells migrate into the necrotic core where they proliferate and secrete fibrous elements of the extracellular matrix forming fibrous plaques. The lesions grow outward thereby obstructing the arterial lumen. Importantly, in advanced lesions, the fibrous plaque is destabilized by matrix modifying enzymes thereby exposing prothrombogenic factors present in the necrotic lipid core to circulating platelets causing clotting and subsequent myocardial infarction or stroke. Foam cells are formed owing to the imbalance in the cholesterol handling pathways of macrophages. For example, proteins involved in lipid transport (adipophilin) and lipid storage (perilipin) are increased in human atherosclerotic lesion (Nuotio et al., 2007). Although the expression of CD36 is enhanced, the protein levels of ABCA1 were drastically reduced in human atheromas (Isoviita et al., 2010). Consistently, knockout of CD36 protected against atherosclerotic lesion development in atherogenic ApoE null mice (Febbraio et al., 2000). Double knockout mice lacking both ABCG1 (ATP binding Gaining a comprehensive understanding of the foam cell development is crucial for devising newer therapeutic strategies against atherosclerotic manifestations. The role of PHLPP1 in metabolic syndrome and particularly in atherosclerosis is speculated but remains largely unexplored (Mathur et al., 2017). The importance of PHLPP1 in cardio-protection is beginning to be understood. PHLPP1 knockdown has resulted in enhanced protective effects in nontransformed cardiomyocytes that are attributed to increased Akt activity Head, trunk, and tail -|343|; Intersegmental blood vessels -10X; Statistical analysis was performed using two-tailed t-test (*p < 0.05 vs WT-HFD). (C) Cholesterol levels in PHLPP1-knockout HFD-fed zebrafish larvae measured by HPLC-based method. Each experimental data point indicates 20 pooled larvae; Statistical analysis was performed using two-tailed t-test (*p < 0.05 vs WT-HFD). (D-F) Pictorial depiction along with linear body width (E) and whole-body length (F) measurement in PHLPP1-knockout zebrafish larvae; Statistical analysis was performed using two-tailed t-test (*p < 0.05 vs WT-HFD). (G and H) Microscopic images (G) and quantification (H) of Oil Red O staining in cholesterol-fed phlp-2 knockdown C. elegans; data presented are from a representative experiment of two independent trials (n = 7-12 per trial); Statistical analysis was performed using one-way ANOVA followed by Bonferroni 0 s post hoc analysis (***p < 0.001 vs N2/control+EtOH; ddd p < 0.001 vs N2/control+cholesterol). (I) TLC analysis of triglyceride levels in cholesterol-fed phlp-2 knockdown C. elegans along with ImageJ quantification (Bottom). Statistical analysis was performed using one-way ANOVA followed by Bonferroni 0 s post hoc analysis (*p < 0.05 vs N2/Ctrl; $ p < 0.05, $$ p < 0.01 vs N2/Ctrl+chol). Data are a representative of at least three independent experiments, unless specified. Numerical data are expressed as mean G SEM. See also Figure S6. iScience Article (Miyamoto et al., 2010). Several studies have emphasized the importance of Akt signaling in cardioprotective effects, potentiated via PHLPP1 knockdown/deletion. PHLPP1/2 knockdown-mediated protection against Ischemia/Reperfusion in leukemia inhibitory factor-induced cardiomyocytes and protection against hypertrophy in PHLPP1 knockout mice points to a significant role of PHLPP1/2 in cardio-protection (Miyamoto et al., 2010;Moc et al., 2015). Following up on our and other studies (Alamuru et al., 2014;Behera et al., 2018;Katsenelson et al., 2019;Lupse et al., 2021), in the current study, we explored the importance of PHLPP1 in the development of foam cells. We observed that lipid accumulation in foam cells in vitro was substantially reduced upon the depletion of PHLPP1 in RAW 264.7 cells. Consistent with the above observation, PHLPP1-knockout zebrafish larvae displayed lower lipid accumulation in intersegmental vessels, emphasizing the crucial role of PHLPP1 in lipid overload which is likely to be evolutionarily conserved.

Preliminary investigations showed no role for PHLPP2 isoform in the accumulation of neutral lipids in macrophages. However, it was recently shown that adenoviral-mediated overexpression of PHLPP2 in mice resulted in the decline of hepatic triglyceride content and plasma triglyceride levels along with the reduction of Fasn and Srebp1c gene expression (Kim et al., 2016). In addition, Huang et al. uncovered the role of PHLPP2 in vascular remodeling, a vital concept in atherosclerotic lesion formation (Huang et al., 2020). Huang et al. observed that PHLPP2 levels were elevated in the human atherosclerotic plaques and blood cells isolated from patients with coronary artery disease. Endothelial-specific deficiency of PHLPP2 reduced neointima formation in carotid artery ligated mice. Furthermore, the hepatic levels of PHLPP2, but not PHLPP1, were reduced in aging and obese mice (Kim et al., 2016). On the other hand, proteins levels of PHLPP1 were elevated in the skeletal muscle and adipose tissue of obese subjects (Andreozzi et al., 2011;Cozzone et al., 2008), skeletal muscle of high sucrose diet or HFD-fed mice (Behera et al., 2018) and OxLDL/7-ketocholesterol-treated mouse macrophages and HFD-fed zebrafish larvae in the current study. The observed differences in the functions of PHLPP1 and PHLPP2 may be owing to context-dependent, tissue/cell-type-specific roles of PHLPP isoforms. In the same lines, while we recently showed that PHLPP1 interacts with and dephosphorylates AMPK in muscle cells and soleus muscle of HFD-fed mice (Behera et al., 2018), Yan et al. reported that PHLPP2, but not PHLPP1, dephosphorylates AMPK in metabolically stressed Jurkat cells (Yan et al., 2021). The expression of PHLPP1 was enhanced during the early stages of foam cell formation but the levels declined at later time points (Figures 1A-1E). Decreased PHLPP1 levels at later stages of OxLDL treatment may amplify the inflammatory program of foam cells. An increase of PHLPP1 protein during the initial period (until 24 h) followed by a decrease in in vitro foam cells may indicate that the protein has a role in the early establishment of foam cell formation. PHLPP1 levels may have been reduced at the later time points as the lipid accumulation in foam cells is likely saturated. Such a reduction in certain lipid metabolism genes was also noticed in advanced human atherosclerotic plaques (Sulkava et al., 2017). In a study by Sulkava et al., atherosclerotic vascular samples from 29 patients subjected to carotid endarterectomy owing to occlusive atherosclerosis were compared with control samples from the left internal thoracic artery (LITA) from 28 patients who underwent coronary artery bypass surgery. Along with changes in several genes such as spp1, ApoE, MMP1, CIDEA, and ACTA1, RNA-seq analysis showed a 50% decrease in PHLPP1 and PHLPP2 transcript levels in human atherosclerotic plaques (Sulkava et al., 2017) ( Figure S1E). Lipid accumulation in foam cells demands the increased expression of several lipid-metabolizing and lipid import/ export-associated genes resulting in the unfolded protein response which in turn stimulates ER stress (Hotamisligil, 2010). As observed in mouse and rat muscle cells, induction of ER stress via sarco/ER Ca 2 ⁺ ATPase inhibition by thapsigargin led to the rhythmic expression of PHLPP1 in RAW 264.7 cells (Behera et al., 2018). Both ChREBP and PHLPP1 showed parallel induction upon 7-ketocholesterol treatment, prompting us to examine the possibility of ChREBP as an upstream regulator of PHLPP1. The recruitment of ChREBP to the Phlpp1 promoter and OxLDL/7-ketocholesterol induced increase in PHLPP1 levels in ChREBP-dependent manner offers a mechanistic basis for PHLPP1 upregulation in in vitro foam cells. Transcriptomics analysis provided the molecular basis for reduced neutral lipid accumulation in PHLPP1ablated foam cells. Gene expression programs driving lipid metabolism, cholesterol biosynthesis and, expectedly inflammation, were perturbed in PHLPP1-deficient macrophages. Among them, CD36 is worthy of mention. Hyperlipidemic mice showed attenuated lesion progression owing to CD36 loss. Also, CD36 binds TLR2-TLR4 commencing the most important TLR4 inflammasome pathway that fuels overall plaque progression (Chá vez-Sá nchez et al., 2014). PHLPP1 augments CD36 expression and is required for CD36mediated lipid uptake. Based on the role of ER stress in CD36 regulation and our findings from our current iScience Article and previous studies, we speculate that PHLPP1 may be one of the mediators of ER stress-induced CD36dependent enhanced lipid uptake (Yao et al., 2014). On the other hand, despite a strong increase in ABCG1 protein levels upon PHLPP1 knockdown, no differential cholesterol efflux was observed upon the perturbation of PHLPP1 levels ( Figures 4C and 4D). A possible explanation is that the reduction in ABCG1 function might be compromised by ABCA1-mediated efflux, because ABCA1 transcript levels and protein levels remained unchanged upon PHLPP1 knockdown ( Figure 4E). Besides, the kit measures BODIPY-labeled cholesterol efflux which is primarily mediated via the ABCA1 transporter (Takata et al., 2019). Passive diffusion-mediated lipid efflux also cannot be dismissed. Transcription factor gene set enrichment analysis using Enrichr identified ChREBP as a target of PHLPP1 ( Figures 5A-5C and S4B). Sarrazy et al. have reported an interesting finding that despite its role in increasing lipid-metabolizing gene expression, ChREBP functions to prevent inflammatory responses and aids in the maintenance of macrophage redox status thereby protecting against atherosclerosis development (Sarrazy et al., 2015). The contribution of ChREBP to in vitro foam cell development is not understood. We observed that forced expression of ChREBP robustly increased neutral lipid accumulation and strongly elevated FFA levels. ChREBP being a potent glucose sensor and a vital transcription factor of various lipid-metabolizing genes is very intricately regulated by various kinases and phosphatases. AMPK-mediated phosphorylation at Ser 568 in ChREBP is one such regulation that inhibits the protein 0 s DNA-binding capacity thereby hindering its transcriptional ability (Liangpunsakul et al., 2013). Also, phosphorylation at Ser 196 of ChREBP, a potential PHLPP1 dephosphorylation site, determines its cellular localization (Sakiyama et al., 2008). Our mechanistic analysis showed that PHLPP1 interacts with ChREBP and modestly enhances its occupancy at Fasn promoter. Furthermore, it dephosphorylated ChREBP in vitro in malachite green-based phosphate release assays, suggesting that ChREBP may be a direct substrate of PHLPP1. However, owing to the unavailability of phospho-Ser 196 -specific commercial antibodies, we could not definitively show that ChREBP is a genuine substrate of PHLPP1. Furthermore, consistent with our previous findings in myoblasts (Behera et al., 2018), we showed that (i) PHLPP1 inactivates AMPK in macrophages as well and (ii) the activation of AMPK prevents PHLPP1-promoted neutral lipid accumulation. Based on our current and past findings (Behera et al., 2018), we speculate that PHLPP1, ChREBP, and AMPK may be a part of an integrated signaling network that regulates lipid metabolism. We propose the following model ( Figure 8): ChREBP is recruited to Phlpp1 promoter in foam cells leading to its transcriptional upregulation. PHLPP1 may subsequently control lipid metabolism through AMPK and ChREBP-mediated mechanisms. First, PHLPP1 directly inactivates AMPK and may indirectly enhance the occupancy of ChREBP on its target promoters. Second, PHLPP1 likely dephosphorylates ChREBP thereby augmenting its nuclear localization and chromatin occupancy. However, the second mechanism needs further validation. Although some of the effects of PHLPP1 maybe mediated through both ChREBP and AMPK (FASN), certain pro-lipogenic effects of PHLPP1 (CD36) are possibly achieved through ChREBP-independent but AMPK-dependent mechanisms. An increase in neutral lipid accumulation upon ChREBP overexpression despite AMPK activation by AICAR or Act-3 intrigued us ( Figure 5N). We previously reported that the complex formation of ectopically expressed PHLPP1a with AMPK was disrupted upon AMPK activation in HEK 293T cells (Behera et al., 2018). Therefore, the increase in lipid accumulation upon ChREBP overexpression independent of AMPK activation status suggests the existence of various other lipid metabolism control pathways that ChREBP/PHLPP1 is a part of. Taken together, our results imply that PHLPP1 promotes neutral lipid accumulation and total cholesterol levels in in vitro foam cells, HFD-diet-fed zebrafish larvae, and HCD-fed C. elegans by augmenting the expression levels of key genes involved in lipid uptake, cholesterol biosynthesis, and lipid metabolism. Collectively, our data position PHLPP1 as an important player in the pathophysiology of atherosclerosis (Figure 8). Limitations of the study We observed a significant reduction of neutral lipid accumulation in cultured foam cells upon PHLPP1 knockdown. Consistent with our in vitro data, knockout of PHLPP1 gene in zebrafish larvae and siRNAmediated phlp-2 (PHLPP ortholog) gene silencing in C. elegans alleviated whole-body lipid accumulation upon HFD and HCD feeding, respectively. Despite foam cells being a hallmark of early atherosclerosis and PHLPP1 having a robust influence on foam cell formation, additional studies are required to unambiguously ll OPEN ACCESS establish the function of PHLPP1 in the pathology of atherosclerosis. Although the functionality of PHLPP1 in the lipid accumulation of zebrafish larval blood vessels is clear, its impact on the formation of atherosclerotic plaque, if any, could not be studied in the HFD-fed zebrafish larval model. Furthermore, the current study falls short of unambiguously establishing ChREBP as a direct substrate of PHLPP1 owing to the unavailability of commercial phospho-specific antibody. In vitro phosphatase assays with purified PHLPP1 and ChREBP proteins followed by immunoblotting with pSer 196 ChREBP antibody may further validate our findings. Foam cell formation is a multi-regulated process. Although our research addresses the effect of PHLPP1 on ChREBP/AMPK controlled lipid metabolism, PHLPP1-mediated regulation of other foam cell mediators is unclear at this stage. Based on our findings, we propose that PHLPP1 may play a role in the initial stages of atherosclerosis and further studies will likely address the above research gaps in the future. STAR+METHODS Detailed methods are provided in the online version of this paper and include the following: iScience Article iScience Article Generation of siRNA-mediated PHLPP1 knockdown in obese model of C. elegans Synchronized (using sodium hypochlorite procedure) L1 stage N2 strains were grown on nematode growth medium (containing 1mM IPTG and 50 mg/mL ampicillin) seeded with HT115 bacterial strain containing either empty L4440 plasmid or phlp-2 RNAi construct. For high fat diet conditions, cholesterol was added during plates preparation at 25mM final concentration at NGM levels. All experiments were conducted with hermaphrodite worms. RNA sequencing, data analysis RNA quality and quantity were assessed using Bioanalyzer 2100 with RNA 6000 Nano Labchips (Agilent Technologies). Samples with RNA integrity number (RIN) higher than 8.0 were used for library preparation using NEBNext Ultra RNA library Prep kit (NEB #E7770S). A total of 300 ng RNA was used for library preparation. The RNA was enzymatically fragmented and then converted to cDNA per the kit protocol (NEB #E7770S). Next, end repair and 3 0 end-adenylation of the fragments was performed and bar-coded adapters were ligated to the cDNA fragments. 12 cycles of PCR were performed to produce the sequencing libraries. The PCR products were purified using AmPure XP beads. Library quality and quantity were measured using Agilent 2100 Bioanalyzer and qPCR with KAPA Library Quantification kit (KAPA Biosystems, Foster City, USA), respectively. Adapter-ligated cDNA fragment libraries were run on Illumina HiSeq. 2500 at a 150-nucleotide read length using the paired-end chemistry. The RAW reads (FASTQ) were subjected to contamination [structural RNA/low complexity sequences, adapters] removal by mapping ll OPEN ACCESS iScience Article with bowtie 2-2.2.1. The dataset after contamination removal was mapped to the Mus musculus GRCm38 using STAR. Read mapping to genes in the Mus musculus GRCm38 gene list [GTF] were counted using feature count module of sub reads package and were normalized in DESeq2-3.5 followed by differential expression analysis. Significantly up-and down-regulated genes (differentially expressed) were selected based on 2-fold change with p value < 0.05 (FDR: 0.05; Student 0 s t-test, unpaired). Hierarchical clustering was performed with the programs Cluster (uncentered correlation; average linkage clustering) and Treeview (Eisen et al., 1998). GO annotation was carried out using the Gorillaweb server [http://cbl-gorilla.cs. technion.ac.il]. Term enrichments of differentially regulated genes were calculated based on mouse genome database as background. GO terms with corrected p < 0.05 were considered to be significantly enriched for pathway analysis. For 'Enrichr' analysis, differentially altered genes (criteria: cut-off: Fold Change >1.5 and FDR: 0.05) were submitted to 'Enrichr' tool. The tool used KEGG pathway 2021 human pathway to analyze the possible disease conditions in all the sets and the possible transcription factors involved in the gene alterations using ChIP-X enrichment analysis (ChEA) gene set library. The disease pathways and enriched transcription factors are calculated based on the following terms: 1) -log 10 p value that are computed based on alterations in the number of genes that are known primary or secondary targets of transcription factors; 2) odds ratio

that indicate measure of association between input and outcome and 3) combined score that signifies deviation from the expected rank and is calculated by log(p) * z where p is p value, and z = deviation. Immunoprecipitation and ChIP studies In immunoprecipitation experiments, precleared 500 mg cell lysates were incubated either with Flag beads (Sigma Aldrich) or PHLPP1 antibody (4 mg) for 2 h or overnight respectively at 4 C. Post incubation, beads were washed with lysis buffer followed by boiling with Laemmli Buffer. The enriched samples were separated and probed with indicated antibodies. RAW 264.7 cells grown in 100-mm dishes in DMEM supplemented with 10% (v/v) FBS or RAW 264.7 cells over-expressing PHLPP1 post 16 h of nucleofection were obtained for ChIP experiments once they attained 80% confluency. Cells were fixed with 1% (w/v) formaldehyde in DMEM for 10 min at room temperature, followed by one wash using ice-cold PBS. The fixation reaction was stopped with the addition of glycine solution for 15 min. Cells were then scraped from flasks with 10 mL ice cold PBS and centrifuged for 5 min at 1,000 rpm. Cells were then lysed using SDS ChIP lysis buffer and the chromatin obtained was sheared by sonication at an amplitude of 45% with a 10-s on/off pulse for 2 min in ice. This optimized sonication conditions resulted in 200-1000-bp DNA fragments. Post lysis, 250 mL of the isolated chromatin was diluted three times using ChIP dilution buffer and incubated with 3 mg of ChREBP antibody or IgG antibody at 4 C overnight. ChREBP antibody and IgG rabbit polyclonal antibody was obtained from Novus biologicals and Merck respectively for ChIP studies. Region akin to ChoRE site present in -488/-473 region of the PHLPP1 promoter were amplified using the following primers Forward: 5 0 CTACAACCCCCACTGTGCAA3 0 Reverse: 5 0 GCATGACGACATTTCTGCGG3 0 A product of 167-bp size was amplified at an annealing temperature of 58 C and visualized on a 1% agarose gel. The ChoRE region present in -349/-333 region of the FASN promoter was amplified using the following primers Total RNA was isolated using Trizol and reverse transcription was performed with 2-4 mg total RNA using random hexamers and oligo dTs in a reaction volume of 20 mL using the Superscript II First Strand cDNA Synthesis System. qPCR reactions were performed using TB Green mix on QuantStudio5 (Applied Biosystems). Data were normalized to b-actin using the Ct method (DDCt). Primer sequences are available upon request. The cDNAs from respective samples were diluted 1:5 for the genes PHLPP, CD36, Fasn, ChREBP, LDLR, etc., and 1:50 for reference genes (b-Actin, GAPDH). The primers were used at a 200 nM concentration in a 20 mL reaction. Relative gene expression was plotted using the control sample set as 1. The statistical significance was calculated on non-normalized and non-transformed DCt values. Colorimetric and fluorescent staining studies Oil red O staining of RAW 264.7 cells. Oil Red O (0.5% (w/v)) was dissolved in isopropanol at 56 C overnight. Pre-warmed stock was diluted at a 3:2 ratio with ultrapure water, filtered and used within an hour. RAW 264.7 cells were fixed in 10% phosphate buffered formalin for 10 min. Post rinsing the cells with 1X phosphate buffered saline (PBS) for 1 min, 60% isopropanol was added to cells for 15 s to facilitate neutral lipid staining. The cells were stained with Oil Red O for 1 min at 37 C in darkness. The cells were then destained with 60% isopropanol for 15 s and washed with 1X PBS thrice for 3 min each. Images were captured using a brightfield microscope (Zeiss primostar, 415500-1501-000). The images were converted to RGB stack in ImageJ software and the threshold was adjusted so that the red droplets were visible as black dots on a white background. The area of the red lipid droplets was measured using ImageJ software. Oil red O staining of C. elegans. Day 1 adult animals were washed from the plates and centrifuged at 1,200 rpm for 30 s. After washing twice with PBS, animals were suspended in 200 mL of 13 PBS. To this an equal volume of MRWB (160 mM KCl, 40 mM NaCl, 14 mM Na2EGTA, 1 mM Spermidine HCl, 0.4 mM Spermine, 30 mM Na PIPES at pH 7.4, 0.2% BME) containing 2% paraformaldehyde was added and incubated for 15 min at room temperature. Worms were centrifuged at 12000rpm for 1 min and pellet was washed once with 13 PBS. An 60% of 5 mg/mL Oil-Red-O solution is added to the worm pellet and incubated with continuous rocking. After an overnight incubation, worms were allowed to settle and dye was removed by washing twice with 13 PBS containing 0.01% Triton X-100. Worms were mounted on glass slide and imaged with Olympus CKX53. Oil-Red-O dye intensity in stained worms was quantified using ImageJ (NIH) software and average integrated density values are plotted. Nile red staining of zebrafish larvae. Control and HFD fed larvae were euthanized at 4 C for 30 min and fixed with 4% formaldehyde overnight and stained with Nile Red at 0.25 mg/mL concentration for 30 min. The Nile red was dissolved in 0.4% acetone. Visualization of Nile red was done using a fluorescence microscope (EVOS cell imaging) under an RFP filter. Quantification of staining was done using ImageJ software as described above. To calculate the Nile Red stain in the inter-segmental vessels, 10X trunk images were cropped using free hand tool in ImageJ and the images were processed as earlier. Data were also analyzed by a researcher blinded to the identity of samples. In vitro phosphatase assay HEK 293T cells were transfected with SFB-PHLPP1a and/or pCMV-Flag-ChREBP. Twenty four hours after transfection, cells were lysed in Phosphatase Lysis buffer (PLB) (50 mM Tris (pH 7.4), 1 mM DTT, 100 mM NaCl, 0.1% NP40 and 5 mM MnCl 2 ). Proteins were enriched using FLAG beads for 3 h. The bead-bound proteins were incubated in the reaction buffer (50 mM Tris (pH 7.4), 1 mM DTT, 10 mM ATP and 5 mM MnCl 2 ) for 45 min with gentle stirring at 30 C. Post incubation, the supernatant was collected and the released phosphate was measured using a malachite green phosphate assay kit (Cayman chemical) per the manufacturer's protocol. The phosphate-containing supernatant was incubated with MG acidic and MG blue solution for the indicated time periods and the absorbance measured at 620 nm. Lipid estimation assays Lipids were isolated from cells using the Bligh and Dyer organic lipid extraction method. Approximately 10 million cells were lysed using lipid lysis buffer [20 mM Tris, 1 mM EDTA]. Post lysis, chloroform and methanol were added at a ratio of 1:2 followed by addition of chloroform and lysis buffer. The samples were incubated at 25 C for 5-10 min and centrifuged at 3000g for 5 min. The organic phase was collected and further ll OPEN ACCESS iScience 25, 103766, February 18, 2022 iScience Article analysed. Total Cholesterol and free fatty acids were measured using kits following manufacturer's instructions. For HPLC based cholesterol estimation in zebrafish larvae, lipids were extracted from larvae using Bligh and Dyer organic lipid extraction method with some modifications as specified (Miyares et al., 2014). HPLC analysis was performed using a Waters Alliance e2695 HPLC system on a SYMMETRY C18, 75*4.6 mm 3.5 mm column. The mobile phase was: Triglyceride estimation in zebrafish larvae was performed using MyBioSource kit following manufacturer's instructions. For Thin Layer Chromatography (TLC) based triglyceride measurements in C. elegans, about 300 Worms were washed with 2 mL of 1X PBS, taken in 2 mL microfuge tubes, centrifuged at 1,000 3 g, and the supernatant was discarded. The pellet containing worms were transferred to glass tubes and resuspended in 3 mL of 2:1 chloroform: methanol solution, briefly sonicated and was shaken vigorously for 30 min, followed by vortexing for 60 s at high speed. The solution was centrifuged at 1,000 3 g, the supernatant was taken and washed with 0.2X volume of water and vortexed at high speed for 60 s. The resulting solution was centrifuged at 1,000 3 g for 60 s and the upper organic phase was transferred to a new glass tube. 0.2X volume of 0.034% of MgCl 2 was added to the organic phase, followed by vortexing for 1 min, centrifuged and the organic phase was transferred into a new tube. The organic phase was dried under nitrogen stream, dissolved into 100 mL of chloroform and loaded on a TLC plate (HPTLC Silica gel 60 F254 AMD extra thin, Merck, HX381899), and TLC was performed using 4:1 hexane: ethyl ether as mobile phase. To visualize the triglycerides, the plates were soaked in copper staining solution (7.5% CuSO4 and 85% ortho-phosphoric acid) and charred at 180 C. The plates were scanned and densitometric analysis was performed using ImageJ software. OxLDL uptake assay and cholesterol efflux assay The OxLDL uptake and cholesterol efflux were quantified in PHLPP1 over-expressed and knocked-down RAW 264.7 cells using cell-based kits as per the manufacturer's protocol. QUANTIFICATION AND STATISTICAL ANALYSIS Numerical data are expressed as mean G SEM unless or otherwise specified. Majority of the data are expressed as non-normalized values except for densitometric protein quantifications where data are shown as mean (fold change) G SEM. q-PCR data are expressed as fold change and their statistical significance is calculated using respective non-normalized and non-transformed DCt values. For comparison between two groups, two tailed t-test was used. One-way ANOVA and a subsequent Dunnett's or Bonferroni multiple comparison post-hoc tests were used to compare differences between multiple group means. p < 0.05 was considered as significant. CoenzymeQ10 and Ischemic Preconditioning Potentially Prevent Tourniquet-Induced Ischemia/Reperfusion in Knee Arthroplasty, but Combined Pretreatment Possibly Neutralizes Their Beneficial Effects Tourniquet (TQ) use during total knee arthroplasty (TKA) induces ischemia/reperfusion (I/R) injury, resulting in mitochondrial dysfunction. This study aims to determine the effects of coenzyme Q10 (CoQ10) and ischemic preconditioning (IPC), either alone or in combination, on I/R-induced mitochondrial respiration alteration in peripheral blood mononuclear cells (PBMCs) and pain following TKA. Forty-four patients were allocated into four groups: control, CoQ10, IPC, and CoQ10 + IPC. CoQ10 dose was 300 mg/day for 28 days. IPC protocol was three cycles of 5/5-min I/R time. Mitochondrial oxygen consumption rates (OCRs) of PBMCs were measured seven times, at baseline and during ischemic/reperfusion phases, with XFe 96 extracellular flux analyzer. Postoperative pain was assessed for 48 h. CoQ10 improved baseline mitochondrial uncoupling state; however, changes in OCRs during the early phase of I/R were not significantly different from the placebo. Compared to ischemic data, IPC transiently increased basal OCR and ATP production at 2 h after reperfusion. Clinically, CoQ10 significantly decreased pain scores and morphine requirements at 24 h. CoQ10 + IPC abolished analgesic effect of CoQ10 and mitochondrial protection of IPC. In TKA with TQ, IPC enhanced mitochondrial function by a transient increase in basal and ATP-linked respiration, and CoQ10 provides postoperative analgesic effect. Surprisingly, CoQ10 + IPC interferes with beneficial effects of each intervention. Introduction Total knee arthroplasty (TKA), an effective surgical treatment for end-stage knee osteoarthritis (OA), successfully provides pain relief and improves the quality of life [1]. Perioperative care of TKA becomes more challenging in patients with increasing age and multiple comorbidities [1,2]. Also, TKA has a trend toward an ambulatory setting [3,4]. Thus, besides a current anesthesia practice, alternative strategies to enhance function at the cellular levels are essential as an adjunct therapy to improve postoperative clinical outcomes. Tourniquet (TQ) application in TKA has been a controversial issue [5]. Previous studies have associated intraoperative TQ use with a reduction in intraoperative blood loss and postoperative blood transfusion [6,7]. However, other studies reported that TQ is related to several disadvantages, including early postoperative pain [8], delayed rehabilitation [9], and ischemia and reperfusion (I/R) injury to the insulted skeletal muscle as well as remote vital organs [10]. Although, the mechanisms underlying the TQ-induced skeletal muscle I/R injury are complex and not thoroughly understood, dysfunction of the mitochondria is considered the main source [10,11]. Mitochondria are multifunctional organelles that play major roles in fundamental cellular processes including cellular bioenergetic, redox status, calcium homeostasis, cell survival, and apoptosis. External stressors, such as the I/R process, impair physiologic regulation of mitochondrial bioenergetics, resulting in mitochondrial dysfunction [11]. The effects of skeletal muscle I/R on cellular respiration have been scantly investigated. In

a murine model of acute hindlimb I/R, the activities of mitochondrial electron transport chain (ETC) complexes of the gastrocnemius muscle were decreased after 3-h ischemia followed by 4-h reperfusion [12]. In addition to changes in the skeletal muscle itself, the maximal respiration of mitochondria isolated from lung parenchyma decreased after mice were subjected to 2-h ischemia of both hindlimbs, followed by 2-h reperfusion [13]. In TKA patients, one previous study has shown a preserved amount and function of the mitochondrial ETC complexes I-III of the vastus medialis muscle at a 60-min ischemic time [14]. To date, mitochondrial bioenergetics of the peripheral blood mononuclear cells (PBMCs) during ischemic and reperfusion periods in TKA patients with TQ application remain to be elucidated. Several pharmacological and interventional strategies to prevent or attenuate the TQ-related I/R injury in TKA have been investigated and a comprehensive review reported that antioxidants and ischemic preconditioning potentially exerted positive effects [10]. Coenzyme Q10 (CoQ10) is an electron carrier in the mitochondrial ETC, transferring electrons from complex I and II to complex III during oxidative phosphorylation (OX-PHOS). It also prevents cellular damage by exerting potent antioxidative activity [15]. Depletion of endogenous CoQ10 content in mitochondria of the skeletal muscle occurs in aging [16]. Fortunately, exogenous CoQ10 supplementation restores plasma CoQ10 concentration [17], enhances mitochondrial respiration [18], and demonstrates protective effects against myocardial I/R injury [19], as well as damage during cardiopulmonary bypass [20]. Ischemic preconditioning (IPC) is an intervention where the application of one or more brief episodes of ischemia and reperfusion generates small amounts of free radicals and subsequently increases tissue toleration to prolonged ischemic stress and reperfusion injury. There are two phases of IPC protection. Acute preconditioning occurs within minutes and disappears within hours, while the delayed phase of protection initiates after 24 h and persists for several days [21,22]. The acute phase of IPC activates cytoprotective processes, such as antiapoptotic, prosurvival MAPK/ERK kinase and PI3K/AKT/PKB pathways, inhibits mitochondrial permeability transition pores (mPTPs) opening, and exerts cardioprotection in myocardial I/R [23,24]. In TKA patients, IPC induces protective genomic responses [25] and potentially improves postoperative pain control [26,27]. Thus, the objectives of this study were (1) to determine the effects of preoperative CoQ10 supplementation on mitochondrial respiration in chronic knee OA patients; (2) to examine the efficacy of the two independent interventions, which were CoQ10 and IPC, during the early phase of TQ-induced I/R; (3) to detect possible interactions between these two factors; and (4) to demonstrate the effects of postoperative CoQ10 on the late phase of TQ-induced I/R. We hypothesized that (1) preoperative CoQ10 improved an impaired mitochondrial function of OA patients at the baseline, and postoperative CoQ10 restored mitochondrial dysfunction related to TQ-induced I/R; and (2) exogenous CoQ10 and IPC, either alone or in combination, were more effective than placebo in preserving mitochondrial oxygen consumption rates (OCRs) and alleviating pain following TKA. Study Design This is a randomized, triple-blinded, placebo-controlled trial with a 2 × 2 factorial design. This study was registered at clinicaltrials.in.th (TCTR20190128001) on 24 January 2019 and approved by the Ethics Committee of Faculty of Medicine, Chiang Mai University on 30 January 2019 (Ethical-Approval number: ANE-2564-05663). This manuscript adhered to the Consolidated Standards of Reporting Trials (CONSORT) statement. Participants After providing a written informed consent, 60-to 80-year-old patients with primary knee OA scheduled for elective unilateral TKA were included in the trial. The exclusion criteria were body mass index > 35 kg/m 2 , any possible high oxidative stress conditions including malignancy or infectious diseases, neurodegenerative diseases, atherosclerotic diseases, chronic kidney disease, active heavy smoking, recent use of antioxidant (within four weeks) or glucocorticoids (within three months), neurocognitive disorders, and allergy to any medications used in the study protocol. Forty-four patients were recruited from 31 January 2019 to the end of August 2020. Study Interventions, Randomization and Blinding The first intervention was CoQ10. Exogenous CoQ10 was administered for one month throughout the perioperative period. To determine the effects of preoperative and postoperative CoQ10 supplements, biochemical outcomes were assessed at two weeks before surgery, on the day of surgery, and at two weeks after surgery. The second intervention was IPC. The IPC was randomly assigned to all patients regardless of CoQ10 supplementation. Consequently, there were four study groups which were control (placebo + sham IPC), CoQ10 alone (CoQ10 + sham IPC), IPC alone (placebo + IPC), and a combined intervention (CoQ10 + IPC). To determine the effects of classical or acute phase of IPC protection, biochemical and clinical outcomes were assessed within the first few hours after IPC application. One month before surgery, a research coordinator screened patients at a pre-anesthesia clinic. After they consented to the study, the patients were randomly allocated to receive either CoQ10 or placebo. The computer-generated random numbers with a block size of four were contained in sealed opaque envelopes. A research coordinator opened the envelope and prepared an unlabeled bottle of study drugs. Both CoQ10 and the identical placebo soft gel were manufactured by MEGA Lifesciences Public Company Limited, Samutprakarn, Thailand. The patients in the CoQ10 group were assigned to ingest a 100-mg CoQ10 soft gel three times per day (300 mg/day) for 14 days before and 14 days after the operation, while those in the placebo group received placebo soft gels with a uniform prescription. At the anesthesia induction room one hour before surgery, a research investigator randomized patients to receive either IPC or sham IPC. The computer-generated random numbers with a block size of four were contained in sealed opaque envelopes. A research investigator opened the envelope and performed the IPC or sham IPC. The TQ was applied on the operated thigh of all patients. Patients assigned for the IPC received the same IPC protocol, which was three cycles of 5/5-min I/R time using a pressure of systolic blood pressure (SBP) +100 mmHg, while those assigned for the sham IPC obtained unpressurized TQ for 30 min. Preoperative, intraoperative, and postoperative data and blood samples were collected by two blinded data collectors. Patients, anesthesiologists, nurse anesthetists, surgeons, ward nurses, laboratory physicians, and data analysts were blinded to the group allocation during the entire study. Anesthesia Protocol and Operative Procedure All patients received spinal anesthesia using 0.5% isobaric bupivacaine 10-15 mg. Four to 6 mL/kg of Lactate Ringer's solution was loaded to prevent hypotension following spinal anesthesia. In addition, femoral triangle block with retained catheter and single-shot IPACK block were performed using 20 mL of 0.25% bupivacaine with 5 µg/mL epinephrine for each injection. Preoperatively, antibiotics and tranexamic acid were administered. Intraoperatively, vital signs were continuously monitored and maintained at the patient's baseline level. Intravenous fluid was provided as a maintenance dose or as needed if there was a significant blood loss. Oxygen was supplemented only if pulse oximetry was less than 94%. If patients required intraoperative sedation, midazolam was administered using the titration technique. Propofol, corticosteroids and local periarticular infiltration were not allowed throughout the perioperative period. The operation was performed by the two orthopedic surgeons (NS and KK) using a standard medial parapatellar approach and cemented TKA. The thigh TQ was inflated at SBP + 100 mmHg before an anterior longitudinal midline incision, and it was released after the closure of parapatellar arthrotomy. Postoperative Analgesic Protocol All patients received standard postoperative pain medications including 1 gm of oral acetaminophen every 6 h and 90 mg of Arcoxia every 24 h for 48 h. Ten mL of 0.25% bupivacaine with 5 µg/mL epinephrine was injected through the femoral triangle catheter every 12 h for 24 h. Intravenous morphine (2-4 mg/dose depending on patient's weight) was administered when patients experienced moderate to severe pain. Biochemical Outcomes The primary outcomes were mitochondrial OCRs of PBMCs. Venous blood samples were collected seven times ( Figure 1). The baseline specimen was obtained two weeks before surgery (T0: baseline) at the preanesthetic clinic. Thereafter, to assess the effect of preoperative CoQ10 supplementation, venous blood was drawn on the day of surgery before the IPC or sham IPC was performed (T1: pre-operation). To compare the effects of the two interventions on the early phase of TQ-induced I/R, blood samples were collected at 30 min after TQ inflation (T2: 30-min ischemia), at the onset of TQ release (T3: onset of reperfusion), and at two hours after reperfusion (T4: 2-h reperfusion). Lastly, to measure the effect of CoQ10 on the late phase of reperfusion, specimens were collected at 24 h (T5: 24-h reperfusion) and two weeks after surgery (T6: 2-week post-operation). Antioxidants 2022, 11, x FOR PEER REVIEW 4 of 15 IPC protocol, which was three cycles of 5/5-min I/R time using a pressure of systolic blood pressure (SBP) +100 mmHg, while those assigned for the sham IPC obtained unpressurized TQ for 30 min. Preoperative, intraoperative, and postoperative data and blood samples were collected by two blinded data collectors. Patients, anesthesiologists, nurse anesthetists, surgeons, ward nurses, laboratory physicians, and data analysts were blinded to the group allocation during the entire study. Anesthesia Protocol and Operative Procedure All patients received spinal anesthesia using 0.5% isobaric bupivacaine 10-15 mg. Four to 6 mL/kg of Lactate Ringer's solution was loaded to prevent hypotension following spinal anesthesia. In addition, femoral triangle block with retained catheter and singleshot IPACK block were performed using 20 mL of 0.25% bupivacaine with 5 µg/mL epinephrine for each injection. Preoperatively, antibiotics and tranexamic acid were administered. Intraoperatively, vital signs were continuously monitored and maintained at the patient's baseline level. Intravenous fluid was provided as a maintenance dose or as needed if there was a significant blood loss. Oxygen was supplemented only if pulse oximetry was less than 94%. If patients required intraoperative sedation, midazolam was administered using the titration technique. Propofol, corticosteroids and local periarticular infiltration were not allowed throughout the perioperative period. The operation was performed by the two orthopedic surgeons (NS and KK) using a standard medial parapatellar approach and cemented TKA. The thigh TQ was inflated at SBP + 100 mmHg before an anterior longitudinal midline incision, and it was released after the closure of parapatellar arthrotomy. Postoperative Analgesic Protocol All patients received standard postoperative pain medications including 1 gm of oral acetaminophen every 6 h and 90 mg of Arcoxia every 24 h for 48 h. Ten mL of 0.25% bupivacaine with 5 µg/mL epinephrine was injected through the femoral triangle catheter every 12 h for 24 h. Intravenous morphine (2-4 mg/dose depending on patient's weight) was administered when patients experienced moderate to severe pain. Biochemical Outcomes The primary outcomes were mitochondrial OCRs of PBMCs. Venous blood samples were collected seven times ( Figure 1). The baseline specimen was obtained two weeks before surgery (T0: baseline) at the preanesthetic clinic. Thereafter, to assess the effect of preoperative CoQ10 supplementation, venous blood was drawn on the day of surgery before the IPC or sham IPC was performed (T1: pre-operation). To compare the effects of the two interventions on the early phase of TQ-induced I/R, blood samples were collected at 30 min after TQ inflation (T2: 30-min ischemia), at the onset of TQ release (T3: onset of reperfusion), and at two hours after reperfusion (T4: 2-h reperfusion). Lastly, to measure the effect of CoQ10 on the late phase of reperfusion, specimens were collected at 24 h (T5: 24-h reperfusion) and two weeks after surgery (T6: 2-week post-operation). The mitochondrial OCR, which is an indicator of OXPHOS, was measured in the PBMCs. Once plasma was collected, PBMCs were isolated using a Ficoll density gradient centrifugation technique. The measurement of OCR was carried out by using the mitochondrial stress test assay (Seahorse XF Cell Mito Stress Test Kit) and the Seahorse XFe 96 Extracellular Flux Analyzer (Agilent Technologies, Santa Clara, CA, USA) according to the manufacturer's protocol. Briefly, the PBMCs were loaded into the XFe96-well plate and the OCR was measured under basal conditions (basal respiration) and in response to subsequently added reagents. One µM-Oligomycin, 2 µM-FCCP, and 0.5 µM-Rotenone/antimycin A were sequentially added to measure proton leak, maximal, and non-mitochondrial respiration rates, respectively. The basal respiration rate is a baseline oxygen consumption used by the ETC complex IV to establish a proton gradient across the inner mitochondrial membrane. To maintain the proton current balance under physiologic conditions, protons re-entry to the mitochondrial matrix dominantly by the adenosine triphosphate (ATP)-generating process through the ATP synthase (complex V) and partly by the uncoupling

protein-mediated proton leaks [28]. Accordingly, after administration of 1 µM-Oligomicin, which is an ATP synthase inhibitor, the ATP-related respiration was suppressed, and the basal proton leak respiration rate was measured. The ATP production (ATP-linked respiration) was calculated by subtracting proton leak respiration from basal respiration. Additionally, the coupling efficiency was calculated by dividing the ATP-linked respiration by basal respiration. Clinical Outcomes The secondary outcomes were postoperative pain scores and morphine consumption. Postoperative pain was assessed using a numeric rating scale (NRS) by the ward nurses every four hours. When a patient experienced moderate to severe pain (NRS ≥ 4), ward nurses administered intravenous morphine and the cumulative morphine requirement was evaluated at 8, 16, 24, and 48 h after surgery. Statistical Analysis Based on our preliminary measurements of ATP-related respiration rate at 30-min ischemia (69.2 ± 19.6 pmol/min) and 2-h reperfusion (77.2 ± 23.2 pmol/min) in TKA patients receiving no study intervention, we estimated the sample size for a one-sample mean test. Ten patients per group were required to identify a 30% difference between groups, with 5% type-1 error and 80% power. Data were analyzed using STATA 16 software (StataCorp LLC, College Station, TX, USA). Continuous variables were examined for normality using the Shapiro-Wilk normality test. Unpaired t-test or Wilcoxon rank-sum test was performed to assess differences between two groups, while one-way ANOVA with Bonferroni multiple comparison test or Kruskal-Wallis rank test was used to assess differences among four study groups. Linear regression with interaction was used to analyze the interaction terms of continuous parameters between the CoQ10 and IPC. Categorical data were assessed for differences using Fisher's exact probability test. Differences associated with p-value < 0.05 were considered significant. Results A total of 92 patients were screened for inclusion. Of these, nine patients were not eligible, and 39 patients were excluded, leaving 44 patients for randomization ( Figure 2). Since the first randomization, all patients received the medication and intervention as allocated and all data were analyzed as intended to treat. Demographic and intraoperative data of patients receiving placebo or CoQ10 were similar, except for the age range where those receiving CoQ10 were significantly older by chance (Table 1). Effects of Preoperative CoQ10 Supplementation on Baseline Conditions of Knee OA Patients Baseline mitochondrial respiration parameters including basal respiration, proton leak, ATP production, and % coupling efficiency were comparable between the patients receiving placebo and those obtaining CoQ10 (Figure 3). After two weeks of CoQ10 ingestion, proton leak decreased (Mean ± SE; placebo 14.0 ± 2.0 vs. CoQ10 8.1 ± 1.4 pmol/min, p = 0.028) and % coupling efficiency increased (Mean ± SE; placebo 80.8 ± 3.8 vs. CoQ10 95.4 ± 5.4 pmol/min, p = 0.012) when compared to the placebo group, indicating that exogenous CoQ10 improved ADP-driven respiration and uncoupling of OXPHOS, which occurred at baseline condition of knee OA patients ( Figure 3B,D). Effects of Preoperative CoQ10 Supplementation on Baseline Conditions of Knee OA Patients Baseline mitochondrial respiration parameters including basal respiration, proton leak, ATP production, and % coupling efficiency were comparable between the patients receiving placebo and those obtaining CoQ10 (Figure 3). After two weeks of CoQ10 ingestion, proton leak decreased (Mean ± SE; placebo 14.0 ± 2.0 vs. CoQ10 8.1 ± 1.4 pmol/min, p = 0.028) and % coupling efficiency increased (Mean ± SE; placebo 80.8 ± 3.8 vs. CoQ10 95.4 ± 5.4 pmol/min, p = 0.012) when compared to the placebo group, indicating that exogenous CoQ10 improved ADP-driven respiration and uncoupling of OXPHOS, which occurred at baseline condition of knee OA patients ( Figure 3B,D). Effects of CoQ10, IPC, and CoQ10 + IPC on Early Phase of TQ-Induced I/R At the 30-min ischemic time, mitochondrial respiration parameters were comparable among the patients receiving no intervention (control), CoQ10, IPC, and CoQ10 + IPC. When compared to the ischemic data, there were no significant changes in mitochondrial respiration throughout the reperfusion periods in patients receiving CoQ10 and the combined intervention (Figure 4). Interestingly, the IPC improved the mitochondrial basal respiration (Mean ± SE; 30-min ischemia 82.9 ± 10.0 vs. 2-h after reperfusion 120.6 ± 21.7 pmol/min, p = 0.036) and ATP production (Mean ± SE; 30-min ischemia 67.8 ± 8.5 vs. 2-h after reperfusion 102.7 ± 18.5 pmol/min, p = 0.016) at 2 h after reperfusion ( Figure 4A,C). These findings indicated that the IPC showed transient protective effects on mitochondrial function. Among patients receiving the IPC, the beneficial effects on the mitochondrial Effects of CoQ10, IPC, and CoQ10 + IPC on Early Phase of TQ-Induced I/R At the 30-min ischemic time, mitochondrial respiration parameters were comparable among the patients receiving no intervention (control), CoQ10, IPC, and CoQ10 + IPC. When compared to the ischemic data, there were no significant changes in mitochondrial respiration throughout the reperfusion periods in patients receiving CoQ10 and the combined intervention ( Figure 4). Interestingly, the IPC improved the mitochondrial basal respiration (Mean ± SE; 30-min ischemia 82.9 ± 10.0 vs. 2-h after reperfusion 120.6 ± 21.7 pmol/min, p = 0.036) and ATP production (Mean ± SE; 30-min ischemia 67.8 ± 8.5 vs. 2-h after reperfusion 102.7 ± 18.5 pmol/min, p = 0.016) at 2 h after reperfusion ( Figure 4A,C). These findings indicated that the IPC showed transient protective effects on mitochondrial function. Among patients receiving the IPC, the beneficial effects on the mitochondrial OCRs were demonstrated only when CoQ10 was not pretreated. We speculated that CoQ10 might interfere with the protective mechanisms of the IPC. However, the interaction between the IPC and CoQ10 on the basal respiration and ATP production at 2 h after reperfusion was not significant (Figure 4E,F). OCRs were demonstrated only when CoQ10 was not pretreated. We speculated that CoQ10 might interfere with the protective mechanisms of the IPC. However, the interaction between the IPC and CoQ10 on the basal respiration and ATP production at 2 h after reperfusion was not significant ( Figure 4E,F). Discussion The present study investigated the effects of CoQ10, IPC, and a combined intervention on TQ-induced I/R during TKA by measuring mitochondrial respiration of PBMCs at the ischemia and various reperfusion periods and assessing postoperative clinical outcomes. Regarding the CoQ10, preoperative 2-week supplement improved baseline mitochondrial uncoupling state of knee OA patients; however, changes in OCRs during the early phase of TQ-induced I/R were not significantly different from the control group. Postoperatively, CoQ10 decreased pain scores and morphine requirements at 24 h. Additionally, the basal and ATP-related OCRs, which were moderately reduced during 2-h reperfusion, were restored at 24 h after surgery. Concerning the IPC, it transiently enhanced mitochondrial basal respiration and ATP production at two hours after reperfusion; however, it showed no clinical benefit. When CoQ10 and IPC were combined, the postoperative analgesic effect of CoQ10 and the protective effects of IPC on mitochondrial function were abolished. These findings suggested a possible interaction between the CoQ10 and IPC. Alteration of mitochondrial function conceivably occurs during the I/R in the skeletal muscle [11,29]. Briefly, during TQ inflation or ischemic phase, cessation of blood and oxygen supply reduces aerobic glycolysis and mitochondrial OXPHOS. Attenuation of activities of ETC complexes I, II, and IV during prolonged ischemia leads to a reduction of ATP synthesis, increase in reactive oxygen species (ROS) production, decrease in antioxidant defense, accumulation of cytosolic H + and Ca 2+ , and formation of the mPTPs. During Discussion The present study investigated the effects of CoQ10, IPC, and a combined intervention on TQ-induced I/R during TKA by measuring mitochondrial respiration of PBMCs at the ischemia and various reperfusion periods and assessing postoperative clinical outcomes. Regarding the CoQ10, preoperative 2-week supplement improved baseline mitochondrial uncoupling state of knee OA patients; however, changes in OCRs during the early phase of TQ-induced I/R were not significantly different from the control group. Postoperatively, CoQ10 decreased pain scores and morphine requirements at 24 h. Additionally, the basal and ATP-related OCRs, which were moderately reduced during 2-h reperfusion, were restored at 24 h after surgery. Concerning the IPC, it transiently enhanced mitochondrial basal respiration and ATP production at two hours after reperfusion; however, it showed no clinical benefit. When CoQ10 and IPC were combined, the postoperative analgesic effect of CoQ10 and the protective effects of IPC on mitochondrial function were abolished. These findings suggested a possible interaction between the CoQ10 and IPC. Alteration of mitochondrial function conceivably occurs during the I/R in the skeletal muscle [11,29]. Briefly, during TQ inflation or ischemic phase, cessation of blood and oxygen supply reduces aerobic glycolysis and mitochondrial OXPHOS. Attenuation of activities of ETC complexes I, II, and IV during prolonged ischemia leads to a reduction of ATP synthesis, increase in reactive oxygen species (ROS) production, decrease in antioxidant defense, accumulation of cytosolic H + and Ca 2+ , and formation of the mPTPs. During reperfusion, ATP production and mitochondrial membrane potential are restored if the function of the mitochondrial respiratory chain is preserved. However, reoxygenation to the ischemia-insulted myocytes may produce additional cell damage. Fortunately, the deleterious apoptotic events after TQ deflation rarely occur because skeletal muscle is tolerant to ischemia and the I/R injury, which is predominantly determined by the ischemic time [30]. The present study demonstrated that mitochondrial OCRs in PBMCs at 30 min ischemia and during reperfusion periods in the control group were not significantly altered from the preoperative parameters. These results are in accordance with the study of Jawhar A. et al., which reported unchanged activities of the ETC complexes I-III at 60-min TQ time [14]. Besides, a previous in vitro human skeletal muscle model showed that 3-h ischemia followed by 2-h reperfusion resulted in a 50% decrease in cell viability [31]. Therefore, the preserved mitochondrial respiration in the present study possibly resulted from a restricted TQ inflation time which is generally limited to less than two hours in a clinical setting. CoQ10 is a potent antioxidant frequently used in clinical practice. It potentially prevents the TQ-induced I/R injury because the CoQ, as part of mitochondrial ETC, regulates ATP and ROS synthesis [12,19,[32][33][34]. In aging and chronic arthritis patients, endogenous CoQ10 levels have been shown to significantly decrease, thus inducing mitochondrial vulnerability to I/R processes [19,35]. Exogenous CoQ10 supplementation effectively reduced mitochondrial oxidative stress and possibly restored the ETC activity [36]. In the present study, preoperative CoQ10 ingestion (300 mg/day for 14 days) improved mitochondrial bioenergetics in PBMCs by reducing proton leak and enhancing mitochondrial coupling efficiency. However, preoperative improvement of mitochondrial function in PBMCs did not show significant protection against an early phase of TQ-induced I/R in TKA patients. During a late phase of reperfusion, a restoration of basal and ATP-related mitochondrial OCRs in PBMCs occurred at 24 h after surgery in patients pretreated with CoQ10. An insufficient benefit of the CoQ10 pretreatment during the initial reperfusion periods might result from a small availability of the CoQ10 molecule at the inner mitochondrial membrane, where the ETC is located. In mitochondria, CoQ10 is likely to be localized in the outer membrane because of its strong lipophilic nature [37]. Moreover, plasma CoQ10 concentration following oral CoQ10 supplement relies strongly on the formulation, dosage, and duration of CoQ10 [17,38]. The protective effect of CoQ10 on skeletal muscle I/R might be different with its various preparations. Concerning the clinical outcomes, the analgesic effect of CoQ10 in a surgical setting has not been particularly elucidated. In the present study, CoQ10 demonstrated better pain control at 24 h after surgery. However, there was a significant negative interaction effect when the IPC was combined. An anti-nociceptive effect of the CoQ10 has been studied in a rat model of knee OA and the mechanism of pain suppression was suggested to be caused by an anti-inflammation [39]. Previous in vivo study has also shown that CoQ10 pretreatment significantly inhibited TNF-α and NF-κβ activation during I/R of skeletal muscle [33]. IPC performed directly on the vessels to be occluded is termed a local IPC. It provides early and late phases of protection [23]. The early phase, as applied in this study, has shown its cytoprotection against TQ-induced I/R injury by improving mitochondrial maximal oxidative capacities and ETC complexes I and II activities, enhancing antioxidant defense, inhibiting mPTP opening related cell death, and reducing expression of genes and proteins involved in apoptosis [25,[40][41][42][43]. Consistently, in the present study, IPC demonstrated an increased basal mitochondrial OCR, which was predominantly driven by H + flow through

the ATP synthase, at 2 h after TQ deflation. This favorable effect was transient and did not last to 24 h after reperfusion. Referring to the clinical endpoints of knee surgery, preoperative IPC (1 episode of a 5-min TQ inflation followed by a 5 to 10-min TQ release) on the operative thigh demonstrated postoperative analgesic effects [26,27,43]. However, this present study failed to validate this benefit of the IPC. The discrepancy might stem from different IPC protocols since the mechanistic effect of the IPC depends on the number of cycles and duration of ischemia and reperfusion time [44,45]. The negative interaction between CoQ10 and IPC on mitochondrial respiration rate deserves discussion. Theoretically, the MAPK/ERK pathway activation, mitochondrial K ATP channel opening, intracellular free radicals production, and lipid peroxidation contribute significantly to the cytoprotective mechanisms of IPC [23]. Therefore, the effects of co-treatment of IPC and antioxidants, such as Vitamin C [46,47], Vitamin E [48], and melatonin [49], have been investigated in animal myocardial I/R models. The results were inconclusive. Vitamin E pretreatment did not prevent myocardial I/R-induced lipid peroxidation and preserved the beneficial effect of IPC. Melatonin maintained the cardioprotective effect of IPC despite inhibiting lipid peroxidation. On the contrary to other antioxidants, vitamin C abolished the reduction in myocardial infarct size induced by IPC, and the proposed mechanism was that Vitamin C inhibited oxidative stress-induced MAPKs activation. In clinical settings, IPC effectively reduced ischemic chest discomfort during percutaneous coronary intervention [50]. CoQ10 supplementation promoted antioxidants in patients with coronary artery disease [51]. In consistent with the effects of Vitamin C, this is the first study showing that the combination therapy normalized the benefits of IPC and CoQ10. In the present study, the enhanced mitochondrial respiration at two hours after reperfusion in patients subjected to IPC was antagonized when IPC was applied in CoQ10-supplemented patients. This finding indicated that CoQ10 potentially prevents the protective effect of the IPC. It is possibly because CoQ10 is effective in preventing lipid peroxidation [15] and inhibiting the MAPK/ERK signaling pathway [52]. This study contains some limitations. Firstly, there are several methods of mitochondrial function assessment, including measurement of respiration rate and membrane potential [28]. In this study, the mitochondrial OCRs of PBMCs collected from systemic circulation were measured by the extracellular flux analyzer. Besides, instead of the systemic PBMCs, mitochondria can be obtained from the PBMCs of local circulation, such as a femoral vein or surgical drainage tube, and directly from the insulted skeletal muscles [10]. From the latter, assessment of mitochondrial alteration related to TQ-induced I/R is likely to be more prominent. Secondly, biochemical parameters indicating oxidative stress, inflammation, and apoptosis should be measured to specify the mechanisms of the two interventions. Thirdly, because of the variable bioavailability of CoQ10 used in supplementation, CoQ10 concentration in plasma or PBMCs should be assessed to determine the actual effectiveness of this CoQ10 formula and dosage. This finding would provide the direct relationship between circulating CoQ10 and its interaction with IPC in patients undergoing TKA. Next, changes in protein or gene expression of the lower extremity muscle during the I/R and postoperative muscle strength should be evaluated because the TQ-induced I/R injury significantly affects the localized skeletal muscle [10]. Lastly, this study contained a small number of subjects, possibly not adequate to detect differences of some outcomes with small effect size. Conclusions When interpreted in the context of a specific dose of CoQ10 and technique of IPC, IPC enhanced mitochondrial function by transiently increasing basal and ATP-linked respiration at two hours after reperfusion and CoQ10 provides analgesic effect by reducing postoperative pain at 24 h. Even though CoQ10 and IPC potentially prevent TQ-induced I/R injury and improve clinical outcomes, a combination of the two interventions possibly interferes with the beneficial effect of each intervention. Further trials investigating the potential mechanisms underlying the CoQ10 and IPC interaction and the mechanism of pain reduction following CoQ10 supplementation are suggested. Data Availability Statement: The authors confirm that the data supporting the findings of this study are available within the article. Conflicts of Interest: P.L., N.C. and S.C.C. received funding sponsors for the research project. Other authors declare no conflict of interest. Audit of Carotid Doppler Sonography: Spectrum of Findings at a Tertiary Hospital in Northwestern Nigeria Introduction: Doppler sonography of the cervical segment of the carotid arteries is becoming a popular tool for evaluating atherosclerosis of the carotid artery. We present the audit of findings on carotid ultrasound examination among patients with clinical suspicion and risks for cerebrovascular disease and possible correlates in Northern Nigeria. Materials and Methods: We performed carotid ultrasound examination on all patients referred for screening and clinical suspicion of cerebrovascular disease within the year 2017. The patients' characteristics, risk factors, presence of atheroma and characteristic of the atheroma, degree of stenotic disease as well as the presence of incidental ultrasound findings were reviewed and documented. Results: Out of the 62 patients, 55 (88.7%) of them had various degrees and types of atheromatous plaques in different segments of the cervical carotid arteries, whereas 7 (11.3%) were normal. The predominant risk factor was smoking followed by diabetes mellitus, whereas the highest indication for the scan was transient ischemic attack. Incidental thyroid lesions such as nodules and cysts were encountered in 14 (22.6%) of the patients. There is a statistically significant difference between sex and age with the side of lesion, degree of stenosis, segment involved, and type of atheromatous plaque. Conclusion: There is a statistically significant difference between sex and age with the side of lesion, degree of stenosis, segment involved, and type of atheromatous plaque. About one-fifth of our patients had incidental thyroid lesions. Therefore, routine screening of population at risk is highly recommended. IntroductIon Doppler sonography of the cervical segment of the carotid arteries is a noninvasive modality for the evaluation of the anatomy and hemodynamics of the carotid arteries. Carotid Doppler ultrasonography is becoming a popular tool for evaluating atherosclerosis of the carotid artery (CA). [1,2] Its ability to measure intimal media thickness (IMT) and characterize the morphology of carotid atheroma makes it a reliable modality of determining the etiology and severity of stroke risk. [3,4] In addition, color Doppler ultrasonography and pulsed Doppler ultrasonography have been used for detecting CA stenosis. [1,4] Stroke is increasingly becoming a challenging public health issue in Africa, and the nonavailability of data has limited research output and consequently the response to this burden. Adeloye [5] estimated that about 3483 new stroke cases among people aged 15 years or more were estimated in Africa in 2009, equivalent to 81.2 (13.2-94.9)/100,000 person-years. Despite the fact that it was recommended to be used as a baseline noninvasive method in the initial evaluation of either asymptomatic or symptomatic patients to define the possible stenosis on CA, [6] the pattern of findings in the local setting remains so far, unknown. This report presents the audit of the spectrum of findings on carotid ultrasound examination among patients with suspected symptoms and risks for cerebrovascular disease and documented possible correlates. This is to generate a database that will benefit clinicians, researchers, industry, and policymakers in confronting the burden of cerebrovascular disease in Nigeria and other developing nations. materIals and methods All patients referred to vascular ultrasound clinic of screening and clinical suspicion of cerebrovascular disease within the year 2017 were included in the study. The following variables were recorded in the dedicated register: 1. Age 2. Gender 3. Indication for examination 4. Number of risk factors 5. Side and segment of the lesion 6. Presence or absence of the atheroma as well as the characteristic of the atheroma 7. Degree of stenotic disease if present 8. Presence or absence of the thyroid lesion and other incidental ultrasound findings. Thereafter, the patients were scanned in supine position on the examination table. The patient's head was turned away from the side and the neck a little extended. The examinations were carried out either from the patient's side or sitting at the patient's head. Coupling gel was applied on each side of the neck. The scan was performed in both transverse and longitudinal dimensions using the 7.5 MHz linear transducer of SONOSCAPE SSI-8000, 2014 digital color Doppler ultrasound system (Shenzhen, China). The transverse scan was performed first, which began from the level of the root of the neck up to the level of the angle of the mandible. The common carotid arteries as well as the level/orientation of bifurcation were located bilaterally. Major areas of abnormality were identified. Longitudinal scans were performed along with color and spectral Doppler interrogation. Carotid IMT was measured as the distance between the leading edges of the two echogenic layers of the wall, using the posterior wall and also measurement, was obtained at the upper CA where the transducer was easily adjusted to make 90° angulation with the vessel wall for more accurate measurement. Magnification was also applied as much as possible to make measurement easier. The IMT measurements were also taken at the level of the carotid bifurcation and the internal CA (ICA). The identified atheromatous plaques were characterized and documented in to five types according to modified Gray-Weale classification as follows: Type 1 plaques were uniformly echolucent, Type 2 predominantly echolucent, Type 3 predominantly echogenic, Type 4 uniformly echogenic, and Type 5 consisted of plaques that could not be classified owing to heavy calcification and acoustic shadows. [7] After capturing a transverse scan of the most stenotic segment of the CA on a B-mode or color Doppler, the original diameter (OD) and residual diameter (RD) were measured using electronic calipers. The RD was defined as the shortest diameter of the residual lumen at the most stenotic segment of CA, and OD was defined as the measured diameter from the outer media to the outer media of the diseased artery on the same plane and at the same direction with the RD. The percentage of carotid stenosis (CS %) grayscale ultrasound was calculated using the following equation: CS% = (1− [RD/OD]) ×100% based on the protocols of the European carotid surgery trial. [8] Color interrogation was used to identify areas of abnormal flow. Spectral Doppler mode was then activated, and measurements of velocimetric indices were done. At least five consistent waveforms were recorded on each segment. (14), Headache (4) normal. The predominant risk factor was smoking, followed by diabetes mellitus (DM), whereas the highest indication for the scan was transient ischemic attack and then screening [ Table 1]. There were nine patients with incidental thyroid nodules, whereas five had simple cysts, as exemplified in Figure 3. Other findings such as arrhythmia were encountered in two patients. However, these incidental lesions do not show evidence of extrinsic indentation on the carotids. The carotid atheromatous lesions were seen mostly bilaterally in both genders (27, 43.5%) and then more on the left side [ Tables 1 and 2]. A male patient in the age group of 60-69 years [ Table 6] had total occlusion of the right ICA in the Type 4 atheromatous plaque category [Tables 7-9]. More patients were generally seen in the age of 50 years and above with a higher degree of stenosis in these age groups. There is a statistically significant difference between sex and age with the measured parameters -side of lesion, degree of stenosis, segment involved, and type of atheromatous plaque (P < 0.05). dIscussIons During the period under review, a total of 62 patients were scanned, with equal distribution of males and females. Their ages ranged from 19 to 84 years with a mean of 61.7 years (+14.5 years). This is in agreement with the practice guidelines developed by the American Institute of Ultrasound in Medicine in collaboration with the American College of Radiology, the Society for Pediatric Radiology, and the Society of Radiologists in Ultrasound, [9] which is also similar to that of Japan Society for Ultrasonics in Medicine. [10] In our study, smokers are the highest risk group (62.9%) of those with plaque. They had plaque (s) mostly at multiple segments and the common CA which also constituted an important independent risk for the development of atheromatous plaque (P < 0.005). This is similar to the findings of Dempsey et al., [11] who studied 790 patients with a history of smoking and concluded that smoking was a statistically significant independent predictor of CA

plaque thickness. However, hypertension (1.61%) and DM (24.19%) scored low, even though hypertension as reported by several researchers [12,13] is a known independent risk for plaque formation and stroke. Joakimsen et al. [14] found more plaques in the walls of the common carotid artery (CCA) bifurcation, followed by the ICA when they studied 3016 men and 3404 women and found that 55.4% of men and 45.8% of women had plaques, with men having more echolucent type of plaque. These findings are similar to our study, which also showed more males with echolucent plaques and more females with predominantly echogenic type. There were more patients with plaques in the CCA and in multiple segments of the CA than other segments of the CA. Regarding the type of carotid atheroma, Type-1 and 2 plaques were seen in 50% of our study population. However, a larger sample size study by Casadei et al. [15] found Type 1 and Type 2 plaques in 160 (21.4%) of the 747 patients they examined. Apart from differences in sample size, the severity of risk factors may also differ between the two study environments. Compared to the findings of Johnson et al., [16] our findings showed that the association between smoking burden and carotid plaque presence was stronger in males than in females. Increasing carotid IMT additionally was associated with male sex. The study also showed that there are more males (19.4%) with a high degree of stenosis or increasing severity of plaque formation than females (12.9%). Furthermore, the severity increased with increasing age, from the age group of 50-59 years and above with a significant positive relationship between increasing age with the severity of carotid atherosclerosis and degree of stenosis (P < 0.005). These findings also agree with those of Yin et al. [17] from the eastern part of China, showing that the prevalence of carotid atheroma and carotid plaque increased gradually with age. These conditions were more prevalent in men than in women. In the course of data collection, we encountered 14 cases of incidental thyroid lesion, constituting 22.6% of the study population. Steele et al. [18] found a lower rate of incidental thyroid lesion (9.4% of 188 retrospective carotid ultrasound Male 15 6 3 2 1 4 31 Female 9 15 3 1 3 31 Total 24 21 6 3 1 7 62 the prevention of cerebrovascular disease. In view of the limitations of this study by small number and other unanswered questions, further prospective studies involving larger patient population and other unexamined risk factors will be needed to validate our findings. Furthermore, routine screening of population at risk is strongly recommended. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest. Immune checkpoint inhibitors in head and neck squamous cell carcinoma: A systematic review of phase-3 clinical trials BACKGROUND The outcomes of patients diagnosed with head and neck squamous cell carcinoma (HNSCC) who are not candidates for local salvage therapy and of those diagnosed with recurrent or metastatic disease are dismal. A relatively new systemic therapy option that emerged in recent years in the treatment of advanced HNSCC is immunotherapy using immune checkpoint inhibitors (ICIs). The safety profile and anti-tumor activity of these agents demonstrated in early phase clinical trials paved the way to the initiation of several promising phase-3 trials in the field. AIM To evaluate the evidence on the effectiveness of ICIs in HNSCC, based on published phase-3 clinical trials. METHODS We searched PubMed, Cochrane Library, Embase, and Scopus to identify published literature evaluating immunotherapy using ICIs in recurrent or metastatic HNSCC (R/M HNSCC) and locally advanced head and neck squamous cell carcinoma (LAHNSCC). We used a combination of standardized search terms and keywords including head and neck squamous cell carcinoma, recurrent, metastatic, locally advanced, immunotherapy, immune checkpoint inhibitors, monoclonal antibodies, programmed cell death protein-1 (PD-1), programmed death-ligand 1 (PD-L1), cytotoxic T- lymphocyte associated protein-4 (CTLA-4), and phase-3 clinical trial. A sensitive search filter was used to limit our results to randomized controlled trials. RESULTS Five phase-3 clinical trials have reported the data on the effectiveness of immunotherapy in HNSCC so far: Four in R/M HNSCC and one in LAHNSCC. In patients with R/M HNSCC, anti-PD-1 agents nivolumab and pembrolizumab demonstrated improved survival benefits in the second-line treatment setting compared to the standard of care (standard single-agent systemic therapy). While the net gain in overall survival (OS) with nivolumab was 2.4 mo [hazard ratio (HR) = 0.69, P = 0.01], that with pembrolizumab was 1.5 mo (HR = 0.80 nominal P = 0.0161). The anti-PD-L1 agent durvalumab with or without the anti-cytotoxic T- lymphocyte associated protein-4 agent tremelimumab did not result in any beneficial outcomes. In the first-line setting, in R/M HNSCC, pembrolizumab plus platinum-based chemotherapy resulted in significant improvement in survival with a net gain in OS of 2.3 mo (HR = 0.77, P = 0.0034) in the overall population and a net gain in OS of 4.2 mo in the PD-L1 positive (combined positive score > 20) population compared to standard of care (EXTREME regime). In patients with PD-L1 positive R/M HNSCC, monotherapy with pembrolizumab also demonstrated statistically significant improvement in survival compared to EXTREME. In LAHNSCC, immunotherapy using avelumab (an anti-PD-L1 agent) along with standard chemoradiation therapy did not result in improved outcomes compared to placebo plus chemoradiation therapy. CONCLUSION Anti-PD-1 agents provide survival benefits in R/M HNSCC in the first and second-line settings, with acceptable toxicity profiles compared to standard therapy. There is no proven efficacy in the curative setting to date. INTRODUCTION Head and neck squamous cell carcinoma (HNSCC) is one of the major causes of cancer-associated morbidity and mortality globally[1-3]. Treatment approaches for HNSCC vary according to the stage of the disease at presentation. Around 40% of HNSCCs present at an early stage and are treated by a single treatment modality, either radical radiotherapy or surgery. The remaining 60% of cases present as locally advanced disease, and treatment options include chemoradiation or surgery followed by adjuvant therapy. However, within 3 years, over 50% of these patients relapse locally or at distant sites. Salvage approaches for the locally recurrent disease include surgery, surgery followed by re-irradiation, or re-irradiation with or without concurrent chemotherapy [4,5]. For a recurrent disease that is not amenable to salvage approach and for metastatic disease, platinum-based chemotherapy was the only available treatment option until recently. While the median survival of recurrent/metastatic HNSCC (R/M HNSCC) patients receiving platinum-based chemotherapy is 7.4 mo, some patients become refractory to platinum and die within a period of 4 mo [6][7][8][9][10][11][12]. Subsequently, the addition of the antiepidermal growth factor receptor (EGFR) targeted agent cetuximab to platinum-based chemotherapy showed improvement in survival compared to platinum-based chemotherapy alone, as demonstrated in a landmark phase-3 trial in 2008[12-16]. A relatively new systemic therapy option that emerged in recent years in the treatment of advanced HNSCC is immunotherapy using immune checkpoint inhibitors (ICIs). The checkpoint pathways in the tumor microenvironment are responsible for immune escape and T cell exhaustion related to the survival of the cancer cells. ICIs are monoclonal antibodies that can block these pathways by inhibiting the binding of checkpoint proteins on the T cells to similar proteins on the tumor cells. Thus, these agents act by reinvigorating the immune cells and re-establishing the anti-tumor immune responses that promote the elimination of cancer cells. Programmed cell death protein-1 (PD-1) receptors, programmed death-ligand 1 (PD-L1) receptors, and cytotoxic T-lymphocyte associated protein-4 (CTLA-4) are the major established targets for cancer immunotherapy with ICIs, and the therapeutic effects of ICIs result from blockade of these receptors [17][18][19][20]. In recent years, many interventional studies have evaluated ICI therapy for the treatment of HNSCC. The objective of this systematic review is to gather the evidence from published phase-3 randomized controlled trials (RCTs) comparing immunotherapy with the standard of care (SOC), among patients with R/M HNSCC or locally advanced HNSCC (LAHNSCC). We aimed to evaluate and synthesize the evidence from the published phase-3 studies investigating immunotherapy in advanced head and neck cancer using checkpoint inhibitors, either alone or in combination with chemotherapy, radiation therapy, or another checkpoint inhibitor. Data sources and literature search The study followed the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines [21]. We systematically searched PubMed, SCOPUS, EMBASE, and COCHRANE Library without any language limit. We used a combination of standardized search terms and keywords including head and neck squamous cell carcinoma, recurrent, metastatic, locally advanced, immunotherapy, checkpoint inhibitors, monoclonal antibodies, PD-1, PD-L1, CTLA-4, and phase-3 clinical trial. A sensitive search filter was used to limit our results to RCTs reported from January 2000 till February 2021. The initial search was conducted in February 2021. We also looked for any updates on the selected studies till April 2021. The search syntax is given in the Supplementary file. Inclusion/exclusion criteria and study selection Studies were included if they were completed phase-3 RCTs conducted among patients with R/M HNSCC or LAHNSCC, in which the intervention patients received ICI either alone or in combination with chemotherapy, radiation therapy, or with another IO and the control patients received SOC. Anatomical sites of primary tumors were oral cavity, oropharynx, hypopharynx, and larynx in the included studies. Early phase trials and observational studies were excluded. Studies involving patients with nasopharyngeal carcinoma were also excluded. Titles generated from the initial search results were exported to EndNote. Duplicates were removed, and the remaining titles were scanned for relevance. Abstracts of articles pertaining to potentially eligible studies were independently reviewed by both authors and uncertainties were resolved through discussion. Potentially eligible studies were further evaluated for relevance, trial status (completed/ ongoing/withdrawn), and availability of results. The following descriptive data were extracted from the included studies: Study design, population, details of the intervention, details of treatment received by the control arm, and the primary and secondary endpoints. Information on adverse events and statistical data on the outcomes were also extracted, which included, overall survival (OS), progression-free survival (PFS), overall response rate (ORR), biomarker effect, and patient-reported outcomes. The flow chart of study selection (PRISMA) is given in Figure 1. RESULTS The original literature search generated 565 titles altogether, of which 100 titles were eventually selected for abstract review for identification of potentially eligible studies. Others were excluded as they were related to phase-1 or phase-2 studies or not precisely relevant to the topic of the review. Through the abstract review, we identified 56 references (including one conference abstract) pertaining to potentially eligible studies. Through full-text review of these references, we selected five original phase-3 RCTs to be included in the systematic review [22][23][24][25][26]. In four of the trials [22][23][24][25], participants were patients with R/M HNSCC, while in one trial, participants were patients diagnosed with LAHNSCC [26,27]. All four studies among patients with R/M HNSCC were open-label RCTs; three of them investigated the effectiveness of ICI as second-line treatment [22][23][24], while in one study [25], ICI was evaluated as first-line treatment. The study among LAHSCC patients was a double-blinded placebo-controlled RCT [26,27]. ICIs assessed in these studies were nivolumab, pembrolizumab, durvalumab, tremelimumab, and avelumab. While nivolumab and pembrolizumab are anti-PD-1 monoclonal antibodies, durvalumab and avelumab are anti-PD-L1 antibodies. The monoclonal antibody tremelimumab is an anti-CTLA-4 agent [28][29][30][31]. We classified the studies into three groups based on the disease status and the treatment setting. The details of these studies in terms of the study population, intervention, comparator, outcomes, and adverse events are given in Table 1. Phase-3 studies evaluating ICI as second-line treatment in R/M HNSCC (three RCTs: CheckMate 141, KEYNOTE 040, and EAGLE) So far, three phase-3 RCTs have compared the effectiveness of ICI against the existing SOC (single-agent systemic therapy with methotrexate, docetaxel, or cetuximab) in the second-line treatment setting [22][23][24] (Table 1). Ferris et al[22] conducted a randomized, open-label, phase-3 study (n = 361) among patients with platinum-refractory recurrent HNSCC (recurrence within 6 mo after platinum-based chemotherapy) to investigate the effectiveness of the anti-PD-1 checkpoint inhibitor agent nivolumab. The intervention arm (n = 240) received nivolumab at a dose of 3 mg/kg body weight every 2 wk, while the control patients (n = 121) received SOC in the form of standard single-agent systemic therapy with methotrexate [40 mg/m 2 intravenously (IV) weekly], docetaxel (30 mg/m 2 IV weekly), or cetuximab (400 mg/m 2 IV once followed by 250 mg/m 2 weekly). OS was

the primary endpoint of the study. Secondary endpoints included PFS, ORR, and biomarker effects on survival, safety, and quality of life assessments. The median duration of follow-up was 5.1 mo (range, 0 to 16.8). ORR: ORR was 13.3% (95%CI: 9.3-18.3) in the intervention arm with nivolumab, whereas it was 5.8% (95%CI: 2.4-11.6) in the control arm (SOC). KEYNOTE 040 (Pembrolizumab vs standard single-agent systemic therapy) In this open-label phase-3 RCT, the investigators tested the efficacy and safety of the immune checkpoint inhibitor pembrolizumab (an anti-PD-1 monoclonal antibody) compared to standard therapy for the treatment of metastatic/recurrent head and neck cancer [23]. This was a multi-center study involving 97 medical centers across 20 countries. There were 247 patients in the intervention arm, while the control arm included 248 patients. Patients with platinum-refractory recurrent or metastatic (or both) HNSCC were included in this study. PD-L1 expression was assessed and categorized according to the tumor proportion score (≥ 50% vs < 50%) as well as the combined positive score (≥ 1 vs < 1) The intervention arm received pembrolizumab 200 mg every 3 wk, while the control arm received investigator's choice of standard doses of methotrexate (40 mg/m 2 IV weekly), docetaxel (75 mg/m 2 IV every 3 wk) or cetuximab (250 mg/m 2 IV weekly following a loading dose of 400 mg/m 2 ). . Among patients with CPS < 1, OS was 6.3 mo (95%CI: 3.9-8.9) with pembrolizumab vs 7.0 mo (95%CI: 5.1-9.0) with SOC (HR = 1.28; 95%CI: 0.8-2.07; P = 08476; P for interaction = 0.07). In terms of PFS, based on the modified RECIST1.1, for patients with TPS ≥ 50%, PFS was longer with pembrolizumab than with SOC, whereas for patients with CPS ≥ 1, PFS was slightly lower (3.6 mo) with pembrolizumab compared to SOC (4.8 mo). Among patients with CPS < 1 and those with TPS < 50%, PFS was longer for SOC compared to pembrolizumab [23]. Adverse events: In KEYNOTE 040, adverse events of grade 3 or more occurred in 13% of patients with pembrolizumab vs 36.1% with SOC. Four patients in the pembrolizumab arm and 2 patients in the control arm had treatment-related death. While hypothyroidism was the most common treatmentrelated adverse event with pembrolizumab (13%), fatigue was the most common adverse event with SOC (18%) [23]. Adverse events: In KEYNOTE 048, 55% patients in the pembrolizumab arm, 85% patients in the pembrolizumab plus chemotherapy arm, and 83% patients in the control arm developed grade 3 or more adverse events of any cause. Of these, treatment-related adverse events consisted of 17% in the pembrolizumab alone group, 72% in the pembrolizumab plus chemotherapy group, and 69% in the control group. While adverse events led to death in 8% of patients in the pembrolizumab arm and 12% of patients in the pembrolizumab plus chemotherapy arm, 10% in the control arm also died of adverse events. Major adverse events (of any grade) in the intervention groups were anemia, fatigue, hypothyroidism, and nausea [25]. Phase-3 studies evaluating ICI for treatment of LAHNSCC (I RCT: JAVELIN head and neck 100 trial) The current SOC for the treatment of LAHNSCC is concurrent chemoradiation therapy (CRT) [33]. So far, only one phase-3 trial has investigated the usefulness of adding an ICI to concurrent CRT. JAVELIN head and neck 100 trial (avelumab plus CRT vs placebo plus CRT) The preliminary results of the study were presented in the 2020 European Society for Medical Oncology annual meeting by Cohen et al [26] followed by a recent journal publication [27]. This study (n = 697) was conducted among patients with previously untreated LA HNSCC who were eligible for definitive CRT with curative intent. The intervention arm (n = 350; median follow-up for PFS 14.6 mo, for OS 16.7 mo) received the PD-L1 inhibitor avelumab (10 mg/kg IV every 2 wk) plus CRT, which consisted of cisplatin (100 mg/m 2 every 3 wk) concurrently with intensity-modulated radiotherapy (standard fractionation of 70 Gy in 35 fractions over 7 wk). The control arm (n = 347; median follow-up for PFS 14.8 mo, for OS 16.8 mo) received placebo plus CRT (Table 1). PFS: Median PFS (primary endpoint) was not reached in the intervention group or the control group. Statistical reports showed that hazard ratio (HR= 1.21; 95%CI: 0.93-1.5; one-sided P = 0.92) did not favor the avelumab plus CRT arm. OS: OS was one of the secondary endpoints in this trial. Median OS was not reached in either study group. Statistical reports showed that the hazard ratio for death (HR = 1.31; 95%CI: 0.93-1.85; one-sided P = 0.937) did not favor the avelumab plus CRT arm. Adverse events: Treatment-related adverse events of grade 3 or more occurred in 80% of patients in the avelumab arm and in 74% of patients in the control arm. Serious adverse events occurred in 36% of patients in the intervention arm and in 32% of patients in the control arm. In the intervention arm, 7% of patients discontinued due to treatment-related adverse events vs 3% in the control arm [27]. DISCUSSION ICIs have emerged as a novel treatment strategy for HNSCC in recent years. The safety profile and antitumor activity of these agents demonstrated in early phase clinical trials paved the way for the initiation of several promising phase-3 trials in the field. Safety profile and clinical activity of pembrolizumab were first reported in KEYNOTE 012, an open-label phase 1b trial among patients with R/M HNSCC [34]. KEYNOTE 055, a phase-2 trial conducted among patients with platinum-resistant R/M HNSCC also reported manageable toxicity and an acceptable safety profile of pembrolizumab [35]. The study demonstrated a clinically meaningful anti-tumor activity of the agent in terms of ORRs and survival. These findings led to the initiation of KEYNOTE 040, the phase-3 trial investigating pembrolizumab for treating patients with platinum-refractory R/M HNSCC, and KEYNOTE 048, the phase-3 trial investigating pembrolizumab as first-line therapy in R/M HNSCC [23,25]. Similarly, two phase-2 trials, the May 24, 2022 Volume 13 Issue 5 HAWK study (a single-arm study investigating durvalumab monotherapy in R/M HNSCC with > 25% tumor PD-L1 expression) and the CONDOR phase-2 trial (an RCT investigating durvalumab with or without tremelimumab in PD-L1 Low/negative R/M HNSCC) served as the rationale for investigating combination immunotherapy regimens in platinum-refractory R/M HNSCC and to initiate the EAGLE study [24,36,37]. Studies on the effectiveness of nivolumab in other solid tumors supported the initiation of CheckMate 141 trial, the first phase-3 trial of nivolumab among patients with platinum-resistant R/M HNSCC [22,38]. Chemotherapy and radiotherapy, alone or in combination, have demonstrated potential synergetic effects when combined with immunotherapy in early phase studies. This phenomenon and the proven effectiveness of the anti-PD-L1 agent avelumab in other advanced solid tumors paved the way to the JAVELIN head and neck 100 trial, the first phase-3 RCT to investigate the effectiveness of combining ICI with chemoradiation in locally advanced head and neck cancer [27,39,40]. In this systematic review, we included the published phase-3 clinical trials evaluating the effectiveness of ICIs in HNSCC. Five studies met our eligibility criteria. Three studies (CheckMate 141, KEYNOTE 040, and EAGLE study) evaluated ICI as second-line treatment for R/M HSCC, one study (KEYNOTE 048) evaluated ICI as first-line treatment for R/M HSCC, while one phase-3 trial (JAVELIN head and neck 100 trial) evaluated the effectiveness of immunotherapy in LAHNSCC [22][23][24][25][26][27]. Effectiveness of ICI for R/M HNSCC in the second-line treatment setting In the second-line treatment setting, nivolumab in CheckMate 141 and pembrolizumab in KEYNOTE 040 demonstrated promising outcomes among patients with platinum-refractory R/M HNSCC [22,23]. In CheckMate 141, the anti-PD-1 agent nivolumab showed a statistically significant 31% reduction in risk of death (HR = 0.69, P = 0.01) and a net gain of 2.4 mo in terms of OS. A 2.3-fold increase in ORR was also reported with nivolumab compared to SOC. A favorable toxicity profile was another finding with nivolumab, with lower rates of treatment-related adverse events of grade 3 or more compared to SOC (13.1% vs 35%). Patient-reported QOL measures remained stable with nivolumab, while a decline in QOL occurred among the control patients. However, the study did not demonstrate any significant PFS benefits with nivolumab (HR = 0.89, 95%CI: 0.70-1.13; P = 0.32). Regarding the impact of biomarkers, survival benefit with nivolumab was found to be irrespective of PD-L1 expression (P for int. = 0.17) in the subgroup analyses based on PD-L1 status, although patients with PD-L1 ≥ 1% had a better magnitude of effect (HR = 0.55) than those with PD-L1 < 1 (HR = 0.89)[22, 41,42]. Similarly, based on the post-hoc exploratory subgroup analysis according to p16 status, the investigators concluded that the longer median OS with nivolumab was irrespective of the p16 status (P for interaction = 0.55). In KEYNOTE 040, the anti-PD-1 agent pembrolizumab demonstrated statistically significant improvement in OS with a 20% reduction in risk of death (HR = 0.80, P = 0.016) compared to SOC in the overall study population [23]. Higher ORR (14.6% vs 10.1%, nominal P = 0.061) and lower rates of adverse events of grade 3 or more (13% vs 36.1%) were also demonstrated with pembrolizumab compared to SOC. At 15 wk, stable GHS/QOL scores were reported with pembrolizumab, while the control patients had a decline in QOL. The study did not, however, demonstrate any PFS benefits with pembrolizumab (HR = 0.96, nominal P = 0.325) compared to SOC. Exploratory subgroup analyses based on PD-L1 expression demonstrated statistically significant interactions between treatment effects and PD-L1 status. For patients with TPS ≥ 50% and CPS > 1, the treatment effects of pembrolizumab vs SOC were found to be higher than in those with TPS < 50% and CPS < 1 [23]. For instance, in terms of OS, patients with TPS ≥ 50% had a net gain of 5 mo with a 47% reduction in risk of death with pembrolizumab compared to SOC (HR = 0.53, nominal P = 00014), suggesting PD-L1 expression may be explored as a predictive biomarker while selecting patients for pembrolizumab therapy. Based on the findings of CheckMate 141 and KEYNOTE 040, nivolumab and pembrolizumab were approved as standard second-line treatment options for platinum-resistant R/M HNSCC[22, 23,43]. The EAGLE study did not detect any statistically significant improvements in OS with durvalumab (HR = 0.88, P = 0.20) or with durvalumab plus tremelimumab (HR = 1.04, P = 0.76) compared to SOC. Again, there were no significant benefits in terms of PFS with durvalumab or with durvalumab plus tremelimumab, compared to SOC. However, investigators of EAGLE have postulated that control patients in the study had an unexpectedly high OS as the data were confounded by discrepancies in performance status favoring the control arm. Option of using paclitaxel as SOC (paclitaxel was not an option in the other two studies in the second-line setting), and subsequent immunotherapy after discontinuation of SOC treatment by control patients were also mentioned as reasons for this finding [24]. Although the primary objectives were not met, one positive finding was that the rates of adverse events of grade 3 or more were lower with immunotherapy compared to SOC. Effectiveness of ICI for R/M HNSCC in the first-line treatment setting In the first-line treatment setting, in KEYNOTE 048, pembrolizumab with platinum-based chemotherapy demonstrated statistically significant improvements in OS (13.0 vs 10.7 mo) with a 23% reduction in risk of death (HR = 0.77, P = 0.0034) compared to cetuximab plus platinum-based chemotherapy (EXTREME) in the total population. Pembrolizumab monotherapy was found to be noninferior to EXTREME (HR = 0.85; 95%CI: 0.71-1.03; P = 0.0456) in terms of OS (11.6 mo vs 10.7 mo) in the total population. No significant impact on PFS was detected with pembrolizumab alone or pembrolizumab with chemotherapy compared to EXTREME in the overall population. Pembrolizumab alone had a lower ORR (17%) compared to EXTREME (36%), while pembrolizumab plus chemotherapy had an ORR (36%) like that of EXTREME. Interestingly, biomarker (PD-L1) based stratified analysis demonstrated superiority in terms of OS in the CPS ≥ 20 and CPS ≥ subgroups with pembrolizumab alone as well as with pembrolizumab plus chemotherapy compared to EXTREME. For instance, within the CPS ≥ 20 population, pembrolizumab monotherapy compared to EXTREME resulted in a net gain of 4.2 mo in terms of OS (14.9 mo vs 10.7 mo) with a highly significant 39% reduction in risk of

death (HR = 0.61, P = 0.0007). In the CPS ≥ subgroup pembrolizumab monotherapy also demonstrated superiority in terms of OS (12.3 mo vs 10.3 mo) compared to EXTREME (HR = 0.78, P = 0.0086), indicating that pembrolizumab monotherapy is a suitable treatment option for PD-L1 positive R/M HNSCC. Similarly, in both subgroups, pembrolizumab with chemotherapy resulted in statistically significant improvements in OS compared to EXTREME. For instance, R/M HNSCC patients with CPS ≥ 20 had a highly significant 40% reduction in risk of death with pembrolizumab plus chemotherapy compared to EXTREME (HR = 0.60, P = 0.0004). Patients with CPS ≥ 1 also had a significant reduction in risk of death with pembrolizumab plus chemotherapy compared to EXTREME (HR = 0.65, P < 0.0001). These findings indicate that tumor PD-L1 expression can be a predictive biomarker for identifying patients who will benefit from pembrolizumab [25,44]. Based on the findings from KEYNOTE 048, pembrolizumab monotherapy was approved as an appropriate SOC for PD-L1 positive R/M HNSCC, and pembrolizumab plus platinum-based chemotherapy became the new SOC for the treatment of R/M HNSCC in the first-line setting [25,43]. In this study, rates of treatment-related adverse events of grade 3 or more were lower with pembrolizumab monotherapy (17%) compared to EXTREME (69%). However, rates of treatment-related adverse events of grade 3 or more were noticeably high (72%) in the combination therapy arm [25]. This finding highlights the importance of weighing up the survival benefits of the pembrolizumab plus chemotherapy regime against its adverse events profile while making treatment decisions for patients with R/M HNSCC. Effectiveness of ICI in LAHNSCC Regarding immunotherapy in LAHNSCC, there is no definite evidence of benefit according to the primary results of the JAVELIN study [26,27]. The combination of avelumab and CRT did not demonstrate any beneficial outcomes in terms of PFS or OS over placebo plus CRT, and based on the modified RECIST 1.1, there were no ORR benefits (74% vs 75%) either. Moreover, avelumab plus CRT resulted in slightly higher rates of adverse events of grade 3 or more compared to CRT plus placebo (80% vs 74%). As an explanation for the absence of PFS benefits, the investigators postulated that the dysfunction of T cells or changes in the tumor microenvironment after radiotherapy might have reduced the ability of the immune system to eliminate the microscopic disease. A recent phase-2 randomized trial of pembrolizumab with radiation therapy against cetuximab with radiotherapy in LAHNSCC also failed to demonstrate significant treatment benefits, although the combination therapy had a favorable toxicity profile [45]. Similarly, a previous randomized phase-2 trial of nivolumab with stereotactic body radiotherapy compared to nivolumab alone did not result in tumor shrinkage in R/M HNSCC [46]. Interestingly, an exploratory subgroup analysis of patients with high PD-L1 expression in the JAVELIN study indicated a potential PFS benefit with avelumab plus CRT compared to placebo plus CRT. Although definite conclusions cannot be made based on this small subgroup analysis, this is a finding that should be explored further to understand the role of biomarker analysis to select patients for immunotherapy. In terms of PFS, none of the studies included in this review demonstrated any beneficial outcomes. A recent meta-analysis by Gyawali et al[47] found no correlation between median OS and median PFS in studies evaluating anti-PD-1 agents. Defining PFS based on the traditional RECIST criteria (developed in the pre-immunotherapy era) that do not properly capture the concept of disease progression with immunotherapy was hypothesized as a probable reason for the finding. While immunotherapy involving anti-PD-1 checkpoint inhibitors resulted in significant improvements in survival, PD-L1 and CTLA-4 blockade did not demonstrate any encouraging outcomes. More studies are needed to build evidence on the role of anti-PD-L1 and CTLA-4 blocking agents in the treatment of advanced HNSCC. Again, in the first-line setting, the evidence on the effectiveness of immunotherapy for R/M HNSCC is based on one single phase-3 trial (KEYNOTE 048), and currently, pembrolizumab is the only ICI approved for treating this group of patients [25]. During our literature search, we identified some of the ongoing phase-3 clinical trials investigating various checkpoint inhibitor agents either alone or as part of combination therapy. Subsequently, we searched the 'clinical trials.org' database and identified the major ongoing clinical trials and confirmed the status of those trials. Studies investigating the combination of two different ICI agents or ICI in combination with another immunomodulatory agent in R/M HNSCC in the first-line treatment setting[48-51]: An ongoing openlabel phase-3 trial (KESTREL) is currently evaluating anti-PDL-1 agent durvalumab alone and in combination with the anti-CTLA-4 agent tremelimumab for R/M HNSCC against the EXTREME regime in the first-line treatment setting [48]. Checkmate 651, another ongoing phase-3 study, is currently evaluating the anti-PD-1 agent nivolumab in combination with the CTLA-4 blocking agent ipilimumab for R/M HNSCC against the EXTREME regime in the first-line setting [49]. In a phase-3 trial among R/M HNSCC, patients with a PD-L1 biomarker expression of CPS ≥ 1, the combination of pembrolizumab and lenvatinib, an anti-vascular endothelial growth factor-multiple kinase inhibitor, is being investigated as first-line treatment against pembrolizumab plus placebo [50]. Similarly, ICI in combination with another immunomodulatory agent is being investigated in the ECHO-304/KEYNOTE 669 study [51]. In this phase-3 trial, the combination of pembrolizumab and epacadostat, an indoleamine 2,3-dioxygenase 1, inhibitor agent is being investigated against pembrolizumab monotherapy, and the EXTREME regime, in R/M HNSCC as first-line treatment [51]. [52,53]: In KEYNOTE 412, the effectiveness of pembrolizumab given concurrently with CRT and as maintenance therapy is being evaluated against placebo plus standard CRT for the treatment of LAHNSCC [52]. In REACH, the superiority of avelumab in combination with RT-cetuximab compared to cisplatin -RT and/or to RT-cetuximab alone is being evaluated [53]. [54,55]: In HN004, durvalumab plus RT is being compared to cetuximab plus RT in platinum ineligible patients [54]. In a recently completed phase-3 trial with no published results (CheckMate 9TM), cisplatinineligible patients received nivolumab plus RT as intervention while control patients received cetuximab plus RT [55]. Studies investigating ICI as neoadjuvant/adjuvant therapy[56-59]: In KEYNOTE 689, pembrolizumab with RT (with or without cisplatin) before and after surgery is compared to RT (with or without cisplatin) given after surgery [56]. Atezolizumab, an anti-PD-L1 agent, is being evaluated as an adjuvant therapy against placebo in the ongoing trial iMvoke010 [57]. In IMSTAR-HN, nivolumab alone or in combination with the anti-CTLA-4 agent ipilimumab is evaluated as follow-up after adjuvant therapy against standard follow-up in surgically resectable LAHNSCC [58]. In NIVOPOSTOP, the efficacy of postoperative adjuvant nivolumab along with CRT is compared to post-operative CRT alone [59]. The details of these ongoing phase-3 studies are given in Table 2. Future directions Novel combination strategies to potentiate and prolong the anti-tumor activity of ICI are being evaluated currently. Thus, several early phase clinical trials (phase 1/2) investigating combination strategies of ICIs and other novel immunomodulatory agents are in the pipeline [60,61]. For example, a randomized phase-2 trial to study the safety and tolerability of nivolumab administered alone or in combination with relatlimab (antibody targeting the novel immunomodulatory receptor lymphocyte activation gene-3) or the anti-CTLA-4 agent ipilimumab is currently ongoing among patients with locally advanced surgically resectable HNSCC [62]. Immune biomarker modulation in response to nivolumab given along with Toll-like receptor 8 agonist motolimod is being analyzed in an ongoing phase-1b pre-operative biomarker trial [63]. Combination of pembrolizumab and the vascular endothelial growth factor-multiple kinase inhibitor lenvatinib demonstrated good anti-tumor activity and manageable toxicity among R/M HNSCC patients in a phase-1b/2 trial, and LEAP 010, a phase-3 trial of this combination strategy is currently ongoing [50,64]. The combination of pembrolizumab and the anti-EGFR agent cetuximab had demonstrated encouraging outcomes in the interim analysis of an ongoing multi-arm phase-2 trial [65,66]. A recently completed study among R/M HNSCC patients investigating pembrolizumab in combination with epacadostat has shown clinically meaningful results, and a larger phase-3 trial (ECHO 304/KEYNOTE 669) of this combination strategy is ongoing currently [51,67]. Combination therapy of pembrolizumab with the EGFR-tyrosine kinase inhibitor afatinib, which also included predictive biomarker analysis, had been evaluated recently in a phase-2 clinical trial (the ALPHA study) in R/M HNSC [68]. The study demonstrated augmentation of the anti-tumor activity of pembrolizumab by afatinib, and the results of biomarker analysis suggested that PD-L1 and EGFR amplification could be predictive biomarkers for cancer immunotherapy. EACH, a randomized phase-2 trial among R/M HNSCC is investigating the superiority of avelumab and cetuximab combination compared to avelumab monotherapy [69]. Another recently completed early phase study on the combination of pembrolizumab with the therapeutic vaccine talimogene laherparepvec demonstrated a tolerable safety profile among patients with R/M HNSCC. However, this investigation did not progress into a phase-3 trial as the efficacy of the combination was found to be similar to pembrolizumab monotherapy [70]. Immunotherapy trials among patients with p16-positive head and neck cancer (oropharyngeal squamous cell carcinoma) are also currently underway. In this group of patients, p16 positivity is a known independent predictive biomarker for survival [71]. The efficacy and tolerability of the combination of ipilimumab (anti-CTLA-4) and nivolumab (anti-PD-1) along with RT in locoregionally advanced human papilloma virus-positive oropharyngeal squamous cell carcinoma are being evaluated in an ongoing phase-2 single-arm trial [72]. Another phase-2 randomized study (KEYCHAIN trial) is investigating RT along with concurrent and adjuvant pembrolizumab against concurrent chemoradiation among p16-positive HNSCC [73,74]. Regarding biomarkers, in addition to p-16 positivity and PD-L1 expression, other biomarkers like microsatellite instability (MSI) and tumor mutation burden were also found to be associated with favorable outcomes with ICI therapy in HNSCC[75]. Tardy et al [76] recently reported a case of complete response to anti-PD-L1 therapy in HNSCC in a patient with high tumor MSI (MSI-H) and a negative PD-L1 histochemical status. Similarly, Hanna et al [77] reported that higher tumor mutation burden predicted response to ICI and better treatment outcomes in virus-negative head and neck cancer. Again, some subtypes of tumor-infiltrating lymphocytes (TILs) such as PD-1 + TIM-3 + CD8 + TILs and PD-1 + LAG-3 + CD8 + TILs have also predicted treatment response to ICIs [75,77]. The data on these emerging predictive biomarkers is still not conclusive; therefore, further research is essential. PRECISION 01, an ongoing prospective observational study is currently evaluating biomarker signatures in tissue samples of platinum-refractory HNSCC patients who received nivolumab monotherapy; the findings may contribute to the knowledge on predictive biomarkers for ICIs [78]. In future studies, patient-reported outcomes like QOL should be evaluated meticulously since such outcomes are very crucial for advanced HNSCC patients and their families [79,80]. Cost-effectiveness is another issue to be considered before including ICIs in the routine treatment guidelines for patients from developing countries and resource-poor settings [81,82]. The impact of factors like age, comorbidities, and performance status on outcomes of patients receiving immunotherapy also needs to be determined [83]. Limitations/strengths There are very few published phase-3 clinical trials evaluating checkpoint inhibitor immunotherapy among patients diagnosed with HNSCC, and the evidence we gathered in this review is based on the five phase-3 RCTs published so far. A previous systematic review on this topic included eight studies, of which two were phase-3 RCTs [84]. Wang et al [85] conducted a systematic review and meta-analysis of nine studies on the effectiveness of checkpoint inhibitors in HNSCC, of which two were phase-3 trials. To our knowledge, this is the first systematic review conducted on the effectiveness of ICIs in HNSCC incorporating phase-3 trials alone. The evidence we presented based on the five studies in this review will help the practicing clinicians to make informed decisions. We further explored the literature and identified a variety of promising clinical studies that are ongoing currently focusing on combination strategies in enhancing and prolonging the anti-tumor effects of ICIs. We also identified the gaps in knowledge on some important issues such as predictive biomarkers and about the identification of patients who will benefit from immunotherapy based on biomarker assessment [86,87]. CONCLUSION ICIs have shown improved survival outcomes with acceptable toxicity profile in R/MHNSCC in the first and second-line treatment settings. The marginal improvement in survival should be weighed against the cost of these therapeutic agents and the QOL of patients. While anti-PD-1 agents demonstrated efficacy, evidence on the effectiveness of anti-PD-L1 and anti-CTLA-4 agents is lacking. There is no proven efficacy in the curative setting to date. The ongoing clinical trials may better define the role of

ICI in R/M HNSCC and LAHNSCC in the future. Research background Head and neck squamous cell carcinoma (HNSCC) is one of the major causes of cancer-associated morbidity and mortality globally, especially in developing countries. Treatment approaches for HNSCC vary according to the stage of the disease at presentation. For recurrent/metastatic HNSCC (R/M HNSCC), platinum-based chemotherapy was the only available treatment option until recently. A relatively new systemic therapy option that emerged in recent years in the treatment of advanced HNSCC is immunotherapy using immune checkpoint inhibitors (ICI). Research motivation Advanced HNSCCs are often associated with significant functional limitations, and aggressive treatment may adversely affect the quality of life of these patients who are already suffering from the effect of advanced cancer. The median survival of R/M HNSCC patients receiving platinum-based chemotherapy is 7.4 mo. Some patients become refractory to platinum and die within a period of 4 mo. The safety profile and anti-tumor activity of ICIs demonstrated in early phase clinical trials paved the way to the initiation of several promising phase-3 trials in the field. Therefore, we decided to gather the current evidence on the effectiveness of these agents in advanced head and neck cancer based on the findings from phase-3 clinical trials of ICI published so far. We also wanted to examine the feasibility of incorporating these agents into routine clinical practice in resource-poor settings. Research objectives The objective of this systematic review was to gather the evidence from phase-3 randomized controlled trials (RCTs) evaluating the effectiveness of immunotherapy among patients with advanced HNSCC. We aimed to synthesize the evidence from the published phase-3 studies that investigated the efficacy and toxicity profile of ICIs administered either alone or in combination with chemotherapy, radiation therapy, or with another checkpoint inhibitor, in advanced HNSCC. Research results Five phase-3 clinical trials have reported the data on the effectiveness of immunotherapy in HNSCC so far: Four in R/M HNSCC and one in LAHNSCC. In patients with R/M HNSCC, anti-PD-1 agents nivolumab and pembrolizumab demonstrated improvement in overall survival (OS) in the second-line treatment setting compared to the SOC. While the net gain in OS with nivolumab was 2.4 mo, that with pembrolizumab was 1.5 mo. However, the study that investigated the anti-PD-L1 agent durvalumab with or without the anti-CTLA-4 agent tremelimumab in the second-line treatment setting did not demonstrate any beneficial outcomes. In the first-line setting, pembrolizumab together with platinum-based chemotherapy demonstrated statistically significant improvement in survival with a net gain in OS of 2.3 mo in the overall population and a net gain in OS of 4.2 mo in the population with a combined positive score of > 20 compared to the SOC treatment. Pembrolizumab monotherapy was found to be non-inferior to EXTREME in terms of OS (11.6 mo vs 10.7 mo) in the total population. In patients with PD-L1 positive R/M HNSCC, monotherapy with pembrolizumab also demonstrated statistically significant improvement in survival compared to SOC. In LAHNSCC, immunotherapy using the anti-PD-L1 agent avelumab along with standard chemoradiation therapy did not result in improved outcomes compared to placebo plus chemoradiation therapy. Research conclusions This systematic review helped us to conclude that anti-PD-1 agents provide survival benefits in R/M HNSCC in the first and second-line settings with manageable toxicity profiles. However, it is important to weigh the marginal survival benefits provided by these therapeutic agents against their cost, especially in resource-poor settings. The review showed that the evidence on the effectiveness of anti-PD-L1 and anti-CTLA-4 agents in advanced head and neck cancer is lacking. To date, there is no evidence on the effectiveness of ICIs in the curative setting either. We believe that the ongoing clinical trials (discussed in the article) will help to define better the role of ICI in R/M HNSCC and LAHNSCC in the future. Research perspectives Novel combination strategies to potentiate and prolong the anti-tumor activity of ICI are being evaluated currently. Gaps in knowledge exist on some important issues such as predictive biomarkers, and about the identification of patients who will benefit from immunotherapy based on biomarker assessment. Future studies should focus on these issues. Kinesiophobia in Breast Cancer Survivors and its Relationship with Quality of Life, Comorbidity and Other Clinical Parameters Meme Kanseri Hastalarında Kinezyofobinin Hayat Kalitesi, Komorbidite ve Diğer Klinik Özellikler ile İlişkisi Objectives: This cross-sectional study aims to determine the frequency of kinesiophobia in breast cancer survivors and evaluate its relationship with mainly quality of life and comorbidities, also fatigue, lymphedema, and depression. Material Methods: This study included 54 women with breast cancer who were followed in remission in Aydın Atatürk State Hospital Medical Oncology Clinic between November-December 2020. Clinicopathological characteristics of the patients were recorded. Kinesiophobia was assessed using the Turkish version of Tampa Scale for Kinesiophobia (TSK). Lymphedema was evaluated with bilateral upper extremity measurements. Depressive status and quality of life were determined using the Beck Depression Inventory (BDI) and European Organization for Research and Treatment of Cancer Quality of Life Questionnaire Version 3.0 (EORTC QLQ-C30 v3). Comorbidities were assessed using the Charlson Comorbidity Index (CCI) while fatigue was evaluated by 10 cm visual analogue scale (VAS). The relationship between the TSK and CCI, BDI, VAS-F, EORTC-30 were investigated. Results: The mean age of patients was 52.11±11.10 years. Of the patients, 36 (66.7%) had kinesiophobia based on having TSK scores above 37. The rate to recieve adjuvant radiotherapy was higher in kinesiophobic group. Kinesiophobic patients had statistically significantly lower global health, physical functioning, and emotional functioning scores, and higher depression and VAS-fatigue scores. Kinesiophobic patients also had higher financial difficulty and higher sypmtom scale scores. Comorbidities and presence of lymphedema did not differ between groups (p>0.05). All EORTC QLQ30 sub-parameter scores except for financial difficulty and symptom severity had negative significant correlations with TSK scores while VAS-fatigue, BDI, EORTC QLQ-30 symptom scale, and financial difficulties showed significant positive correlations. TSK score was not correlated with CCI. Conclusion: Kinesiophobia is rather frequent in breast cancer survivors and has associations with worse quality of life and higher depression and fatigue scores. These patients should be trained and encouraged for an active life style. Introduction Breast cancer is the most common cancer type and leading cause of cancer-related death among women in Turkey [1]. Cancer survivors may experience depression, anxiety, and poor health-related quality of life [2]. Furthermore, fatigue and pain are frequent in cancer patients and survivors and might lead to low level of physical activity and sedentary lifestyle [3][4][5][6]. Kinesiophobia is defined as the fear and anxiety of movement and activity due to belief of fragility or susceptibility to injury [7]. The Cognitive Fear Avoidance Model suggests that pain-related fear leads to escape mechanisms resulting in avoidance of movement and activity. Thereafter, in prolonged avoidance, they run a vicious circle occurring due to non-use, disability, and depression [8]. Although there are several papers on kinesiophobia and chronic back pain, musculoskeletal disorders, and cardiovascular diseases, there are few studies about breast cancer and kinesiophobia [9][10][11]. Previous studies showed that approximately 40% of pain-related disability could be attributed to kinesiophobia [12]. Also, it was propounded that kinesiophobia increased the risk for lymphedema, depression, and poorer upper extremity function in breast cancer survivors [13]. A previous study assessing the relationship between kinesiophobia and global health status on 1236 cancer survivors concluded that fear of movement was significantly related to global health status. The paper further reported that kinesiophobia decreased significantly after rehabilitation with graded activity in high TSK scorers [14]. Approximately 60%-96% of the cancer patients are reported to have high levels of fatigue during or after cancer treatment, which often leads to diminished quality of life [15]. In a study evaluating the quality of life using the Short form-36 (SF-36) in breast cancer survivors, SF-36 physical component score (PCS) and mental component score (MCS) were lower in kinesiophobic patients, where only SF-36 PCS difference was statistically significant. Kinesiophobic patients had also significantly higher mean scores of depression and significant correlations between presence of lymphedema, depression scores, and the TSK score were noted [13]. In this study on 81 breast cancer survivors, a significantly higher rate of lymphedema was also reported in patients with kinesiophobia [13]. We consider that the desire to engage in physical movement may be hampered in breast cancer survivors along with depression and fatigue particularly in case of comorbid diseases. Therefore, we aimed to determine the frequency of kinesiophobia in breast cancer survivors and investigate its relationship with mainly quality of life and comorbidities, secondly lymphedema, fatigue, and depression in the current study. Material Methods: This cross-sectional study included 54 patients with breast cancer who were followed in remission in Aydın Atatürk State Hospital Medical Oncology Clinic beetwen November and December 2020. The inclusion criteria were having breast surgery at least a year ago and being followed in remission with diagnosis of breast cancer, at least 6 months since last adjuvant chemotherapy cycle or radiotherapy, being over 18 years, and being able to read and write in Turkish. The exclusion criteria were having metastatic disease, co-existent second malignancy, chronic inflammatory diseases, and cognitive or psychiatric disturbances that may impede with fulfilling the questionnaires. The demographic characteristics including age, marital and educational status, body mass index, stage of disease, side and type of surgery, hormon receptor status, history of adjuvant chemotherapy, radiotherapy and hormonotherapy were recorded. Lymphedema of the arm was evaluated by circumferenial measurements of bilateral upper extremities. Measurements were performed at four levels of metacarpal, wrist, 10-cm below and above the lateral epicondyle using a standard retractable fiberglass tape in sitting position with 90° shoulder flexion, elbow extension, and forearm pronation. Assessments were fulfilled by the same physician and noted in cm unit in cm unit. A difference of ≥2 cm at any single location in the affected arm led to the diagnosis of lymphedema. Kinesophobia Kinesophobia was evaluted using the Turkish version of Tampa Scale for Kinesophobia (TSK) [16,17]. This scale is a 17-item scale widely used for chronic musculoskeletal disorders. A 4-point likert questionnaire ranging from "I fully disagree" to "I fully agree" is applied for each item. After reversing the 4th, 8th, 12th, and 16th items, total score of 17-68 is reached. The higher scores indicate higher kinesiophobia level. While the use of total score is advocated in studies, the cut off point of 37 is determined to define severe and mild kinesiophobia [8]. Comorbidity The presence and intensity of comorbidities was assessed using the Charlson Comorbidity Index (CCI). The index predicts 10-year mortality for patients with various comorbid conditions depending on their strength of association with mortality. Total score is counted summing the scores of 16 available clinical conditions. Higher scores indicate higher mortality risk [18]. Quality of Life The quality of life was evaluated using the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire Version 3.0 (EORTC QLQ-C30 v. 3.0). The EORTC QLQ C-30 is a cancer-specific QOL questionnaire with fourpoint likert scale with items ranging from 1:not at all to 4: very much. This 30-item selfadministered questionnaire evaluates five functional scales physical (PF), role (RF), cognitive (CF), emotional (EF) and social (SF) functioning, and three symptom scales (fatigue, pain, and nausea/vomiting), a global Quality of Life scale, and some single items for evaluation of other complaints of cancer patients (e.g. dyspnea, loss of appetite, sleep disturbances, constipation, and diarrhea), and financial effects of the disease and treatment. All functional scales and individual item scores are indicated on a 0-100 scale. Higher functional scale values indicate better functioning while increased scores of symptom scales and financial difficulties suggest worse symptoms and poorer financial status [19]. Beck Depression Invantory The depressive status was assessed using the Beck Depression Invantory (BDI). It is a 21item scale scored 0-3 for each item. Higher scores indicate more severe depressive state, namely 10 refers to minimal depression; 10-18 to mild to moderate depression, 19-29 to severe depression; and above 30 is determined as severe depression [20]. Fatigue severity according to visual analogue scale (VAS 10 cm) was noted where increased scores corresponded to more intense fatigue. The ethics committee aproval was obtined from Adnan Menderes University Clinical Researches Ethics Committee. All patients were provided with information about the study and were given informed consent forms. Statistical analysis The SPSS version 11.5 software (SPSS Inc., Chicago, IL, USA) was used for analyses. Histogram and p-plots were

examined and Kolmogorov Smirnov test was used to assess data normality before statistical analyses. Continuous variables were presented as mean ± standard deviation and median (interquartile range) and tested by student's t test or Mann Whitney U test. Categorical variables were described as numbers and percentages, and tested by Chi-square or Fisher's Exact tests. Pearson correlation coefficient was used to determine the relationship between continuous variables. Patients were determined to be kinesiophobic if TSK scores were above 37 [8]. Patients with and without kinesiophobia were compared in terms of clinical parameters. p<0.05 was considered statistically significant. Results The mean age of patients was 52.11±11.10 years. The mean disease duration of 54 patients was 51.61±5.91 months. The clinical characteristics of patients were given in Table 1. Of the patients, 36 (66.7%) had kinesiophobia based on having TSK scores above 37. The mean TSK score was 39.31±5.89 in the whole group. When patients were further divided into as kinesiophobic patients and others, the rate to have recieved adjuvant radiotherapy was higher in kinesiophobic group (p=0.038). Furthermore, kinesiophobic patients had statistically significantly lower global health, physical functioning, and emotional functioning scores and higher depression and VAS-fatigue scores (p=0.014, p=0.009, p=0.015, p=0.006, p=0.040 respectively). Kinesiophobic patients also had higher scores of financial difficulty and sypmtom scales indicating worse financial and somatic symptom burden (p=0.034, p=0.013 respectively). Comorbidities, presence of lymphedema, and Table 3 shows the correlations between fear of movement, fatigue severity, and comorbidities in patients with breast cancer. Discussion In this study, the frequency of kinesiophobia and its association with quality of life and comorbidities were investigated in breast cancer survivors. This study revealed that almost two third of breast cancer survivors suffer from kinesiophobia. Also we have observed that patients with kinesiophobia had poorer quality of life, higher depression, and fatigue scores. Comorbidity and lymphedema were not found to have significant relationship with kinesiophobia in breast cancer survivors. In our study group, 66.7% of patients had kinesiophobia which seems higher than previous studies [13]. The socioeconomic differences of groups assessed might have led to this discrepancy. In the prior studies, it was stated that cancer survivors with increased fear of movement might benefit from rehabilitation programs with graded activity [14]. Therefore, we consider these programs may be offered to all patients with suspected fear of movement. In a study on 1236 cancer survivors among whom 615 had breast cancer, which evaluated the relationship between kinesiophobia and global health status, kinesiophobia was reported to be inversely associated with global health status determined by the EORTC QLQ-30, similar to our results [14]. Similarly, another study on 62 women with breast cancer evaluating the quality of life using the functional assessment of cancer therapybreast (FACT-B+4), a significant negative relationship between kinesiophobia and quality of life was reported [21]. In the current study, we also detected significant relationship between kinesophobia and physical and emotional functioning sub-scales of the EORTC QLQ-30. The breast cancer patients with upper extremity lymphedema may have a belief that their arms might become swollen if they move them and avoid daily activities which can contribute to kinesiophobia [21]. The frequency of lymphedema is 53.7 % in our study group. While previous papers suggest distinct results, a report remarks a cumulative incidence of 41.1% [22]. The study by Gencay et al. noted a significant association between presence of lymphedema and the TSK scores whereas we did not observe such a relationship [13]. Similar to their results, our patients with kinesiophobia had worse quality of life and higher depression scores [13]. The relationship between depression and fear of movement was also revealed in other patient groups [23,24]. It is known that up to 40% -80% of cancer patients undergoing active treatment suffer from cancer related fatigue which can result in decreased quality of life [25]. In a study evaluating the relationship between kinesiophobia and fatigue, fatigue severity was found to be associated with kinesiophobia in patients with chronic obstructive pulmonary disease [26]. Similarly, we observed a positive correlation between fatigue and kinesiophobia scores in breast cancer survivors and statistically significant difference in terms of fatigue between patients with and without kinesiophobia in our study group. In that study by Vardar et al., kinesiophobia was strongly associated with multisystemic comorbidities in patients with chronic obstructive pulmonary disease [26]. However, we detected no relationship between kinesiophobia and comorbidity in breast cancer survivors in contrast to the case in aforementioned study. As far as we are concerned, this is the first study to evaluate the relationship between kinesiophobia and comorbidity in breast cancer survivors. Another strong element of the current study is assessing the quality of life with a cancer-specific scale. The limitation of our study is the small sample size. Further studies are required to assess the role of comorbidity on kinesiophobia in these patients. In this study, we found that kinesiophobia is rather frequent in breast cancer survivors and has relationship with quality of life, depression, and fatigue. These patients should be encouraged for an active life style. Within- and across-frequency temporal processing and speech perception in cochlear implant users Objective Cochlear implant (CI) recipient’s speech perception performance is highly variable and is influenced by temporal processing abilities. Temporal processing is commonly assessed using a behavioral task that requires the participant to detect a silent gap with the pre- and post-gap stimuli of the same frequency (within-frequency gap detection) or of different frequencies (across-frequency gap detection). The purpose of the study was to evaluate behavioral and electrophysiological measures of within- and across-frequency temporal processing and their correlations with speech perception performance in CI users. Design Participants included 11 post-lingually deafened adult CI users (n = 15 ears; Mean Age = 50.2 yrs) and 11 age- and gender-matched normal hearing (NH) individuals (n = 15 ears; Mean Age = 49.0 yrs). Speech perception was assessed with Consonant-Nucleus-Consonant Word Recognition (CNC), Arizona Biomedical Sentence Recognition (AzBio), and Bamford-Kowal-Bench Speech-in-Noise Test (BKB-SIN) tests. Within- and across-frequency behavioral gap detection thresholds (referred to as the GDTwithin and GDTacross) were measured using an adaptive, two-alternative, forced-choice procedure. Cortical auditory evoked potentials (CAEPs) were elicited using within- and across-frequency gap stimuli under four gap duration conditions (no gap, GDT, sub-threshold GDT, and supra-threshold GDT). Correlations among speech perception, GDTs, and CAEPs were examined. Results CI users had poorer speech perception scores compared to NH listeners (p < 0.05), but the GDTs were not different between groups (p > 0.05). Compared to NH peers, CI users showed increased N1 latency in the CAEPs evoked by the across-frequency gap stimuli (p < 0.05). No group difference was observed for the CAEPs evoked by the within-frequency gap (p > 0.05). Three CI ears showing the longest GDTwithin also showed the poorest performance in speech in noise. The within-frequency CAEP increased in amplitude with the increase of gap duration; while the across-frequency CAEP displayed a similar amplitude for all gap durations. There was a significant correlation between speech scores and within-frequency CAEP measures for the supra-threshold GDT condition, with CI users with poorer speech performance having a smaller N1-P2 amplitude and longer N1 latency. No correlations were found among GDTacross, speech perception, and across-frequency CAEP measures. Conclusions Within- and across-frequency gap detection may involve different neural mechanisms. The within-frequency gap detection task can help identify CI users with poor speech performance for rehabilitation. The within-frequency CAEP is a better predictor for speech perception performance than the across-frequency CAEP. Introduction Temporal resolution, the ability to resolve rapid changes within the acoustic signal over time, is critical for speech understanding, music appreciation and sound localization [1][2][3]. A gap detection paradigm is a standard measure of auditory temporal resolution, which measures the shortest gap of silence between two acoustic markers that an individual can detect [4][5][6][7][8][9][10]. Previous researchers have commonly examined gap detection with the pre-and post-gap markers that are spectrally identical or similar, which is referred to as within-frequency gap detection [4,11,12]. Alternatively, gap detection can be assessed using the pre-and post-gap acoustic markers that are spectrally distinct, which is referred to as across-frequency gap detection [13,14]. Gap detection thresholds (GDTs) are poorer in hearing-impaired listeners relative to normal hearing (NH) listeners due to the damage in the auditory nervous system [10]. In hearingimpaired listeners using a cochlear implant (CI), a prosthetic device converts sounds into electrical pulses that directly stimulate the auditory nerve. GDTs can be evaluated using different presentation modalities including acoustic stimulation and direct electrical stimulation (by passes the CI speech processor). Acoustic GDTs reflect the temporal processing impairments within the auditory system and signal distortion associated with the speech processing strategy and individual map parameters (stimulation rate, maxima, preprocessing strategies). Therefore, acoustic GDTs are expected to be poorer than those measured with direct electrical stimulation and are more representative of temporal processing abilities in everyday situations. Examining CI users' GDTs using acoustic stimulation may provide valuable information on why the clinical outcomes of cochlear implantation are highly variable across individuals. CI users display a wide range of within-frequency gap detection threshold (GDT within ) with some individuals performing comparable to their NH peers (several ms) and others displaying much larger GDT within [15][16][17][18][19]. The correlation between GDT within and CI users' speech perception performance has demonstrated that larger GDT within is linked to poorer speech performance. Tyler et al. [20] reported the GDT within in CI users with a smaller GDT within (< 40 ms) displayed better speech perception performance than individuals with a larger GDT within (> 40 ms). Muchnik et al. [21] reported that CI users with open-set speech recognition had lower GDT within (Range:12-46 ms) than CI users without open-set speech recognition (GDT Range: 41-72 ms). The authors of the present study reported that CI recipients with smaller GDT within had better speech perception scores for words in quiet and sentences in noise [18]. In our everyday environment, sounds before and after silent gaps are rarely identical in frequency, therefore across-frequency gap detection thresholds (GDT across ) might be a measure that is more correlated to speech perception than the GDT within . The GDT across has been reported to be higher than the GDT within [8,13,[22][23][24], and it is highly variable even in NH listeners [23,25]. The GDT across increases as the pre-gap marker duration is shortened [13,17] and as the frequency separation is increased [7,26,27]. The GDT across in CI users has been examined in a limited number of studies, only using direct electrical stimulation [14,28,29]. For instance, van Wieringen and Wouters [28] reported that the GDT across was greater than 50 ms, which was much larger than the GDT within (< 5 ms). Hanekom and Shannon [14] reported that the GDT across increased when the channel separation between the markers increased and the GDT across (10-200 ms) was much larger than the GDT within (Range:1-4 ms). Although behavioral GDTs are easily measured in adults, the tasks are not feasible for individuals who cannot perform behavioral tasks reliably, e.g., young children or adults with cognitive deficits. Thus, an objective electroencephalographic (EEG) measure that does not rely on behavioral cooperation but correlates well with behavioral GDTs would be beneficial to evaluate temporal processing in difficult-to-test populations [30]. Moreover, the combination of EEG and behavioral measures will explain possible neural mechanisms underlying gap detection and temporal processing, which is impossible if only behavioral measures alone are used, as in most previous studies [8,13,22,28,31]. The cortical auditory evoked potential (CAEP) recorded using EEG techniques has been examined using stimuli containing silent gaps in young adults with normal hearing [32][33][34][35][36][37][38], older adults with normal hearing or minimal hearing loss [36,39,40], individuals with auditory neuropathy [30,32,41] and cochlear implant recipients [41,42]. When evoked by a stimulus containing a gap, the CAEP peaks, N1 and P2, occur at approximately 100 and 200 ms, respectively, after the onset of the pre-and post-gap markers (pre-gap CAEP and post-gap CAEP). In CI candidates with auditory neuropathy, He et al. [30] reported a significant negative correlation between open-set word recognition and the minimum gap in noise that can evoke a CAEP. In CI users, Mussoi and Brown [42] did not show a correlation between behavioral GDTs or CAEP evoked by electrical stimuli containing gaps and speech in noise performance. Gap duration has a differential effect on the post-gap CAEP response amplitude and latency. With regard to

the effect of gap duration on amplitude, it has been consistently reported that the post-gap CAEP amplitude increases as a function of the saliency of the acoustic change [32,33,35,37,38,43]. In addition, gap durations that are inaudible or not behaviorally perceived (sub-threshold) generally do not elicit a post-gap CAEP response, and suprathreshold gap durations generate repeatable CAEP responses. Lastly, as the gap duration decreases towards threshold, the amplitude of the post-gap CAEP waveform subsequently decreases as well [37,38]. With regard to the effect of gap duration on latency, the results are not as conclusive. Some studies report that gap duration had a significant effect on post-gap CAEP latency [33,37], while other studies reported no effect [32,38], or a non-monotonic effect on latency [35]. Pratt et al. [33] reported that the post-gap CAEP latency decreased as the gap duration increased. In contrast, Michalewski et al. [32] examined the effect of gap duration (2, 5, 10, 20 and 50 ms gaps) on post-gap CAEP latencies. They observed that while the 50 ms gap resulted in the longest latency compared to shorter gap duration conditions, results did not reach statistical significance. Palmer and Musiek [38] reported no change in post-gap CAEP latency as gap duration changed. Lastly, He et al. [35] reported that for gap durations up to 20 ms the post-gap CAEP latency tended to decrease, however, for longer gap durations the post-gap CAEP latency increased. Taken together, results suggest the post-gap CAEP amplitude is a better indicator of auditory discrimination abilities and can be used as an objective indicator of the neural encoding of an acoustic change at the level of the auditory cortex. In summary, the CAEP evoked by within-frequency gap stimuli is a promising objective tool to examine neural aspects of temporal processing. However, limited research on the CAEP has been conducted using across-frequency gap stimuli. Given the importance of temporal processing to speech perception, the present study was undertaken to: (1) examine the behavioral GDT within and GDT across and speech perception; 2) examine the CAEP evoked by within-and across-frequency gap stimuli; and (3) explore the relationship among behavioral GDTs, speech perception performance, and CAEP measures. The results would not only help understand the large variability in CI speech outcomes and but also result in the use of nonlinguistic and objective measures to predict CI speech outcomes. Participants Eleven adult CI users (Mean Age = 50.2 yrs; Age Range = 25.2-68.3) and eleven age-and gender-matched NH adults (Mean Age = 49.0 yrs; Age Range = 24.7-68.5) were enrolled in the study. Four CI users were bilaterally implanted and the rest were unilaterally implanted, with a total of 15 CI ears being tested individually. The NH listeners were tested in the same ear as their matched CI users, with a total of 15 NH ears being tested individually. All CI users were post-lingually deafened (onset of bilateral severe-to-profound hearing loss > 3 years of age), implanted with Cochlear Americas Implant System, (Cochlear Americas Ltd, New South Wales, Australia) and had a minimum of 1 year CI experience. CI users reported full time use of their device during all waking hours. Relevant demographic and device information is displayed in Table 1. All NH and CI participants were right-handed (Edinburgh Handedness Inventory; [44]), native speakers of American English, and did not report a history of neurological or psychiatric disorders, or brain injury. The research study was approved by the Institutional Review Board of the University of Cincinnati. Written informed consent was obtained from all participants and they were paid for participation. Stimuli The stimuli used for behavioral gap detection tasks and EEG recordings were pure tones, consisting of a pre-gap marker, a silent gap, and a post-gap marker (Fig 1). The pure-tone markers (1 and 2 kHz) were created using Audacity software (version 1.2.5; open source, http:// audacity.sourceforge.ent) with a sampling rate of 44.1 kHz. Pure-tone stimuli instead of more complex stimuli were used because: 1) stimulus complexity can affect GDTs [19,45], and 2) the GDT assessed with pure tones is a more accurate measure of temporal resolution [4]. All stimuli were shaped with a cos 2 window and included a 10 ms rise-fall time. A silent gap was inserted between the pre-and post-gap markers and included a 1 ms risefall time around the silent gap, as in several previous studies [37,43,46]. Furthermore, the marker frequencies (1 and 2 kHz) were selected to facilitate comparison to studies that used a similar paradigm in adults with normal hearing or minimal hearing loss [37,39]. The pre-gap marker was a 2 kHz tone for all stimuli. The post-gap marker was a 2 kHz tone for the withinfrequency stimuli and a 1 kHz tone for the across-frequency stimuli. For the behavioral gap detection task, the pre and post-gap markers varied in duration from 250 to 350 ms to prevent the participant from using duration cues to aid in behavioral gap detection [7,37]. For EEG testing, the pre and post-gap marker duration was fixed at 300 ms to allow averaging of the time-locked neural responses across stimulus presentations. An advantage of using a relatively long duration for pre-and post-gap markers (300 ms) was that the CAEPs following the preand post-gap markers were less likely to overlap. General study procedures All testing was completed in a double-walled sound treated booth (Industrial Acoustics Company, North Aurora, Illinois) over the course of one or two sessions depending on participant's preference and if one ear (3 hours of testing) or both ears were tested (6 hours of testing). All participants completed testing in the following order: audiological assessment, speech perception testing, behavioral GDT assessment, EEG testing. CI users completed all testing with their speech processors turned on and adjusted to their everyday settings (volume, sensitivity, and program) which were held constant throughout the entire test session(s). For NH individuals and unilateral CI users, testing was completed with the contralateral ear occluded with an E-A-R disposable foam ear plug. For bilateral CI users, testing was completed with the contralateral speech processor removed. All testing was completed at the Most Comfortable Level (MCL; 7 on a 0-10 loudness scale) [47]. Testing at the MCL has been commonly used in CI research and it allows easier comparison of test results between NH and CI users [4,48,49]. Audiological assessment. All NH listeners completed otoscopy and 226 Hz tympanometry to ensure normal outer and middle ear status. Next audiometric thresholds were measured (GSI-61 or AudioStar Pro; Grason Stradler, Eden Prairie, MN). For NH listeners, thresholds were measured at octave test frequencies from 0.25 to 8 kHz with pulsed pure-tones using insert ER-3A (Etymotic Research, Elk Grove Village, IL) earphones. For CI users, audiometric thresholds were measured using frequency-modulated tones from 0.25 to 6 kHz using sound field speakers (LSR305, Sweetwater, IN). The purpose of measuring hearing thresholds was to ensure the presentation level for the GDT assessment was at least 25-35 dB above threshold, which is necessary for optimal gap detection [6]. Speech perception assessment. Speech perception performance was measured with the following commonly used clinical speech tests for adult CI Users [50][51][52]: 1) the Consonant-Nucleus-Consonant Word Test (CNC) assesses open-set monosyllabic word recognition in quiet. Two CNC word lists per test ear were administered to each tested ear; 2) the Arizona Biomedical Sentence Recognition Test (AzBio) assesses open-set sentence recognition in quiet and noise (AzBio-Quiet, AzBio-Noise). Participants were presented with two lists in quiet and two lists in noise (+10 dB signal-to-noise ratio; SNR) [52]; and 3) the Bamford-Kowal-Bench Speech-in-Noise Test (BKB-SIN) measures open-set sentence recognition with SNRs that ranged from +21 to -6 dB using an adaptive procedure. Two BKB-SIN word list pairs were administered [50]. All stimuli were presented at the MCL through either an insert earphone (NH PLOS ONE Temporal processing and speech perception in cochlear implant users listeners) or a sound field speaker (LSR305, Sweetwater, IN) placed at 0 degrees azimuth 1 meter from the participant (CI users) with the contralateral ear occluded with an ear plug. Participants were instructed to verbally repeat what they heard and did not receive feedback based on their responses. Gap detection tasks. Two separate gap detection tasks (within-frequency and across-frequency gap detection) were administered with a random order to assess the GDT within and GDT across . Stimuli were presented using APEX software [53] on a laptop computer (Altec Lansing with SRS premium integrated soundcard) using an adaptive, two-alternative forcedchoice procedure with an one-up one-down stepping rule. The gap durations ranged from 2 to 120 ms (2-20 ms in 2 ms increments, 20-120 in 5 ms increments). For each trial, the participant was presented with two sounds, one containing a silent gap (target) and the other no gap (reference). The no gap stimuli was a continuous 1 or 2 kHz tone that was constructed similarly to the target stimuli (e.g., cos 2 window and also included a 10 ms rise-fall time). The presentation order of the target and reference stimuli was randomized and the interval between presentations was 0.5 seconds. All stimuli were presented at the MCL through a loudspeaker placed at 0 degrees azimuth 1 meter from the participant with the contralateral ear occluded with an ear plug. For each trial, the participant was instructed to select the sound that contained a silent gap. All participants completed a practice trial first to ensure comprehension of the task. They were instructed to just focus on the gap and ignore the frequency change that might exist (in across-frequency gap detection). Visual feedback was provided, and testing continued until five reversals were completed and the mean of the last three reversals was recorded as the GDT. EEG recordings. EEG recordings were conducted using a Neuroscan TM recording system (SCAN software version 4.3, Compumedics Neuroscan, Inc., Charlotte, NC) with a NuAmps digital amplifier. A 40-channel Neuroscan Quik-Cap (Compumedics Neuroscan, Inc., Charlotte, NC) was placed over the participant's scalp according to the 10-20 International system. The reference electrode was placed on the earlobe contralateral to the test ear, which has been found to reduce stimulus artifact in CI users [54,55]. Continuous EEG activity was recorded with a sampling rate of 1000 Hz. Electrooculography was recorded to document eye movement activities, which were removed later during offline analysis. For CI users, approximately 1 to 3 electrodes surrounding the transmission coil were not used during recording, but the activities at these electrodes were interpolated during the offline analysis. A total of eight EEG recordings were collected from each participant under four withinand four across-frequency conditions with gap durations that were customized based on the participants' GDT within and GDT across . The four gap duration conditions included the following: (1) sub-threshold (GDT/3), (2) threshold (GDT), (3) supra-threshold (GDTx3), and (4) reference (no gap). Calculated gap durations were rounded to the nearest whole integer. The order of stimulus conditions was randomized across participants. Continuous EEG recordings were collected with a minimum of 200 and 400 stimulus trials collected from NH and CI users, respectively. The inter-stimulus interval was fixed at 0.9 seconds and triggers were time-locked to the onset of the pre-gap marker. For all participants, the EEG stimuli were presented at the MCL through a loudspeaker placed 1 meter from the test ear. During EEG testing, participants were seated in a comfortable chair, instructed to ignore the stimuli, relax, and read books of their choice. Data analysis Speech perception and gap detection data. The CNC word (CNC-Word and CNC-Phoneme) and AzBio sentence (AzBio-Quiet, AzBio-Noise) tests were evaluated in terms of percent correct. For the BKB-SIN, the SNR-50 was calculated to determine the SNR necessary to understand 50% of the key words contained in the sentence. The GDT within and GDT across were calculated as the mean gap duration (ms) of the last three reversals in the gap detection task. EEG data. Initial EEG analysis was completed within Neuroscan software version 4.3. The EEG data was digitally band-pass filtered (0.1 to 30 Hz with a 6 dB/octave roll-off), segmented from 100 ms prior to the stimulus onset (-100 ms) to 200 ms beyond the offset of the post-gap marker, and baseline corrected using the pre-stimulus window (-100 to 0 ms). Subsequent analysis was completed in EEGLAB 13.6.5b (http://sccn.ucsd.edu/eeglab) running under Matlab R2017b (The Mathworks, Natick, MA), as in previous studies from our lab and other researchers [56][57][58][59]. Briefly, the data

were visually checked to remove approximately 10% of the segments that contained non-stereotyped artifacts. Independent Component Analysis (ICA) was then used to decompose the EEG data into mutually independent components including those from neural and artifactual sources. Independent components that represented artifacts arising from ocular movement, electrode, or CI artifact were identified and removed after visual inspection of component properties including the waveform, 2-D voltage map, and the spectrum [60,61]. The remaining components were then constructed to form the final EEG dataset. The unused channels were deleted and then interpolated. Data were rereferenced to the average reference [60,62], baseline corrected using the pre-stimulus window (-100 to 0 ms), and filtered from 0.1 to 30 Hz using a band-pass Fast Fourier Transform linear filter. Because the CAEP was most easily identified along the central/midline electrodes (Fz, FC3, FCz, FC4, Cz), data from these five electrodes were averaged together to form one final averaged waveform [32,36]. For each tested ear, there were 8 averaged waveforms (4 gap conditions/stimulus type x 2 stimulus types). For each averaged waveform, the presence of pre-and post-gap CAEPs was determined by visual inspection of the waveform morphologies by two reviewers (Blankenship and Zhang). CAEP peak components (N1 and P2) were identified for both the pre-and post-gap markers in their specific latency ranges reported in previous studies [36,63]. Specifically, the N1 was the maximum negative peak between 75 and 150 ms, and P2 was the maximum positive peak occurring between 150 and 220 ms after the onset of the markers. For convenient comparison, the post-gap CAEP peak latencies were adjusted so that the post-gap marker onset corresponded to time zero. The dependent variables obtained from EEG data were the N1-P2 peak-to-peak amplitudes (a commonly used measure for the CAEP amplitude [35,38,40] and N1 and P2 peak latencies. Statistical analysis. Descriptive statistics were calculated for all outcome variables. Linear mixed effect models were used to examine group differences in behavioral and CAEP data. These models were selected because it is an appropriate method for analyzing non-independent data and allows the inclusion of both fixed and random effects. For behavioral measures, multiple linear mixed effect models were used to examine differences in speech perception and GDTs between the NH and CI users with participant group and test ear included as fixed effects. Age at test was included as a covariate and participant ID as a random effect to account for participants that were tested in both ears separately. For the CAEP data, multiple linear mixed effect models were conducted with participant group, gap duration condition, and test ear included as fixed effects. Age at test was included as a covariate and participant ID as a random effect. The Satterthwaite approximation method was used to estimate degrees of freedom and the Holm's step-down adjustment method was applied for pairwise comparisons. Bivariate scatter plots were used to evaluate the assumptions of linearity, multivariate normality, and homoscedasticity. Non-parametric spearman rank correlation coefficients were used to assess the strength of pairwise correlations between 1). GDT across and speech perception measures (CNC, AzBio, BKB-SIN) and 2.) within-and across-frequency CAEP (N1-P2 amplitude, N1 and P2 latency) for the supra-threshold condition and speech perception (CNC-Word, AzBio-Noise, SNR-50). Bonferroni corrections were applied to account for multiple comparisons. Data were analyzed using SPSS statistical software (IBM Corp. Released 2019. IBM SPSS Statistics for Windows, Version 26.0. Armonk, NY: IBM Corp.). Audiometric thresholds NH and CI group mean audiometric thresholds are displayed in Fig 2. NH individuals had thresholds � 25 dB from 0.25 to 8 kHz. CI users had thresholds that ranged from 15 to 45 dB HL across all test frequencies. Obtaining thresholds ensured that the presentation level for GDT assessment was at least 25-35 dB above threshold [6]. Speech perception and GDTs For non-adaptive speech perception tests (CNC and AzBio), all NH listeners' scores were � 97%; CI users showed variable performance (Range = 38-99%, with some CI users' showing much worse scores compared to NH listeners' scores while some were comparable to NH listeners. On the adaptive BKB-SIN test, the mean SNR-50 was -0.7 dB for NH listeners and much higher in CI users (+8.2 dB). On the gap detection tasks, all NH listeners had a GDT within of 2 ms, which was the smallest gap duration on the task. All CI users also had a GDT within of 2 ms, except for 3 CI ears (Sci36-Left = 51.7 ms, Sci36-Right = 26.7 ms, Sci43-Left = 41.7 ms). The GDT across had a similar range between NH (14.6-120.0 ms) and CI users (25.0-120.0 ms) but NH listeners showed a lower mean GDT across . Group mean speech perception scores and GDTs are displayed in S1 Table. Linear mixed effect model results (S2 Table) showed CI users performed significantly poorer (p<0.001) on all individual speech perception measures compared to NH individuals but lacked a group difference in GDT within or GDT across (p > 0.05). Age at test was a significant factor that negatively affected GDT across (p < 0.001). EEG results Group mean CAEP waveforms for within-frequency pre-and post-gap markers in NH listeners and CI users are displayed in Fig 3. The CAEP mean and standard deviation of the peak amplitude and latency are displayed in Table 2. It can be seen that CI users' CAEPs for both pre-and post-markers displayed smaller amplitudes and longer peak latencies compared to NH listeners. The post-gap marker CAEP has a smaller amplitude than the pre-gap marker CAEP for both groups. For NH listeners, there is a trend that the CAEP for the post-gap marker has a progressively decreased amplitude for shorter gap durations (Fig 3 top right panel). Such trend was not seen in CI users (Fig 3 bottom right panel). Multiple mixed effect models were used to evaluate the effect of group, gap duration condition, test ear, and age at test on within-frequency post-gap CAEP amplitude and latency, which is the CAEP of interest in this study (S3 Table). For the N1-P2 amplitude, no group effect was observed. Instead, a significant effect of age at test indicated that older individuals (NH and CI) displayed smaller N1-P2 amplitudes (Mean = 1.2 μV) than younger adults (Mean = 0.5 μV, p = 0.041). There was no main effect of gap duration condition or test ear. For N1 and P2 latency, no effect of group, gap duration condition, test ear, or age at test was observed (p > 0.05). Group mean across-frequency pre-and post-gap CAEP waveforms are displayed in Fig 4. The CAEP mean and standard deviation of N1-P2 amplitudes and peak latencies are displayed in Table 3. It can be seen that CI users' CAEPs for both pre-and post-markers displayed longer peak latencies compared to NH listeners, but the N1-P2 amplitude was similar for both groups, which did not differ for different gap duration conditions (Fig 4 top and bottom right panels). Moreover, the post-gap marker CAEP has a slightly larger amplitude than the pre-gap marker CAEP for both groups. Multiple linear mixed effect models were used to evaluate the effect of group, gap duration condition, test ear, and age at test on the across-frequency post-gap CAEP amplitude and latency values (S3 Table). No significant effects of group, condition, test ear, or age at test were observed (p > 0.05) for N1-P2 amplitude. CI users displayed increased N1 latencies (Mean = 117.1 ms) compared to NH listeners (Mean = 107.8 ms, p = 0.017), and right ear latencies were increased (Mean = 116.7 ms) compared to the left ear (Mean = 108.2 ms, p = 0.012). Age at test was also significant for N1 latency (p = 0.010) and P2 latency (p = 0.003) with younger adults (NH and CI) displaying shorter latencies. GDTs vs. speech perception The correlation between the GDT within and speech perception was not examined because all CI ears except 3 demonstrated a GDT within of 2 ms, the lowest gap duration in the within-frequency gap detection task. However, it was noted that these 3 CI ears with a much longer GDT within (Sci36-Left = 51.7 ms, Sci36-Right = 26.7 ms, Sci43-Left = 41.7 ms) have poorer speech performance, especially in noise (Median: AzBio-Noise = 51%, SNR-50 = +11 dB) compared to other CI ears (Median: AzBio-Noise = 76%, SNR-50 = +7 dB). Non-parametric spearman rank correlation coefficients were used to assess the strength of pairwise correlations between GDT across and speech perception measures (CNC, AzBio, BKB-SIN) for the NH and CI group collectively. Bonferroni corrections were applied to account for multiple comparisons, and therefore a p-value of 0.01 was considered statistically significant. None of the correlations reached significance following a Bonferroni correction (p > 0.01). CAEPs vs. speech perception Non-parametric spearman rank correlation coefficients were used to assess the strength of pairwise correlations between within-and across-frequency supra-threshold post-gap CAEP (N1-P2 amplitude, N1 and P2 latency) and speech perception (CNC-Word, AzBio-Noise, SNR-50) for the NH and CI group collectively. Bonferroni corrections were applied to account for multiple comparisons, and therefore a p-value of 0.004 was considered statistically significant. Within-frequency correlations revealed a significant positive correlation between CNC-Word score and N1-P2 amplitude, a significant negative correlation between AzBio-Noise and N1 latency, and a significant positive relationship between the BKB-SIN SNR-50 and N1 latency. For across-frequency CAEPs, none of the correlations reached significance following a Bonferroni correction (p>0.004). The correlations between CAEP measures for the within-frequency supra-threshold gaps and speech perception are displayed in scatterplots in Fig 5. Discussion The purpose of the present study was to examine the behavioral GDT across and GDT within and CAEPs evoked by within-and across-frequency gap stimuli; furthermore, the correlation between these measures and speech perception were explored. CI users had significantly PLOS ONE Temporal processing and speech perception in cochlear implant users poorer speech perception scores, but their GDT across and GDT within were not significantly different from NH listeners. Age at test negatively affected GDT across and within-and across-frequency CAEPs. Additionally, across-frequency CAEP results displayed a significant group effect for N1 latency and an ear effect for N1 latency. Speech perception NH listeners performed near ceiling level in speech tasks while CI users performed significantly poorer on all speech perception tasks. On the CNC and AzBio, NH listeners displayed mean scores that were > 99%. CI users that displayed mean scores that ranged from 66 to 89% correct, which is on average 11 to 34% poorer compared to NH participants. On the BKB-SIN, CI users required an 8.9 dB SNR increase compared to NH participants to understand 50% of the target words in the sentence. NH and CI group mean speech perception scores in this study are consistent with those reported in the literature [18,50,64]. In a previous study by the authors, a mean CNC-Word score of 64.9%, CNC-Phoneme score of 79.1%, and AzBio-Quiet of 75.5% was reported in a group of post-lingually deafened CI users [18]. Gifford et al. [65] reported mean speech recognition scores for post-lingually deafened adult CI users for the unilateral listening condition were: 55.7% for CNC-Word, 72.1% for AzBio-Quiet, and +11.4 dB for BKB-SIN SNR-50. Donaldson et al. [48] reported a median BKB-SIN SNR-50 value of +11.9 dB in a group of adult CI users. Slight differences observed in speech perception scores across studies could be due to several factors including differences in stimulus presentation levels, participant age, and the number of unilateral and bilateral CI recipients in the study. PLOS ONE GDTs GDT within . NH listeners performed at floor level on the within-frequency gap detection task with all participants displaying a GDT within of 2 ms, the shortest gap included in the task. It is possible that the true GDT within is less than 2 ms in some participants. Previous studies have shown that the GDT within in NH listeners is affected by stimulus factors such as stimulus frequency, spectral complexity, and marker durations [18,37,46]. For instance, using 1 and 2 kHz pure-tone with the marker duration of 10-20 ms, Heinrich and Schneider [19] reported a GDT within that ranged from 0.9 to 1.6 ms for younger and 1.2 to 2.5 ms for older participants. Using a marker duration of approximately 20 ms, Blankenship et al. [18] reported that 2 kHz pure-tone GDT within ranged from 2 to 15 ms (Mean = 5.8 ms) for young NH listeners. In contrast, using

narrow-band noise of 400 ms in duration at a center frequency at 2 kHz and 1 kHz, Lister et al. [37] reported that GDT within ranged from 7 to 15 ms (Mean = 9.8 ms) in young NH adults. Using a 2 kHz narrowband noise of approximately 300 ms as the markers, Alhaidary and Tanniru [46] reported that the same group of NH listeners displayed a mean GDT within of 3.33 and 8.33 ms using different testing programs that incorporated stimuli with only subtle spectral differences. Previous studies have shown that GDTs increase with stimulus complexity and shorter marker durations which could explain the slightly higher GDTs reported by previous studies [18,37,46]. CI users displayed a GDT within of 2 ms in 12 CI ears and a much longer GDT within (Range = 27-52 ms) in 3 CI ears. The GDT within from CI users did not differ statistically from that in NH peers. In our previous study using a shorter marker duration (~20 ms), 2 kHz pure-tone GDT within ranged from 5 to 70 ms in CI users (Mean = 23.2 ms; [18]). Using noise centered at 500 Hz as the markers, Tyler et al. [20] reported CI users displayed a GDT within that ranged from 7.5 to 300 ms. Using narrowband noise and various stimulus durations, Muchnik et al. [21] reported GDT within in CI users that ranged from 12-72 ms, with smaller GDTs as the stimulus duration increased. Using electrical stimulation, Shannon [66] reported a GDT within in CI users that is comparable to the reported GDT within in NH listeners when the stimulus intensity was comparable. The authors suggested that CI users as a group may not necessarily have poorer temporal resolution. Mussoi and Brown [42] reported a median GDT within in a group of CI users that was less than 3 ms. The slightly higher mean GDTs reported in the literature could be due to several factors including the complexity of the signal, overall stimulus duration, and inclusion of pre-lingually deafened CI recipients. Overall, our NH and CI GDT within results are consistent with the literature and further supports the variability seen in CI recipients. GDT across. The difference between NH and CI users in GDT across scores (NH = 58.8 ms, CI = 82.4 ms), regardless of the wide variability was not found to be significant. GDT across in NH participants have been reported in several previous studies [8, 13, 22-25, 37, 39]. For example, using 2 kHz to 1 kHz pre-to post-gap narrowband noise markers, Lister et al. [39] reported that the GDT across has a mean value of 29 ms in young adults and 56 ms in older adults. These studies have documented wide variability in GDT across and that the GDT across can be ten times poorer than the GDT within [13,46]. The difference in GDT across obtained in the current study and compared to the literature could be due to differences across studies regarding participant age, instrumentation, stimulus, and presentation levels. There are only a few GDT across studies in the literature involving CI users, which were conducted with direct electrical stimulation. Hanekom and Shannon [14] measured GDT across using 200 μs/phase biphasic pulses with a stimulation rate of 1000 pps in three adult CI users as a function of electrode separation between the pre-and post-gap markers. The GDT across systematically increased as the channel separation increased with thresholds varying significantly across participants (Range = 10-200 ms). Upon examining the GDT across using electrode pairs corresponding to the frequencies used in the current study, GDT across was approximately 7, 25, and 50 ms in these 3 participants (Mean = 48.7 ms). Using a similar approach, van Wieringen and Wouters [28] examined the GDT across in four post-lingually deafened CI users. While one participant displayed a low GDT across (<10 ms), the other participants had higher GDT across (10 to 50 ms), with a longer GDT across for a larger electrode separation. In the current study, the mean GDT across in the CI group was 82.4 ms (Range = 25-120 ms), which is higher than most individual GDT across reported in previous studies using electrical stimulation [14,28]. However, the current study used acoustic which are expected to be poorer than those measured with direct electrical stimulation. Age effect on the GDT across has been observed in the current study, with a much higher GDT across in older (> 50 years of age) than younger participants for both CI group (Younger = 65 ms, Older = 98 ms) and NH group (Younger = 31 ms, Older = 83 ms). This negative effect of participant's age is consistent with studies by previous studies, with more obvious effect for the GDT across [12,19,37,39,67]. In summary, the GDT within and GDT across reported in the literature and in this study showed some differences, which may be attributed to different factors in the stimuli used and participants' age. The stimulus factors include stimulus level, marker bandwidth, marker duration and frequency, difference in the pre-and post-gap markers in spectral composition and intensity, and stimulus presentation mode [11,12,15,16,31,37,68]. Both GDT within and GDT across show a large variability, with some CI users displaying GDTs comparable to those in NH listeners while others had much higher GDTs [4,28]. CAEPs CAEPs to within-frequency gap stimuli. NH listeners displayed a trend of increased N1-P2 amplitude for longer gap durations (sub-threshold = 1.2 μV, threshold = 1.5 μV, suprathreshold = 1.8 μV, see Fig 3 and Table 2). This finding is similar to that reported in previous studies [37,39]. Compared to NH listeners, CI users' CAEPs showed the following differences: 1) both pre-gap and post-gap markers displayed poorer morphology and smaller N1-P2 amplitude relative to NH listeners. This might be due to a reduced number of neurons that are able to respond synchronously to the markers as a result of deafness effects in CI users. 2) CI users did not display a trend observed in NH listeners that N1-P2 amplitude increases with the increase in the gap duration. This may be explained with the neural recovery mechanism involved in gap detection [37][38][39]. Specifically, for within-frequency gap stimuli, the neurons initially respond to the pre-gap marker and then are stimulated again by the post-gap marker. In NH listeners, the neural activities are more likely to recover from the pre-gap stimulation if the gap duration is longer. However, CI users have compromised neural firing and synchronization and their neurons may have a large variability in the recovery time. Therefore, their post-gap CAEP is typically poor and does not show the progressive change with the change of the gap duration. CAEPs to across-frequency gap stimuli. Compared to the CAEP to the within-frequency gap stimuli, the CAEP evoked by the across-frequency gap stimuli show the following differences: 1) all post-gap CAEPs have similar amplitude for all gap durations in both NH and CI groups (see Fig 4 and Table 3). 2) the post-gap CAEP appears to be slightly larger than the pregap CAEP in both NH and CI groups. 3) the NH vs. CI difference in the post-gap CAEP is less dramatic in its amplitude for the across-frequency gap stimuli than for the within-frequency gap stimuli (see Fig 3 vs. Fig 4). This is because the across-frequency gap stimuli elicit responses of different groups of neurons as the pre-and post-gap frequencies are different and thus the post-gap CAEP is similarly synchronized compared to the pre-gap CAEP. A limited number of studies have examined the CAEP to across-frequency gap stimuli, all of which were conducted with individuals with normal or minimal hearing loss [37,39]. Lister and colleagues [37,39] examined the CAEP to across-frequency gap using narrowband noise stimuli (2 kHz pre-gap and 1 kHz post-gap marker) in a group of young and older adults. The authors included different gap durations conditions including threshold, sub-threshold, and supra-threshold conditions, and a reference condition (a standard 1 ms gap). Results showed the CAEP amplitude was similar for all gap durations. Similar to the findings from Lister and colleagues, the current study did not find significant effects of gap duration condition on CAEP amplitude to the across-frequency gap stimuli. Age effect on the CAEP has been observed in this study. Individuals younger than 50 years of age displayed shorter N1 and P2 latencies (N1 = 95.3 ms, P2 = 151.1 ms) compared to older individuals (N1 = 99.2 ms, P2 = 171.4 ms). Similarly, Lister and colleagues [37,39] reported an age effect on the P2 latency. Different mechanisms for within-and across-frequency gap detection Based on the CAEP and GDT findings in this study, we propose that different mechanisms may be involved in detecting within-and across-frequency gaps. For the within-frequency gap detection, the listeners rely on the comparison of sensory encoding of the pre-and post-gap markers of the same frequency. When the gap is long enough, the post-gap neural activity will be strong enough [68], the sensation for the post-gap marker will be recovered [15], and the participant can detect the gap. For the across-frequency gap detection, the post-marker CAEP is nearly similarly strong for all gap conditions because the post-marker CAEP is mainly evoked by the acoustic changes consisting of the frequency change (from 2 kHz to 1 kHz in this study) and the gap. Therefore, this post-gap CAEP is a type of the acoustic change complex (ACC) that increases the amplitude with the increase of salience of the sound change in gap and frequency [42,59,63,69]. Although the participants were instructed to just pay attention to the silent gap and ignore the frequency difference between the pre-and post-gap markers, the latter may have acted as a distractor so that the participants cannot fully engage their cognitive attention to perform the across-frequency gap detection and produced a much larger value for GDT across than GDT within . Therefore, the across-frequency gap detection involves more central mechanism than the within-frequency gap detection. The involvement of cognitive function and allocation of attentional resources in gap detection has been suggested in previous behavioral studies [13,28]. With the CAEP and behavior measures, this study provided more evidence that there may be different neural mechanisms underlying within-and acrossfrequency gap detection, with the later involving more cognitive mechanisms. Correlation analysis GDT within vs. speech perception. Previous studies have shown a correlation between the GDT within and speech perception in CI users, with a much shorter stimulus duration compared to that used in this study [18,21,70]. Our data using pure-tone markers of approximately 300 ms showed that 3 CI ears with the longest GDTs (Sci36-Left = 51.7 ms, Sci36-Right = 26.7 ms, Sci43-Left = 41.7 ms) had a much worse AzBio-Noise performance compared to other CI users. Examining demographic and device data in Table 1, Sci36 is the oldest participant in the study with the longest duration of auditory deprivation (51 yrs.). Sci43 is also an older participant (60 yrs.) with 23 years of auditory deprivation. With regard to speech perception, Sci43 displayed the poorest performance in the CNC-Word (43%) and Phoneme (62%) and Sci36 displayed relatively high performance on the CNC-Word (Left = 80%, Right = 74%) and Phoneme (Left = 93%, Right = 86%). On the AzBio sentences in quiet, their scores ranged from 82% to 89%. However, on the AzBio sentences in noise and the BKB-SIN, these two participants were among the poorest performers with AzBio-Noise scores ranging from 40 to 56% and SNR-50 scores of +10.8 to +12.8 dB. Our results indicate that adequate speech understanding in quiet may be achieved with GDT within greater than 25 ms, but smaller GDT within (i.e., better temporal processing) is needed to perform speech-in-noise tasks. The lack of variability in the GDT within and behavioral speech perception partially undermined correlation analysis between these two measures. A shorter marker duration in gap detection is more likely to separate individuals with normal and abnormal temporal resolution [68]. Using tone stimuli of durations of approximately 20 ms in CI users, Blankenship et al. [18] reported GDTwithin ranged from 2 to 100 ms and the GDTwithin were significantly correlated with speech perception performance in quiet and background noise. Tyler et al. [20] reported that CI recipients with GDTs > 40 ms displayed poorer speech and environmental noise identification abilities compared to CI recipients with GDTs < 40 ms. Similarly, Muchnik

et al. [21] reported mean narrowband noise GDTs of 12. GDT across vs. speech perception. Temporal cues in natural speech such as the voice onset time and silent gaps occur between sounds of various spectral compositions. Therefore, it is reasonable to assume that the GDT across is more related to behavioral speech perception than the GDT within . Previous studies examining the correlation between the GDT across and speech perception performance have reported mixed results. Elangovan and Stuart [71] showed a significant correlation (ranging from approximately 0.55 to 0.75) between voice onset time phonetic boundaries and the GDT across rather than the GDT within . In contrast, Mori et al. [72] did not find a significant correlation between voice onset time boundaries or slope (/ba/ to /da/) and the GDT across in a group of Japanese listeners. In the current study, our results did not reveal a correlation between clinical measures of speech perception and the GDT across (p >0.01). Our results suggest that the GDT across is not a good indicator of clinical measures of speech perception. CAEP vs. speech perception. This study found a significant correlation between withinfrequency CAEP measures and speech perception in quiet (CNC words) and noise (AzBio and BKB-SIN sentences). Individuals with poorer speech performance had a smaller N1-P2 amplitude and longer N1 latency. No correlations were found between across-frequency CAEP measures and speech perception performance. Overall correlation analyses indicate that the CAEP evoked by the within-frequency gap stimuli is a promising objective tool that can be used to predict CI speech outcomes. Only a few other studies have explored how the CAEP to gaps may relate to speech perception using bivariate correlations or descriptive summaries [30,32,41]. Michalewski et al. [32] reported that individuals with auditory neuropathy (9-60 yrs) had elevated behavioral GDTs and poorer speech understanding (sentences in quiet), compared with normal hearing young adults (18-30 yrs). Additionally, CAEPs were presented for conditions with longer gap durations. The GDT measured with behavioral and electrophysiological methods were similar in both auditory neuropathy group and normal hearing group. The researchers did not examine the correlation between CAEP measures and speech perception performance. In pediatric CI recipients with auditory neuropathy, He et al. [41] reported a significant negative correlation between electrophysiological GDTs (800 ms biphasic electric pulses with silent gaps) and PBK word scores. Furthermore, in CI candidates with auditory neuropathy, He et al. [30] reported a significant negative correlation between electrophysiological GDTs and word scores. However, it is important to note that the CAEP values used in the correlation analyses were electrophysiological GDT (ms) and not P1-N1-P1 peak amplitude and latency values. Most studies examining the correlation between CAEP and speech perception have focused on the CAEP evoked by stimulus onset (onset-CAEP) rather than the CAEP evoked by the gap (a change in a longer stimulus). These studies have shown a relationship between speech perception performance and P1-N1-P2 response amplitude and latency. Specifically individuals with individuals with poorer speech identification scores showed a significantly smaller CAEP amplitude [73,74] increased N1 latency [75,76] and poorer CAEP morphology [74,77] compared to CAEP responses from individuals with good speech perception performance. Differences in the CAEP responses between individuals with good vs. poor speech perception performance have been proposed to be due to a combination of factors surrounding neural integrity (density, synchronization, adaptation, and refractory periods) [36,[78][79][80]. For example, smaller CAEP amplitude and longer latencies may be due to a reduced number of neurons that are able to respond to acoustic stimuli due to auditory deprivation. This explanation can also be used to explain the current finding that individuals with poorer speech perception performance had poorer within-frequency gap-evoked CAEPs. Additionally, the cortical neurons have an increased refractory period due to long-term deafness and are not able to fire as quickly following the pre-marker stimuli. Implications and limitations Our results indicate within-frequency gap detection, which is related to speech perception performance, is a better approach than the across-frequency gap detection to examine temporal resolution. CAEP to within-frequency gap stimuli could be used an objective measure of temporal resolution. However, there are several limitations of this study. The first limitation is that there was a floor effect observed on the GDT within in both NH listeners and CI users, the smallest possible value on the task. Future behavioral gap detection studies should shorter marker durations, thereby increasing the difficulty level of the task and avoiding the floor effects of the GDT within . Similarly, there was a ceiling effect observed for NH listeners for several of the speech perception tasks. Future studies should be designed so that the task difficulty between the NH and CI listeners is equivalent. Additionally, the rise/fall time around the silent gap was 1 ms, as in previous studies [37], which may cause acoustic transients and energy spread [81] that affect the GDT and CAEP results. However, such a spectral spread should be the same for both within-and across-frequency gap stimuli used in this study. Therefore, the different trends of the post-gap CAEPs for the within-and across-frequency stimuli still enabled us to infer different neural mechanisms involved in the within-and across-frequency gap detection. Future studies will use approaches suggested in previous studies, e.g., by using notched noise to mask the transient cue, for gap detection tasks [7,10]. Conclusions CI users had significantly poorer speech performance compared to age-matched NH listeners but corresponding group difference in behavioral GDTs was not observed. The 3 CI ears (two CI users) with the worst GDT within performed poorer in AzBio-Noise than other CI users whose GDT within was comparable to that in NH listeners. The CAEP to within-frequency gap stimuli in NH listeners showed the anticipated trend of increased N1-P2 amplitude with the increase of the gap duration. This trend was weak in CI users, likely due to compromised neural firing and increased neural recovery time. The CAEP to across-frequency gap stimuli displayed a similar amplitude for all tested gap durations in both NH and CI groups. This is because the post-gap stimuli elicits a different group of neurons compared to the pre-gap stimuli. Older individuals displayed larger GDTs and had poorer CAEP responses (smaller amplitude and delayed latencies). Additionally, within-frequency CAEP responses were significantly correlated with open-set word and sentence recognition. Therefore, our findings support that GDT within rather than the GDT across is a better indicator for temporal resolution and the CAEP to within-frequency gap stimuli can serve as an objective tool to predict speech perception performance. Supporting information S1 Comparison of Different Culture Conditions for Mesenchymal Stem Cells from Human Umbilical Cord Wharton’s Jelly for Stem Cell Therapy Address for Correspondence/Yazışma Adresi: Ersin Töret, M.D., Altınbaş University Faculty of Medicine, Medicalpark Bahçelievler Hospital, Department of Pediatric Hematology-Oncology & Bone Marrow Transplantation Unit, İstanbul, Turkey Phone : +90 505 799 42 34 E-mail : [email protected] ORCID: orcid.org/0000-0002-6379-8326 Received/Geliş tarihi: August 6, 2019 Accepted/Kabul tarihi: November 12, 2019 To the Editor, Many recent studies have demonstrated that the umbilical cord is an excellent source of mesenchymal stem cells (MSCs) [1,2,3]. However, in order to use human umbilical cord Wharton's jellyderived mesenchymal stem cells (hUC-MSCs) in clinical therapy, a suitable culture procedure for good manufacturing practicecompliant production is mandatory. Nutritional deficiency is the major pathophysiological situation in an ischemic microenvironment in the clinic [4]. Thus, the development of serum-free culture systems is needed [5]. Furthermore, hypoxia is common in vivo in mammals [6]. The average oxygen tension falls to 1% in some cases of pathological ischemia, including fracture hematoma, and in cases of myocardial ischemia [7]. Hence, the investigation of biological characteristics of hUC-MSCs exposed to hypoxic and/or serum-free conditions is of great interest. In our study, we conducted parallel assays by using four cell groups. For the hypoxic controls, cells from group A (n=10) and group B (n=10) were exposed to 5% CO 2 and 94% N 2 in an airtight modular incubator chamber (Billups-Rothenberg Inc., Del Mar, CA, USA). The final oxygen tension was 1%-3% as measured by an oximeter (Oxybaby M+, Witt Technology, Solza, Italy). For the normoxic controls, cells from group C (n=10) and group D (n=10) were placed in an incubator at 37 °C, 5% CO 2 , and 21% O 2 . Cells from group A and group C were expanded in a mixture of Dulbecco's modified Eagle's medium and nutrient mixture F-12 (GIBCO, USA) supplemented with 10% fetal bovine serum (GIBCO, USA). Cells from group B and group D were expanded in StemPRO MSC serum-free medium (StemRD, USA). Flow cytometric analysis, differentiation potential, proliferative activities, cell cycle analysis, and apoptosis analysis of these four cell populations were evaluated. We repeated all these experiments 3 times. Flow cytometry analysis of MSC-specific surface marker expression showed that hUC-MSCs cultured under four experimental conditions for six passages were positive for CD44, CD73, CD90, CD105, CD29, and HLA-ABC (BD Pharmingen, USA) and negative for CD34, CD45, CD14, and HLA-DR (BD Pharmingen, USA); no significant differences were detected between the four cell populations ( Figure 1). This finding indicates that culturing cells under hypoxic and/or serum-free conditions did not induce significant variations in the typical MSC marker expression profile. hUC- 5 MSCs maintained their multilineage differentiation potential in vitro after expansion under various conditions [8]. MSCs from all groups demonstrated a similar osteogenic phenotype, as evidenced by positive staining for alizarin red S (Sigma-Aldrich, USA) and deposits of calcified matrix (Figure 2). In the case of adipogenic differentiation, cells from all groups formed lipid vacuoles detected by oil red O (Sigma-Aldrich, USA). There were no significant quantitative changes among the groups (Figure 3). Thus, the results presented in this report indicate that hypoxic and/or serum-free conditions do not affect the biological characteristics of hUC-MSCs. Under hypoxic and serum-free conditions, hUC-MSCs have higher proliferation according to their growth curves (BD Pharmingen, USA) and MTT assays (BD Pharmingen, USA) than cells grown under normoxic and serum-containing culture conditions, but without more apoptosis (Figures 4 and 5). Taken together, our data indicate that hypoxic and serum-free culture conditions do not influence the major properties of hUC-MSCs. Under hypoxic and serum-free conditions, hUC-MSCs showed higher proliferation, while their apoptosis rate did not increase. This finding is consistent with that of previous reports, which demonstrated enhanced proliferation of bone marrow-derived MSCs (BM-MSCs) under hypoxic or serumfree conditions [9,10]. Therefore, the availability of optimized in vitro conditions, including hypoxia and serum-free media, for hUC-MSC manipulations may have a substantial scientific and clinical impact. Conflict of Interest: The authors of this paper have no conflicts of interest, including specific financial interests, relationships, and/or affiliations relevant to the subject matter or materials included. Financial Disclosure: The authors declared that this study received no financial support. Roles of macrophage migration inhibitory factor in polymyositis: Inflammation and regeneration Objective To elucidate the clinical significance of macrophage migration inhibitory factor (MIF) serum concentration in patients with polymyositis. Methods Thirty-six patients with polymyositis were enrolled. Serum samples were obtained and stored to detect MIF and interleukin (IL)-6 using commercially available enzyme-linked immunosorbent assay kits. The relationships between these cytokines and clinical data were analyzed. Results The serum MIF concentration was significantly lower in patients in remission (34.74 ± 17.75) and in healthy controls (38.87 ± 9.30 ng/ml) than that in patients with active polymyositis (50.04 ± 23.84 ng/ml). There were no significant differences between healthy controls and patients in remission. The serum IL-6 concentration in patients with active polymyositis (19.67 ± 7.16 pg/ml) was significantly higher than that in patients in remission (15.81 ± 4.00 pg/ml) and controls (8.14 ± 3.71 pg/ml). The serum IL-6 concentration was negatively correlated with the serum MIF concentration (r = −0.283). No relationship was found between the serum MIF concentration and glucocorticoid dose. The MIF concentration peaked twice during treatment when the creatine kinase concentration was decreasing. Conclusion MIF and IL-6 play important roles in the inflammation associated with polymyositis. MIF might also be involved in the early stage of regeneration in polymyositis. MIF may thus serve as a biomarker of disease activity and outcome. Introduction Polymyositis (PM) is an autoimmune inflammatory myopathy. It is mainly characterized by proximal and symmetrical muscle weakness and often has internal organ involvement. The main pathophysiological mechanism of PM is infiltration of CD8þ T cells and macrophages into muscle fibers, leading to lysis of the muscle fibers. 1

In clinical practice, we still depend on the physician's clinical assessment, mainly including manual muscle strength testing and measurement of serum muscle enzyme levels, to evaluate the disease activity and prognosis. The serum levels of many inflammatory factors are reportedly increased in patients with inflammatory myopathies, 2,3 and IL-6 may be a biomarker of disease activity or outcome. Macrophage migration inhibitory factor (MIF) is a potent and pleiotropic cytokine that is secreted by activated T cells and macrophages and plays a critical role in inflammatory and autoimmune diseases. It has anti-apoptotic, proproliferative and pro-inflammatory effects and may act as a modulator in cytokine responses. 4 MIF plays important roles in many inflammatory diseases. Interestingly, MIF can induce the cytokines IL-6 and tumor necrosis factor-a (TNF-a), which are known as the most useful biomarkers in patients with PM. 3 MIF inhibitor can significantly decrease IL-6 and TNF-a production. 5 Moreover, MIF is a modulator of glucocorticoid sensitivity and may thus improve the status of high-dose glucocorticoid therapy in patients with PM. The concentrations of MIF in skeletal muscle are reportedly higher in patients with inflammatory myopathies. 6 However, the significance of the serum MIF level in patients with PM is still unknown. In this study, we explored clinical significance of the serum MIF level in patients with PM. Patients In total, 36 inpatients with PM were enrolled in Zhejiang Provincial People's Hospital from August 2010 to December 2014. Patients with PM satisfied the criteria proposed in 1975 by Bohan and Peter 7,8 as well as the consensus guidelines for PM in China. We excluded patients with sporadic inclusion body myositis, muscular dystrophy, and other myopathies. All patients underwent muscle magnetic resonance imaging. Patients with overlapping syndromes were excluded. According to the established disease activity tools described by the International Myositis Assessment and Clinical Studies Group, we used the measures of global activity, muscle strength, physical function, and laboratory assessment. Twenty-five patients had an active disease status and 12 patients were in remission when serum samples were obtained (two serum samples from one patient were obtained: one during active disease and one during remission). Detailed clinical and laboratory data were collected. Patients with active disease were initially administered 1.0 to 1.5 mg/kg/d of glucocorticoids, and the glucocorticoids were gradually tapered according to the clinical assessment. Two patients were followed during treatment, and their serum samples were obtained when related laboratory data were checked every week. Serum samples were voluntarily obtained from 10 healthy age-and sex-matched controls. All serum samples were stored at À80 C prior to analysis. The study was approved by the local ethics committee, and informed consent was obtained from patients and controls. Methods Enzyme-linked immunosorbent assay (ELISA) for MIF. The serum MIF concentration was detected by a commercially available ELISA kit (Quantikine ELISA for human MIF; R&D Systems, Minneapolis, MN, USA). According to the protocol, all of the reagents, working standards, and serum samples were prepared before detection. A total of 100 ml of Assay Diluent RD1-53 was added to each well of a 96well microplate, and 50 ml of standard, control, or serum samples was then added to each well. The microplate was incubated for 2 hours at room temperature on a horizontal orbital microplate shaker. The microplate was then washed four times with wash buffer. MIF conjugate (200 ml) was added to each well of the microplate, which was incubated for another 2 hours. Another four washes were repeated. Next, 200 ml of substrate solution was added to each well, and the microplate was incubated for 30 minutes at room temperature. Finally, 50 ml of stop solution was added to each well, and the optical density was determined within 30 minutes using a microplate reader set to 450 nm. ELISA for IL-6. The serum IL-6 concentration was detected by a commercially available ELISA kit (Quantikine ELISA for human IL-6; R&D Systems). According to the protocol, all of the reagents, working standards, and serum samples were prepared before detection. A total of 100 ml of Assay Diluent RD1W was added to each well of a 96-well microplate, and 100 ml of standard, control, or serum samples was then added to each well. The microplate was incubated for 2 hours at room temperature on a horizontal orbital microplate shaker. After washing four times with wash buffer, 200 ml of IL-6 conjugate was added to each well of the microplate, and incubation was performed for another 2 hours. The microplate was washed another four times. A total of 200 ml of substrate solution was added to each well, and the microplate was incubated for 20 minutes at room temperature. Finally, 50 ml of stop solution was added to each well, and the optical density was determined within 30 minutes using a microplate reader set to 450 nm. Statistical analysis The serum MIF concentration is expressed as the mean AE standard deviation. Statistical differences between two groups were evaluated by an independent-samples t test. Spearman's correlation test was used to evaluate the relationship of the MIF concentration and laboratory data or glucocorticoid dose. A p value of <0.05 was considered statistically significant. All data were analyzed with the statistical software package SPSS 17.0 (SPSS Inc., Chicago, IL, USA). Results In total, 36 patients with PM were enrolled, comprising 29 women and 7 men with a mean (AEstandard deviation) age of 56.13 (AE11.10) years. The mean serum MIF concentration in 25 patents with active PM was 50.04 AE 23.84 ng/ml. The mean concentration in the 12 patients in remission was 34.74 AE 17.75 ng/ml, which was significantly lower than that in the patients with active disease (p ¼ 0.037). The mean serum MIF concentration in the healthy controls was 38.87 AE 9.30 ng/ml, which was significantly lower than that in patients with active PM (p ¼ 0.045). However, there was no significant difference between the healthy controls and patients in remission (Figure 1). The mean serum IL-6 concentration was 19.67 AE 7.16 pg/ml in the patients with active PM and 15.81 AE 4.00 pg/ml in patients in remission. These concentrations were significantly higher than those in controls (8.14 AE 3.71 pg/ml, p < 0.001). Furthermore, the mean IL-6 concentration in patients with active PM was significantly higher than that in patients in remission (p ¼ 0.043). The serum MIF concentration was negatively correlated with the serum IL-6 concentration (r ¼ À0.283, p ¼ 0.046) (Figure 2). When the serum samples were obtained, no significant difference in the serum MIF concentration was found between patients with active PM who were and were not receiving glucocorticoid therapy. Additionally, no relationship was found between the serum MIF concentration and glucocorticoid dose. Other parameters, such as the erythrocyte sedimentation rate and serum levels of creatine kinase (CK), lactate dehydrogenase, aspartate transaminase, alanine transaminase, and C-reactive protein, showed no significant correlation with the serum MIF concentration. Two patients were followed up during treatment. Both patients were newly diagnosed and given 1 mg/kg/d of a glucocorticoid for treatment. Patient 1 had muscle weakness for 1 month when the treatment began, while Patient 2 had muscle weakness for 6 months. In Patient 1, serum samples began to be obtained after nearly 1 month of treatment. In Patient 2, the first serum sample was obtained before treatment. We observed two peaks in the serum MIF concentration during treatment when the CK levels were descending, and the levels decreased when the CK levels normalized (Figures 3 and 4). Discussion MIF acts as both a potent cytokine and a hormone-like molecule. Many researchers have found that MIF can induce TNF-a in monocytes, IL-6 and IL-12 in peritoneal macrophages, and IL-6 and IL-8 in synovial fibroblasts. 9-12 However, Kudrin et al. 13 found no induction of TNF-a, IL-6, or IL-12 release by synovial fibroblasts or macrophages using very highly purified MIF. MIF can be produced by many cell types, mainly by activated T cells and macrophages. 14 Furthermore, MIF is involved in the processes of autophagy and autophagic cell death. 15 The pathogenesis of PM involves cytotoxic CD8þ T cells attacking skeletal muscle fibers followed by invasion of non-necrotic fibers by these T cells and macrophages. 1 Additionally, macrophages are important cells in the regeneration of inflammatory disorders of skeletal muscle, while the random migration of macrophages can be inhibited by MIF. Few studies have focused on MIF in relation to skeletal muscle. In the present study, we found that the MIF concentration was higher in patients with active PM. However, there were no significant differences between patients in remission and healthy controls. Reimann et al. 6 also found that the MIF concentrations in protein lysates were higher than in controls. This may suggest that MIF plays a role in patients with active PM. The serum IL-6 level may be a sensitive biomarker of disease activity in dermatomyositis 16,17 and is reportedly an important proinflammatory cytokine in the inflammatory process of PM. Therefore, we evaluated the serum IL-6 concentration in the present study. Although IL-6 and MIF are both inflammatory factors, we unexpectedly observed that the serum IL-6 concentration was negatively correlated with the serum MIF concentration in patients with PM. A previous study showed that MIF was detected not only in inflammatory cells but also in muscle fiber membranes, suggesting that MIF is also involved in the response to muscle fiber damage. 6 Whether a balance in MIF exists between inflammation and regeneration in patients with PM requires further research. In an in vitro experiment, glucocorticoids affected MIF production in a bimodal way: low concentrations of glucocorticoids induced MIF production, and high concentrations of glucocorticoids inhibited MIF production. 9 Another in vivo study showed that MIF could be up-regulated by endogenous glucocorticoids in rats with adjuvantinduced arthritis. 18 In humans, the serum MIF concentration is influenced by exogenous glucocorticoids even after adjusting for disease activity variables. 19 We found no significant differences in the serum MIF concentration between patients with active PM who were and were not receiving glucocorticoid therapy as well as no relationship between the serum MIF concentration and glucocorticoid dose. Exogenous glucocorticoids are not the main factor impacting the MIF concentration in patients with active PM. Consequently, unlike IL-6 as a proinflammatory factor, MIF may play another role in PM. We also followed up two patients during their treatment. In these patients, glucocorticoids were given at 1 mg/kg/d, and the dose was not changed until the CK level had normalized to exclude the effects of the glucocorticoids. While the CK level was descending, we observed two peaks in the MIF concentration. MIF, which has anti-apoptotic, pro-proliferative, and macrophage-attracting functions, 4 may have another effect in the regeneration after muscle injury. MIF was detected at muscle fiber membranes, at the borders of infiltrates, or in necrotic fibers in the skeletal muscle of patients with PM. Focal sarcoplasmic reactivity was also observed, especially in fibers showing sarcolemmal MIF immunoreactivity. 6 The MIF concentration decreased when the CK level normalized. These data suggest that MIF plays a role not only in the inflammatory process of PM but also in the early stage of the regeneration response. Further studies involving more patients should be performed to confirm this result. In conclusion, MIF and IL-6 play important roles in the inflammation associated with PM. MIF acts as a potent and pleiotropic cytokine and may also be involved in the early stage of regeneration in PM. MIF may be a biomarker of disease activity and outcome. Declaration of conflicting interests The authors declare that there is no conflict of interest. Promising effects of 33 to 36 Fr. bougie calibration for laparoscopic sleeve gastrectomy: a systematic review and network meta-analysis The standard size of bougie for laparoscopic sleeve gastrectomy (LSG) is not yet established. Therefore, a systematic review and network meta-analysis were conducted to assess the weight loss effects and associated complications of LSG for patients with morbid obesity, based on different bougie sizes. A total of 15 studies were reviewed in this systemic review and network meta-analysis (2,848 participants), including RCTs and retrospective studies in PubMed, and Embase until September 1, 2020. The effectiveness of different bougie calibration sizes was assessed based on excess weight loss (EWL), total complications, and staple line leak. Within this network meta-analysis, S-sized (≤ 32 Fr.) and M-sized (33–36 Fr.) bougies had similar effects and were associated with

the highest EWL improvement among all different bougie sizes (S-sized: standardized mean difference [SMD], 10.52; 95% confidence interval [CI] − 5.59 to − 26.63, surface under the cumulative ranking curve [SUCRA], 0.78; and M-sized: SMD, 10.16; 95% CI − 3.04–23.37; SUCRA, 0.75). M-sized bougie was associated with the lowest incidence of total complications (M-sized: odds ratio, 0.43; 95% CI, 0.16–1.11; SUCRA, 0.92). Based on our network meta-analysis, using M-sized bougie (33–36 Fr.) is an optimal choice to balance the effectiveness and perioperative safety of LSG in the clinical practice. www.nature.com/scientificreports/ Not only changes in gastric motility and related hormonal secretion but also dominant mechanisms of gastric restriction, based on the essential intraoperative bougie calibration with longitudinal gastric transection of the fundus, body, and antrum along the lesser curvature, lead to limited eating volume of about 100 mL, subsequent dietary habit modification, and eventually weight reduction 9,10 . Although the surgical standardization and relative contraindications for LSG have been well documented, the standard bougie size used to calibrate the gastric sleeve still remains to be established, based on the weight loss effects and related complications, such as staple line leak (SLL), gastric stenosis, or de novo reflux esophagitis [11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26] . Although several systemic meta-analyzes, based on pooled data from case series and randomized controlled trials (RCTs), have been published, the standard bougie size for LSG is not yet established 16,17,21,[23][24][25][26][27][28] . Herein, we performed a network meta-analysis and cataloged results of these controlled trials into a comprehensive systematic review and meta-analysis of available data to determine the standard bougie size for calibration during LSG, based on excess weight loss (EWL), associated complications, and SLL percentage. Materials and methods Current network meta-analysis was performed after establishing guidelines from the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Network Meta-Analyses (PRISMA-NMA) 29,30 and Meta-analyzes Of Observational Studies in Epidemiology (Table S1.1 and Table S1.2) 31 . Registered protocol was available in the Open Science Framework (https:// osf. io/ drhsb). Data sources and search strategy. A systematic publication review without language restrictions was performed and retrieved from PubMed and Embase from inception until September 1, 2020. Gray literature and manual searches for potentially eligible articles from review articles were reviewed. The US Government Clinical Trials database (www. Clini calTr ials. gov) was searched for ongoing clinical trials. The search terms comprised "laparoscopic sleeve gastrectomy, " "bougie calibration, " and "bougie size, " along with a list of all interventions and possibly relevant keywords. (Table S2). Inclusion and exclusion criteria. Both RCTs and observational cohort studies were included in the study design. Targets of comparison were patients with morbid obesity (aged at least 18 years) received LSG with bougie calibration to create the neogastric tube. Morbid obesity is defined as body mass index (BMI) of ≥ 40 kg/m 2 or 35 kg/m 2 generally associated with comorbidities, except for lower BMIs among East Asians 20,32 . The comparison included two or more different bougie sizes; studies reporting the relationship between bougie sizes and weight loss were enrolled in this study. The exclusion criteria included (1) studies that evaluated adolescents or pediatrics participants, (2) those in which interested outcomes were not reported, (3) those not specific to patients with morbid obesity, and (4) single-arm studies without comparators. Two authors (KH Chen, TW Chang) independently selected trials that met the inclusion criteria, and another author (KH Chen) adjudicated differences. In case of disagreement, the same authors consulted with another one (PC Chang) to achieve decisions after a deliberate group discussion. Data extraction and bias assessment. Two reviewers (HJ Jhou, TW Chang) independently screened the studies, extracted relevant data, and assessed the risk of bias among included studies using the Cochrane risk of bias tool (Table S3.1 and Table S3.2) 33 . Data extraction was performed using a special designed sheet obtained from reports of a previous meta-analysis 28 . Study information regarding studies, participants, and treatment characteristics was obtained. If available data are lacking, corresponding authors were contacted for data collection. Outcome definition. 1. Percentage of EWL (% EWL): (weight loss/baseline excess weight) × 100, where weight loss = preoperative weight -the initial weight; baseline excess weight = initial weight − ideal weight (X) where X was calculated using an ideal BMI, and the ideal BMI cutoff point has been used differently in enrolled studies (Table S4). 2. Overall complications: All complications related to LSG, such as de novo gastroesophageal reflux disease (GERD), postoperative bleeding, nonspecific abdominal pain, nausea, dehydration, surgical site infection, portal vein thrombosis, or as defined by the study authors, were obtained (Table S4). 3. SLL: Postoperative neogastric tube leak or as defined by study authors (Table S4). We used the frequentist network meta-analysis (NMA) model to compare effect sizes among studies with the same interventions. All frequentist approach network meta-analyzes were performed using the statistical package Netmeta (Version 1.2-1) 34,35 in R Project 3.6.1 (R Core Team, Vienna, Austria) and Stata version 16 (Stata Corp, College Station, Texas). The symmetry and geometry of the evidence were examined by producing a network plot with nodes for the number of study participants and connection sizes corresponding to the number of studies 36 . For continuous data, summary standardized mean differences (SMDs) with 95% confidence intervals (CIs) were calculated using a random-effects model (21) www.nature.com/scientificreports/ ratios (ORs) with 95% CIs were estimated with 0.5 zero-cell correction 37 . All comparisons were set as two-tailed, and a p value statistical significant cutoff point was set at 0.05. For generating high resolution figure, we applied Adobe Photoshop CC 2018 and Sketch Version 45.1 software. The rank of treatment within defined groups measured was measured using the surface under the cumulative ranking curve (SUCRA) 38 , which is the relative percentage of probabilities as the best treatment, in a scale from 0 (worst) to 1 (best) 39 . SUCRA can be clinically used to compare treatment effects of all treatments for the target outcomes. Potential inconsistencies between direct and indirect evidences were compared within the network model. Moreover, global inconsistencies were examined using a design-by-treatment interaction model, whereas local inconsistencies between the included comparators were examined using a node-splitting method 40 . Statistical significance was set at 5% for analyzes. If the inconsistency existed, sensitivity analysis was performed to determine possible reasons. The assumption of network transitivity was examined by visually inspecting tables with patient's population across included studies, study methodologies, design intervention details, and outcome measurement differences 41 . Comparison-adjusted funnel plots and Egger's test regression were used to assess possible publication bias or potential small study effects for available interventions 42 . Ethical statement. There was no human trial in this systematic review and network meta-analysis, and the approval from the ethics committee does not apply. Consent statement. Because this study is a systematic review with network meta-analysis, the informed consent does not apply. Results Systematic literature review. Figure 1 presents the whole flowchart of the current NMA. After the initial screening procedure, a total of 25 articles were considered for full-text review; 10 of which were excluded for various reasons. These 15 trials were included in our study, and a total of 2848 participants receiving LSG using different bougie sizes calibrated as XL, L, M, and S were included. Study characteristics are summarized in Table 1. The sample size of enrolled studies ranges from 24 to 1395 patients with morbid obesity. The total numbers of RCTs and observational cohort studies are 687 and 2161 patients respectively. Among these participants, the mean age of enrolled patients was 37.46 years. The mean BMI and body weight of participants were 46.97 kg/m 2 and 121.16 kg, respectively. The bougie calibration sizes ranged from 27 to 60 Fr. Inconsistency and sensitivity analysis. In the EWL outcome, global inconsistency existed with statistical significance between design inconsistency in design-by-treatment interaction model. Local inconsistency was also observed with statistical significance in the node-splitting model. Thus, a sensitivity analysis was performed to evaluate possible reasons of inconsistency. As heterogeneity may exist in the follow-up period between studies, Hady et al. 's study was excluded with a 6-month follow-up 23 . The remaining seven included studies had constant follow-up period of at least 1 year and enrolled in our sensitivity analysis. In the sensitivity analysis, S-, M-, and L-sized bougies were associated with significantly better EWL than that in the XL-sized. According to the SUCRA value, the S-sized bougie was associated with the greatest EWL among all of the different sizes of bougie. (S-sized: SMD, 7.76; 95% CI 1.75-13.77; SUCRA, 0.84). Furthermore, the L-and M-sized bougies had similar effect and were ranked as second and third, respectively (L-sized: SMD, Figure S1). In outcomes of total complication or SLL, no inconsistency was observed between evidence derived from direct and indirect comparisons, including either global inconsistency as assessed using the design-by-treatment interaction model or local inconsistency as assessed using the node-splitting model (Table S6). Risk of bias, inconsistency, and publication bias. We found that 57.1% (100/175 items), 27.4% (48/175 items), and 15.4% (27/175 items) of included studies were assessed as low, unclear, and high risk of bias, respectively. Funding sources and concealing procedure after randomization mainly contributed to the high and unclear risk of bias, respectively (Table S3.1 and Table S3.2). The assumption of network transitivity was established by visually inspecting tables with patient population across included studies (Table 1). Formal assessments of funnel plots across included studies were conducted for all outcomes and revealed general symmetry without publication bias. Results of the Egger's test indicated were not statistically significant, which also suggested no publication bias in the present NMA ( Figure S2.1 to S2.3). Discussion To the best of our knowledge, this is the first network meta-analysis that analyzes the effectiveness and safe range of bougie sizes to achieve reduced weight loss and lower complications for LSG. The network meta-analysis presented herein not only shows the relative treatment effect (EWL) and associated complications from all pairwise comparisons but also offered ranking of different bougie sizes 34,36,[39][40][41][42] . We comprehensively reviewed the major database and included only high-quality articles. Based on currently available evidences, our results suggest that S-and M-sized bougies both have the greatest EWL, and the latter was associated with the lowest www.nature.com/scientificreports/ incidence of total complications, including SLL. Within our network meta-analysis, using the M-sized bougie (bougie size between 33 and 36 Fr., including 36 Fr.) for intraoperative calibration is an optimal choice to balance the effectiveness and safety during LSG. LSG itself is considered as a purely restrictive bariatric surgical procedure and also has impacts on gastrointestinal motility, hormonal regulations, and gut microbiota. LSG has been demonstrated to increase the rate of gastric emptying and intestinal transit. Studies also found increased glucagon-like peptide 1 and peptide YY levels and increased endocrine functions for bile acids after LSG 9,43,44 . In regard to clinical effects, LSG can produce efficient weight loss and improve obesity-related comorbidities accordingly, such as type 2 diabetes mellitus, hypertension, dyslipidemia, or obstructive sleep apnea [43][44][45] . Therefore, an ideal gastric sleeve with the proper size should be created to strike the balance between the acceptable weight loss and occurrence of complications. Intraoperative bougie calibration is an essential part for LSG via different bougie tube sizes to assist bariatric surgeons and to determine the expected gastric tube. Although the clinical significance of an ideal gastric sleeve cannot be overemphasized, the ideal bougie size used in LSG remains to be established. In 2013, a literature review to discuss the ideal bougie size reported the L-sized bougie could decrease the incidence of SLL with the similar EWL effect as the S-sized bougie 46 . Series of RCTs or retrospective studies to compare thinner and bigger size of bougie calibration were conducted in recent decades, and conflict results were presented 12,13,[17][18][19][20][21][22] . In 2018, Wang et al. conducted a meta-analysis that discovered thinner-sized bougie in LSG was more effective in augmenting weight loss, and overall complications were not increased 28 . With new evidences enrolled in this study, our results are in agreement with that of www.nature.com/scientificreports/ previous meta-analyzes, and we further determined the ideal range of bougie size balancing the effect of weight loss and safety in clinical practice. In addition to bougie sizes, some studies advocated that related surgical manipulation of the distance from the pylorus with associated

antral resection/preservation might influence LSG outcomes [47][48][49][50] . For the restrictive purpose, antral resection with shorter distance from the pylorus limits more gastric volume to create a smaller www.nature.com/scientificreports/ gastric tube and is hypothesized to increase intragastric pressure and decrease the distention ability after eating, which results in early satiety theoretically [47][48][49][50] . However, some surgeons suggested antral preservation to prevent possible gastric outlet stenosis and to decrease intragastric pressure in order to reduce the SLL risk [51][52][53][54] . In a 2018 meta-analysis, antral resection is associated with better effect of weight loss and without increased risk of surgical complications as compared with antral preservation 54 . Some adverse events reported to be associated with LSG were bleeding, nausea, wound infection, SLL, or de novo GERD 11,61 . However, the advantages and disadvantages of LSG for GERD remains controversial. In 2019, a systematic literature review found LSG is associated with an increased incidence of de novo GERD. Those with mild GERD might have improved symptoms after LSG; however, patients with morbid obesity, severe reflux and erosive esophagitis may have high possibility of persistent GERD thereafter 55 . Moreover, SLL is a catastrophic complication after LSG, and previous studies have postulated that larger bougie size may decrease the SLL risk 46,56 . Nevertheless, surgeons' personal experience might play a vital role in decreasing and even preventing this undesirable complication 57 . In 2018, Demusy et al. analyzed the nationwide data in the United States, which disclosed that the bougie size was not associated with postoperative leak rate, and the risk of bleeding and reoperation was decreased via concomitant staple line reinforcement intraoperatively 58 . In our network meta-analysis, overall complication rates or SLL did not significantly increase in the group with bougie size of < 32 Fr. (S-sized group), whereas the M-sized group was associated with the lowest incidence of total complications. Herein, we have categorized different bougie sizes into four groups for the following reasons. First, we routinely used a 32-Fr. oral gastric tube during vertical gastric sleeve stapling to format the gastric tube in our institution. Accordingly, we choose 32 Fr. as the first cut point. Moreover, we set 36 Fr. as the second cut point because it was found to be the optimal bougie size to augment EWL in recent studies 18,28,59 . A retrospective multicenter cohort study conducted by Sánchez-Santos et al. in 2016 concluded that a bougie size of > 40 Fr. had a protective effect to minimize the overall complication rate 60,61 . Therefore, 40-Fr. bougie was used as the third cut point to investigate the potential protective effect among the group with larger bougie sizes. Some modifications have been made to increase the strength of the evidence in our network meta-analysis. First, we strictly followed standardized guidelines based on the PRISMA statement to improve reporting of systematic reviews 29,30 . Second, inconsistency and sensitivity analyzes were performed to evaluate possible reasons of inconsistency, and factors that could increase inconsistency were successfully identified and excluded. Nevertheless, the present network meta-analysis has three limitations. First, to increase patient numbers in our studies, some prospective and retrospective cohort studies, which may decrease the strength of the evidence were enrolled. Second, these four groups of different bougie sizes were categorized artificially, which may lead to some potential bias. Third, the criteria of reported complications in enrolled studies may be different, which may result in inaccurate complication rates. Therefore, results of this network meta-analysis should be cautiously interpreted. Conclusion Based on our network meta-analysis and current evidences, S-and M-sized bougies had similar effects of EWL, with the latter being associated with the lowest incidence of total complications, including SLL. Intraoperative calibration with M-sized bougie (33)(34)(35)(36) is an optimal choice to balance the effectiveness and safety for patients with morbid obesity undergoing LSG. In vivo two-photon microscopy of the human eye Two-photon (2P) microscopy is a powerful tool for imaging and exploring label-free biological tissues at high resolution. Although this type of microscopy has been demonstrated in ex vivo ocular tissues of both humans and animal models, imaging the human eye in vivo has always been challenging. This work presents a novel compact 2P microscope for non-contact imaging of the anterior part of the living human eye. The performance of the instrument was tested and the maximum permissible exposure to protect ocular tissues established. To the best of our knowledge, 2P images of the in vivo human cornea, the sclera and the trabecular meshwork are shown for the very first time. Acquired images are of enough quality to visualize collagen arrangement and morphological features of clinical interest. Future implementations of this technique may constitute a potential tool for early diagnosis of ocular diseases at submicron scale. The eye is the sense organ responsible for vision. Among its physiological complex elements, two of them are rich in collagen: the cornea and the sclera. The former is transparent and contributes to approximately two-thirds of the eye's refractive power 1 . Within the cornea, the stroma is mainly composed of stacked collagen lamellae and it makes up ~90% of its thickness. Changes in the corneal shape and/or in its mechanical properties might lead to a loss of transparency and in consequence, to vision loss or sight impaired. Then, understanding the morphology of the collagen structural distribution is of great importance in disease diagnosis. On the other hand, the sclera is an opaque structure (known as the white of the eye) that protects and provides rigidity and structural integrity to the ocular globe. Different clinical instruments are used to image corneal structures. These include specular microscopy (or reflectance confocal microscopy) and optical coherence tomography (OCT) 2,3 . OCT and confocal microscopy are effective when imaging the microstructures of the different corneal layers (epithelial and endothelial cells, keratocytes,…). However, the visualization of the collagen fibers of the stroma has been challenging. In confocal imaging, the performance is limited by the reduced image contrast mechanism from the transparent stromal fibers. Stromal structure can be imaged with OCT, but micrometric resolution is not achievable. These constraints have been overpassed by using two-photon (2P) microscopy 4 . This is a well-established imaging technique used to study non-labeled biological tissues. It consists on a quasi-simultaneous absorption of two infrared photons, followed by the emission of a unique photon in the visible range. This nonlinear process only occurs when the photon flux of the excitation light is in the range of 10 20 -10 30 photons/(cm 2 ·s). This energy density can be reached if a pulsed (~80 MHz) ultra-short (∼100 fs) laser system is used 4,5 . Since this excitation occurs just within the focal plane, the technique provides inherent confocality. Apart from this intrinsic sectioning capability, penetration depth, minimized photo-bleaching and photo-toxicity effects are also important advantages of this type of microscopy. 2P microscopy includes two-photon excitation fluorescence (TPEF) and second harmonic generation (SHG), among others. TPEF signal arises from flavins and intrinsic chromophores such as NAD(P)H 6 . On the other hand, SHG allows the visualization of non-centrosymmetric structures such as type-I fibrillar collagen 7 . Since its invention more than 25 years ago, 2P microscopy has been used in many different fields of life sciences [7][8][9][10] . However, its successful application to visualize ocular structures made scientists and ophthalmologists increase their interest in this tool (see ref. 11 as a general reference). Its capability to image the eye's structures with sub-micron resolution provide potential advantages compared to other "classical" clinical imaging techniques, such as OCT, fluorescence microscopy or confocal reflectance. Different ex vivo human ocular structures such as the retina 12,13 , the sclera 14 , and the trabecular meshwork 15,16 have been investigated through 2P microscopy. However, much effort and interest were put on the cornea since its early SHG visualization 17 . Although SHG images of the healthy human corneal stroma have been reported by different authors [18][19][20][21][22] , the analysis of pathological corneas raised special attention. In human donor corneas, these pathologies include keratoconus [23][24][25][26][27][28][29] , keratitis 29,30 , bullous keratopathy 31 and fibrosis 32 . More recently, multiple ex vivo 2P experiments included evaluation of the feasibility for transplantation of human corneas 33 , analysis of stromal striae 34 and scars 29 , control of limbal epithelial stem cells in cultured human donor corneas 35 and objective analysis of corneal structures using non-reference image quality assessment methods 36 . Although 2P microscopy is useful to understand the changes suffered by corneal collagen under different circumstances, all those previous experiments used exclusively ex vivo samples (from human donors). Some previous experiments involved living animal models [37][38][39][40] , however not all them were successful providing SHG images of non-labeled corneas. To the best of our knowledge, 2P imaging microscopy of the living human eye has never been performed before. In this work we present a new compact clinically-adapted 2P microscope to record images of the in vivo human eye in a non-contact manner (i.e. using a non-immersion objective). The performance of this customized instrument was tested. Images from different in vivo ocular tissues of the anterior segment of human eye were acquired (sclera, trabecular meshwork and cornea). Moreover, the safety limits and the threshold to avoid ocular tissue damage while 2P imaging were also established. Results Safety considerations for the cornea and the retina during 2P imaging. The main goal of the instrument here described has been to image in vivo human ocular tissues. Then, to protect those biological structures, the corresponding irradiance limits have to be established. A high-repetition (or mode-locked) infrared laser system was used as illumination source (see Methods section). This might cause thermal damage in the tissues when the laser power density is above the maximum permissible exposure (MPE) for a given exposure time. Due to this, safe 2P imaging performance of the living human eye requires a well-defined and accurate calculation of the damage threshold. To establish these safety limits for the ocular structures, both the International Commission on Non-Ionizing Radiation Protection (ICNRP) 41 and the ANSI Z136.1-2000 42 standards were considered. The calculation of these limits was performed considering the worst possible scenario: a static focal point on the sample, that is, a confined light spot illuminating the same spatial location during the exposure time corresponding to the acquisition of one image. This assumption benefits the security levels, as the position of the laser beam during imaging recording is not static, but dynamic (that is, the beam scans the ocular tissue during that time). Figure 1a depicts the MPE at the corneal plane (i.e. maximum corneal irradiance) as a function of the exposure time, according to the ICNRP. Then, if we consider a static beam focused on the sample during a time equivalent to the image acquisition time (0.42 s), the MPE will be 3.49 W/cm 2 . It must be remembered (see Methods and Suppl. Material) that 0.42 s is the exposure time used to acquire a 2P image of 200 × 200 pixels with a resolution of 1.5 μm/pixel. Since the incident average laser power density at the corneal plane used here was 20 mW/cm 2 (see Suppl. Material for details on this), the experimental settings lead to conditions well below the MPE. For a better understanding, this power density is represented with a dotted blue line in Fig. 1a. Although the instrument has been designed to acquire 2P images of the anterior surface of the eye, it is also important to consider that the laser beam, after focusing on the cornea, it propagates across the ocular media, reaching the retina. This light spot on the retina must also be within the safety limits. To calculate the retinal projection, a simplified model of both the microscope optics and the eye has been simulated using an optical design software (Zemax LCC, Kirkland, WA, USA). The eye model was composed www.nature.com/scientificreports www.nature.com/scientificreports/ of cornea, aqueous humor, lens, vitreous and retina. Data of dimensions, refractive indices and curvatures were taken from the literature 43 . Ocular absorption was considered to be null. Considering once again a static laser focused on the cornea, the size of the retinal projection was 7.6 mm 2 . For this retinal area, and according the ANSI Z136.1-2000 standards, the absolute maximum laser power (MP) to avoid retinal damage as a function of

the exposure time is shown in Fig. 1b. For an exposure time of 0.42 s, this MP corresponds to 306 mW. This threshold corresponds to a corneal power density of 4 W/cm 2 , a value which is two orders of magnitude higher than the 20 mW/cm 2 used in this experiment. In addition, for this corneal illumination value, the total laser power delivered to the retinal projection is 1.54 mW (200 times lower than the MP above computed, 306 mW). For the sense of completeness, this value has also been plotted in Fig. 1b. All these calculations corroborate that the experimental conditions here used are below the damage threshold, and both the cornea and the retina are protected during 2P imaging operation. Figure 2a-2d show SHG images of the living human cornea from two healthy subjects (subject #1, 2a and 2c; subject #2, 2b and 2d). Images were acquired at the corneal apex and correspond to two different stroma locations: anterior ( Fig. 2a,b, ~50-µm depth) and posterior stroma (Fig. 2c,d, ~350-µm depth). A visual inspection reveals the interwoven arrangement of the collagen fibers and the presence of two preferential orientations (PO). SHG images of living human ocular structures. A quantitative analysis of the stromal structure can be carried out by using the structure tensor (further information on this tool can be found in ref. 44 ). This is a mathematical algorithm providing objective information on the degree of organization of the corneal stroma and the existing POs. As expected, the analysis of the SHG images of these healthy corneas presents a low structural dispersion (SD) which is associated to an organized distribution of collagen 44 . In particular, for the SHG images shown in Fig. 2a-d the SD values ranged between 14° and 22°. For the sense of completeness Fig. 2e,f compare SHG images of the corneal stroma acquired in in vivo and ex vivo human eyes. It can be observed that both samples present a similar overall lamellar distribution, in particular a crosshatched pattern 21 , that is, an arrangement with collagen fibers running along two preferential directions. This simple observation is quantitatively shown in the PO histograms of Fig. 2g,h, computed using the structure tensor 44 . These two structural directions are associated to the two peaks of each plot. The structure tensor used here has been reported to be useful to compare the spatial distribution of structural features in pairs of 2P images 45,46 . In addition, another comparison using the Mean Image Gradient (MIG) has also been calculated. MIG serves as edge detection operator to extract information from digital images 47 . The computed MGI values for the SHG images in Fig. 2e,f are 0.10 ± 0.06 and 0.09 ± 0.07 respectively. These values confirm that, although ex vivo 2P images seem much better resolved, the structural information provided by both types of images in terms of collagen structure is similar. www.nature.com/scientificreports www.nature.com/scientificreports/ SHG images of the sclera of both volunteers are depicted in Fig. 3a-d. Despite the opacity of the scleral tissue, two depth locations spaced 50 μm were imaged. The image resolution and the acquisition time were the same as in Fig. 2. Despite a fixation target was used to minimize ocular movements (see Methods), the stability of the eye during measurements and the repeatability of the recorded images are crucial to test the accuracy and performance of the clinically-oriented instrument here developed. Figure 3e-h presents two pairs of in vivo SHG images (cornea and sclera) recorded 2 s apart. It can be observed how the recording operation is rapid enough to provide images with hardly noticeable movements. The differences in grey levels between pairs of SHG images were 2.4% and 1% for the top and bottom panels respectively. At this point it is interesting to note that all the images here shown are individual frames (i.e. "raw" images) without post-image processing procedures. 2P images of the in vivo trabecular meshwork and the juxta-canalicular tissue are depicted in Fig. 4. As an illustrative example, the trabecular meshwork of subject #1 is shown for two different scanning areas (Fig. 4a,b). Figure 4c corresponds to subject #2. The juxta-canalicular tissue can be seen in Fig. 4d. To record these images, no filter was used in front of the detection unit. This means that the signal reaching the detector combines TPEF and SHG signals. Discussion To the best of our knowledge, we have demonstrated 2P imaging in the intact human eye for the first time. A prototype was developed to successfully perform in vivo 2P images of different non-stained living ocular structures. This custom-made instrument is easy to operate and has a reduced size, an incident beam in a horizontal configuration and a fast scanning procedure. It might also be adapted to a clinical environment. Unlike clinical confocal instruments, our setup uses a long working-distance dry microscope objective. This avoids the contact with the eye and the use of ocular local anaesthetics, what greatly increases the subject's comfort. Our results demonstrate that 2P images (individual frames, no image post-processing) can be obtained within the safety limits. The performance of our instrument was tested by comparing living and ex vivo SHG images of the human cornea. The results showed a similar interwoven arrangement of the stromal lamellae. This is in agreement with previous findings 21 and confirms an appropriate in vivo image acquisition. Although with both experimental conditions lamellar domains were retrieved and similar structural information is obtained, living images seem to be "less sharp", with features less outlined. The main reason is thought to be based on "motion artefacts". Despite the exposure time and the image size are optimized to minimize these movements (apart from using the fixation test and the control cameras) there still exist remaining non-controlled ocular micro-movements that might contribute to this. It is worth to remember that 2P images here shown are "raw" images, that is, neither frame averaging nor image registration were employed here. Zhang et al. reported 2P imaging of the cornea in the living mouse stained with injected fluorescent viable dyes 39 . Mice were anesthetized and a head-holding adapter was used to immobilize the animals. Moreover, a plastic eye cup filled with saline solution minimized heartbeat and breathing induced eyeball movement and fit the experimental requirements of the water-immersion microscope objective used. Since they only explored TPEF signals, the stroma fibers couldn't be visualized and only cellular layers and some nerves were observed. Hao and co-authors couldn't get 2P signal (neither TPEF nor SHG) from normal non-labelled in vivo rabbit corneas. The reason was thought to be the low laser power of the commercial instrument used, since they could SHG imaging of anesthetized rat corneas has been reported by Latour and co-authors 38 . The use of a custom-built aplanation device minimized eye movements. In addition, an ophthalmic gel maintained the optical contact between the immersion objective and the eye. They did not provide precise data on the image acquisition time, but it was assumed that the main limitations referred to vital movements. To solve this issue, they performed image processing operations using small imaged regions (13 × 13 μm 2 ) with a size larger that the cornea movement. Final images were of enough quality for the collagen lamellae to be visualized. However, despite those results, the authors stated that a few improvements in the acquisition conditions were required. In particular, they proposed a reduction of the acquisition time by increasing the pixel size (with the corresponding decrease in resolution), an improved immobilization of the animal's eye and additional image processing to better compensate for eye motion. Although no alteration of the stromal collagen was observed during their experiments, they also claimed that further studies to calculate the maximum permissible power were necessary. Most of the constraints of those previous studies involving animals were minimized or over passed in the present work. A precise evaluation of the femtosecond laser infrared light levels to avoid any tissue damage and ensure cell viability was presented herein. This ensured that our experimental conditions were well below the exposure limits. Although a test on the feasibility of safe in vivo TPEF imaging has been recently reported, that was exclusively centred on high resolution retinal imaging 48 . Some of the issues associated with image stability have also been minimized here by establishing an appropriate imaging protocol. First, the subject's head was immobilized by means of a customized chin-rest similar to that used in clinical instruments. The accurate positioning device of this chin rest, the control cameras and the collaboration of the subjects helped the operator to acquire images under optimal conditions. The subjects involved in the experiment felt comfortable during the entire image recording operation also due to the use of a non-immersion objective. Second, the fixation test made the subjects stare at a point to avoid large ocular movements. Finally, different experimental parameters involving safety limits, pixel dwell time, pixel-based image dimensions and imaged region size set an optimized acquisition time (see also Suppl. Material) to routinely record 2P images with minimal impact from ocular movements. Despite some involuntary lateral eye movements might take place, 2P images were good enough to directly visualize collagen fibers. Axial ocular movements are expected to be marginal 38 and do not disrupt the imaging procedure. Useful information on structural and morphological information can be extracted from the images. In particular, by combining SHG microscopy and the structure tensor analysis, is was possible to objectively identify www.nature.com/scientificreports www.nature.com/scientificreports/ the orientations of the collagen fibers. The size of the collagen fibers could also be computed and compared with previous ex vivo studies 21 . Apart from collagen-based ocular structures, we were also able to image the trabecular meshwork. Images revealed small features (limbus ultra-structure and individual cells within the juxta-canalicular tissue) that might be important in the regulation of the intraocular pressure and glaucoma diagnosis. The in vivo appearance is in good agreement when compared to ex vivo images previously reported 15,16,49 . An increase in contrast and/or resolution leads to better image quality. However, this implies image averaging (more sequential frames for the same imaged area) or an increase in the exposure time for each scanning position (i.e. higher photon integration time at detection unit). For ex vivo samples this is not an issue since the specimen is static. However, for in vivo measurements these are not appropriate approaches. Moreover, this would break the balance of the experimental acquisition conditions here reached. 2P images here shown are representative of normal healthy eyes. However, it has been shown that SHG microscopy is able to reveal structural changes produced by different corneal pathologies [23][24][25][26][27][28][29][30][31][32] , which seriously compromise its transparency and biomechanical properties. Effects of abnormal levels of intraocular pressure in the cornea, the sclera and the limbus could also be tracked with this technique. The characterization of the trabecular meshwork morphology as a function of the intraocular pressure will also help in glaucoma diagnosis. Since high myopia affects the scleral structure at different locations 50 , the visualization of spatial distribution of the collagen fibers would also be of great interest. It is well-known that SHG signal predominantly occurs in the forward direction and the backward-emitted component is weaker 20,51 . As a result, SHG images acquired in both directions of propagation are qualitatively different 18,19 . Despite this, it has been reported that the retrieved lamellar orientation maps are the same for both modalities 11,38 . It is obvious that only the backward direction can be used with in vivo measurements. In conclusion, real-time 2P image recording of the living human eye has been reported. A new clinically-oriented prototype was successfully developed for this goal. Images here showed enabled structural information of ocular structures without the use of markers. This minimally invasive method will open the door to in vivo investigations and would be crucial in ophthalmological environments. Diagnoses and follow-ups of different pathologies and surgeries will benefit from this approach in the next future. Methods Compact 2P microscope for living ocular measurements. A 2P microscope was built to fit a reduced space with dimensions 35 × 25 × 25 cm 3 . The design allowed living measurements in human eye. The prototype was composed of two platforms: the upper one contained

all the optical elements (but the illumination laser system) and the bottom one included all the electronic components (data acquisition card, power supplies, connection board, etc). A picture and a schematic diagram of the device are shown in Fig. 5a,b. The light source was a tunable Ti:Sapphire femtosecond laser (Mira 900 f, Coherent, St. Clara, CA, USA) set to a wavelength of 800 nm, with a repetition rate of 76 MHz. The pulse duration at the focal position of the microscope objective was 395 fs as measured with an auto-correlator (Mini, APE, Berlin, Germany). The light beam entered the microscope through an electro-mechanical shutter (S; SHB1, Thorlabs Inc., Newton, NJ, USA) and reached a Galilean telescope (BE) used to expand (2x magnification) and re-collimate the laser beam. The aperture AP fit the size of the beam to the back-aperture of the objective. The XY scanning was performed by means of a pair of mirrors attached to a dual-axis galvo unit (GVSM002, Thorlabs Inc., Newton, NJ, USA). After passing a telescopic system composed of two achromatic doublets (L1 and L2) and a dichroic short-pass mirror (DM, 69-218, Edmund Optics Ltd., York, UK), the beam entered the microscope objective OB (PLAN APO 20×, NA = 0.42, Edmund Optics Ltd., York, UK) and reached the living human eye under study. DM separated the excitation light from the generated 2P signal coming back from the eye. The objective was coupled to a motorized actuator (Z806, Thorlabs Inc., Newton, NJ, USA) to correct the focus position and provide optical sectioning along the Z direction through the human cornea (i.e. three-dimensional imaging). This microscope objective was chosen to have a long working-distance (WD = 20 mm). It provides non-immersion focusing, what avoids the uncomfortable eye-contact required in the actual clinical instruments such as corneal confocal microscopes 2 . The average laser power at the eye position was controlled by means of a neutral density filter (NDF) located behind the BE. 2P signal emissions (TPEF and/or SHG) from the eye were collected via the same microscope objective. They passed the corresponding spectral filter (SF) placed in front of the photomultiplier (PMT; H9305-02, Hamamatsu Photonics, Hamamatsu City, Japan) used as detector unit. The 2P signals were isolated by means of two different spectral filters: a band-pass for SHG (FB400-10, Thorlabs Inc., Newton, NJ, USA) and long-pass for TPEF (FELH0450, Thorlabs Inc., Newton, NJ, USA). An operational amplifier converted the output current from the PMT into voltage differences. A data acquisition card (DAQ; NI 6363, National Instruments, Austin, TX, USA) was used to synchronize the signal acquisition, the speed of the scanning unit, the shutter and the axial focus positioning. The microscope system was fully controlled via C ++ custom-written software. Similar to a slit-lamp clinical device, this compact 2P microscope incorporates a 3-axis adjustable chin rest to ensure correct positioning and alignment of the eye, as well as comfort for the subject. In addition, to minimize ocular movements during imaging a Maltese cross was used as fixation test (FT). This test was retro-illuminated with a white diode and allowed the subject to stare at it with the non-measured contra-lateral eye. The setup also included a dual camera system (DMM 42BUC03-ML, Imaging Source, Bremen, Germany) for the operator to control the correct position of the eye (as described below). One of the cameras (not shown in the figure) was placed on top of the objective and in front of the eye. The other control camera was set in a lateral position. To correctly visualize the position of the eye with respect to the incident beam, this was illuminated with infrared LEDs (950 nm). Once the position was reached, the LEDs were automatically turned off before the scanning imaging operation starts. www.nature.com/scientificreports www.nature.com/scientificreports/ surface, as shown in the bottom left panel. As an example, the panel on the right depicts a typical SHG image of the corneal stroma. Subjects and imaging protocol. Two healthy volunteers were involved in the present study. They were aged 34 (subject #1) and 47 (subject #2). 2P images corresponding to three different locations of the ocular structures were recorded (Fig. 6): corneal apex (labeled as location I), corneal limbus and trabecular meshwork tissue (location II) and sclera (location III). In this experiment, the maximum average laser power density at the corneal surface was 20 mW/cm 2 (see more details on Suppl. Material). The image resolution was 1.5 µm/pixel (200 × 200 pixels) and the acquisition time 0.42 s. The entire protocol was approved by the Ethical Committee of the Universidad de Murcia, Spain. All experiments were performed in accordance with guidelines and regulations. Participants read the study participation informative document and signed the corresponding informed consent. Ex vivo human corneas from healthy donors (not suitable for transplantation) were provided by the Hospital Universitario Virgen de la Arrixaca, Murcia, Spain. This study and the experimental procedure were approved by the Ethical Review Board of the Hospital. Samples were treated following the instructions of the World Medical Association's Declaration of Helsinki. Once the appropriate laser power was set (by means of the NDF) and the imaging parameters established, the subject's eye was centered and aligned using the chin rest, the control cameras and the Z-motor. Then, the user proceeded to initialize the image recording procedure through the interface. The synchronization of all the elements of the instrument makes the mechanical shutter close just after the image recording to avoid extra-laser exposure. If a 3D-scanning microscopy modality is chosen, the shutter is closed as the Z-motor moves between two consecutive axial locations (i.e. while the Z-motor moves from a plane to the next one). Data Availability The datasets and codes used within this paper are available from the corresponding author upon reasonable request. Figure 6. Simplified schematic indicating the locations of the imaged areas. It should be notice that this figure is intended for illustration, as the objective is unique and the eye rotates within its own orbit for the incident beam to be always perpendicular to the imaged region. For simplicity the eye's rotation was omitted here. Social Support and Loneliness Among Black and Hispanic Senior Women Experiencing Food Insecurity The impact of social determinants of health (SDOH) is understudied and until recently not a focal point in nursing education. The new Essentials coupled with the impact of the coronavirus (COVID-19) pandemic deem it necessary to address the intersection of SDOH and population health. The impact of COVID 19 disproportionately affects Black and Hispanic families. Couple the disproportionate numbers of COVID 19 among these groups with the growing incidence of food insecurity, and there is a need to explore intersecting links. Emerging research link the lack of social support systems and loneliness to food insecurity. In alignment with addressing competency-based education, it is critical to assess factors such as social support systems and loneliness and the intersection of its effects on such determinants as food insecurity. The article provides an overview for its readers in examining the incidence of food insecurity in older ethnic minority women along with postulated social attributes as contributing factors to the growth rates of food insecurity. The incidence of food insecurity among older ethnic minority women has grown exponentially amid the pandemic. The authors illustrate the role nurses can play in addressing primary, secondary, and tertiary interventions using Neuman’s Theory. The intervention pathways are delineated through the lens of nursing theoretic framework created by Betty Neuman Systems Model. INTRODUCTION The impact of social determinants of health (SDOH) on vulnerable populations, especially during the COVID-19 pandemic, is understudied. Additional research in this area is needed. However, while research is being carried out, interventions to improve health outcomes for vulnerable populations can be considered. The aim of this article is to provide a theoretic framework for nurses to identify pathways for nursing interventions to minimizing the influence of food insecurity, perceived social support, and loneliness on stress and client health. The intervention pathways are delineated through the lens of The Neuman Systems Model. 1 The vulnerable population of interest is Black and Hispanic senior women because, when compared with White senior women, Black and Hispanic senior women are more likely to have low or moderately low levels of social support and to experience food insecurity, which may be exacerbated by loneliness. Background Poverty Historically in the United States, poverty rates among Blacks and Hispanics have been higher than poverty rates for Whites. 2,3 Since the onset of the COVID-19 pandemic in February of 2020, overall poverty rates in the US have increased from 15.3% to 16.7%; without federal stimulus payouts afforded by the CARES Act, the increase would have been greater at 18.0%. 4 When compared with Whites (0.08%), Blacks (1.4%) and Hispanics (2.1%) experienced the greatest increases in poverty rates during the pandemic. 4 Considering these inequities, it is no surprise that compared with White (30.7%) women over the age of 65, Black (50.2%) and Hispanic (48.7%) women in that same age group are more likely to live 200% below poverty. 5 Supplemental poverty measure data showed even higher rates and greater disparity, with 41.4% of White women over the age of 65 living 200% below poverty compared with 64.1% of Black women and 67.4% of Hispanic women in the same age group. 5 Food Insecurity Because household income is linked to food insecurity 6,7 -defined as food scarcity 8 it is not surprising that Blacks and Hispanics are more likely to be food insecure when compared with Whites, and that Black and Hispanic women are more likely to be food insecure when compared with men. Specifically, Blacks (11.5%, 7.6%) and Hispanics (10.7%, 4.9%) are more likely to have low and very low food security, respectively, when compared with Whites (4.6%, 3.3%). 9 Women are also more likely to have low (19.1%) or very low (9.6%) food security when compared with men (9.5%, 5.9%, respectively). 9 Similar conditions exist for Black women. In 2017, 79.4% of non-Hispanic White women but only 9.0% of non-Hispanic Black women were food secure. 10 Additionally, non-Hispanic Black women were more likely to be food insecure (22.8%) than to be food secure (9.0%). 10 Furthermore, Black (15.1%) and Hispanic (14.8%) seniors age 60 and older are more likely to be food insecure when compared with White (6.2%) seniors in the same age group. 11 Loneliness and Social Support For women, food insecurity has been linked with social support 10 and the social capital those supports provide. 12 Most of the women who are food insecure have low (59.1%) or moderate (31.0%) levels of social support. 10 Marital status, 13 participation in a government assistance program, 13,14 household income (ie, poverty), 6 education, 15 employment, 16 and loneliness 13,14 are additional factors of food insecurity. During the COVID-19 pandemic, seniors have reported experiencing increased loneliness. 17,18 RATIONALE Loneliness has been linked to negative health outcomes, including depression for those who developed closer relationships within their social networks during the pandemic. 18 Food insecurity has been linked generally to overall poorer selfreported health 7,19 as well as to prediabetes, 6,20-22 diabetes, 7,21,23 high blood pressure, congestive heart failure, heart attack, asthma, 7 obesity, 24 and nonalcoholic fatty liver disease. 25 Comorbidities may negatively mediate the influence of food insecurity on health outcomes. 26 Comorbidities include cardiovascular disease, 27,28 cancer, 27,29 chronic fatigue syndrome, 30 musculoskeletal injury, 31 and depression, 27,32 and health-related behaviors include smoking, substance abuse, and poor 27 and disturbed eating habits. 33 Enrollment in a nutrition assistance program may mediate the influence of low food security on overall physical health outcomes. 19,34 Additionally, social support 35potentially in the form of nursing prevention interventions 35 -can reduce the influence of comorbidities and health-related behaviors 27 and thus act as a barrier against the adverse effects of stress on patient health. 35 The specific focus on Black and Hispanic populations is warranted not only because of the greater potential for those populations to be socioeconomically challenged 2,3 and food insecure 9,11 but also because Blacks and Hispanics have been found to have higher incidence of stress when compared with Whites. 27 Sources of stress disparity include (a) greater exposure to incidents of discrimination 36 and violence, (b) greater exposure to barriers to occupational advancement, 37 and (c) the cultivation of resources useful for overcoming these sources of

stress 38 have been exacerbated with the COVID 19 pandemic. THE CLIENT SYSTEM The Neuman systems model is based on the concept of the client system, which can be considered a single client, a group, or multiple groups, and is focused on how those systems interact with their environments 39 in response "to actual or potential environmental stressors, and the use of primary, secondary, and tertiary nursing prevention interventions for retention, attainment, and maintenance of optimal client system wellness." 35(p67) Types of environmental stressors vary 39 and can be "intrapersonal, interpersonal, and extrapersonal" and characteristically "physiologic, psychological, sociocultural, developmental, and spiritual." 35(p67) Examples of stressors include "loss, pain, sensory deprivation, [and] cultural change". 39(p20) Levels of Energy In the client system model, available energy for resisting stressors exists in 3different capacities and in addition to basic bodily functions such as genetic structure, organ strength and weakness, and body temperature regulation. 39 Those levels of energy within client systems are referred to as lines of resistance, normal lines of defense, and flexible lines of defense. In all cases, a client's levels of energy are supported by coping mechanisms, cultural and spiritual belief systems, and lifestyle factors. A client's normal line of defense refers to the client's usual state of wellness and is developed and shaped over time through client behaviors. 39 The client's usual state of wellness, defined as the stable condition of the client system, serves as a baseline for assessing deviances from that condition. When a client's normal line of defense is disrupted, the client's lines of resistance are activated whereby the client's internal and external resources (both known and Social Support and Loneliness unknown) engage to protect the client against the identified encroaching stressor . 39 The client's lines of resistance include major biological protection systems such as the immune system's activation of white blood cells in response to injury or infection. Ideally, the client's lines of resistance will be sufficient enough to return the client system to a stable condition. The alternative is the depletion of system energy and client death. A client's flexible line of defense is their primary protective element against environmental stressors that disrupt the client's normal line of defense (ie, stable health). 39 The client's flexible line of defense is dynamic and can fluctuate rapidly. A simplified graphic of the Neuman systems model is presented in Fig. 1. Client Perceptions of Health In addition to client system reactions to stressors, the ways in which clients perceive their health 35 and the way they cope with stressors 39 influences the strength of their lines of defense and resistance. Essentially, stress is a neutral concept that only gains the capacity for positive or negative influences on health outcomes to the degree that the client perceives the stressor will have positive or negative outcomes and to the extent that the client perceives they are capable of coping with the stressor. 35,39 The mere exposure to a pandemic is a stressor for the client. Couple this with systemic racism and lack of resources, this potentiates stress affecting lines of defense. The concepts of client perceptions and coping are rooted in Lazarus and Folkman's theory of stress and coping. Stress and Coping Unlike traditional, and dichotomous, perspectives of stress that characterize stress as either a stimulus or response, Lazarus and Folkman 35 considered stress a factor related to the characteristics of both the person and the environment for which the person functions. This relationship between stress and the characteristics of both the person and the environment for which the person functions is similar to the way illness cannot be considered solely a function of external influences but a combination of those influences and a person's behavior and susceptibility to illness. Central and critical components to managing stress are appraisal and coping, whereas the degree to which a person perceives a situation to be stressful (appraisals) and that that person can successfully cope with that stressor is due, in part, to that person's perceived level of social support. Subsequently, the person's perceived the stressfulness of a situation and their capacity to cope with the stressful situation influences the person's ability to adapt to the stressful situation. Finally, a person's ability to adapt to a stressful situation mediates the influence of stress on a person's health outcomes. In this way, social support encourages a relationship perspective that can protect people from the negative health outcomes associated with stress. 40 Additionally, from this perspective, people with greater levels of perceived social support will be less likely to experience stress-related health problems because they will be less likely to judge their situations as stressful. Primary, Secondary, and Tertiary Roles of the Nurse Because the Neuman systems model can be applied to a variety of populations and conditions, it is uniquely adaptable to a range of health care concerns in nursing and can be used to delineate the primary, secondary, and tertiary roles of the nurse within the client system. 39 Primary preventions are those that occur before a client encounters and reacts to a stressor. Nursing actions that can function as primary preventions are associated with general nursing knowledge used to identify and assess potential client stressors and implement interventions to mitigate or alleviate those stressors. Additionally, primary preventions may be focused on increasing a client's flexible lines of defense. Secondary preventions are those that occur after a client encounters and reacts to a stressor. 39 Nursing actions that can function as secondary preventions are associated with symptomology related to client reactions to stressors. Actions in this category of prevention include prioritizing interventions and implementing interventions focused on reducing the negative effects of clients' reactions to stressors. Interventions of this nature might include early screening and detection, and treatment of symptoms. Tertiary preventions are those that occur after a client has reacted to a stressor and received treatment. 39 Nursing actions that can function as tertiary preventions are those that help clients adjust and adapt to changing health conditions and move clients closer to system stability. Interventions of this nature might include education meant to prevent future susceptibility to a particular stressor. CONCEPTUAL APPLICATION OF THE NEUMAN SYSTEM MODEL In this article, the Neuman systems model is used to consider food insecurity as a source of stress for the client system, in this case, the Black or Hispanic female senior patient who is food insecure. Perceived social support and loneliness are considered factors of food insecurity. The nurse is conceptualized as a source of primary, secondary, and tertiary interventions. A graphic representation of the relationships among the theoretic concepts presented to this point and the associated covariates and mediating factors is presented in Fig. 2. The holistic perspective of the Neuman systems model makes it not only "timeless [but] expansive in being adaptable to all client care situations." 41(p112) Because the client system is dynamic, nurses may effectually help transform those systems to promote better health outcomes for clients. Using the Neuman systems model to examine the relationships between perceived social support, loneliness, and food insecurity Social Support and Loneliness provides an effective means not only for considering the patient as a complex system but for understanding how the environment influences that system and the varied ways in which nurses can serve as sources of support for patient wellness. As initiators of preventions in the food insecure client senior minority model, nurses may direct clients to sources of emotional and instrumental support and companionship. They also may function to help clients improve their perceptions about their experiences with food insecurity and to better adapt to outcomes of the food insecurity they are experiencing, including stress. Additionally, nurses may help promote client engagement in positive health-related behaviors while also addresses comorbidities that may be having additional negative influences on the client system. In these ways, nurses have the capacity to contribute to improved client wellness, in this case, specifically Black and Hispanic senior women who are food insecure. APPLICATION IN PRACTICE A list of suggested primary, secondary, and tertiary preventions and their conceptualized applications in practice as they relate to perceived social support, loneliness, and food insecurity is presented in Table 1. These suggestions are not all inclusive, and nurses are encouraged to generate other potential means of prevention. Nurses conduct health assessments and obtain data related to the psycho-social-cultural being which allows one to examine potential risks for allostatic load and burden of diseases. Early detection is key and providing a toolkit for vulnerable populations in particular to have during unplanned circumstances such as pandemics and or natural disasters may be key elements in attaining good health outcomes. Table 1 The nurse as source of primary, secondary, and tertiary prevention Model Component Application in Practice Primary prevention Identify and assess potential client stressors Identify food insecurity as a potential stressor to the client system Assess the degree to which food insecurity has the potential to negatively affect the client system Identify social support and loneliness as potential influences on client's perceived severity of food insecurity Assess the degree to which social support and loneliness are potential influences on client's perceived severity of food insecurity RECOMMENDATIONS Some of the preventions suggested in Table 1, in particular those that require identification and assessment, necessitate the observation and/or measure of social support, loneliness, and food insecurity. Nurses are urged to consider the various definitions of these terms as they plan potential interventions. A list is provided in Table 2, although this list is not inclusive. Because of the conceptual complexity of social support and food insecurity, those terms are discussed in more detail. Social Support Over the last 4 decades, researchers and theorists have proposed various definitions of social support in response to their explorations of its connection to psychological and physical manifestations of health. 45,50 Lack of agreement on the definition is due to its multidimensionality in the way it operates, 43 whereas support can be (a) both given and received, (b) considered from the perspectives of both availability and use, and (c) considered from the perspective of the origin of the support. 51 However, social support also can be informal or formal 52 and emotional or instrumental. 42 Social support also can be considered with respect to clinical utility and health outcomes, 52 whereas social support can be a direct influence on health outcomes or a buffering agent such that social support lessens the negative outcomes associated with stressful events. 53 Food Insecurity The United States Department of Agriculture (USDA) typically refers to food security, which they define as "access by all people at all times to enough food for an active, healthy life" 9(p2) Murillo and colleagues 8 referred to that USDA definition for food security when they defined its opposite, food insecurity. As shown in Table 2, Wright and colleagues 22 and Nagarajan and colleagues 49 defined food insecurity using similar language. However, unlike Murillo and colleagues who included reasons for the lack of access to or availability of healthy foods in their definition of food insecurity, neither 47(p1257) Food insecurity measured by the 4 domains of the Four Domain Food Insecurity Scale: "shortage of food (quantitative), unsuitability of food and diet (qualitative), preoccupation or uncertainty in access to enough food (psychological), and alienation or lack of control over their food situation (social)" Conceptual Anderson 48(1560) "Food insecurity exists whenever the availability of nutritionally adequate and safe foods or the ability to acquire acceptable foods in socially acceptable ways is limited or uncertain" Murillo et al. 8(p428) "Lack of access or availability to healthy foods due to scarce resources or money" Nagarajan et al. 49 Lack of availability of healthy foods Wright et al. 22(p130) "Limited access to a sufficient quantity of affordable, nutritious food" Social Support and Loneliness Wright nor Nagarajan and colleagues do so. The USDA also does not reference reasons for having access to enough food in its definition of food security. The most comprehensive definition of food insecurity includes quantitative, qualitative, psychological, and social factors associated with food insecurity. 47 Because it is broad in scope, that definition is suggested for use in future research. SUMMARY A disciplinary focus in the nursing field is social justice; as nurses, we are morally obligated to act in ways that promote immediate social change in the form of

improved patient care and outcomes. 54 Nurses are in an ideal position to examine SDOH and to implement strategies to rectify health inequities among vulnerable groups such as the aging population. Additionally, nurse educators are well-situated to raise awareness of the importance of nurses to act in this capacity. Such efforts are encouraged and could help the United States move closer to the 2030 goal of eradicating food insecurity championed by the Food and Agriculture Organization of the United Nations and colleagues 55 and reducing the incidence of health inequities among Black and Hispanic senior women. In addition to improving patient outcomes, reduction of cost negative health-related outcomes of food insecurity could be reduced. According to Berkowitz and colleagues, 56 median annual state-and county-level health care costs are $687,041,000 and $4,433,000, respectively. For food insecure adults, additional health care costs amount to $1834 annually. At the national level, those numbers are even more astounding at $77.5 billion and $1,863, respectively. 57 Saved monies could be reallocated for additional interventions to further reduce the incidence of health inequities among Black and Hispanic senior women. The timing of this article correlates with the updated recommendations put forth in The Essentials: Core Competencies for Professional Nursing Education which include SDOH as one of the 8 featured concepts for professional nursing education programs. 58 The 8 concepts are intertwined among 10 domains of competence; together, they represent what the American Association of Colleges of Nursing describes as a new model for nursing education that is competency based, structured for application across levels of education, and adaptable to accommodate a future change in the field of nursing. The inclusion of SDOH as an essential concept of learning for nurses underscores the important role nurses can serve in addressing SDOH and health inequities that contribute to inequity in health outcomes. Through health and needs assessments, health promotion, patient education, and improved access to care, nurses may have a direct impact on the health care of community members from vulnerable populations. 58 In this article, we argued that these very actions be taken by nurses acting as primary, secondary, and tertiary interventions to improve health outcomes for Black and Hispanic senior women experiencing food insecurity. CLINICS CARE POINTS Screening for SDOH during the patient intake process can be expedited using The Hunger Vital Sign screening tool (2 items), 59 The Three-Item Loneliness Scale, 60 and The Social Support Questionnaire (SSQ3; three-item short form). 61 Referral to food support programs is associated with decreased Loneliness, 62 Medication nonadherence, 63 Admissions to nursing homes, 64 and Overall health care costs. 65 Knowledge, attitudes and self-reported practices toward children oral health among mother’s attending maternal and child’s units, Salé, Morocco Background The occurrence of severe dental caries is particularly prevalent and harmful in children. A better understanding of parental factors that may be indicators of children’s risk of developing dental caries is important for the development of preventive measures. This study was conducted to assess knowledge, attitudes, and practices (KAP) of mothers in Salé, Morocco regarding oral health and their predictors. Methods A cross-sectional KAP study was conducted of Mother and Child units in Salé, Morocco. Mothers attending the selected units from November 2014 to 29 January 2015 were recruited. Data were collected using a semi-structured questionnaire, administered by face-to-face interviews, to record socio-demographic factors and KAPs. The main outcome measures included knowledge about oral health diseases and preventive measures, and attitudes and practices related to oral health prevention measures and dental care. KAPs scores were then recoded based on responses and scores were determined for each KAP domain. Linear regression analysis was conducted to assess predictors of KAP scores. Results Among 502 mothers included, 140 (27.8%) were illiterate and 285 (60.9%) were aware that fluoride has a beneficial effect in caries prevention. Mothers’ own practices about dental care were statistically related to their children’s use of dental care services (p < 0.001). Multiple linear regression analysis revealed that the knowledge score was associated with mother’s age (β = 0.05; 95% CI; p < 0.001), education level, and median income (β = 0.38; p = 0.04). Significant predictors of oral health-related practices were mother’s education level and children’s health status. Conclusions Limited KAP scores were observed among the studied population. A great emphasis on oral health education and some risk factor modifications are recommended. Electronic supplementary material The online version of this article (10.1186/s12889-018-5542-2) contains supplementary material, which is available to authorized users. Background Appropriate infant oral health attitudes and practices are of fundamental importance for preventing chronic oral diseases [1]. Oral health patterns are consolidated during childhood, and some attitudes may increase a child's risk of caries development [2,3]. Morbidities due to dental caries are particularly harboured in children from families of low socio-economic level [4], whose nutrition [5,6] and quality of life may be consequently impaired [3,[7][8][9]. Potential risk factors of dental caries include biological and behavioural factors, all of which may be modulated by environmental factors [10,11]. Parents play an important role in promoting positive attitudes and strategies toward oral health behaviours [12,13]. Mothers are the immediate and reliable caregivers of children in many countries, and they have a central role in providing effective guidance and positive attitudes toward oral health [14,15]. Despite improvements in oral health measures in high-income countries, the literature notes the persistence of an imbalance in caries prevalence in certain countries [16,17]. Moreover, most KAP-related surveys concentrate on parents from high-income countries, and less is known from countries with high prevalence of dental caries such as Morocco [17]. Morocco has recently commenced an expanded program on immunisation. Mother and Child units (MCUs) serve as basis of this program, which is administered on a routine and outreach basis. The MCUs in Morocco offer vaccines free of charge to children from birth. They also offer counselling for mothers regarding infantfeeding practices and general health. The aim of the present study was to assess knowledge, beliefs, attitudes, and practices (KAPs) of mothers attending MCUs regarding infant oral health and preventive measures. Participants This study was a cross-sectional survey among mothers attending early childhood services between 12 November 2014 and 29 January 2015, including childcare and mother's health services in selected MCUs in Salé, in the Kingdom of Morocco. Salé is a city situated on the opposite bank of the Bouregreg River from the capital, Rabat. The total population of Salé in 2014 was 890,403 [18]. The study population was mothers aged 18 years and above who had at least one borne child. Mothers who were less than 18 years old, unable to respond to questions, or declined to participate in the study were excluded. All mothers who attended the selected units during the period of survey were approached and invited to participate in the survey; this includes either mothers with newborn children attending the units for vaccination or for paediatrics consultation. Participation was strictly voluntary; no subject-identifying data were collected and no incentives were provided. The institutional review board and the regional health authority approved the study prior to its implementation. Data collection A structured questionnaire to investigate knowledge, attitudes, and practices was developed for the purpose of the study. The questionnaire drew on previous research in this area [13-15, 17, 19, 20], with necessary modifications made considering lifestyle and cultural factors related to the Moroccan population. The pre-final questionnaire was then pre-tested with 102 mothers. This pre-final form served to test the face validity of the questionnaire and determine how meaningful it was to the studied population. After discussion with participants and interviewer, the variability in responses, understanding of items, and ambiguity were evaluated. The questionnaire was then refined and the final version was used for this study. The entire survey took place from 12 November 2014 to 29 January 2015, during different periods in each unit. For each unit, the interviewer approached all mothers during the study period to determine eligibility. All eligible mothers were asked if they were interested in participating in the study. The information was collected by face-to-face interview. The questionnaire was filled out anonymously on site. The interviewer was previously trained to standardise approach to the participants for consent and in application of the questionnaire. The final version of the questionnaire covered the followings areas (Additional file 1): -Family socio-economic and demographic variables, including parent's education level (none, primary, 6 < EL < 12, university), number of children, household income (low, moderate, and high), mother's employment status (working/non-working), and medical coverage (yes/no). The monthly family income was measured relative to the Moroccan minimum wage during the period of data gathering. -Mother's general health rating (good global health/ chronic illness), children's health status (good global health/chronic illness). -Knowledge was measured using a 10-item questionnaire relating to general knowledge of oral health, including the value of fluoride in dental caries prevention, recommended fluoride dose for children, preventive measures regarding oral bacterial acquisition, relation between primary teeth and permanent teeth, importance of oral health, age to begin teeth brushing, frequency of brushing, nutrition behaviour and dental caries, and importance of preventive dental health. Responses were rated and recoded yes/ no/don't know). The following ranking scale was used to record responses: 1 point per correct answer, 0 for incorrect or don't know). Scores ranged from 0 to 10, and a higher score indicated a higher level of knowledge. -Attitudes were measured using a six-item questionnaire. Recorded items included attitudes about oral bacterial acquisition, prevention measures, counselling, and control methods (parental support of infant oral health and nutrition behaviour). The attitudes were first recorded and then classified as positive or negative attitudes. -Practice domains, including practice patterns regarding oral hygiene measures (frequency and regularity of tooth brushing, fluoridated toothpaste use, sugar consumption, age child commenced tooth brushing, brushing motion), dental attendance (mother's own dental attendance pattern, child's dental attendance pattern), and nutritional practices. The practices were first recorded and then classified as positive or negative patterns. The section regarding attitudes and practices included multiple-choice as well as open-ended questions. The practice patterns section was completed prior to the knowledge and attitudes section to avoid leading answers. Data analysis Characteristics of the study population were presented as median and interquartile ranges for continuous data with skewed distribution or with mean and standard deviation for variables with normal distribution. The normality of distribution of quantitative variables was tested by Kolmogorov-Smirnov test. The Mann-Whitney test was applied for comparison of non-normally distributed variables. Categorical variables were expressed as percentages and tested by the chi-squared test. Data were analysed by developing themes related to the specific objectives. These themes were knowledge, attitude, and practice levels. Each participant was assigned a knowledge score, an attitude score, and a practice score based on the number of correct responses for knowledge, good practice, and positive attitude. The scores were further analysed using linear regression analysis to determine independent effects of explanatory variables or predictors of each score (knowledge, attitudes, and practices) in both univariate and multivariable analysis. Knowledge, attitude, and practice were the primary outcome variables. Explanatory variables included: age, number of children, mother's educational level, mother's marital status, children's health status, and income. Predictors with p value ≤0.25 in univariate analysis were introduced in multivariable analysis. Assumptions of residuals being normally distributed were fulfilled for all scores using graphical method. Significance was predetermined at a probability value of 0.05 or less. Resulting associations were reported as regression coefficients β and their confidence intervals (95% CI). Data analysis was conducted using SPSS 13.0 software (Chicago, IL, USA). Results A total of 503 mothers were included during the study period. Ten mothers refused to respond to the all items of the questionnaire. The socio-demographic characteristics of the included mothers are represented in Table 1. The mean age of the studied mothers was 30.3 ± 6.7; 461 (91.8%) of the mothers were non-employed. The median number of children was 2(1-3). Regarding educational background, 140 (27.8%) of the mothers were illiterate and 129 (25.6%) of the mothers were primary-school level; 372 (74%) had medical coverage. The median knowledge score was 5(4-6). A total of 370 (95.9%) of the mothers were unaware that teeth should be cleaned beginning at eruption. The median age of tooth-brushing inception declared by the mothers was 3.5 (3)(4)(5) of the mothers studied, 253 (53.3%) believed that primary teeth are

not necessary and that more care should be taken for permanent teeth, 285 (60.9%) of the mothers were aware that fluoride has a beneficial effect in prevention, and 409 (86.3%) of the mothers were unaware that medication intake during pregnancy and childhood may affect teeth development. The main sources of information cited by mothers were family members and other mothers. Paediatricians and dentists were cited in less than 10% of cases. Previous children's use of dental services was reported by 230 (45.7%) of the mothers. Emergency dental care was reported by 46.9% of the mothers, and 69 (13.7%) of the mothers reported that they sought treatment for their children from a dental practitioner for a toothache. Mothers' own practices about dental care were statistically related to their children's use of dental care (p < 0.001). Among mothers who had never had a previous dental consultation, 121 (77.1%) declared that their children had never sought a dentist's care while 78 (41.7%) of mothers who had received dental care had sought a previous consultation for their children. The findings of the linear regression analysis, examining predictors of knowledge, attitude, and practices score, are presented in Tables 2, 3 and 4. The adjusted linear regression model showed that the knowledge score was positively related to mother's age (β = 0.05; p < 0.001), education level, and median income (β = 0.38; p = 0.04). Significant predictors of oral-health-related practices were mother's education and children's health status. The predictors of oral-health-related attitudes were income, mother's employment status, and education level. Discussion The findings of the present study indicated a low level of knowledge and unfavourable attitudes and practices related to oral health. Socio-economic factors were the main predictors of KAP scores. Parents' knowledge and attitudes predicted a range of negative oral health practices. Less knowledge of parenting strategies to promote healthy teeth may be a predictor of more unhealthy teeth [10][11][12][13][14][15]. Regular use of fluoride to control dental caries was reported in many studies [16,17,20,21]. However, fluoride use in the present study was not widespread. These results may be explained by low knowledge of fluoride's role in prevention of dental caries. Studies in various countries showed that low adherence to preventive measures was associated with high prevalence of dental caries [11,[22][23][24]. The development of positive and negative patterns in children (oral health behaviour, hygiene attitudes, sugar consumption) are a function of parental feeding practices [25,26]. The present study revealed a lower rate of adequate oral-health-related attitudes and practices among both mothers and children. Parents constitute an important social model in delivering health skills to their children [26]. Many studies have shown that tooth-brushing behaviour established during infancy is often maintained during early childhood, adolescence, and adulthood [23,27]. The family, and particularly mothers, provide the child's proximate home environment [25,26]. Being unaware of or neglecting the importance of oral health and common preventive measures against dental caries and periodontal diseases are important inhibiting factors to achieving an acceptable level of children's oral health [16,17,28]. Thus, lack of awareness and corollary attitudes and practices likely contribute to a high prevalence of common oral health diseases (dental caries and periodontitis), and the associated costs impose a substantial burden on societies [16,17,28]. As role models, parents can encourage oral health preventive measures by regularly encouraging children to brush and brushing themselves [26]. Parents should also set eating examples by encouraging less use of sweetened beverages (particularly sweetened milk) [29][30][31]. As the results of this study revealed, there are still a considerable number of mothers who prefer other options to taking the child to the dentist, even when child has an emergency need. To some extent, this may predispose the child to greater dental health complications and impact nutritional status. A possible reason for this low prevalence may be attributable to economic limitations. Studies have shown lower rates of dental care use among low socio-economic groups [32,33]. The present study showed that the level of healthrelated knowledge increased with years of schooling. Mothers with higher education are thought to have better opportunity to obtain information about childcare than mothers with lower education levels [34,35]. Mothers with less education may not have basic knowledge about the impacts of potential risk factors on the occurrence of preventable oral diseases [34,35]. Therefore, health education and promotion are warranted to be developed among mothers with low educational background [36][37][38]. Dental caries and periodontal diseases are common and frequently occurring oral diseases among the Moroccan population [16]. To control these diseases, it is not enough to focus on the surgical treatment, as subjective factors such health literacy and self-management efficacy also have great impact [38]. Within families, social and cultural norms are known to shape many attitudes towards health [39,40]. In the present study, the mothers followed the advice of their circle of mothers and those in their community. This attitude reaffirms traditional values independently of their relevance, and highlights the role of social network transmission of health behaviours [40][41][42]. In the context of the present study, 140 (27.8%) of the 502 mothers included were illiterate. To a certain extent, therefore, greater attention to barriers to seeking and using medical advice should be considered [42]. Health-related behaviour is affected by different aspects of knowledge, attitude, and practices [25,26,38]. This highlights the role of educational interventions affecting the health condition of one individual that could also affect the health condition of others in his/her environment [25,26,38,40,42]. Mother's educational level and income were significantly inversely related to children's oral health practices and attitudes. Greater income among families with working mothers and greater educational level could explain findings results regarding attitude scores. These socio-economic predictors are associated with higher levels of oral diseases [11,43,44]. The level of family income was identified as significant predictor of KAP scores among the studied population. Low socio-economic status, poor oral health knowledge and attitudes, and poor oral health behaviours are risk factors for dental caries and periodontal diseases [28]. For this, socioeconomic (SES) inequality remains an important focus in health research because economic inequality is associated with a variety of negative health outcomes [45]. Hobdell MH et al. [46] stated that socio-economic variables account for approximately 50% of the differences in the prevalence of dental caries. The influence of SES on health outcomes may vary depending on the indicator used [47]. The present study used two indicators of SES: mother's education level and income. These indicators were chosen because it is assumed that they are associated with material factors, psychological factors, and health behaviour [48]. The present study found that having at least one child with a health concern is associated with increased level of positive attitude. This result may be explained by the fact that having a child with a health concern may lead to better parental awareness of and attention to overall health [49]. Even if immense efforts are made to control caries prevalence, it should be borne in mind that monitoring and evaluation of knowledge, attitudes, and practices in a community have a major role in creating sustainable control interventions [2,28]. When moving from the clinical to population level, characteristics of the target population for prevention strategies should be taken into account [36,37]. A dynamic solution to this complex problem will introduce challenges for the use of knowledge of preventive measures and its dissemination [50] among groups at high risk for oral diseases. From this point of view, early and accurate selection of high-risk groups for prevention and intervention is of great importance. Coordinating action across health services may be a useful tool. Mother and Child units seem to be one of the most important targeted units for dissemination of primary prevention knowledge. Because vaccination programs are promoted in these units, every mother can access important knowledge and not be disadvantaged from attaining it because of class, socioeconomic status, or determined circumstance. Educational protocols need to be established to advise all mothers attending these units regarding many aspects of health, including oral health. Mothers can then receive counselling for their children from a professional who is taught to have codified knowledge. However, ensuring optimal dissemination of the best knowledge on oral health promotion presents an ongoing challenge and requires improved awareness among health professionals working in the MCUs. Some limitations should be addressed for the present study. Its cross-sectional design limits the ability to conclude a causal inference. However, the predictors included (mother's education, income, and employment status) are unlikely to change over time. Another limitation that should be taken into account is the high rate of mothers in the low and medium income category. This could influence the results, and a repetition of the study design should take into account this limitation. However, this study also has some strengths. Firstly, the findings add data to the limited literature on oralhealth-related KAPs among mothers from countries with a high oral disease burden. Moreover, these findings have important implications for future research. Acknowledgement of cultural practices that reflect inappropriate attitudes and practices should be addressed for more contextual preventive measures. Research to improve public perceptions, as well as the perceptions of mothers specifically, regarding oral health may lead to greater improvement of oral health indicators. Better understanding of the importance of dental health in daily life and better oral hygiene could enhance oral health indicators and reduce the burden of dental caries among specific groups. Health professionals, together with health workers, should perform counselling activities to address the risk factors associated with the disease. Mothers attending MCUs are a key target population for education regarding all aspects of health. In this context, health education should be planned to increase knowledge and minimise misconceptions of the mothers as well as the community as a whole. Conclusion The results of this study indicated that socioeconomic factors such as mother's educational level and income were predictors for oral health practices and attitudes. These findings call for primary care services, particularly maternal and child health services, to increase oral health knowledge. Oral health education and promotion workshops should be organised in maternal and child health units to educate and inform mothers regarding dental caries prevention measures. Additional file Additional file 1: Knowledge, attitudes and self-reported practices toward children oral health among mother's attending maternal and child's units, Salé, Morocco. (DOCX 18 kb) Acknowledgments We are grateful to all those who participated in this study. Funding The research program supported by Mohammed V University Rabat, Morocco funded this study. The university support program had no role in the design of the study, analysis and interpretation of the data, or in writing the manuscript. Availability of data and materials Data are available upon request from the first and third authors. Authors' contributions SC was involved in the study design, prepared the data, performed the statistical analyses and drafted the manuscript. FA was involved in the design of the study and coordination. RA participated in the design of the study, advised on the statistical analyses, the interpretation of the analyses and helped to draft the manuscript. SH helped design the study, acquired and analyzed the data. All authors read and approved the final manuscript. Ethics approval and consent to participate The study protocol was performed in accordance with ethics principles. Verbal informed consent was gained from the participants before enrollement because of the cultural stigma about signing papers. The proposal was approved by the institutional review board, University Mohammed V Rabat. Analysis was conducted on anonymized data. Chronic arthritis in children and adolescents in two Indian health service user populations Background High prevalence rates for rheumatoid arthritis, spondyloarthopathies, and systemic lupus erythematosus have been described in American Indian and Alaskan Native adults. The impact of these diseases on American Indian children has not been investigated. Methods We used International Classification of Diseases-9 (ICD-9) codes to search two Indian Health Service (IHS) patient registration databases over the years 1998–2000, searching for individuals 19 years of age or younger with specific ICD-9-specified diagnoses. Crude estimates for disease prevalence were made based on the number of individuals identified with these diagnoses within the database. Results Rheumatoid arthritis (RA) / juvenile rheumatoid arthritis (JRA) was the most frequent diagnosis given. The prevalence rate for JRA in the Oklahoma City Area was estimated as 53 per 100,000 individuals at risk, while in the Billings Area, the estimated prevalence was

nearly twice that, at 115 per 100,000. These rates are considerably higher than those reported in the most recent European studies. Conclusion Chronic arthritis in childhood represents an important, though unrecognized, chronic health challenge within the American Indian population living in the United States. Background As a group, the rheumatic diseases of childhood represent one of the most common chronic disease conditions in children [1]. These illnesses have global distribution [2][3][4][5], but little information exists regarding either prevalence or phenotypic expression of these diseases in children in any population other than North American and European whites [6]. Aggarwal and colleagues [7] have reported their experience with juvenile rheumatoid arthritis (JRA) on the Indian subcontinent, and their findings suggest that the patterns of disease reported in Europe and North America are not seen in that population. Most conspicuous in Aggarwal's study was the relative rarity of the pauciarticular form of JRA, in contrast to European and North American populations, where that subtype accounts for 50 to 75% of the cases [3,[8][9][10][11]. The relative rarity of pauciarticular JRA in non-European populations has been documented in children of Kuwait [5], Turkey [12], Thailand [13], Japan [14] and South Africa [15] as well as in African American children in Detroit [16]. Further evidence that examining prevalence rates of rheumatic diseases in specific populations may be informative comes from studies of systemic lupus erythematosus (SLE). Studies from Great Britain, for example, indicate that the prevalence rates for SLE in people of Afro-Caribbean descent may be 4-8 times higher than that in Caucasians [17,18]. Prevalence rates of rheumatic disease in North American Indian/First Nations populations have been reported in small studies from single tribes. From these studies, significantly higher prevalence rates for rheumatoid arthritis (RA) have been reported in adults from tribes living in the Great Lakes region [19], the Pacific Northwest [20], the Southwest [21] and Canada [22][23][24]. To our knowledge, a comprehensive survey of rheumatic diseases affecting children, adolescents, and young adults has not been reported in any non-Caucasian population. Because both our clinical experience here in Oklahoma suggests that rheumatic diseases in children may also be more prevalent in the American Indian population compared with Caucasians, we undertook a search of the Oklahoma City Indian Health Service (IHS) user population databases in order to develop prevalence estimates of rheumatic diseases in American Indian children and adolescents. We performed the same queries from the database in the Billings Area IHS office as a basis of comparison. Populations in the billings and Oklahoma City areas The Oklahoma City Area IHS serves a population of 291,288 individuals, most of whom reside in Oklahoma, with small numbers living in the neighboring states of Kansas and Texas [25]. IHS services are limited to members of federally recognized tribes, 39 of whom have tribal headquarters in Oklahoma. The 39 federally recognized tribes [26] represent people from multiple Native cultures including Eastern Woodlands tribes (e.g., Cherokee, Delaware, Seneca), Southeastern tribes (e.g., Creek, Choctaw, Chickasaw, Seminole), Southwestern tribes (e.g., Apache), as well as tribes who have long been resident on the southern Great Plains (e.g., Kiowa, Comanche, Southern Cheyenne). Tribal membership is determined by the tribes themselves and may or may not include specific blood quantum requirements for membership. Historical factors, the absence of reservations, and the proximity of European-descended people in Oklahoma has resulted in significant admixture between Native and Caucasian populations in many parts of the Oklahoma City Service Area. The Billings Area IHS serves a population of 72,591 individuals. The tribes in this service area, a significant proportion of whom live on 8 reservations located in Montana and Wyoming, consist largely of northern plains tribes (e.g., Crow, Sioux, Blackfeet). In both Areas, the population is younger than the population of the United States as a whole, with 40 percent of the population 19 years of age or younger [27]. Database search International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9) codes were used to search the Oklahoma City and Billings Area IHS National Patient Information Reporting Systems and Patient Registration user databases [28] over a three-year period (1998)(1999)(2000) to identify individuals with rheumatic diseases. Outpatient data over this three-year period was gathered for the user population ≤ 19 years of age and included: patient chart number (with the chart number scrambled to protect patient identity), date of birth, sex, date of visit and diagnosis (the IHS database allows at least 9 diagnoses to be recorded on any given patient). The codes used and the diagnoses denoted by those codes are listed in Table 1. Data were downloaded into a standard database (Microsoft Excel), which was then searched to eliminate duplicate records and to sort patients for each year on the basis of age, sex, and diagnosis. A case was defined as any person who was 19 years of age or younger on January 1,1998, January 1, 1999 and January 1, 2000 and whose diagnoses included at least one of the entities listed in Table 1. The population at risk was defined as the number of individuals ≤19 years of age within the IHS user population (i.e., eligible individuals who have used the IHS facilities at least one time in three years) [29]. The user population (i.e., people who actually used IHS services) may differ from the IHS service population, which includes all individuals who are eligible to receive IHS services. Estimation of disease prevalence was based on three assumptions: (1) that the diagnoses recorded were, in fact, accurate; (2) that the population at risk did not change significantly over the three-year period; (3) that the three-year mortality rate for the diseases of interest was no greater in the IHS user population than for all races in the United States. For the Oklahoma City Area, accuracy of the IHS database information was assessed by matching the IHS identification number of known JRA cases followed at the Children's Hospital of Oklahoma with the same number in the IHS patient databank to be sure that known patients were identified and coded accurately. Rheumatoid arthritis/juvenile rheumatoid arthritis Rheumatoid arthritis and juvenile rheumatoid arthritis (RA/JRA; #ICD-9 # 714.0, 714.30) were the most frequent rheumatic disease diagnoses recorded in individuals ≤ 19 years of age in both IHS areas. In the Oklahoma City Area, we identified 62 individuals (45 females and 17 males) with these diagnoses. Assuming a population at risk of 117,409 (i.e., individuals 19 years of age or younger; source: IHS Headquarters office, data processing services unit, Albuquerque, NM), this gives a crude prevalence rate of 53 cases per 100,000 at risk. The prevalence rate for Billings was considerably higher. The 714.0 and 714.3 codes identified 33 individuals in the Billings Area. Based on a population at risk of 28,724, the prevalence estimate for Billings was 115 per 100,000 at risk (see Table 2). The age distribution of affected children, adolescents, and young adults in both areas differed from what has been reported in studies from predominantly European or Europeandescended populations (Figures 1A and 1B). There is a distinct biphasic distribution of JRA prevalence by age in Caucasians, with peaks in the late preschool years and in early adolescence [30,31]. Data from both IHS databases show a distinct peak at age 5-12 years with proportionately smaller numbers of patients in the preschool and early adolescent age groups. In both Areas, peaks in late adolescence and early adulthood are observed, consistent with our observation that rheumatoid arthritis is a disease of young adults in this population (Mauldin et al, manuscript in preparation). The absence of a prevalence peak in the preschool years may reflect the almost complete absence of children with monoarticular or pauciarticular JRA in the IHS user population, consistent with previous studies in non-European populations [5,7,8,13,15,32,33]. We found no individuals in the database with the ICD-9 code commonly used to denote pauciarticular-onset JRA (#ICD-9 # 714.32). In the Oklahoma City Area we found a single child (a 12 year old female) diagnosed with monoarthritis (#ICD-9 # 714.33). Included in the above analyses are individuals who may not fit established criteria for a diagnosis of JRA. Since we did not examine age at onset, it is impossible to know whether a 19 year old identified as having rheumatoid disease was age 7 or age 17 at disease onset. JRA diagnosis criteria stipulate that patients must have disease onset at 15 years of age or younger [34]. When data were analyzed to include only children 15 years of age or younger, we identified 35 patients (23 females, 12 males) in the Oklahoma City Area and 21 patients in the Billings Area (9 females, 12 males) with a diagnosis of JRA. Based on the populations at risk of 87,936 (Oklahoma City) and 21,777 (Billings) this yields an estimated prevalence rate of 40 per 100,000 in the Oklahoma City Area and 96 per 100,000 in the Billings Area. Both of these estimates are more than twice the prevalence derived from a recent European study (14.8 per 100,000) [35]. In order to test the integrity of the IHS database in the Oklahoma City Area, we matched known cases of JRA followed at the Children' Hospital of Oklahoma (CHO; n = 15) by IHS identification number with patients in the database. All 15 of the children followed at CHO were identified within the database and correctly identified by subtype. We did not have access to patient records in the Billings Area, but do have access to databases at other IHS facilities. At a large facility in the Aberdeen Area (comprising the states of North Dakota, South Dakota, Nebraska, and Iowa and serving a patient population very similar to that served in the Billings Area) we identified 20 children with JRA using the same search strategy as that used in for the Billings and Oklahoma City Areas. Subsequent chart review demonstrated that 17 of these children had strong clinical evidence to confirm the diagnosis of JRA, while diagnosis could not be supported or excluded in the other three. The estimated prevalence for JRA in the population served by this facility calculates to 236/100,000, within the same order of magnitude by considerably higher than the Billings Area estimate. Sacroilitis, not elsewhere classified 720. 8 Other inflammatory spondyloarthopathies 720.89 Other spondyloarthropathies We identified 20 patients (12 females, 8 males) with spondyloarthropathy in the Oklahoma City Area IHS database, giving an overall crude prevalence rate of 17 per 100,000 at risk. Included in this group are three patients (all female) with psoriatic arthritis (ICD-9 Code # 696.0). In the Billings Area, we identified 12 individuals (7 females and 5 males) with ICD-9 codes used to identify patients with spondyloarthopathy, yielding a prevalence rate of 42 per 100,000 at risk. These included one patient (a 12 year old male) with ankylosing spondylitis (#ICD-9 # 720.0), one (a 5 year old male) with psoriatic arthritis (#ICD-9 #696.0), and 10 patients (6 females and 4 males) with nonspecific sacroiliitis (#ICD-9 # 720.2). The Oklahoma and Billings Area estimates were within the range previously reported for childhood spondyloarthropathies (e.g. ankylosing spondylitis) in the United States and United Kingdom (12 to 33 per 100,000) and Mexico (13 to 65 per 100,000) [48], but slightly lower than previously estimated prevalence rates of 29 per 100,000 (all spondyloarthropathies) for First Nations children in western Canada [22]. The female-to-male preponderance in both Areas was unusual and has not been, to our knowledge, reported with any previous population. Discussion Although rare individually, the rheumatic diseases, taken together, are among the most common chronic health conditions affecting children [1]. Exact prevalence rates among children living in the United States are difficult to obtain, owing, in large part, to the de-centralized delivery of health care in this country. The IHS represents an exception to that decentralization and is, arguably, the closest representation to a nationalized health care delivery system currently functioning within the United States. Thus, records and data available through the IHS represent a unique opportunity to assess population-wide health needs not otherwise available to child health researchers in this country. This report provides a first-ever population-wide estimate of the prevalence of chronic arthritis in American Indian children living within the United States. While the prevalence rate for the Oklahoma City Area was within the same order of magnitude as the most recent reports from Europe [35], the prevalence

rate in the Billings Area was nearly 10 times this recent European estimate. These findings are consistent with earlier studies of rheumatoid disease in American Indian adults, where prevalence rates 10 times higher than the general population were reported [49,50]. It should be pointed out, however, that prevalence estimates of JRA vary widely, ranging between 16 to 113 per 100,000 [30,31,[51][52][53][54][55]. The reasons for the discrepancy in prevalence estimates between the Oklahoma City and Billings Areas are not clear. One possibility is that the northern plains tribes are particularly susceptible to rheumatoid disease in ways that other groups (e.g., Eastern Woodlands or Southwestern tribes) are not. It is also possible that the difference in ethnic composition of the two populations accounts for this difference. While many Oklahoma tribes require at least a 25% blood quantum of tribal ancestry (e.g., the Kiowa tribe [56]), other tribes require only proof of descent from an individual on the original Dawes rolls of 1893 [57]. Thus, the Oklahoma City Area includes many individuals whose degree of American Indian ancestry is 1/4 or less and may include individuals with less than 1/ 64 American Indian ancestry. In contrast, there has been less intermingling between Caucasian and American Indian populations on the northern plains, and a larger percentage of the Billings Area population includes individuals with full-blooded American Indian ancestry. Our study once again points out the rarity of the pauciarticular form of JRA in non-European populations. In studies of European and European-descended populations, pauciarticular JRA is the most common form of chronic childhood arthritis [3,8]. The Oklahoma City Area database listed a single child with ICD-9 codes #714.32 (pauciarticular JRA) or 714.33 (monoarticular arthritis), the codes used to identify such children. These findings are consistent with reports from the Indian subcontinent, [7] Kuwait [5], Turkey [12], Thailand [13], Japan [14], South Africa [15], and with our experience with African American children in Detroit [16]. The slight female-to-male preponderance for spondyloarthopathy is also worth noting. High prevalence rates for spondyloarthopathies have been noted in both Northwestern and Southwestern tribes [24,[41][42][43][44]. However, in these studies, a strong male preponderance was noted. Whether the findings from Oklahoma City and Billings represent a novel finding or inaccuracies in the ICD-9 coding await confirmatory studies, as we discuss below. An important limitation to this study is the fact that we did not have the means to verify every individual case listed in Oklahoma City database and were unable to confirm any diagnosis in the Billings Area database. However, our limited test of the accuracy of the Oklahoma City data provided surprising confirmation of the accuracy of coding for known cases. While we could not confirm any of the Billings cases, our search of the database of a single IHS facility in the neighboring Aberdeen Area corroborated the prevalence statistics we derived from the Oklahoma City and Aberdeen databases. Indeed, our experience suggests that a search strategy like the one we used is likely to under-estimate rather than over-estimate the prevalence of rheumatic disease in the IHS user population. We are aware that there are many factors that might overestimate disease prevalence using this type of database search. The first is the possibility that a given ICD-9 code might have been used to designate a "working" diagnosis that was never established by the patient's clinical course. The second opportunity for overestimation of prevalence would occur if children were systematically misdiagnosed. This could occur easily if physicians use serologic data as the sole criterion for diagnosis. For example, many physicians routinely screen children with musculoskeletal complaints using antinuclear antibody (ANA) tests. However, the prevalence of low-titer positive ANA tests is extraordinarily high in the pediatric population [58]. Thus, if ANA-positive children with musculoskeletal pain [59] are listed as having "JRA," then there would be a gross overestimation of the actual prevalence. Similarly, there are factors that might have led to underestimation of JRA prevalence by relying solely on a threeyear database search. Children or adolescents with wellcontrolled JRA may not have seen an IHS physician during the relevant time period, and thus would have been excluded. Similarly, physicians who rely on rheumatoid factor tests as a diagnostic criterion for JRA might fail to diagnose the disease in a child, since only a small number of children with JRA have detectable IgM rheumatoid factor [60]. The ideal method for obtaining true disease prevalence rates would include rigorous, pro-active case finding in a known population at risk. This approach was taken by Manners and Diepeveen in a study of school children in Australia [61]. Using such an approach, these authors reported a prevalence rate of 4 per 1,000 for JRA, significantly higher than any previous estimates. We are now preparing a similar project involving American Indian communities on the northern plains. Conclusion We conclude that the rheumatic diseases of childhood may represent a significant burden of morbidity in these two IHS user populations. More detailed studies with rigorous case ascertainment are required to follow up these preliminary data. Hypoglycaemia combined with mild hypokalaemia reduces the heart rate and causes abnormal pacemaker activity in a computational model of a human sinoatrial cell Low blood glucose, hypoglycaemia, has been implicated as a possible contributing factor to sudden cardiac death (SCD) in people with diabetes but it is challenging to investigate in clinical studies. We hypothesized the effects of hypoglycaemia on the sinoatrial node (SAN) in the heart to be a candidate mechanism and adapted a computational model of the human SAN action potential developed by Fabbri et al., to investigate the effects of hypoglycaemia on the pacemaker rate. Using Latin hypercube sampling, we combined the effects of low glucose (LG) on the human ether-a-go-go-related gene channel with reduced blood potassium, hypokalaemia, and added sympathetic and parasympathetic stimulus. We showed that hypoglycaemia on its own causes a small decrease in heart rate but there was also a marked decrease in heart rate when combined with hypokalaemia. The effect of the sympathetic stimulus was diminished, causing a smaller increase in heart rate, with LG and hypokalaemia compared to normoglycaemia. By contrast, the effect of the parasympathetic stimulus was enhanced, causing a greater decrease in heart rate. We therefore demonstrate a potential mechanistic explanation for hypoglycaemia-induced bradycardia and show that sinus arrest is a plausible mechanism for SCD in people with diabetes. Introduction Blood glucose levels are elevated in both type 1 and type 2 diabetes. Insulin is used to therapeutically lower high glucose in individuals with type 1 and long-standing type 2 diabetes; however, this commonly results in iatrogenic hypoglycaemia, one major barrier in achieving and maintaining normal glucose control in diabetes [1]. Insulin causes glucose to fall, but in individuals with diabetes, insulin levels are reduced due to auto-immune damage to the pancreatic β cells (type 1 diabetes) or a combination of impaired insulin secretion and action (type 2 diabetes). In both types of diabetes, exogenous insulin is used to lower glucose therapeutically but due to the limitations of current insulin delivery, insulin replacement is unphysiological. It is challenging to achieve stable basal concentrations and those with insulin-treated diabetes are subject to both hyper-and hypoglycaemia. Attempts to mimic normal glucose levels through intensive insulin therapy increase the risk of severe hypoglycaemic episodes. The effects of hypoglycaemia on the cardiovascular system have only recently emerged and severe hypoglycaemia has been associated with increased mortality in randomized trials [2,3]. In one of the trials, intensive glucose therapy resulted in a sixfold increase in severe hypoglycaemia [2]. Over one-third of deaths were attributed to sudden cardiac death (SCD), raising the possible link between hypoglycaemia and cardiac arrhythmias. Indeed, there is growing evidence in recent clinical and animal studies that hypoglycaemia might contribute to malignant cardiac arrhythmias leading to SCD. Severe hypoglycaemia in response to glucose therapy was associated with an increased risk of arrhythmic death in individuals with type 2 diabetes [3]. Spontaneous nocturnal hypoglycaemia was also implicated in the deaths of young people with type 1 diabetes. The term 'dead in bed syndrome' was introduced to describe these unexpected deaths [4], which were attributed to SCD. Subsequent studies reported a relatively large proportion of unexplained deaths in this population [5,6]. Although rare in absolute numbers, SCD is 10-fold higher in young people with type 1 diabetes compared to those without diabetes [7]. In observational studies, we and others reported an increased risk of arrhythmias, including slow heart rate, bradycardia, during spontaneous hypoglycaemia versus normoglycaemia (NG) [8][9][10][11]. These arrhythmias could be related to abnormal cardiac repolarization observed in spontaneous [12] and experimental hypoglycaemia studies, where hypoglycaemia was induced by an infusion of insulin [13][14][15]. Experimental hypoglycaemia in rats can cause lethal cardiac arrhythmias, although at glucose levels significantly lower compared to those in human studies [16]. Possible hypoglycaemia-induced proarrhythmic mechanisms include the direct effect of low extracellular glucose on the myocardial cell, which can block potassium channels that carry an outward current during repolarization [17]. This causes a prolongation in the cardiac action potential (AP) as well as lengthening of the QT interval duration in the ECG through a mechanism identical to that in proarrhythmic medications [18]. In addition, the counterregulatory sympathoadrenal activation in response to hypoglycaemia causes a release of adrenaline [19], which in turn lowers serum potassium concentration [20] and increases intracellular calcium concentration [21]. Combined, these can cause further QT interval prolongation and calcium overload and lead to proarrhythmogenic electrical instability [21]. Abnormal autonomic responses which are common in diabetes could cause further risk in these individuals. The autonomic response to falling glucose levels includes sympathetic and parasympathetic activation, with the latter limiting the increase in heart rate during hypoglycaemia. However, bradycardia was also observed during spontaneous [22] and experimental hypoglycaemia [23]. In our observational studies, we detected increased rates of bradycardia during hypoglycaemia compared to NG in some individuals with diabetes [8,10,11]. Bradycardia was also common during experimental hypoglycaemia in rats [24]. These observations suggest that bradycardia might be an additional risk factor during hypoglycaemia, possibly by the development of early afterdepolarizations leading to ventricular tachycardia, particularly in the presence of hypokalaemia [25]. Our focus in this study was therefore on the effect of hypoglycaemia on the sinoatrial node (SAN), the heart's natural pacemaker. Hypoglycaemia studies present challenges to investigators. While hypoglycaemia and cardiac arrhythmias are relatively common, SCD is rare, suggesting the involvement of additional factors, including variable and dysfunctional autonomic responses and impaired awareness of hypoglycaemia. Observational clinical hypoglycaemia studies are limited by an inability to measure electrolytes and catecholamines during spontaneous episodes of hypoglycaemia. Experimental studies on the other hand while more controlled require supra-physiological concentrations of insulin to induce hypoglycaemia, rarely encountered in clinical practice. In addition, a greater fall in potassium in these studies is not representative of what happens in people with diabetes in the 'free living' condition. Animal models offer an alternative approach but species differences in the hormonal responses to hypoglycaemia and ensuing cardiac electrophysiological responses mean there are limitations in drawing conclusions from these data. In this study, we have therefore used a computational model to overcome some of these limitations. There are currently no computational models that describe the electrophysiological responses to hypoglycaemia, and in this study, we adapted a biophysically detailed computational model of a human SAN cell electrophysiology to examine the effect of hypoglycaemia on pacemaker rate. Model of the human sinoatrial node action potential We used a computational model of the human SAN AP, developed by Fabbri et al. [26]. This model was in turn adapted from a computational rabbit SAN model, developed by Severi et al. [27] by incorporating experimental data from human SAN cells. These included updated cell capacitance and dimension data, a new formulation of the I f current, new steady-state activation and updated conductances of delayed rectifier K + currents (I Kr and I Ks ) and added I Kur current formulation. Parameters, for which human experimental data were not available, were fitted using automatic parameter optimization. The modulation of the pacemaker rate in response to autonomic stimulation was similarly adapted from the rabbit model, where acetylcholine-and isoprenaline-induced variations were introduced to mimic the activation of the sympathetic (increased rate) and parasympathetic (decreased rate) nervous system, respectively. Further details about the model are provided in the electronic supplementary material. We ran simulations in MATLAB (The MathWorks, Inc.,

Natick, MA, USA) vR2017a, using automatically generated code for the Fabbri model obtained from the CellML platform (www. cellml.org). For solving numerical differentiation, we used the 'ode15s' function with relative error tolerance 10 −7 , absolute error tolerance 10 −6 and a maximum step size of 10 −3 s. We ran the simulations for 60 s and used the last beat to calculate the features of AP waveforms and for figures. We confirmed the stability of the main features of the AP waveform (electronic supplementary material, figure S1). Features of sinoatrial action potential waveform We characterized the sinoatrial AP waveform using a set of parameters presented by Fabbri et al., and we calculated them based on the guidelines for standardization of AP parameters [28]. The following parameters were calculated: CL-cycle length of the pacemaker activity, defined as the duration between the peaks of two consecutive APs; PP-peak action potential, peak (maximum value) of the AP; APA-AP amplitude, defined as the difference between the maximum and minimum AP potentials; MDP-maximum diastolic potential preceding the PP, defined as the most negative repolarization potential; DDR-diastolic depolarization rate; DDR 100 -diastolic depolarization rate over the first 100 ms of diastolic depolarization; APD 90 -AP duration at 90% repolarization (figure 1). The main parameter of our interest was CL. We explored the changes induced by hypoglycaemia and the main features of AP responsible for these changes. Effect of low extracellular glucose-experimental data In an experimental study, Zhang et al. demonstrated a direct effect of low extracellular glucose concentration (Gluo) on the rapid delayed rectifier potassium current (I Kr ) [17]. Using a whole-cell patch clamp, they showed an impairment in the function of the corresponding human ether-a-go-go-related gene (hERG) channel, proportional to the reduction in Gluo. Zhang et al. used human embryonic kidney cells (HEK293 cells), and in this study, we assume that electrophysiological properties and responses of the hERG channels are similar in the heart, where hERG is highly expressed. Zhang et al. demonstrated two separate mechanisms by which reduced Gluo alters the function of the hERG channel: a reduction in I Kr current density and a shift in its current-voltage (I-V) relationship. We digitized their graphs to numerically quantify the extent of changes. We further describe the effect size for both mechanisms and how we incorporated them in our model. Reduced I Kr current density Zhang et al. presented experimental data for I Kr I-V relationships at various Gluo: Gluo = 5 (NG), 2.5, 1 and 0 mmol l −1 . They demonstrated a 10-20% reduction in current density at Gluo = 2.5 mmol l −1 compared to 5 mmol l −1 , a 20-30% reduction at 1 mmol l −1 and a 50% reduction at 0 mmol l −1 . The current density decreased proportionally with decreased Gluo, but the reduction was not constant across the voltage range. In our model, we incorporated the drop in current density by decreasing the conductance (g Kr ) of I Kr . For easier interpretation, g Kr in our model represents normalized conductance, a multiplier of maximum conductance (4.2 nS in the Fabbri model). We decreased g Kr from 1 (maximum conductance, no effect of glucose) to 0 to investigate the overall response of the model. We then varied g Kr within the [1, 0.7] range (100 to 70% of maximum conductance) to consider only physiologically relevant glucose concentrations, based on the above data. Shift in the I Kr current-voltage relationship (IVshift) Zhang et al. demonstrated that the I Kr I-V relationship shifted towards more negative potentials with reduced Gluo. Specifically, they reported a −10 mV shift at Gluo = 0 versus 5 mmol l −1 . There was a similar −10 mV shift in normalized conductance. I-V relationships were not shown for other Gluo, but they performed additional experiments under conditions with corrected osmolarity at Gluo = 1 mmol l −1 . In these experiments, they measured a −3 mV shift in normalized conductance at Gluo = 1 versus 5 mmol l −1 and for the purpose of our study, we assumed that the shift in I-V curves is similar. This is since the shifts in the I-V relationship and conductance were similar in the experiment without correction for osmolarity. To model physiologically relevant effects of hypoglycaemia, we varied IVshift within [−3, 0] mV range but we explored shifts below −3 mV to test the overall response of the model. We extracted all relevant data by digitizing the graphs, and we measured the shifts at 50% I-V curve or conductance amplitude. In the Fabbri model, I Kr is formulated as a function of membrane potential by combining the activation and inactivation kinetics. To model the shift in the I-V curve, we incorporated IVshift into the membrane potential variables of all equations describing the activation and inactivation relationships and their corresponding time constants. Latin hypercube sampling We used Latin hypercube sampling for simulations where two or more input parameters were randomly selected from a predefined range. The sampling was performed by using the 'lhsdesign' algorithm in MATLAB. The variable input parameters included g Kr , IVshift and extracellular potassium concentration (Ko). Outline of the study We investigated the individual effect of each parameter g Kr , IVshift and Ko on CL and features of the AP waveform. For each parameter, the modelling was performed for physiologically relevant values as well as outside this range to test the behaviour of the model and to identify input values that produce meaningful APs. We also looked at the combined effects of these parameters when their values were within physiologically relevant ranges. Using Latin hypercube sampling, we varied g Kr (g Kr is a multiplier of maximum conductance) between 0.7 and 1, IVshift between −3 and 0 mV and Ko between 3.0 and 4.0 mmol l −1 . Ko at 4.0 mmol l −1 represents normal Ko while Ko = 3.0 mmol l −1 indicates moderate hypokalaemia. Finally, we checked the combined effect of these parameters with additional activation of either the sympathetic or parasympathetic autonomic system. The current model allows only one of them to be activated at any time. Reduced conductance of the hERG channel We reduced g Kr (g Kr is a multiplier of maximum conductance) from 1 (maximum conductance) to 0 in steps of 0.05. The model failed to generate spontaneous AP activity at g Kr < 0.11. The relationship between g Kr and CL is presented in figure 2a. CL shortens with decreased g Kr and the relationship is approximately linear down to g Kr = 0.4, followed by a plateau. The minimum CL at g Kr = 0.11 is 0.47 s (−49%). The vertical blue lines indicate the [0.7, 1] range, which represents our estimated reduction in g Kr at clinically relevant reduced glucose concentrations. CL at g Kr = 0.7 is 0.75 s (−18%), APD is prolonged (+21%), the MDP potential is elevated (+8%) and the diastolic depolarization rates DDR and DDR 100 are increased (+32% and +27%) (table 1). Examples of the corresponding AP waveforms are presented in figure 2b. Negative shift in I Kr current-voltage relationship We introduced a negative shift (towards more negative potentials) in the I-V relationship (IVshift), starting from 0 mV in steps of −0.1 mV. CL exponentially increases with negative IVshift (figure 3a) and the longest valid CL is obtained at −5.4 mV (CL = 5.94 s). The blue vertical lines indicate the [−3, 0] mV range which represents our estimated shift at clinically relevant glucose concentrations. CL at IVshift = −3 mV equals 1.2 s (+30%) and the prolongation is mostly caused by decreased MDP (−8%) and depolarization rates DDR (−43%) and DDR 100 (−15%) (table 1). The corresponding AP waveforms are presented in figure 3b. Combined effects of reduced g Kr and negative IVshift We combined the effects of reduced g Kr and negative IVshift by simultaneously sampling g Kr from [0.7, 1] and IVshift from [−3, 0] mV ranges. We used Latin hypercube sampling to generate 300 pairs of random and independent input variables for 300 evaluations of the model. The resulting CLs are presented in figure 4. Red data indicate CL prolongation and blue data CL shortening versus baseline (CL = 0.92 s) (figure 4a). CLs are longest at baseline g Kr (g Kr = 1) combined with maximum IVshift (IVshift = −3 mV) and are shortest at g Kr = 0.7 combined with baseline IVshift (0 mV). The grey dashed line indicates a simultaneous linear change in both parameters. CLs along this line are approximately constant and are slightly longer than baseline CL (approx. 0.93 s, +1%). In figure 4b, a projection of CL values to the g Kr axis shows a gradual shortening of CL with decreased g Kr , and for each g Kr , CL is prolonged with the negative IVshift. The maximum and minimum CLs obtained with these parameters are 1.2 s (+30%) and 0.75 s (−18%), respectively (table 1). Features of the action potential waveform The individual effects of reduced g Kr and negative IVshift produce CL changes in opposite directions (figures 2 and 3) and their combined effects cancel each other out to some degree, producing a slight CL prolongation. A similar opposite effect is obtained with other AP features, including APA, MDP and DDR. For example, baseline MDP equals −62.0 mV (table 1) and the combined effects of g Kr = 0.7 and IVshift = −3 mV slightly raise MDP to −60.3 mV (+3%). Individual changes in g Kr and IVshift, however, produce a bigger difference compared to baseline with minimum/maximum MDP values −66.8 mV (−8%) and −56.9 mV (+8%), respectively. On the other hand, PP, APD 90 and DDR 100 are predominantly affected by one of the input parameters but not by the other. For these AP features, the combined changes in g Kr and IVshift produce values outside the range of their individual effects. For example, APD 90 is prolonged at g Kr = 0.7 (APD 90 = 0.17 s, +21%) but with added IVshift = −3 mV, there is a further prolongation (APD 90 = 0.18 s, +29%). Maximum values of PP, Table 1. Features of the AP waveform (rows 4 and onwards) calculated using a combination of input parameters (rows 1 to 3) representing NG, LG and a combination of LG and low potassium. Baseline values of input parameters (italics), their extreme values and their combinations are considered. We changed the baseline Ko = 5.4 mmol l −1 in the Fabbri model to Ko = 4.0 mmol l −1 . NG LG LG + low Ko Effect of low extracellular potassium With the other input parameters at baseline (NG), we lowered Ko from 5 mmol l −1 in steps of −0.1 mmol l −1 . The relationship between Ko and CL is presented in figure 6a in blue colour. The lowest Ko that produces periodic pacemaker activity is 0.6 mmol l −1 (CL = 1.54 s, +67%). CL is slightly prolonged with the decrease in Ko from 4 mmol l −1 to 3 mmol l −1 (blue vertical lines) and reaches a local maximum of 0.94 s (+2%) at 3.3 mmol l −1 . The prolongation is mainly caused by more negative MDP (−6%) and decreased DDR 100 (−12%) (figure 5c and table 1). Combined effects of low glucose concentration and potassium concentration We first combined the effects of low Gluo and Ko by fixing g Kr and IVshift at the maximum of their ranges: g Kr = 0.7, IVshift = −3 mV and by reducing Ko in a stepwise manner. Second, we randomly sampled all three parameters (g Kr , IVshift and Ko) from their predefined ranges. By combining the effects of randomly sampled input parameters, CL stays within the range obtained from parameters at the extreme ends of their predefined ranges (table 1). Autonomic modulation We investigated the effect of autonomic modulation by separately including in the model either sympathetic or parasympathetic activity. In both scenarios, we looked at the combined responses of autonomic activity, LG and low potassium. To model LG, we first fixed g Kr and IVshift at the extreme ends of their ranges. Second, we explored all possible outcomes by randomly sampling all input parameters (g Kr , IVshift and Ko). Fixed g

Kr and IVshift At NG and Ko = 4 mmol l −1 , an addition of sympathetic activity shortens the CL from 0.92 s to 0.69 s (−25%). In LG, Ko = 4 + Symp.Act. LG, Ko = 3 + Parasymp.Act. The AP waveforms at Ko = 3, 3.5 and 4 mmol l −1 are presented in figure 7b. CL is prolonged with reduced Ko, but the prolongation is greater at LG versus NG (red data above blue), with the biggest difference at Ko = 3 mmol l −1 (CL = 1.73 s versus 1.50, +15%). Some combinations of input parameters with added parasympathetic activity do not produce periodic pacemaker activity (data not presented, see below). To identify input parameters which in combination with parasympathetic activity produce long CL or no pacemaker activity, we divided the parameter space as shown in figure 9a. Red data indicate CL prolongation and blue data CL shortening compared to NG with added parasympathetic activity (CL = 1.27 s). Green data indicate CL > 2 s and purple data indicate invalid CL with no periodic pacemaker activity. At Ko = 4 mmol l −1 , all simulations result in valid periodic activity, with some CLs longer than 2 s (see also figure 8b). These are roughly confined to the area with g Kr > 0.9 and IVshift < −2 mV. At Ko = 3 mmol l −1 , this area extends to g Kr > 0.8 and IVshift < −1 mV (green circles). At Ko = 3.7 mmol l −1 , the model stops producing pacemaker activity around g Kr = 1 and IVshift = −3 mV. At Ko = 3 mmol l −1 , this area extends to g Kr > 0.88 and IVshift < −1.8 mV. Figure 9b shows examples of APs over 40 s simulations. Baseline AP (CL = 0.92 s) is shown in the top row, followed by CL = 2 s in the second row and sinoatrial pauses in rows 3 to 5. The penultimate row shows the longest CL (around 25 s), and the last row shows an example with no periodic pacemaker activity. Discussion In this first study investigating the effects of hypoglycaemia on the SAN electrophysiology, we demonstrate the following findings. First, in response to hypoglycaemia, the reduced conductance of the hERG channel and the shift in its I-V relationship produce opposite effects on the pacemaker rate. When combined, they cancel each other out, resulting in a small decrease in heart rate. Second, mild hypokalaemia causes a slight decrease in heart rate. When combined with hypoglycaemia, there is a marked further decrease in heart rate. Third, hypoglycaemia combined with sympathetic activity enhances the effect of sympathetic activity by further increasing the heart rate compared to NG. However, when hypoglycaemia is combined with hypokalaemia and sympathetic activity, the effect of sympathetic activity is diminished, causing a smaller increase in heart rate compared to that at NG with added sympathetic activity. Fourth, hypoglycaemia enhances the effect of parasympathetic activity by causing a greater decrease in heart rate compared to NG. Hypoglycaemia combined with hypokalaemia and parasympathetic activity causes a further marked reduction in heart rate, resulting in sinoatrial pauses, sinus arrest and block of pacemaker activity. Combined, the above findings indicate that hypoglycaemia tends to decrease the pacemaker rate of the SAN. During experimental hypoglycaemia, where glucose was reduced in a controlled environment using insulin infusion, hypoglycaemia counterregulation caused a significant increase in circulating adrenaline and noradrenaline concentrations together with a significant decrease in potassium in healthy people [15] and in people with type 2 [15] and type 1 diabetes [14]. Despite the significant increase in adrenaline and noradrenaline, there was a relatively small (about 3 bpm) increase in heart rate during hypoglycaemia compared to NG [14,15]. It is therefore possible that decreased glucose and potassium levels counteract the effects of the autonomic nervous system at the level of the SAN. Robinson et al. have reported the results of experimental studies in healthy subjects where they restored normal concentrations of potassium during both NG and hypoglycaemia [14]. They have shown that heart rate did not change during NG after potassium has been added (Ko = 3.6 versus 4 mmol l −1 ). On the other hand, there was an approximate 7 bpm increase in heart rate (61 versus 68 bpm, +11%) when potassium was added (Ko = 3.2 versus 3.9 mmol l −1 ) during hypoglycaemia. These data are in line with our findings that decreased potassium does not significantly affect the heart rate during NG but causes a decrease in heart rate when combined with decreased glucose. For comparable values of Ko, our model shows a 2% higher heart rate during NG and a 9% higher heart rate during hypoglycaemia. In our study, we demonstrate a potential mechanistic explanation for hypoglycaemia-induced bradycardia, which has been observed during clinical episodes in people with type 1 [30] and type 2 diabetes [22]. We have undertaken several observational studies, where we compared rates of cardiac arrhythmia during spontaneous hypoglycaemia versus NG. We observed increased incidences of bradycardia in people with type 1 diabetes [10], with type 2 diabetes [8] and with type 2 diabetes after discharge from ICU [11]. However, increased rates of bradycardia were only detected in a few individuals. We speculated that these effects are idiosyncratic and that mechanisms other than those involved in hypoglycaemia counterregulation might contribute to the development of bradycardia [10,11,31]. These could include autonomic dysfunction and/or abnormal and variable autonomic responses such as those observed during experimental hypoglycaemia, where the early parasympathetic withdrawal was followed by parasympathetic reactivation during sustained hypoglycaemia [15]. Nocturnal hypoglycaemia could also increase the risk as sleep and prone position diminish the sympathoadrenal responses during the night [32]. Our findings that decreased blood glucose and potassium diminish the effect of the sympathetic activity and enhance the parasympathetic activity might further exacerbate the risk of bradycardia in the above cases. To validate the responses of our model during hypoglycaemia, we used the heart rate differences between matched spontaneous hypoglycaemia and NG episodes from our observational study in people with type 1 diabetes [10]. We chose only nocturnal episodes due to diminished sympathoadrenal responses which in turn also cause a smaller reduction in Ko. Our computational model shows an approximate −1% decrease in heart rate during hypoglycaemia at Ko = 4 mmol l −1 . In our observational study, there was a nonsignificant +2% increase in heart rate during hypoglycaemia versus NG with a mean difference of 1.5 bpm (95%CI −1.3, 4.4 bpm). In a similar study, Koivikko et al. [33] reported a +2% increase. Our computational data are within the confidence intervals of those from the above observational studies and the difference in trends could be partly due to counterregulatory sympathetic activation. There are several conditions which could combine with the effects of hypoglycaemia to further promote decreased pacemaker rate at the SAN. Channelopathies, such as mutations in the HCN4, SCN5A and KCNQ1 genes, which encode ion currents I f , I Na and I Ks , can lead to reduced pacemaker rate. In addition, remodelling of the I f depolarization current in trained athletes, which causes decreased resting heart rate and potential training-induced bradycardia [34] could lead to increased risk during hypoglycaemia. These cases could be further explored in the computational model of the human SAN as well as in observational or experimental hypoglycaemia studies. In our model, we show a gradual increase in pacemaker rate with decreased g Kr (figure 2). This contradicts experimental data from animals, where selective blockage of I Kr resulted in decreased rates in rabbit [35] and guinea pig [36] SAN. Due to the lack of human experimental data to confirm our findings, we explored the reduction of g Kr in a computational model of the rabbit SAN. We chose the Severi model [27], which is the parent model of the one used in this study. The AP waveforms and the relationship between g Kr and CL are presented in figure 10. CL is proportionally prolonged with reduced g Kr (figure 10a) and the pacemaker activity is abolished with complete block of I Kr (g Kr = 0, figure 10b). CL prolongation is accompanied by APD prolongation, the elevation of MDP and decreased DDR. This is in line with experimental data from the rabbit SAN [35]. In our human model, a reduction in g Kr is accompanied by APD prolongation and elevation of MDP, as in the rabbit model. In contrast with the rabbit model, DDR increases in the human model (table 1). One of the reasons for differences between the models is the roughly 2.5 times higher resting pacemaker rate in rabbits (CL = 0.92 s in human versus 0.36 s in rabbit). While APD prolongation is similar in both models, this makes up a much bigger proportion of CL in the rabbit AP. In figure 10b, APD prolongations are similar to the prolongations of CL, while the opposite effects of elevated MDP and decreased DDR on CL seem to cancel each other out. By contrast, the duration royalsocietypublishing.org/journal/rsif J. R. Soc. Interface 18: 20210612 of diastolic depolarization is much longer in the human model, making the CL more susceptible to changes in DDR. To further investigate the differences between DDR responses in both models, we compared the major currents responsible for diastolic depolarization at g Kr = 1 versus g Kr = 0.5. In the rabbit model, the amplitude of I f current is relatively high versus the other currents and its amplitude decreases by about 50% with reduced g Kr (electronic supplementary material, figure S2). Similarly, the amplitude of I NaCa markedly decreases. By contrast, the amplitude of the I f current is much smaller in absolute terms as well as relative to the other currents in the human model (electronic supplementary material, figure S3). In addition, the amplitude of I NaCa slightly increases with reduced g Kr . These differential responses in currents, combined with the higher resting rate in rabbits could explain the differences between the g Kr /CL relationship in humans versus rabbits. It is also possible that the decreased pacemaker rate in response to reduced g Kr in rabbits contributes to the reduction of heart rate and occurrence of bradycardia during experimental hypoglycaemia in rats [24]. The key strength of our study is that we explored for the first time the electrophysiological effects of hypoglycaemia on the SAN. Furthermore, to mimic the physiological conditions in humans, we added the stimuli of hypokalaemia as well as sympathetic and parasympathetic stimulation. This approach overcomes some of the technical, biological and ethical constraints of in vivo human and animal studies. One of the limitations of the study is the lack of detailed experimental data on the effect of Gluo on cardiac cell electrophysiology. Zhang et al. demonstrated the main effects of LG on the hERG channel [17], but experimental data at finer glucose increments would greatly improve the understanding of the balance between increased heart rate, caused by reduced conductance of the channel versus slowing of the rate caused by the shift in its I-V properties and how these relate to clinically relevant glucose levels. To the best of our knowledge, there is no evidence of any other cardiac ion channel being affected by hypoglycaemia. Fabbri et al. optimized their model to reproduce baseline electrophysiological characteristics from a limited number of human experimental data and validated it against baseline heart rates observed in common ion channel mutations. Given the lack of experimental data, the model is not optimized and validated for variations in heart rate and altered extracellular ion concentrations. Loewe et al. [37] extended the Fabbri model to work at variable Ko concentrations by including a new formulation of I Kr , where g Kr is dependent on Ko. Their model showed a positive relationship between heart rate and Ko and they found that this was in disagreement with data from the general population, which showed a negative linear relationship. Our model, as the extended Fabbri model, shows a positive relationship between heart rate and Ko. This could mean that our model overestimates the CL at Ko = 3 mmol l −1 during NG as well as with LG and in combination with parasympathetic activity. This behaviour of our

model could be inherited from the parent model. Indeed, existing human atrial models show divergent and sometimes unphysiological responses to changing extracellular Ko [38] and the corresponding electrophysiological conditions remain to be explored. Finally, the Fabbri model offers selective activation of either sympathetic or parasympathetic activity. In reality, both branches of the autonomic system are continuously active, and they complement each other to modulate the pacemaker rate. The responses of their combined activities and how these relate to their individual effects, observed in this study, remain to be investigated. Conclusion Computational modelling has the potential to clarify electropathophysiological effects of hypoglycaemia. We have shown that sinus arrest, a case of extreme bradycardia, is a plausible mechanism for SCD, possibly more probable than tachycardia during the night. Investigation of confounding factors that lead to the development of abnormal cardiac responses, as well as modelling of different cardiac cells at different levels may help to identify individuals with diabetes who are at an increased risk of sudden death due to sinus arrest during hypoglycaemia. This has important clinical implications because equivalent studies are not feasible or are challenging to perform in a clinical setting. Data accessibility. Numerical data, source code of the model and the code to reproduce the figures are available in a Github repository: https://github.com/AlanBernjak/Bernjak_etal_Human_SAN_model_ hypoglycaemia. Modification of Attentional Bias to Emotional Faces Following Mindfulness-Based Cognitive Therapy in People with a Current Depression Mindfulness-based cognitive therapy (MBCT) is an evidence-based treatment to prevent relapse in individuals with recurrent major depressive disorder (MDD). It is not clear if MBCT is an effective therapy for current depression, and it is not clear what mechanisms are responsible for the effectiveness of MBCT. Theoretically, MBCT is believed to modify the processing of emotional information and reduce cognitive vulnerability to depression; however, it is not clear if MBCT leads to normalization of attentional biases in depressed individuals. The aim of the current study was to determine if MBCT can modify some of the attentional biases underlying depression in MDD. Participants were 53 individuals with diagnosis of current MDD. They were randomly assigned to either MBCT (n = 25) or wait list control group (n = 28) condition. Before and after the 8-week MBCT intervention participants completed the Center for Epidemiological Studies Depression scale (CESD), they viewed slides presenting sad, angry, happy, and neutral facial expressions, and their eye movements were recorded during the viewing task. As expected, compared to participants in the control group, the CESD scores of participants who received MBCT decreased following treatment, their attention to sad faces decreased, and their attention to happy faces increased. Moreover, cross-lagged analysis suggested a causal link from changes in attentional bias to changes in depression. We found that MBCT can modify the attentional processing of emotional facial stimuli and that attentional bias modification may translate into clinical improvement in currently depressed individuals. depression (Teasdale et al. 2000) with mindfulness training , and MBCT has been found to reduce relapse in patients with recurrent MDD, defined as 3 or more previous episodes (e.g., Godfrin and van Heeringen 2010;Ma and Teasdale 2004;Teasdale et al. 2000). For a similar conclusion, see meta-analyses by Piet and Hougaard (2011) and Kuyken et al. (2016). Based on this evidence, MBCT has been recommended as a prophylactic treatment for recurrent MDD (i.e., National Institute for Clinical Excellence 2009); however, it is not clear if MBCT is effective in treating current MDD. Interestingly, despite the fact that MBCT was developed to prevent the recurrence of depression, patients who were currently depressed were excluded from the first seminal studies (Teasdale et al. 2000;Ma and Teasdale 2004). Several researchers have raised concerns that depressive symptoms such as difficulty with concentration and intensity of negative thinking that may preclude the acquisition of attention regulation skills that are essential to MBCT (e.g., Segal et al. 2002). Regardless, a growing body of research has found that MBCT can reduce depressive symptoms per se (Finucane and Mercer 2006;Kenny and Williams 2007). Recently, several attempts have been made to use MBCT to treat the currently depressed (e.g., Van Aalderen et al. 2012), and this research suggests that MBCT could also be an effective treatment for current depression (Goldberg et al. 2018;Strauss et al. 2014). MBCT was designed to target the cognitive reactivity , e.g., the ease with which dysphoric mood can reactivate constellation of negative thoughts and feelings that may lead to a depressive episode. Several laboratory studies have found that cognitive reactivity is a marker for vulnerability to relapse and recurrence of depression . For example, Segal et al. (2006) found that recovered depressed individuals tended to decline into depressogenic thinking patterns following the induction of sad moods, and those who exhibited greater reactivation of dysfunctional thinking styles in response to the sad mood induction were at the greatest risk of relapse over the following 18 months. Importantly for the current study, research has also found that attentional bias for sad faces is associated with significantly greater impairments in mood recovery following mood induction among individuals with MDD than it is among nonclinical subjects (Clasen et al. 2013). These findings suggest that change in depressive attentional bias may be linked to mechanisms that are the focus of MBCT. Cognitive theories assume that attentional biases to negative emotional stimuli play important roles in the development and maintenance of depression (e.g., Clark and Beck 1999;Williams et al. 1988), and this assumption has been supported by a wealth of research (Armstrong and Olatunji 2012;Foland-Ross and Gotlib 2012). For example, eye tracking studies have consistently found that depression is associated with prolonged maintenance of gaze on negative information (Caseras et al. 2007;Eizenman et al. 2003;Kellough et al. 2009), and greater neglect of positive information (e.g., De Raedt and Koster 2010, see Peckham et al. 2010). Prolonged processing of negative information leads to more persistent sad mood (e.g., Disner et al. 2017) and has been found to be related to greater trait rumination (Donaldson et al. 2007;Owens and Gibb 2017) and increased daily rumination in depressed individuals (Holas et al. 2019). "Mindfulness is defined as awareness that emerges through paying attention on purpose, in the present moment, and nonjudgmentally to the unfolding of experience moment by moment" (Kabat-Zinn 2003, p. 145). Mindfulness thus involves awareness of the direction of attention and the ability to switch attention between thoughts, feelings, or bodily sensations and present stimuli (Bishop et al. 2004). Learning to stabilize attention through mindfulness exercises could be a way for people to recognize their cognitive reactivity and provide a basis for decentering and avoiding negative patterns of thinking (Williams and Kabat-Zinn 2011). In MBCT, this is achieved by instructing people to redirect their attention to the present moment, usually their own breathing. In turn, this can increase the voluntary deployment of attention . Research on MBCT has found that decreases in dysfunctional thinking (rumination and worry) mediated treatment outcomes and relationships between cognitive reactivity and relapse risk (for meta-analysis see Gu et al. 2015 andvan der Velden et al. 2015). Nevertheless, modifying attention to emotional stimuli as a therapeutic mechanism of MBCT as a treatment for current depression has not been examined that thoroughly (Alsubaie et al. 2017). Biases in attentional processing are a hallmark characteristic of depression and a vulnerability factor of depression (Joormann and Gotlib 2007;Koster et al. 2011), and so research on modifying attention within the confines of MBCT is clearly warranted. Some research has been done on the impact of mindfulness training on attentional processes. This research has focused on processing at the level of specific components of attention (alerting, orienting, and executive attention), general attentional processing following short-term (like MBCT), and long-term meditation training in healthy participants (e.g., Anderson et al. 2007;Chambers et al. 2008;Chan and Woollacott 2007;Jha et al. 2007;Moore and Malinowski 2009;Slagter et al. 2007; Van den Hurk et al. 2010). Only a few studies have examined the effects of MBCT on attention in clinically depressed patients (e.g., De Raedt et al. 2012;Van den Hurk et al. 2012;Verhoeven et al. 2014). In addition, recent research suggests that mindful attention can lead to decentering by disengaging the self from imagined situations so that negative affective reactions do not develop (Lebois et al. 2015). To our knowledge, only two published studies have evaluated changes in attentional biases to emotional information following MBCT (De Raedt et al. 2012;Verhoeven et al. 2014). These studies examined attentional processes in people who were formerly depressed but were in remission, and they measured reaction times manually. The majority of research on attentional biases in emotional disorders has measured attention using reaction time (RT) (e.g., Bar-Haim et al. 2007 for meta-analysis). Although studies using RT to measure attention are informative, as discussed by Armstrong and Olatunji (2012), RTbased measures of attention are limited in potentially significant ways. For example, when measuring attention using RT, there is a possible confound between the effects of a stimulus and the time it takes to press a key to record a response. Another limitation is that RT is based on measures of individual points in time ("a snapshot"), which makes it problematic to capture the dynamics of attention. Eye tracking is an alternative to RT that measures attention dynamically and provides accurate measures for assessing the time course of visual attention with no delay between attending to a stimulus and the measurement of attention. Because of these advantages, the use of eye-tracking to measure attention in studies of relationships between attentional processes and emotions has increased in recent years (e.g., Armstrong and Olatunji 2012). Given these advantages, in the present research, we measured attentional processes using eye-tracking. The aims of the current study were to determine if MBCT can reduce depressive symptoms and if it can modify attentional biases to negative information underlying depression in currently depressed individuals. In addition, we wanted to examine possible causal relationships between the severity of depressive symptoms severity and attentional biases. We expected that MBCT would reduce depressive symptoms in MDD individuals and lead to normalization of attentional biases. Specifically, we expected a reduced attention (gaze duration) to sad and angry faces, and increased gaze duration to happy faces after MBCT, and no such effects in our control group. Furthermore, we hypothesized that modification of attentional bias to negative faces precludes decrease in depressive scores. We expected therefore, that gaze duration at pre-test would be related to CESD scores at post-test but that CESD scores at pre-test would not be related to gaze duration at post-test. Participants Participants were recruited via open calls posted on popular Internet portals. The calls invited people to participate in a study that involved MBCT training for 8 weeks and concerned how people feel and how they process information in their daily lives. A total of 692 people completed an online screening that consisted of background data and the Center for Epidemiological Studies Depression scale (CESD, Radloff 1977). A total of 249 individuals were excluded on the basis following criteria: their CESD scores were below 20 (n = 66), or that they were younger than 25 (n = 150) or that they were taking psychiatric medications (n = 73) (some of them had multiply criteria). We then phone interviewed 273 people in order to qualify them in terms of 2 major criteria for depression (sadness or anhedonia for at least 2 weeks) to establish higher likelihood of meeting DSM criteria for current major depressive episode. Based on the phone interview, 154 people were excluded if they were currently receiving psychotherapy (n = 14); if they were suffering from a current or lifetime psychotic disorder, bipolar disorder, substance abuse problem; if they reported current suicidal tendencies (n = 15); or if they did not meet criteria for major depression (n = 63). Sixty-two were not interested in participation. We qualified 119 people who were then invited for a clinical interview in the laboratory. A total of 84 people came to a structured clinical interview (the Mini-International Neuropsychiatric Interview (MINI) version 5.0.0; Sheehan et al. 1998) conducted by a trained clinician. Three people were excluded from the study because they had comorbid diagnoses based on the MINI. Details how the subjects flowed through the study are presented in the CONSORT flow chart (Fig. 1). The final sample consisted of 81 participants meeting DSM-IV criteria for MDD. In terms

of ethnicity, all participants were white, which was not unusual given the demographics of the country in which the study was conducted. Participants were assigned to either MBCT group or wait list control condition. Twenty-four participants dropped out of the study after the pretesting. The final sample consisted of 25 participants in the MBCT group (20 Female, M age = 34.96, SD = 8.01) and 28 in control condition (19 Female, M age = 34.25, SD = 8.75). The groups did not differ in terms of mean age and gender (ps = .76, .32, respectively). Summary statistics of previous episodes of depression and of comorbidity are presented in Table 1. Procedure The study consisted of the following components: a laboratory session during which participants performed the eye-tracking task, followed by the maintenance of an online diary for 7 consecutive days (these data are not relevant to the current hypotheses and are not discussed in this paper), 8 weeks of MBCT training for experimental group or a waiting period for the control group, the maintenance of an online diary for 7 consecutive days (these data are not relevant to the current hypotheses and are not discussed in this paper), and then followed by a laboratory session during which post-training measures were collected, including the eye tracking task and the CESD. Eye Tracking Task Four-face task. In this task, based on Leyman et al. (2011), a series of 12 slides with four facial expressions (neutral, sad, angry, and happy) was shown to participants. We used expressions of 6 female and 6 male faces taken from the Karolinska Directed Emotional Faces database (Goeleven et al. 2008). The four facial expressions of the same model were simultaneously presented for 10 s in one of 4 quadrants of the slide on a black background (see Fig. 2). The positions of expressions within each slide were randomized and fixed for all participants, and the order of the slides was random for each participant. Each trial began with a fixation cross in the middle of the screen for 1000 ms. Participants were told to look at each of the slides. Intervention We used a MBCT protocol developed by Segal et al. (2002). It consisted of 8 weekly sessions of 2.5 h each, with a 5-h day-long retreat between the 6th and 7th sessions. The MBCT training was provided by two trainers with extensive experience in meditation and MBCT. One was a psychiatrist and cognitive behavioral therapist with 8 years of experience guiding MBCTand MBSR groups. The second was a psychologist with 8 years of mindfulness-based intervention experience. Training groups consisted of up to 16 participants. Measures Eye Tracking Presentation of the stimuli was controlled by SMI Experiment Center software with an eye-tracking device. The stimuli were presented on a 19′ monitor at a resolution of drawn around each face in each trial. For each of the face, the total gaze duration (GD), the dwell time spent fixating on the face, was calculated. Self-Report of Depressive Symptoms Before and after training, participants completed the Center for Epidemiological Studies Depression Scale (Radloff 1977). The CESD is a 20-item scale measuring symptoms of depression. Using a 4-point scale, with end points anchored 0 (rarely or none of the time, less than 1 day) and 3 (most of the time, 5-7 days) participants reported how often they experienced each of the symptoms during the previous week. The reliabilities for the scale were satisfactory at both pre-and post-testing (αs = .82 and .94 respectively). Copies of all measures and the raw data are available at: https://osf.io/dq64j/ Data Analysis Recording eye movements during the four-face task allowed us to examine if MBCT training modified participants' attention to positive and negative faces. Our general expectation was that MBCT training would lead to a reduction of gaze toward negative faces and an increase in gaze duration to positive faces. We carried out a series of 2 × 2 × 4 mixed ANOVAs with training (MBCT vs. control) as a betweensubject factor, testing session (pre-training vs. post-training) and face type (angry vs. happy vs. neutral vs. sad) as withinsubject factors. We also expected to find significant reduction in depression symptoms. We examined such changes using a 2 × 2 ANOVA with training (MBCT vs. control) as a between-subject factor and testing session (pre-training vs. post-training) as a withinsubject factor. When appropriate, post-hoc tests of differences were performed using Bonferroni correction, and the results are reported with Greenhouse-Geisser correction whenever the sphericity assumption was violated. Finally, we tested cross-lagged relationships between gaze duration and depressive symptoms. We examined if relative gaze duration at pre-training session predicted depressive symptoms measured at time post-training and if depressive symptoms at pre-training predicted relative gaze duration at the post-training. If one such lagged relationship is significant and the other is not, this suggests that one construct is a cause and the other is an effect. We expected that gaze duration at pre-test would be related to CESD scores at post-test but that CESD scores at pre-test would not be related to gaze duration at post-test. Eye-Tracking Measures of Focus on Emotional Facial Expressions We measured visual attention using gaze duration (i.e., how long, in total, a person dwelled on a face) separately for angry, happy, neutral, and sad faces. Before the analyses, we examined distributions of raw gaze durations. As a result, we excluded the data from five participants from the analyses. The mean gaze durations of two participants from the MBCT group and two participants from the control group were 3SD or more above their group's mean, and the mean gaze duration of one participant from the control group was 3SD below the group mean. Similar to Owens and Gibb (2017) and Krejtz et al. (2018), we measured attentional bias by calculating gaze duration for specific facial expressions controlling for total gaze duration for all facial expressions. We used the relative score not an absolute gaze duration because it is a more direct measure of an attentional bias reflecting the difference between processing of happy, sad, angry, and neutral faces, and it is currently the standard procedure in studying attentional biases. We calculated relative scores for each of the facial expressions using a standard formula (e.g., Everaert et al. 2014;Owens and Gibb 2017). The total gaze duration for each type of face (target face in the formula below) was divided by the total viewing time for all facial expressions. GD relative ¼ Gaze Duration target face GD positive þ GD angry þ GD neutral þ GD sad This adjusted the gaze duration for a specific type of face for total of gaze durations on all four types of faces. A score of .25 indicates equal gaze distribution between the target face and other faces, scores above .25 indicate a preference for the target face (longer total viewing time), and scores below .25 indicate less attention toward the target face relative to other facial expressions. Relative Gaze Duration Descriptive statistics for gaze duration and relative gaze are presented in Table 2. We should note that analyses of absolute gaze duration produced results that were functionally equivalent to analyses of relative gaze duration. We present the results of analyses of relative gaze duration because the relative measure is reported more often than the absolute measure. The analyses of relative gaze duration produced a significant interaction effect among training, testing session, and face type (F (3, 138) = 4.02, p = .009, ɳ p 2 = .08). This interaction can be understood in two ways (see Table 2). First, the interaction can be interpreted as a result of the fact that for participants in the MBCT group the relative gaze time at happy faces significantly increased from pre-to post-training (F (1, 46) = 13.36, p < .001, ɳ p 2 = .225), whereas the relative time spent looking at sad and angry faces significantly decreased from pre-to post-training (F (1, 46) = 4.79, p = .034, ɳ p 2 = .094; F (1, 46) = 7.35 p = .009, ɳ p 2 = .138, respectively). In contrast, for the control group, there were no significant differences between pre-and post-training sessions for these measures (ps > .5). Other Effects These analyses also produced other significant effects (e.g., a main effect for face types). Given that these effects did neither test nor were they relevant to any of our hypotheses and many of the effects were qualified by the triple interaction, these effects are not described. They are presented in the online, supplemental materials. Cross-Lagged Analyses We examined possible causal relationships between depressive symptoms (CESD scores) and gaze duration using cross-lagged panel analyses. In these analyses, CESD scores at post-test were regressed on gaze duration at pre-test, while simultaneously, gaze duration at post-test was regressed on CESD scores at pre-test. The logic of these analyses is that temporal precedence is a necessary (although not sufficient) condition for causes: Causes precede effects. A causal link from gaze duration to depression would be suggested by a significant relationship between duration at pre-test and CESD scores at post-test, whereas a causal link from depression to gaze duration would be suggested by a significant relationship between CESD scores at pre-test and duration at post-test. These analyses were limited to participants in the MBCT condition, and each measure of gaze duration was analyzed separately. The analyses were conducted using MPlus (Muthén andMuthén 1998-2012), and the model used for these analyses is presented in Fig. 3. As can be seen from this figure, each post-test measure was regressed on both pre-test measures, and the error covariances between the two measures at pre-test and at post-test were modeled. The lagged coefficients from these analyses are summarized in Table 3. As can be seen from these coefficients, the lags from duration of gaze on sad faces to CED scores and from duration of gaze on happy faces to CESD scores were significant. In contrast, neither of the lags from CESD scores to gaze duration on sad faces and on happy faces was significant. Such a pattern of results is consistent with a casual sequence in which changes in attentional focus lead to CESD T1 CESD T2 Fig. 3 The model used for cross-lagged analyses changes in depressive symptoms. There was no evidence for a causal sequence in which changes in depressive symptoms lead to changes in attentional focus. Discussion Attentional biases are believed to play a key role in the onset, maintenance, and recurrence of depression (e.g., De Raedt and Koster 2010). Research has documented that attentional biases in depression consist of greater attention to negative information such as sad faces and a reduction in attention to positive stimuli such as happy faces (Armstrong and Olatunji 2012). In the current study, we evaluated the effectiveness of an 8-week MBCT training on modification of attentional biases to emotional faces in a sample of currently depressed individuals. As expected, we found a significant decrease in CESD scores in the MBCT group, whereas there were no changes in the control group. This finding is consistent with the results of several studies that have shown that individuals who are currently depressed can benefit from MBCT (e.g., Van Aalderen et al. 2012), and it is consistent with the conclusions of meta-analyses showing the effectiveness of mindfulnessbased interventions (MBIs) in treating current depression (Goldberg et al. 2018;Strauss et al. 2014). Further, as expected, we found reduced gaze duration to negative faces, measured by a decrease in time spent viewing sad and angry faces for the MBCT group, whereas there were no such changes in the control group. Moreover, relative gaze time at happy faces significantly increased from pre-to post-training in MBCT group, and following MBCT individuals spent more time looking at happy faces than at angry and neutral faces. These findings are consistent with how mindfulness and MBCT have been conceptualized. Mindfulness training teaches people to rely less on preconceived ideas, beliefs, and biases and to rely more on paying attention to all available information in the present moment (Bishop 2002). The enhancement of selective attention to attend to only one particular component of the present moment (such as breathing sensations) and the enhancement of sustained attention have been proposed as central practices in MBCT (Teasdale et al. 1995). They have also been proposed as essential elements

of why mindfulness training changes people's thinking and increases their well-being (Holas and Jankowski 2013). MBCT emphasizes attentional training and a focus on participants' moment-to-moment experiences irrespective of affective valence or the presence of physical discomfort (Kocovski et al. 2015). This may help people to develop the ability to decenter from mental events such as negative thoughts or memories . Surprisingly, very little work has been done to determine if MBCT leads to the modification of attentional biases to dysphoric information in current depression. The only published study of which we are aware that examined the effects of mindfulness training on attention to emotional faces (De Raedt et al. 2012) found that after MBCT, previously depressed individuals exhibited reduced facilitation of attention for negative faces. These findings were interpreted as indicating of an open attention toward all emotional information (De Raedt et al. 2012). The results of the present study suggest more specific changes than those discussed by De Raedt et al., i.e., a general decrease in attention to negative faces and an increase to positive faces following MBCT. Our findings are consistent with the results of Verhoeven et al. (2014) who found that response times on an emotional Stroop task decreased after MBCT in the previously depressed. Such results suggest information related to depression may attract less attention following MBCT. In addition to finding that MBCT led to changes in means of gaze duration and depressive symptoms among participants who received MBCT, we also found lagged relationships from gaze duration to depressive symptoms, whereas there were no significant lagged relationships from symptoms to gaze duration. Such a pattern is consistent with the assumption that changes in attentional focus lead to changes in depressive symptoms, or more broadly, with the assumption that attentional focus is more of a cause of depression (as posited by numerous theories and models, (e.g., Clark and Beck 1999;Koster et al. 2011) rather than a result of it. Attentional improvement has been suggested as an important reason why MBCT is effective in depression in addition to change in mindfulness, rumination, worry, and selfcompassion which have been more frequently investigated for mediation effects (Alsubaie et al. 2017;van der Velden et al. 2015). Attentional improvement posits that improved abilities to regulate attention and to disengage from depressogenic cognition translate into better treatment outcomes (van der Velden et al. 2015). Within the context of Limitations and Future Directions Although the present study generally met our expectations, there were some limitations. First, there was no follow-up assessment, so we cannot know if the changes we found were maintained over time. In addition, we examined a limited set of emotions. Although happiness and sadness are central to the experience of depression, it would be useful to study attention to other emotions such as fear. Finally, the causal relationship between attentional biases and symptoms should be examined in experimental settings that provide a better basis for drawing casual inferences and with larger samples that would provide more statistical power than we had. Although MBCT has been shown to be effective, an important limitation of the majority of studies, including the present study, is the absence of an active control condition (ACC). The results of studies comparing MBCT to ACCs provide mixed support for the effectiveness of MBCT in terms of depressive relapse and related symptoms (e.g., Williams et al. 2014;Shallcross et al. 2015). In addition, several authors, including founders of MBCT , have warned that certain characteristics of depression, such as aversive content of depressive thoughts and feelings and ruminative processing, may make it difficult for depressed people to engaging in and benefit from mindfulness based interventions. Clearly, more randomized control trials with ACCs are needed to draw firmer conclusion about the efficacy of MBCT to treat current depression. Current models of depression suggest that cognitive biases reflect regulatory deficits in attentional control (e.g., Joormann 2010; Koster et al. 2011), the ability to exert topdown control of attention. Future studies are needed to determine if MBCT not only modifies processing tendencies but also modifies dysfunctions in attentional control that may be responsible for negative interpretations of emotional information. Author Contributions PH: designed and executed the study and was primarily responsible for writing the paper. IK: helped with the study design, conducted data analyses, and helped write the paper. KW: assisted with the data analyses and helped write the paper. MR: helped with the study design and data collection. JN: helped with the study design, data analysis, and writing of the paper. All authors approved the final version of the manuscript for submission. Funding Information Support for this research was provided by the National Science Center, Harmonia Grant: 2012/04/M/HS6/00/470 to the 2nd author and Internal Grant (BST, no 0177700-41) of Psychology Faculty University of Warsaw to the 1st author. Compliance with Ethical Standards Conflict of Interest The authors declare that they have no conflict of interest. Ethics Statement This study was approved by the institutional review board for research involving human subjects of SWPS University of Social Sciences and Humanities. Informed Consent Statement Informed consent was obtained from all participants. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Ballistic Tests of Alumina-UHMWPE Composites Submitted to Gamma Radiation The energy absorption in ballistic tests of alumina-ultra high molecular weight polyethylene (UHMWPE) composites with 60, 80 and 90 wt% alumina submitted to gamma radiation doses of 25, 50 and 75 kGy was investigated. The ballistic tests were carried out at subsonic speed using a compressed air system. The results showed that the composite with 80% alumina irradiated with 50 kGy yields the best ballistic results. Gel content, Differential scanning calorimetry (DSC) and X-ray diffraction (XRD) results showed that this composite is the one with the highest concentration of crosslinks and the lowest volume fraction of amorphous UHMWPE. Scanning electron microscopy (SEM) images of the same composite showed a high pullout, suggesting that gamma irradiation increases the adhesion between alumina and UHMWPE. Introduction In the beginning of the 21st century, despite the absence of international conflicts on the scale of the two World Wars, local and regional conflicts involving different tribes, ethnic groups, militias, gangs and heavily armed drug dealers have become a serious threat in several parts of the world. Standard bulletproof personal protection vests use aramid fabric as a single layer of shielding. This protection is limited to relatively low impacts, up to 9 mm ammunition. Protection against high impact projectiles requires a multiagent system (MAS) 1 . A conventional MAS has, beside the aramid fabric, a ceramic front layer which absorbs most of the impact energy, eroding the projectile tip. However, such protection increases the cost and compromises the soldier mobility due a significant increase of vest weight. In addition, the front layer may be fragmented by the first impact, compromising its resistance to subsequent shots. The main ceramic materials used for ballistic protection are alumina (Al 2 O 3 ), silicon carbide (SiC) and boron carbide (B 4 C). Alumina has been suggested for ballistic protection due to good physical and chemical properties. However, the low flexural strength and low fracture toughness mean that the use of pure alumina for ballistic protection may lead to catastrophic failure. Moreover, the high density, about 4 g/cm 3 , limits its use in applications where weight is crucial, such as bulletproof vests 2,3 . According to Figueiredo et al. 4 , alumina-UHMWPE composites may yield a good compromise between high energy absorption and low density, the only problem being low adhesion between the alumina particles and the polymer matrix 5 . The adhesion may be increased by exposure of the composites to gamma radiation 6,7 , that has other favorable effects, such as increased stiffness and stability, as reported by Shafiq. et al. 8 . On the other hand, according to Hobbs et al. 9 , high radiation doses generate microcracks that weaken the alumina component. The purpose of this work was to investigate the properties of gamma irradiated alumina-UHMWPE in order to determine the best combination of alumina concentration and radiation dose for ballistic protection applications. The UHMWPE is used to decrease the density and increase the flexural strength, making the shield more suitable for personal protection and avoiding fracture after the first shot 10,11 . There is also an economic factor involved, since the composite is prepared at a relatively low temperature, 230 o C, while pure alumina must be prepared sintering alumina powder at high temperatures, of the order of 1400 o C, a more expensive procedure 12-15 . Gamma Irradiation of the composites was performed using a Gammacell 220 Excel irradiator with a Co-60 source. Ballistic tests For the ballistic tests, an air rifle Gunpower SSS was used with a noise suppressor Padrão Armas. The projectile was a 22 gauge lead shot with an estimated mass of 3.3 g. An Air Chrony ballistic chronograph model MK3, with a precision of 0.15 m/s, was used to measure the impact speed, and a ProChrono ballistic chronograph model Pal, with a precision of 0.31 m/s, was used to measure the residual speed. The air rifle was positioned 5 m away from the target, consisting of the composite disk attached in an aluminum frame, secured by a bench vise and aligned perpendicularly to the rifle. One ballistic chronograph was positioned 10 cm from the nozzle of the noise suppressor and another was placed 10 cm behind the target. The energy absorbed by the target was calculated using the equation where m p is the projectile mass, v i is the impact speed and v r is the residual speed 16 . Taking into account the fact that in the case of personal protection weight may be an important factor, one may define a figure of merit given by the following equation: where m c is the composite mass. Gel content measurements The gel content shows the amount of crosslinking in the samples. It was determined according to ASTM D2765 using a Soxhlet device 15 . The samples were extracted using boiling xylene for 6 h and were subsequently washed with acetone and dried at 140 o C. The percent gel content was determined using the following expression: where W o and W 1 are, respectively, the weight of sample before and after extraction. Differential scanning calorimetry Differential scanning calorimetry (DSC) analysis was performed with a Netzch DSC 404 F1 Pegasus calorimeter. The measurements were carried out with a heating and cooling rate of 10 o C min -1 under nitrogen (50 mL min -1 ). The samples were heated from room temperature to 180 o C and cooled to 50 o C and then reheated to 180 o C. The melting temperature and the percent crystallinity was determined only for the first run. The percent crystallinity Xc was calculated using the following expression: where ΔH o and ΔH m are, respectively, the enthalpy of melting of crystalline polyethylene (290 Jg -1 ) and the enthalpy of melting of the sample. X-ray diffraction The sintered samples were characterized by X-ray Diffraction (XRD) in an X'PERT PRO PANalytical diffractometer, with monochromatic radiation (Cu Kα, λ = 1.5406 ˚A), step size of 0.05 o s -1 , time per step 150 s and 2θ between 10 o and 90 o . The XRD patterns were refined using the Rietveld method, with the help of the TOPAS Academic version 4.1 software. SEM images After the ballistic tests, images of composites with 80% alumina were taken in a FEI Quanta FEG 250 SEM. Ballistic tests All shots completely penetrated the disks. Two shots were made in each

experiment and eight experiments were performed for each composition. Representative samples are shown in Figure 1 after the first shot. No significant decrease in energy absorption was observed after the first shot. Since the test was performed with the samples in the cast aluminum holders, lateral movement was suppressed, reducing the damage to the composite 17,18 . Table 1 shows the average values of the composite mass (m c ). projectile mass (m p ), average impact speed (v i ), average residual speed (v r ), absorption energy (E abs ) and merit factor (FM) for each composition. As expected, m c increases with increasing alumina concentration. Figure 2 shows the FM of composites with different alumina concentrations and radiation doses. One can see that the best compromise between high energy absorption and low weight was displayed by the composite with 80% alumina irradiated with 50 kGy. Table 2 shows that the gel content of UHMWPE, nonirradiated and gamma irradiated with doses of 25 kGy, 50 kGy and 75 kGy. The highest gel content was for the samples irradiated with a dose of 50 kGy, suggesting the concentration of crosslinks is maximum for this radiation dose. The decrease of gel content for a higher radiation dose is attributed to degradation of UHMWPE, forming free radicals. DSC The DSC analyses were performed during the first fusion and crystallization. Table 3 shows that the samples irradiated with 50 kGy had the highest values of crystallinity, and among them the largest value was that of the sample 80/50. X-ray diffraction The samples with 80% alumina were investigated further due to the good ballistic performance displayed by the sample with 80% alumina irradiated with 50 kGy. The purpose was to determine why the ballistic performance increased with radiation dose up to a radiation dose of 50 kGy, but decreased when the radiation dose was increased to 75 kGy. Figure 3 shows the XRD patterns of samples A80/00, A80/25, A80/50 and A80/75. Except for a broad line at small angles attributed to amorphous UHMWPE, only the diffraction lines of α-alumina and UHMWPE are seen, showing that there was no phase transformation due to gamma irradiation. In Figure 4, one can see that the amplitude of the broad line goes through a minimum for a radiation dose of 50 kGy. This suggests that the crystallinity of UHMWPE increases with increasing radiation dose up to a radiation dose of 50 kGy, but decreases for larger radiation doses. several cracks whose interaction creates subsequently a conical damage zone with ductile and fragile failure, similar to that observed in ceramic tiles 19 . Figure 7 show the conical damage zone in the distal face of an A80/50 sample. Figure 8 shows microcracking on the distal faces of A80 and A80/50 samples after the first impact. The main difference between Figures 8a and 8b is that there is evidence for pullout in the irradiated samples as a consequence of increased adhesion between alumina and UHMWPE, as also shown in Figure 9. Figure 4. In the case of the UHMWPE line, there is a significant displacement to the left for the sample submitted to a radiation dose of 50 kGy. This is attributed to a relaxation due to increasing crystallinity and is consistent with the decrease in the amplitude of the broad line in Figure 4. In the case of the alumina lines, there is a shift to the right in the samples irradiated with 25 and 75 kGy. In the first case, the compression may be due to loss of water and in the second to radiation-induced microckacks. Images of alumina powder, sample A80/00 and sample A80/50 Figure 6 is an SEM Image of the alumina powder, showing the irregular shape of the grains, which improves the stiffness of the composites. When impact occurs, compressive shock waves propagate along the thickness of the sample and are responsible for Conclusions Ballistic tests were performed on gamma-irradiated alumina-UHMWPE composites. The composite with 80% alumina submitted to a radiation dose of 50 kGy was the one that had the best ballistic properties. Gel content, DSC, XRD and SEM results suggest that this is due to three effects of gamma radiation: an increased extent of crosslinking, a decrease of the volume fraction of amorphous UHMWPE and an increased adhesion between alumina and UHMWPE. The results also showed that increasing the radiation dose above 50 kGy has adverse effects, probably due to the production of a large number of microcracks that leads to a weakening of the alumina particles and to UHMWPE degradation, generating free radicals. Complete Remission of Lupus Nephritis Following Chemoradiotherapy of Carcinoma Cervix: An Association Case Report Lupus nephritis (LN) is a serious complication of systemic lupus erythematosus (SLE). Despite modern induction therapy of proven efficacy, LN can still be associated with treatment failure. SLE is associated with a higher incidence of both hematological and solid organ malignancies, including cervical carcinoma which creates a paradox on how malignancies must be addressed therapeutically in the context of autoimmunity.[1,2] Introduction Lupus nephritis (LN) is a serious complication of systemic lupus erythematosus (SLE). Despite modern induction therapy of proven efficacy, LN can still be associated with treatment failure. SLE is associated with a higher incidence of both hematological and solid organ malignancies, including cervical carcinoma which creates a paradox on how malignancies must be addressed therapeutically in the context of autoimmunity. [1,2] Here we report an unexpected complete remission (CR) of a class IV LN following chemoradiotherapy of cervical carcinoma allowing the long-term discontinuation of immunosuppressive medications. This is an open access journal, and articles are distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 License, which allows others to remix, tweak, and build upon the work non-commercially, as long as appropriate credit is given and the new creations are licensed under the identical terms. Her Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI-2K) score was 22. She was initiated on hemodialysis and three intravenous (IV) methylprednisolone pulses of 1 g each followed by oral steroids of 1 mg/kg were given. Following normalization of platelet counts, renal biopsy was performed. On light microscopy, there were 20 glomeruli, of which one was globally sclerosed. Four glomeruli showed cellular crescents, four fibrocellular crescents, and four fibrous crescents. Underlying tufts showed segmental variable mesangial hypercellularity and matrix expansion. Five glomeruli showed segmental endocapillary proliferation. Tubules showed patchy interstitial fibrosis and tubular atrophy of 30%. Patchy acute tubular injury, interstitial edema, patchy tubulitis, and mild peritubular capillary dilatation and margination were also seen. Blood vessels showed hypertensive changes. Direct immunofluorescence showed granular C1q 2+ positivity in mesangial area and capillary loops. A diagnosis of class IV LN according to ISN/RPS classification was made with a modified NIH activity score of 8/24 and chronicity score of 6/12 [ Figure 1]. Treatment and follow-up After initial IV methylprednisolone pulses followed by oral steroids of 1 mg/kg, she was started on high dose monthly IV cyclophosphamide (IV CYC; 1000 mg/m 2 ), along with hydroxychloroquine 300 mg and angiotensin receptor blockers. At the end of 6 months, she achieved a partial remission with a S. Cr of 1.9 mg/dL, proteinuria of 890 mg/24 hours, S. alb of 3.5 g/dL. Maintenance therapy with mycophenolate mofetil (MMF) 1 g in two divided doses and low dose steroids (wysolone 5 mg) was started. At the 8 th month, she had a proteinuric relapse with a proteinuria of 1.7 g/24 hours, S. alb of 3.3 g/dL, and a S. Crt of 1.4 mg/dL for which MMF dose was increased to 1 g twice daily and wysolone was continued at 10 mg per day. Patient continued to have proteinuria. At 18 months, she had a S. Cr of 1.2 mg/dL, S. alb of 3.2 g/dL, and proteinuria of 1.2 g/24 hours; was planned to start on tacrolimus in addition to MMF and steroids. Meanwhile, she developed bleed per vagina for which she underwent a cervical biopsy and radiological evaluation: she was diagnosed with squamous cell carcinoma of cervix stage IIB. MMF was withheld in view of the underlying carcinoma and low-dose-steroids (5 mg) were continued. She received 4 cycles of radiotherapy @ 46 Gy/23 fractions for 4.5 weeks with concurrent chemotherapy with a radio-sensitizing dose of cisplatin 40 mg/m 2 *4 weeks followed by 2 cycles of Intracavitatory brachytherapy of 9 Gy/fraction. After two months of concurrent chemotherapy and radiotherapy, the patient went into CR with a proteinuria <500 mg/24 hours, S. Cr 1.04 mg/dL, S alb 4.2 g/dL, negative anti-DsDNA, and normal complements. The patient was continued on low dose wysolone (5 mg) along with hydroxychloroquine and angiotensin receptor blockers. She is currently 5 years post-diagnosis and continues to be in CR with low dose steroids with a SLEDAI-2K score of 0, normal complements, negative anti-ds DNA and also has no active disease of carcinoma cervix. The complete clinical course is depicted in Figure 2. Discussion Here we report a patient with SLE with Class IV LN who received induction with NIH regimen followed by maintenance with MMF. She developed a proteinuric relapse for which she re-induced with MMF to which she failed to achieve remission. Post-diagnosis of carcinoma cervix IIB, she received concurrent chemoradiotherapy following which she went into CR of LN despite being on just low dose steroid. SLE patients have increased risk for both hematological malignancies like non-Hodgkin's lymphoma, and solid tumors in lung, liver, vulvar/vaginal, and thyroid. [1] SLE patients with long-term use of immunosuppression have also been observed to have a higher prevalence of low-grade and high-grade cervical intraepithelial lesions (CIN) in comparison with those without long-term use of these agents; one series reported an overall 3-year incidence of CIN of 25% in IV CYC treated patients. [2] A dose relationship was observed between cumulative IV CYC exposure and CIN; every 1 g corresponded to a 13% increased risk of CIN. The cumulative dose of IV CYC received in our patient was 6 g which is less likely to have increased risk of carcinoma cervix. Initially, the patient did not go into CR despite being on adequate immunosuppression. However, subsequently with the chemoradiotherapy for carcinoma cervix, she achieved a CR that was sustained for the next 5 years. We hypothesize that the remission of the lupus activity can be explained by the effect of either radiotherapy or cisplatin-based chemotherapy, or a combination of both. f a e SLE patients have lower serum TGF-β1 levels than healthy individuals which are associated with increased disease severity, especially renal damage. [3,4] TGF-β1 is an important negative regulator of B cell differentiation and proliferation, and it also inhibits cytotoxic T lymphocyte, while promoting peripheral T Regulatory cells. [5] Hence, TGF-β1 is a powerful immunosuppressive cytokine that plays a dual role during the development and progression of immune-mediated inflammatory diseases, including SLE. [5] In fact, it is shown that addition of TGF-β1 and IL2 to peripheral blood mononuclear cells from SLE patients reverses the upregulated IgG production. [6,7] Radiotherapy is known to increase TGF-β1 levels. [7] For cancer patients with SLE, the release of TGF-β1 as a result of radiation therapy could be beneficial to the underlying SLE. [8] However, there have been no reports in the literature of remission of SLE with radiotherapy. Furthermore, there has been a concern of use of radiotherapy in SLE patients as these patients have increased risk of developing acute and chronic toxicities albeit moderate, presumably due to increased radiosensitivity leading to the additive damage to the microvasculature. [9] Cisplatin is an alkylating agent that directly damages DNA thereby disrupting its replication and transcription, inducing cancer cell apoptosis. In addition, cisplatin is seen to have immunological effects as well. Platinum compounds can modify the immune response during both the induction and the effector phase leading to immunogenic cell death. [10] We hypothesize that the use of cisplatin could have contributed to the remission. Topoisomerase I inhibitor irinotecan was shown to reverse LN and prolonged survival of NZB/WF1 mice by inducing single-stranded DNA breaks, inhibiting renal cell apoptosis, and preventing subendothelial deposits of IgG. [11] A study in murine lupus model also found that paclitaxel, an antimicrotubule agent, was beneficial in the suppression of autoimmunity by reducing the anti-dsDNA antibody titer and prolonging survival. [12] Animal experiments indicated that murine LN treated with fludarabine, a fluorinated purine

analog that inhibits DNA synthesis had a more significant reduction in renal pathology compared with LN treated with CYC. [13] A patient with chronic lymphocytic leukemia and steroid-resistant SLE achieved durable remission after administrating fludaribine. [14] There have been case reports of long-term remission of SLE activity and disappearance of clinical symptoms following removal of coexisting ovarian dysgerminoma along with post-operative melphalan for six weeks. [15] Thus, chemotherapy with cisplatin may have played a role in controlling SLE activity, though it may be difficult to explain a sustained remission with just low dose steroids in our case. In conclusion, we report an unexpected association between CR of a class IV LN and chemoradiotherapy of cervical carcinoma allowing the discontinuation of immunosuppressive medications, the effect of which could be attributed to either the radiotherapy or cisplatin, or a combination of both. 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. Nutritional Status of Lusitano Broodmares on Extensive Feeding Systems: Body Condition, Live Weight and Metabolic Indicators The present research aimed to evaluate the effects of foaling season and feeding management in extensive systems on the nutritional status of Lusitano broodmares throughout the gestation/lactation cycle, by assessment of body condition (BC), body weight (BW), and some blood metabolic indicators. Four groups of Lusitano broodmares (A, B, C, D) were monitored during four years, in a total of 119 gestation/lactation cycles. All mares were kept on pasture, and A and B mares were daily supplemented. Monthly, mares were weighed and BC evaluated. Suckling foals from these mares were also monitored for BW and withers height. Glucose, non-esterified fatty acids, urea and albumin concentrations were determined in blood. BW changes were influenced by reproductive stage and foaling season (P<0.001), reflecting also pasture availability. Changes on BC were observed (P<0.05), although with small amplitudes within each group. Higher scores were reached at the end of spring, decreasing 0.25 point until late summer. Early foaling had also a marked effect, hindering the recovery of BC along the cycle. Glucose values decreased from late gestation to early lactation (P<0.05) and lower levels were recorded during the summer months. Uremia was mainly influenced by the reproductive stage (P<0.05). Under nutrition was not detected. Foals born in February-March had higher average daily gain than those born in April-May (P<0.05), probably reflecting differences in milk production of the mares. BC and BW changes and, particularly, blood indicators showed an overall balanced nutritional status, reflecting an adaptation to feed availability and climate. Introduction As in other livestock females, and particularly in pasture based systems, mares store and mobilize body reserves during their reproductive cycle according to breeds and goal of production: draft and leisure breeds vs race and sport breeds (Bergero et al., 2006). This mobilization can be observed during the winter, at the end of pregnancy and sometimes in early lactation periods. In these periods, nutritional requirements may exceed nutrients provided by the diet, especially when low quality forages are used. When mares are fed limited energy, the mobilization of adipose tissue is also in part directed towards the production of the lipid fraction of mares' milk (Doreau et al., 1993;. But mares can restore body reserves shortly as far as feed allowances are high after foaling without any detrimental effects for its reproductive efficiency and their foals (Guillaume et al., 2006;. The most common method reported for the evaluation of body reserves, mainly regarding body fat tissue, is the body condition scoring (BCS). This method has been designed for horses (Henneke et al., 1983;INRA, 1990;Arnaud et al., 1997). In field conditions, BCS is a cheap, practical and accurate tool for monitoring energy balance, reflecting the abundance or shortage of nutrients on the animal recent past (Caldeira et al., 2007b;Martin-Rosset et al., 2008). In addition, BC evaluation on a regular basis allow to determine the pattern of changes (deposition or mobilization of reserves), which is fundamental for diets adjustment to different requirements in each productive phase. This method could be complemented by the assessment of some metabolites in body fluids, allowing for an early detection of nutritional unbalances. In general, serum or plasma concentrations of glucose and non-esterified fatty acids (NEFA) could provide information about the energy status of the animal, while albumin and urea are good indicators of protein status in mares (Doreau et al., 1981) as in other farm animals (Caldeira et al., 2007a). Throughout the gestation-lactation cycle, and considering different breeds, the weight of the broodmare may vary between 13% to 20%, reaching the highest and lowest values before and after foaling, respectively Lawrence et al., 1992;Cassil et al., 2009). These body weight (BW) changes could reflect variations on three different main components: weight of the foetal-placental unit (in particular at the end of gestation), digestive contents and body reserves . Changes in broodmare BW and body condition (BC) along the year are also linked to seasonal and management factors. Regardless the stage of gestation/lactation, BW and BC rise during the spring, which is probably related with pasture quality and availability Pagan et al., 2006). Nowadays, there is an increasing worldwide interest on Lusitano breed as a sport and leisure horse, due to its functional and behaviour characteristics. Although raised in several countries, most Lusitano stud farms in Portugal are traditionally based on extensive feeding systems. On these systems, mares are often bred outdoors throughout the year, being pastures an important part of their diets. Most of these pastures (natural or sown) are under Mediterranean influence and herbage production is commonly limited by summer dryness . When grass resources are scarce, supplementary feeds are generally used, but farm practices vary widely. To the best of our knowledge, few longitudi-nal studies were performed in order to access BC changes across the gestation/lactation cycle in the sport broodmare (Lawrence et al., 1992;Pagan et al., 2006), but none in Mediterranean conditions. Thus, further information on nutritional status and body reserves management in the Lusitano breed will contribute for better decisions on the most appropriate feeding plans and foaling seasons, in order to improve the efficiency and profitability of these production systems. The present research aimed to evaluate the effects of foaling season and feeding management in extensive systems on the nutritional status of Lusitano broodmares throughout the gestation/lactation cycle by the assessment of BC, BW, and some blood metabolic indicators. Materials and methods The protocol of this study was approved by the Ethical committee of the Faculty of Veterinary Medicine, University of Lisbon, Portugal. All the animals were handled with care during the experimental procedures. Animals and study design This experiment was conducted in four stud-farms located at the main region of Lusitano breeding, the southern area of Portugal. The broodmares of each stud-farm, hereafter designated by groups A, B, C and D, were monitored for BCS and BW over a period of four years (A and B: 2006 to 2008; C and D: 2008 and 2009), in a total of 119 gestation/lactation cycles. Average age at the beginning of the study was 11.0±3.4 (mean±SD) years for A mares (n=17), 8.4±2.9 years for B mares (n=15), 6.5±3.8 years for C mares (n=6) and 11.0±4.7 years for D mares (n=15). The suckling foals from these mares were also monitored for BW and height at withers (WH), through early lactation months. A standardized herd health schedule with routine vaccination and deworming programs was practiced in the four stud-farms. Feeding plans All mares were maintained on pasture all year and had free access to water. The studfarms are located between latitude 38º88' to 39º29' N and longitude 07º67' to 08º87'W and, according to Köppen classification, are under the influence of a Mediterranean climate (Csa). The annual rainfall is 652 mm (B, C and D) and 836 mm (A). The annual mean temperature is 17ºC (B, C and D) and 16ºC (A) (Instituto de Meteorologia, I.P., Lisboa, Portugal). Temperatures range, in winter, from -3ºC to 25.2ºC for B, C and D stud-farms and from -4.5ºC to 21.9ºC for stud-farm A and, in summer, from 11.9ºC to 45.2ºC for B, C and D stud-farms and from 9.4ºC to 41.3ºC for studfarm A. Pastures are settled mainly on fluvisols and regosols (B and C), cambisols (A) and podzols (D) (FAO, 2009). Floristic composition was typical of natural rain fed pastures of these areas, with a high biodiversity. Besides a mixture of native grasses and legumes, some forbs and weeds were also present. Among grasses (Poaceae), the main genera includes Lolium spp., Phalaris spp., Bromus spp., Agrostis spp. and Poa spp. In the legume family (Fabaceae), plants from the genera Trifolium spp., Vicia spp., Melilotus spp., Ornithopus spp. and some annual species of Medicago were identified. A and B mares were daily supplemented in group with commercial compound feeds (ranging from 1.5 kg head -1 d -1 to 5.5 kg head -1 d -1 ), and with grass hay or cereal straw (ranging from 2 to 6 kg head -1 d -1 ), according to animals' physiological stage and pasture availability. C and D mares were only supplemented with grass hay (ranging from 5 to 10 kg head -1 d -1 ), in periods when pasture was scarce. Dry matter production in similar pastures of the same regions ranged from 40 kg ha -1 d -1 in February to 90 kg ha -1 d -1 in April, and decrease to 50 kg ha -1 d -1 in May (Paço and Fradinho, 2011). In all but B stud-farm, rotational grazing was practiced. Samples of supplementary feeds used in stud-farms A and B were regularly collected for nutritional assessment. Pasture samples (one compound sample per pasture) were also collected on the spring of 2006 (stud-farms A and B) and on the spring of 2008 (stud-farms C and D). Chemical composition analyses were made in a reference nutrition laboratory. Samples were dried for dry matter (DM) determination in a forced-air oven, to a constant weight, at 104ºC. Ground samples were analyzed for ash by burning overnight at 550ºC (Instituto Português da Qualidade, 1983). Crude protein (CP) was measured by Kjeldahl method (ISO, 2005). Crude fibre (CF) analyses were conducted according to the official procedure for feed analysis (ISO, 2000). Neutral detergent fibre (NDF) and acid detergent fiber (ADF) fractions were analyzed by sequential detergent fiber determination according to Van Soest et al. (1991). Digestible energy (DE), net energy (NE) and digestible protein (DP) were estimated using the French feed evaluation system (INRA, 2012). In addition, protein value of feeds, expressed in MADC (Horse Digestible Crude Protein) was calculated according to the new prediction equations (INRA, 2012). Body condition scoring and weight measurements Body condition of each mare was monthly evaluated (from the 9 th month of gestation to the post-weaning period) according to the BCS method of Arnaud et al. (1997) by a single observer, blinded to previous data. This BCS system (0 to 5 points scale) is based on five palpable areas of the horse's body (along the neck, along the withers, behind the shoulder, over the rib cage, between the 10 th and the 14 th ribs and the area adjacent to the tail head) and on two visual appraisals (the top line of the back and the croup). A quarter of point division was used for a better accuracy of the method. On the same occasion, body weight was determined (without restriction of feed or water) using a portable electronic scale (Iconix, FX15, New Zealand), which accuracy (0.5 kg) was regularly checked. Foals measurements In order to indirectly assess mares' milk production, growth and development of the suckling foals were monthly evaluated throughout the first months of lactation because, at least until two months of age, average daily gain (ADG) is linearly related with milk intake . This information allowed for a better understanding of mares' BC and BW changes on this period, reflecting the balance between the requirements for a higher or lower milk

production and the abundance or shortage of feed. Foals' BW was assessed with the same electronic scale used for broodmares. Height at withers was measured with an aluminum standard measuring stick from the ground to the highest point of the withers. Blood metabolites On the days of BC and BW assessments, between 8.00 h and 11.00 h and before any compound feed was distributed, blood samples (≈18 mL) were collected from 10 mares of group A, 10 mares of group B, six mares of group C and 10 mares of group D, by jugular venipuncture into plain and heparinized tubes (Monovette Serum and Monovette Li-Heparin, Sarstedt AG & Co., Nümbrecht, Germany) for determination of glucose, NEFA, urea and albumin concentrations. Blood was collected into plain tubes for NEFA determination and was allowed to clot at room temperature. Heparinized tubes were immediately placed on ice until centrifugation. All blood samples were transported to the laboratory on ice and centrifuged at 670 g, at 4ºC, for 15 min. Plasma and serum samples were [Ital J Anim Sci vol.12:e71, 2013] stored at -20ºC until analysis. Plasma glucose, urea and albumin concentrations were measured by colorimetric methods in an autoanalyzer (Kone Optima Analyzer, Therma Clinical labosystems, Vantaa, Finland) with commercial kits (Albumin Monofluid, Glucose HK and Urea UV, Bradford, Kemia Cientifica S.A., Madrid, Spain). Serum NEFA concentrations were determined by an enzymatic colorimetric method with a commercial kit (NEFA-HR(2), WAKO Chemicals GmbH, Neuss, Germany). Statistical analyses Each group of mares with their foals was kept in different environmental conditions which impairs the direct comparisons among them. Therefore, independent but identical statistical analysis was conducted for data obtained in each group. Data were analyzed using the MIXED procedure of SAS (SAS Inst. Inc., Cary, NC, USA). The model considered the foaling season, the physiological stage (gestation/lactation month), and the interaction between them as fixed factors, and mares as random effect. The physiological stages (from 9th month of gestation to post-weaning) were treated as repeated measures and an autoregressive covariance matrix was used. Two foaling seasons were considered for groups A, B and C: February-March and April-May. Due to the concentration of foaling dates in group D, only one foaling season was considered: February-March. For this group the model considered the physiological stage as fixed factor and mares as random effect. When significant differences were detected, the differences among means were evaluated using the Tukey-Kramer test. Statistical significance was assumed when P<0.05. Foals data (BW and WH) were also analyzed with a mixed linear model allowing for repeated measures on time. The effects of foaling season, time (expressed in days of age), time×time and their interactions were included in the model for foals A, B and C. For D foals, only the effects of time and time×time were studied. All results are presented as Lsmeans±SEM, unless stated otherwise (Table 1). When variables did not follow the normal distribution, transformed data (log10 or square root) were used in the analysis. For those variables the means were back transformed and the standard error of means were replaced by back transformed confidence intervals. To evaluate relationships between variables, Spearman's correlation coefficients were calculated. Results and discussion Only mares that had a successful delivery and nursed a foal until weaning were included in the study. Globally, foaling took place between January and May. Weaning occurred in early October for A and B groups, early November for group C and middle September for group D. In each year and each group, foals were separated from their dams on the same day, regardless of foals' age. Body weight and body condition Variations on BW throughout the year were tightly associated with mares reproductive stage (i.e. gestation or lactation month) confirmed by the effect of the physiological stage on BW (P<0.001), in the four groups of mares. As expected, changes in BW occurred either before foaling (increasing in the two last months of gestation) or after foaling (P<0.01) (Figure 1 a-d). Besides seasonal factors (e.g., grass abundance in the spring), major BW changes in the broodmare were normally Fradinho et al. linked with the conceptus gain (foetus + adnexa and uterus) during gestation and its reduction associated with foaling Heidler et al., 2004;Pagan et al., 2006). In the present study, and considering BW after foaling, mean weight gain observed throughout pregnancy ranged between 7.6% and 13.9%, while in Thoroughbred mares was 13.9±3.2 % (Cassill et al., 2009). But the major increase took place in the last three months of pregnancy and could represent 7% to 9% , which coincide with the average value of 8.8% found among the four groups of mares throughout this stage. As BCS did not change significantly during this period, it could be considered that the observed BW gain accounted for the gain of conceptus. After foaling, it was reported for the Lipizzaner mare a weight reduction of 64.8±2.4 kg, which represents about 12% of mares BW after foaling (Heidler et al., 2004). Considering the four groups together, the weight reduction observed in the present study was 11.6%. Also in heavy breeds, a reduction of 12% of BW after foaling was observed . Body weight changes were also influenced by foaling season and an interaction between foaling season and physiologic stage was found in groups B and C (P<0.05). The effect of foaling season in BW was very clear in mares of group A with consistent higher values observed in Apr-May foaling mares when compared to Feb-Mar foaling mares (P<0.01) (Figure 1a). Mares of group D showed, in gen-eral, the lowest values of BW. In this group, and besides the changes observed in the two last months of gestation and after foaling, an increase in BW occurred from the 1 st to the 4 th month of lactation (P<0.01), reflecting probably an influence of pasture spring production (Figure 1d). Throughout the year, changes on BCS were observed in the four groups of mares (P<0.05), although with small amplitudes within each group. Generally, higher scores were reached at the end of spring, decreasing then until the end of summer (which is coincident with late lactation period), when the lowest scores were observed (Figure 1 a-d). On group A, a clear effect of foaling season on BC was observed (P<0.01) with consistent higher values throughout the year in Apr-May foaling mares. BCS varied between 2.59±0.08 and 2.88±0.08 on Feb-Mar mares and between 2.96±0.09 and 3.23±0.09 on Apr-May mares. Mares of Group B showed the smallest variation in BCS, with values between 3.18±0.08 and 3.29±0.08 on Feb-Mar foaling mares and between 3.20±0.10 and 3.44±0.05 on Apr-May mares. In this group an interaction between foaling season and physiological stage was observed (P<0.05), with the Apr-May foaling mares showing a BCS peak in the 11th month of gestation, which was absent in Feb-Mar foaling mares (Figure 1b). Although with changes throughout the year (P<0.01), BCS on C mares was not influenced by foaling season. In this group, an increase in BCS was observed from late lactation (3.09±0.07) to post-weaning period (3.28± 0.07) (Figure 1c). Mares of group D presented the lowest BCS and the largest amplitudes, between 2.21±0.9 and 2.77±0.10. Higher BCS values were observed in late gestation months and there was a steady decrease during lactation period until the last months of lactation and post-weaning (P<0.01) (Figure 1d). Considering data from all mares, a positive Scorrelation was found between BC and BW (r=0.57; P<0.0001). The higher correlation coefficient was obtained for mares of group A (r=0.62; P<0.0001) and the lower was obtained for mares of group D (r=0.34; P<0.0001). The important role of an adequate amount of body reserves on reproductive and productive performances of the broodmare is generally recognized (Guillaume et al., 2006). The effects of nutritional status (assessed by regular BCS) on reproductive performance of the mare was highlighted in several studies (Henneke et al., 1984;Godoi et al., 2002;Guillaume et al., 2002). It was also shown that mares' intake and milk yield and composition depends on BCS (Doreau et al., 1992;1993) and, consequently, the growth of nursing foals could be influenced by the amount of body reserves of their dams (Martin-Rosset and Young, 2006). The current study provides an overview of BC and BW changes along the year in light broodmares reared on pasture based systems of southern Europe. Like in Thoroughbred mares (Pagan et al., 2006), in our field study BC changes appear to be strongly influenced by the time of year, reflecting seasonal pasture production and quality. Regardless of gestation or lactation stage and feeding regime, higher BC were generally found at the end of the spring (about 3.03 in earlier foaling mares and 3.33 in latter foaling mares) reflecting pasture abundance in this season. During the summer, herbage production in these regions is limited by climate conditions (high temperature and scarce rainfall). Therefore, the decrease in BC in groups A, C and D until the end of the summer (about 0.25), when the lowest values were observed, suggests a mobilization of body reserves to match the nutritional needs throughout the last stage of lactation. According to the prediction model proposed by Martin-Rosset et al. (2008), a decrease of 0.25 point in BCS represents a variation of 3% of body weight. Thus, considering the average body weight of A, C and D mares (first weight evaluated after foaling) as 529.4 kg, 528.9 kg and 486.3 kg, the amount of mobilized adipose tissue would have represented 15.9 kg, 15.9 kg and 14.6 kg, respectively. As the estimated net energy variation is -19.78 MJ kg -1 BW (Martin-Rosset et al., 2012), a deficit of 6.4% could be identified in the period between May and September regarding the maintenance requirements (INRA, 2012). Besides herbage scarcity, this deficit could also have been the consequence of some heat stress resulting from an increase of environmental temperature out of the thermo neutral zone, which is accepted to range from 5C° to 25C° in mild cli-mate (Morgan et al., 1997;Morgan, 1998). For example, maintenance requirements of adult standardbred geldings in mild climate were 9% higher in summer than in winter (Martin-Rosset and Vermorel, 1991). In contrast, the absence of BCS changes in mares of group B during the summer months (Figure 1b) was probably related to feeding practices. In fact, and besides pasture, A and B mares were daily supplemented with variable amounts of other feeds, which nutritive value is presented in Table 1. But the proportion of the requirements that was covered by compound feeds used in group B was, on average, 6% higher for energy and 10% higher for protein than in group A in the period between the 2 nd and the 7 th months of lactation. Unfortunately, it was not possible to collect and analyze pasture samples in the summer period. However, the results concerning the last samples taken in stud-farms A and B (May samples) indicate a better nutritive value for the pastures of stud-farm B (Table 2). Considering global BCS throughout the study, mares of group D presented the lowest values (2.21) and the biggest amplitudes (0.56). Foaling season occurred very early in the year and, because these mares were highly dependent on grazing resources, a steady mobilization of body reserves was observed from February in order to cope with higher nutritional requirements of early lactation (Figure 1d). In addition, it is quite clear that the nutritive value of pastures of stud farm D (April samples) is the lowest when compared with the nutritive value of pastures from stud-farms A and C in the same month (Table 2). Regardless the absence of a significant effect of foaling season on BCS of group C, mares that foaled later (Apr-May for A, B and C groups) showed, on average, higher BCS suggesting a better utilization of grazing resources. Overall, mares reached at key points of the reproductive cycle (e.g., foaling and weaning) an average BCS of 3.12 (2.67 to 3.36) and 2.96 (2.21 to 3.27), respectively. This level of BC at foaling is similar to the recommended (INRA, 2012) in order to optimize fertilization during the first month after foaling and to support the first months of lactation. Blood metabolites The assessment of blood parameters for monitoring metabolic and nutritional status in livestock species has been widely used for a long time because of the quality of the information that

could provide and the simplicity of collection and determination (Doreau et al., 1981;Caldeira, 2005). In the present study, values of blood metabolites were, generally, within the reference ranges described in literature for horses and particularly for pregnant and nursing mares (Harvey et al., 2005). Small changes (P<0.05) in plasma glucose concentrations were observed in groups A, C and D throughout the gestation/lactation cycle. Only in mares of group B, glucose concentrations differed with foaling season (P<0.01). Blood glucose is under a powerful homeostatic Fradinho et al. control that keeps it within narrow limits. However, low values may indicate decreased energy intake or gluconeogenesis rate (due to lack of glucose precursors) in periods of greatest needs of glucose (e.g., early lactation). In the present study, glucose concentrations decreased from late gestation to early lactation and lower values were also recorded during the summer months (P<0.05) (Figure 2). Basal lower glucose concentrations in early lactation were also observed in Thoroughbred and Lipizzaner mares when compared with values found in late gestation (Hoffman et al., 2003;Heidler et al., 2004). It is likely that the increased use of glucose during early lactation was influenced by mammary gland demands for lactose synthesis in milk, because mares' milk is highly concentrated in this disaccharide (Doreau et al., 1993;Santos and Silvestre, 2008). The observed decrease on glucose concentrations from spring to summer months and the steady levels found during this season probably reflect the influence of pasture quality and availability, when the nutritional requirements linked to the lactation stage are still important. Throughout the year, changes in NEFA concentrations were observed in A and B mares and an interaction between foaling season and physiological stage was found (P<0.01). As observed for glucose, NEFA concentrations were only influenced by the physiological stage in groups C and D (P<0.001). Serum concen-trations of NEFA have been used in several studies as a metabolic predictor of energy status. Feed deprivation or restriction leads to fat mobilization and a consequent rise in NEFA concentrations (Sticker et al., 1995;Caldeira, 2005;Dugdale et al., 2010). Overall, higher concentrations of NEFA were observed during periods when fat mobilization could be expected in order to cope with energy needs, due to either a specific physiologic state or a lower availability of feed. In groups A, B and D, lower NEFA concentrations were recorded in spring months, when pasture was abundant, and peaked in June-July, when pasture becomes dry. Concerning the metabolites associated with the protein status, some changes were found in groups A, B and C (P 0.05). Urea plasma concentrations were influenced by foaling season in groups B and C (P<0.05) and also by physiological stage in groups A, B and C (P<0.05) ( Figure 3). As for other species, previous studies in the horse have shown that uremia is sensitive to high protein dietary levels (Miller-Graber et al., 1991). In contrast, dietary protein restrictions (50% of protein requirements) appear to decrease urea concentrations in blood (Sticker et al., 1995). The results obtained in the present study showed an influence of physiological stage on urea concentrations, which could be related with time of year. In general, uremia was higher during spring months and decreased in the summer, suggesting a relation to protein levels of pasture. Albumin is the most abundant protein in blood and in situations of nutritional deficiency may function as an important pool of labile protein. Considering the contribution of dietary protein from compound feeds in groups A and B, it would be expected that higher values of albumin would be found in these groups. However, albumin values tended to be relatively constant along the year, and in a similar range to non-supplemented groups (C and D). In winter months, grazing is quite limited and supplementation with low quality hays could eventually justify the significant decrease in albumin values found in group C. Foals' growth and development Significant interactions between foaling season and time (P<0.05) were observed on BW changes for A and C groups, indicating a different pattern of growth between foals born in Feb-Mar and foals born in Apr-May (Table 3). At 90 days of age, estimated BW varied between 140.7 kg (group B) and 160.8 kg (Feb-Mar foals of group A). Concerning WH, the influence of foaling season was only observed in group C, with an interaction between foaling season and time (P<0.05) ( Table 3). Estimated WH at 90 days of age varied between 120.2 cm (group B) and 123.8 cm (Feb-Mar foals of group C). Overall, higher growth performances (ADG) through the first three months of life were observed in early born foals (Feb-Mar) of groups A and C for BW and in early born foals of group C for WH. Previous research in other geographical regions referred a clear influence of month of birth and season of year on suckling foals' growth rate, with lower values for winter born foals, when access to pasture was limited (Hintz et al., 1979;Pagan et al., 2006). Considering the growth performances of the early born A and C foals in our study, mares that foaled in Feb-Mar have, apparently, higher milk production, reflecting the influence of spring pasture quality and availability. However, the shift of nutrients for milk production at this stage caused a less effective recovery of BC during the spring, in comparison with that observed in the mares that foaled in Apr-May, implying that Feb-Mar mares reached the summer with less body reserves. Nutritional status of Lusitano broodmares Regardless of different feeding practices in the four groups of mares (supplemented vs. non-supplemented), growth performances of later born foals may suggest that other supplementation strategies (namely in what concerns some limiting nutrients) should be implemented, when pasture growth is depressed by summer dryness. Conclusions Results show that changes in BW and BC in the Lusitano broodmare, managed on grazing systems, are mainly influenced by pasture availability and quality and the time when foaling season occurred in the year. In fact, Mediterranean pasture cycle leads to a general increase in body reserves in spring and their mobilization until autumn and winter, although this change does not represent more than half point of BC. The quality of pastures and supplementary feeds has a strong effect on the mean annual BC among stud-farms, determining almost one point of BC variance. Early foaling in the season had also a marked effect, hindering the recovery of BC and decreasing the level of BC throughout the whole cycle. BC and BW changes and, particularly, blood indicators showed an overall balanced nutritional status and an apparent metabolic welfare without any evident signs of under or over nutrition, reflecting an adaptation to feed availability and climate. The association of these data with further studies on mares' fertility and foals' growth until weaning will contribute for better management decisions on the most suitable foaling season and the most appropriate feeding plan, in order to improve the efficiency and profitability of the Lusitano production system. A Breath of Knowledge: Overview of Current Adolescent E-cigarette Prevention and Cessation Programs Purpose Adolescent use of electronic cigarettes (e-cigarettes) has risen rapidly, which is concerning given the health effects of e-cigarettes and youth susceptibility to nicotine addiction. It is critical that efforts to educate, prevent, and reduce adolescent use of e-cigarettes are developed and evaluated. The purpose of this paper is to review available current prevention and cessation programs. Findings A web-based search of currently available e-cigarette prevention and cessation/treatment programs was conducted using Google in May of 2020. Programs were then reviewed on whether they included theory- and evidence-based practices of effective adolescent prevention and cessation programs. Eight prevention programs, seven cessation programs, and one program that addressed both prevention and cessation were identified and included in this review. Most prevention programs included the importance of understanding flavored e-cigarette products, addressed industry-targeted marketing, included social learning activities to develop refusal skills, delivered free-of-cost, available online, and explicitly stated their incorporation of theory. Five prevention programs and two cessation programs had empirically evaluated their e-cigarette-related components. Conclusions Although the programs reviewed largely incorporated theory and included key components known to be effective, there are some gaps in the programs’ overall ability to prevent and stop adolescents from using e-cigarettes, such as lack of dedicated e-cigarette materials. More evidence-based tools, resources, and evaluations are needed to best inform adolescent e-cigarette cessation. Addressing the gaps that existing prevention and cessation programs present requires intervening at multiple systematic levels, conducting more rigorous program evaluations, and bolstering the availability of cessation programs. Introduction Although the prevalence of adolescent combustible cigarette use has declined greatly, adolescents' use of alternative tobacco products has risen rapidly, owing largely to their use of electronic cigarettes (e-cigarettes) [1]. The 2019 Youth Risk Behavior Survey found that a total of 50.1% of US high school students had ever used an e-cigarette product, and 32.7% were past 30-day e-cigarette product users [2]. In the past 5 years, there has been an enormous proliferation of new types of e-cigarettes on the market, likely contributing to this surge in e-cigarette use. Since Juul came on the market in 2015, which comprised nearly three quarters of the US ecigarette market share in 2018 [3], many pod-based e-cigarette products have been developed and marketed. In 2019, a new group of e-cigarette products, disposables such as Puff Bar, came on the market. All of these newer products have high amounts of nicotine, with an average of about 60 mg of nicotine per pod/device [4]. These newer products contain saltbased nicotine, a type of nicotine that is combined with benzoic acid to change the PH balance of the nicotine, making it less acidic and harsh and therefore easier to use especially among nicotine naïve adolescents [5]. These patterns of use, newer products, and high concentrations of nicotine are concerning given the health effects of e-cigarettes and that adolescents are highly susceptible to nicotine addiction [6,7]. Given the high rates of use and known health consequences of using e-cigarettes, it is critical now more than ever that adolescents are educated about e-cigarettes so that they can make informed decisions regarding their health, and that efforts to prevent and reduce adolescent use of e-cigarettes are developed, implemented, disseminated, and evaluated. In this paper, we elaborate on the health effects and reasons for youth use as the rationale for prevention and cessation programs, and importance of theory-based programs, and then discuss in detail the various prevention and cessation programs available to prevent and reduce adolescent use of e-cigarettes. Health Effects Studies have shown that adolescent use of e-cigarettes has deleterious health effects, such as resulting in cardiac and respiratory problems, secondhand aerosol exposure, and impacts on brain development and nicotine addiction [8]. A study found that a validated nicotine dependence scale for ecigarettes (PROMIS-E) identified more than half of the study sample of adolescents as experiencing some level of nicotine dependence from using e-cigarettes [9]. Nicotine's impact on adolescent brain development can lead to additional mental health-related negative outcomes, such as depression and substance use disorder [10]. Furthermore, recent research shows a relationship between adolescent e-cigarette ever-use, dualever-use and past 30-day dual-use of e-cigarettes and cigarettes and COVID-19-related outcomes [11, 12••]. In light of recent attention brought to youth e-cigarette, or vaping, product use-associated lung injury (EVALI) [13] and associations between youth e-cigarette use and COVID-19-related outcomes [12••], it is important that public health efforts continue to inform adolescents about emerging health harms and prevent or stop adolescent e-cigarette use. Why Adolescents Use E-cigarettes Despite e-cigarette health risks, adolescents are susceptible to using e-cigarettes for a number of reasons, including the appeal of flavors, social influences, marketing, and perceptions of reduced harm of e-cigarettes [14•, 15-17]. Here, we briefly discuss some of these well-researched factors about why adolescents use e-cigarettes. A primary reason for youth appeal of e-cigarettes is the numerous available flavors. Adolescents have reported that they think flavors, especially fruity, candy or sweet, menthol, and mint, are for them, and this direct appeal of flavors leads to increased acceptability and use of e-cigarettes [15]. In addition to flavors driving youth e-cigarette use, adolescence is a particularly unique time in the developmental process, as youth are transitioning to adulthood and learning to discover and develop their personal identities [18]. During this developmental phase, adolescents are particularly influenced by social pressures and the

rapidly evolving world of social media and technology [19,20]. Studies have found that peer pressure is an important reason why youth use e-cigarettes [21,22]. Adolescents are constantly being exposed to both messaging around the flavors and positive aspects of e-cigarettes, both from social media and official advertising from the tobacco industry itself [23,24]. Adolescents have reported feeling targeted by such content and e-cigarette advertising [25]. E-cigarette messaging results in perceptions of reduced risk and greater social acceptability of these products [20,25]. Data show that adolescents think e-cigarettes are safer than combustible cigarettes, with some adolescents believing that e-cigarettes contain just water vapor without nicotine or other chemicals [14•, 26-28]. Several qualitative studies have found that adolescents do not fully understand the severity and negative consequences of nicotine addiction, compared with other types of addictions [29,30]. Coupled with these perceptions, recent studies show other proximal levels of influence, whereby adolescents are using e-cigarettes as a means to manage stress, anxiety, and depression [31][32][33]. Flavors, social influences, marketing, and misperceptions of the true risk of e-cigarettes as well as the misinformation that they are exposed to by social media and marketing have driven this widespread epidemic of adolescent e-cigarette use. As such, these critical components for adolescent e-cigarette use should be considered when designing, implementing, and evaluating any e-cigarette prevention and cessation program. Importance of E-cigarette Prevention and Cessation Programs E-cigarette prevention and cessation programs are important to help educate adolescents on how to make the best decisions for themselves as they navigate through their adolescence and the social pressures that come with it. It is important to continually adapt tobacco prevention and cessation programs to include and address emerging novel products that can quickly become popularized by adolescents, such as e-cigarettes. Given that e-cigarette prevention and cessation programs are relatively new, we have drawn upon the cigarette prevention and cessation program theories and best practices as a framework for evaluating e-cigarette prevention and cessation programs. However, it is important to address e-cigarettes separately from tobacco and combustible cigarette education, as adolescents do perceive e-cigarettes as new and different products [34,35], which is why there is a currently urgent need to expand e-cigarette prevention and cessation programs for adolescents. Schools and classrooms have traditionally been the settings for delivering tobacco prevention efforts because a large part of adolescents' social lives take place in and revolve around a school setting and with their peers [36][37][38]. When taught in a classroom or other group setting, these prevention curricula help adolescents with decision-making around using tobacco, help teach and learn refusal skills strategies, and help adolescents develop a sense of collective confidence in learning these skills alongside their peers [39]. The evidence on effectiveness of school-based tobacco prevention programs is mixed in regard to helping change the norms around tobacco use and addressing the misperceptions that exist about tobacco products, and mixed results in changing actual behavior [37,[40][41][42][43]. However, effective components of such school-based tobacco prevention programs include interactive curricula, activities around refusal skills, and content addressing targeted marketing and health effects [44, 45•], which if applied collectively in prevention curriculum may lead to decreases in youth intentions to use and actual use. In addition to needing effective e-cigarette prevention programs, evidence-based e-cigarette cessation programs are needed to help youth who are experiencing addictive behavior with respect to e-cigarettes [46]. Unfortunately, there are few validated cessation programs for adolescents experiencing nicotine dependence or addiction to e-cigarettes. The scientific literature is currently lacking in studies on nicotine replacement therapies for adolescents trying to quit e-cigarettes, and there are only a few studies on cognitive behavioral therapy and other therapies for adolescent cessation [47,48]. Alternative-to-suspension programs, designed for students caught using e-cigarettes in school, are increasingly being used in school settings to address e-cigarette use among adolescents in a more supportive way, rather than through punishment [49]. Text messaging and other online chat-based cessation platforms are also becoming popular [50]; however, additional research is needed on the efficacy of this form of textbased counseling and intervention. Although evaluated ecigarette cessation programs for adolescents are lacking, addressing nicotine addiction early is important in preventing adolescent occasional e-cigarette-users from escalating to more frequent use, as well as preventing adolescent e-cigarette users from later using combustible cigarettes. We next review theories and best practices for prevention and cessation programs, then review specific programs, discussing what these programs are doing well and illustrating the gaps that still exist. Importance of Theory Prevention and cessation programs are most effective when based on valid theories of youth development, behavior change models, and how youth learn [51]. Theory lays the foundation to understanding why risk behaviors, such as e-cigarette use, happen and how to design better resources with strategies that have been validated through past programs. The most commonly used and cited theories for tobacco prevention and cessation programs include the Theory of Planned Behavior, Social Cognitive Theory, and the Transtheoretical Model of Behavior Change [52][53][54]. The Optimal Health Framework and Positive Youth Development are frameworks that are now being applied in many tobacco prevention programs to build engagement and foster participatory learning [51]. The following presents further information on the core components of these theories and frameworks that are most effective in prevention and cessation programs. The Theory of Planned Behavior explains how variables such as intention, attitudes, subjective norms, and perceived behavioral control can influence behavior, and has been widely used in understanding tobacco use intentions and designing interventions [52,55]. Self-efficacy is an important aspect of the theory of planned behavior, to help build confidence to feel personally empowered to make decisions and resist risk behaviors, such as e-cigarette use [39]. Social Cognitive Theory explains how individuals imitate modeled behaviors and highlights how educating adolescents about social norms can help build refusal skills, resist peer pressure, and shape behavior change [52]. The Transtheoretical Model of behavior change posits that health behavior changes progress through a number of stages and has been used in many tobacco interventions, especially smoking cessation [54,56]. The Optimal Health Framework posits that influences of physical and social environments, such as parents, peers, schools, communities, and media can increase and reduce the likelihood of adolescents engaging in unhealthy risk behaviors [57••]. The Optimal Health Framework also explains how risk-taking behavior results from neurobiological development during adolescence, which is why adolescents tend to seek novel experiences and experimentation to help develop independence and self-identity [57••]. The Optimal Health Framework can help inform programs in understanding why adolescents use e-cigarettes and how to best address this issue in the environments adolescents find themselves in. Positive Youth Development is another approach that encourages educators to engage adolescents in learning about the process of development and decision-making during the journey into adulthood [51]. Elements of Positive Youth Development in tobacco prevention programs include supportive environments with adult mentorship and encouraging community involvement and leadership [51], such as through youth-led projects [51,58,59]. Core components of ecigarette programs and interventions should thus include elements of social-emotional learning, Positive Youth Development, and involve adolescents' families and communities [51, 57••]. Best Practices in Prevention and Cessation Effective tobacco prevention programs include normative education and interactive content. Normative education is defined as informational and behavioral interventions aimed at addressing beliefs around a certain behavior, such as decreasing e-cigarette use among adolescents [60]. Overall, it is important for prevention programs to address reasons why adolescents use e-cigarettes and beliefs around what constitutes nicotine addiction or susceptibility to marketing as a means to enable youth to reflect on perceptions underlying their decisions to use e-cigarettes. Thus, in preventing adolescent ecigarette use, programs should address flavors, peer influences, perceptions of harm, and other key reasons why adolescents are particularly attracted to these more novel tobacco products. Specifically for school-based tobacco prevention, programs with the potential for long-term impact also include education around commitments and intentions not to use, training around refusal skills, and use of peer leaders [61]. Interactive content is defined as curricula that encourage discussions and activities, and fosters games, role playing, and practicing of refusal and other skills [61]. Under no circumstances should any youth tobacco prevention program be developed and implemented by the tobacco or e-cigarette industry, as such program content and delivery can be adversely affected by bias towards commercial interests [44,62]. Unfortunately, there is less scientific literature around best practices for tobacco cessation programs for adolescents. Research has evaluated the efficacy of overall strategies of adolescent cessation, such as individual or group counseling, messaging interventions, and pharmacological interventions, but have not identified what aspects of such programs or specific components were effective at a more granular or content level [47,63,64]. Thus, we do not have as exhaustive of a list of best practices for adolescent tobacco cessation programs as we do for prevention programs. Overall, programs that are adaptable and accessible, ideally being fully online or web-based and completely free-of-cost, help with wider dissemination and implementation of programs. Effective programs address both middle and high school levels, which is around the age when most adolescents are known to try e-cigarettes for the first time [65]. Next, we review the current landscape of e-cigarette prevention and cessation programs available and assess whether or not they are theory-informed and meeting the identified best practices. Review of Programs: Methods To identify available e-cigarette prevention and cessation programs, we conducted a web-based search of currently available e-cigarette prevention and cessation programs using Google in May of 2020. The Google Boolean search terms used were "e-cigarette prevention program," "e-cigarette prevention," "e-cigarette education," "vaping prevention, "e-cigarette cessation," "youth e-cigarette cessation," "e-cigarette treatment," and "e-cigarette alternativeto-suspension." The purpose of using Google as the main search platform was to identify programs that were more widely known and searchable not just by an academic audience but also by parents, adolescents, and educators. This is especially important since many e-cigarette prevention and cessation programs have likely not yet been evaluated and therefore might not show up in searches on PubMed, PsychLit, or other academic search sites. The results of these Google searches were reviewed to identify programs that focused on adolescent e-cigarette prevention or cessation. Inclusion criteria for both prevention and cessation programs were that the program included content addressing ecigarettes was created and used in the USA, and that the stated intended audience was adolescents or educators who delivered program content to adolescents. Exclusion criteria for both prevention and cessation programs were not having working webpages and/or evidence of not having been updated within the last 5 years, when e-cigarettes became most popular among youth. Programs around vaping marijuana were not included to focus this review on nicotine e-cigarette prevention and cessation. Inclusion criteria specific for prevention programs were school-based education curricula, mass media campaigns, and educational games. Inclusion criteria specific for cessation programs were websites with quitting resources, quitlines, text messaging services, and alternative-to-suspension programs. With this inclusion and exclusion criteria, eight prevention programs [59, 66-72], seven cessation programs [73-79], and one program combining prevention and cessation [72] were identified as focusing on adolescents and included in this review. Table 1 provides detailed information on each of the nine prevention programs and Table 2 provides detailed information on each of the eight cessation programs. Data in both tables are presented to summarize information on the following program aspects: source of the program, duration of exposure, component of program dedicated to e-cigarettes, date the program was created, date of last update of the program, intended audience, cost, channel of delivery, whether or not the program explicitly uses theory, and whether the program has been empirically evaluated. Additional components evaluated for prevention and cessation programs differ, as the prevention programs identified were mostly educational or media campaigns, whereas the cessation programs were mostly individual counseling or alternative-tosuspension programs, and thus these prevention and cessation programs are not directly comparable to one another. What Programs Are Doing Well As shown in Table 1, all of the prevention programs discuss the importance of understanding flavored e-cigarette products, and all but one [80] of the prevention programs address industry-targeted marketing and advertising to adolescents. Six of the prevention programs [59, 66, 68, 69, 72, 81] included social learning activities to develop refusal skills, and the three programs that did not include refusal skills were mass media campaigns [67,70,71]. Of the nine prevention programs examined, all but

one program [68] was free-of-cost, and all but one program [68] was fully available online or required a simple enrollment process by individual educators/professionals. Of the eight cessation programs examined, all were free-of-cost, and all but two programs were fully available and accessible online [75,79]. Two of the cessation programs explicitly stated their incorporation of theory in their program descriptions and materials [76,79]. Three of the cessation programs included texting features, which indicated that these programs were trying direct-to-youth strategies using platforms and technologies that adolescents frequent [73, 74,77]. Four of the cessation programs explicitly stated incorporation of theory in their program descriptions and published materials [72,76,78,79]. Of the cessation programs, one explicitly cited Positive Youth Development [76], one explicitly cited Social Cognitive Theory [79], and one explicitly cited Theories of Behavior Change [78]. Overall, studies show that some programs have been evaluated and show promising results. Of the nine prevention programs, five programs have evaluated and published their content containing e-cigarette education [59, 66, 69-71]. These evaluated prevention programs show that the programs are successful at changing adolescent perceptions and behavior, including reducing e-cigarette use and increasing knowledge around these products [45•, 82-86]. Of the eight cessation programs, two have evaluated and published their content regarding e-cigarette cessation [73,75]. These evaluated cessation programs have promising results showing youth engagement with the program content and motivating adolescents to quit e-cigarettes [87,88]. Many programs also included normative education and interactive content around how ecigarettes are not completely harmless, such as the four messaging campaigns [67,70,71,81]. Gaps in Programs Despite the prevention and cessation programs largely including key components known to be effective, there are some gaps in the programs. There were only a few programs solely dedicated to or with separate modules on e-cigarettes. Three programs combined the topic of e-cigarettes with their general tobacco curricula [66, 68,72]. Overall, we found eight programs with separate modules or stand-alone programs for ecigarettes [59, 67, 69-71, 73, 77, 81]. Thus, varying durations of exposure to prevention content make it challenging to compare the efficacy of programs with one another. Ideally, e-cigarette prevention and cessation should be separated from general tobacco prevention and cessation. What works to reduce the use of other tobacco products may not work for e-cigarettes due to differences in social norms and social desirability by peers, ability to use e-cigarettes discreetly and without odor, an overwhelmingly aggressive advertising and social media landscape, appealing flavors, and perceived lower health risk of e-cigarettes compared to combustible cigarettes [14•, 89, 90]. E-cigarette product types and components are also fast-evolving, unlike other traditional tobacco products. Our review showed that several programs were not up to date or did not include the latest products that youth are using. Only one prevention program [59] includes education around newer disposables like Puff Bar and add-on flavor-enhancers like Puff Krush. Specifically for prevention programs, components of all programs should include a discussion of contents of ecigarette flavors, explanation of negative health effects, exposure and identification of targeted marketing, and practicing refusal skills to understand social influences and allow greater resistance to use. It was also unclear whether the level of interactivity and esthetic appeal of prevention content achieved is what adolescents are used to seeing in other marketing content from the industry, which presents a potential gap in appeal and engagement. Only one prevention program was designed as an interactive game platform to educate an adolescent audience about tobacco and e-cigarette prevention [66]. Our search identified more prevention than cessation programs, and there is less information specifically around ecigarette cessation programs. More effective and evidencebased tools, resources, and programs are needed for adolescent e-cigarette cessation. Most e-cigarette specific cessation programs were started an average of 2 years ago. This could be due to the fact that e-cigarettes are evolving rapidly in the market and there is growing evidence of adolescent addiction to e-cigarettes, resulting in cessation programs being developed more recently, and older tobacco prevention programs needing to adapt e-cigarette content into their materials and models. Of these e-cigarette specific cessation programs that have only been developed in the past year or two, only one alternative-to-suspension program has published pilot data on their outcomes and found that the program did influence students to make a plan to stop using e-cigarette products [88], and one texting program has published 3-month observational outcomes and found promising results for using texting to reduce the barriers of accessing a cessation intervention [87]. Furthermore, there are no FDA-approved nicotine replacement therapies identified for adolescent e-cigarette use. Additional research is needed to assess psychosocial needs and particular withdrawal symptoms related to e-cigarette cessation. More physiological studies are also needed to design nicotine titration plans for varying levels of e-cigarette use by adolescents of different ages for pharmacological treatments and to test effective strategies to enable medication adherence, acceptability, and sustained cessation. Ultimately, more evaluation is needed to move from evidence-informed to evidence-based practices in e-cigarette prevention and cessation. Some prevention programs have evaluated their entire curricula or tobacco prevention efforts, but not specifically their e-cigarette components. Our review of e-cigarette programs showed a dearth of published studies evaluating existing prevention and cessation programs for adolescent e-cigarette use with pre-post data or randomized controlled trials. When studies are conducted, we rarely know the long-term impact of the program or unintended consequences such as e-cigarette education leading to decreased youth ecigarette use but increased use of other products. Published data across programs are inconsistent with regard to information about program development, formative research, theoretical foundations, key principals and aims, modes of delivery, qualitative justification, reach, partnerships, evidence used in designing and updating content, measures of success, implementation research on acceptability and feasibility, and emphasis placed on the e-cigarette content. On the contrary, ecigarette counter-marketing campaigns set a good example of publishing details on how they are developed and designed [91,92]. Few educational programs describe the details of how they were developed, and thus it is difficult to distinguish which are most rigorous with respect to one another. Due to wide variations in content, duration of exposure, and mode of delivery, it is also difficult to tease apart which are the crucial components that work from each of these programs. There is also a need for program implementation research on barriers and facilitators in the school context that influence the probability of greater reach, impact, acceptability, and scalability. Recommendations Addressing the gaps in the existing prevention and cessation programs requires incorporating a theory-based approach, expanding focused interactive content, and conducting more rigorous program evaluations, in order to develop effective and successful long-term programs. Overall, current programs are supported by theory and follow best practices, but some gaps still exist. Programs should strive for a holistic approach or multi-modal approach involving schools, parents, teachers, and the greater community, which a systematic review of prevention education suggested is the way forward [93]. Programs should continue to bolster their digital accessibility and interactive components in order to keep up with the advancing technology and media that surround adolescents. Most importantly, programs need to focus on addressing the factors most likely to influence adolescent use of e-cigarettes, include more information around e-cigarettes' impact on mental health, describe their programs for ease of replication, and evaluate their programs. Conclusion Unabated adolescent use of e-cigarettes, misinformation about the harms of e-cigarettes fueled by marketing, appealing flavors, other product characteristics such as their sleek design, high nicotine content causing adolescents to become addicted, and other health effects of e-cigarettes are why both e-cigarette prevention and cessation programs for adolescents are critical [94]. This review highlighted that there are few adolescent-focused e-cigarette prevention programs and very few cessation programs. Furthermore, few have been evaluated to determine the most effective components for helping adolescents to make informed decisions regarding their health. Most programs are available online, with many free and few requiring cost. Few programs contain an updated stand-alone e-cigarette curriculum, not all programs utilize a theoretical framework, and not all programs address key influences on adolescent e-cigarette use: marketing, flavors, nicotine addiction, and peer pressure. The development, evaluation, and dissemination of prevention and in particular cessation programs is needed in order to mitigate adolescent use of e-cigarettes. Funding Research reported in this paper was supported by the Taube Research Faculty Scholar Endowment, and by a grant from the Tobacco-related Disease Research Program (TRDRP, grant number 27IR-0043). The funders had no role in the design or conduct of the study, including the data collection, data management, analyses, interpretation of the data, or the manuscript preparation. Compliance with Ethical Standards Conflict of Interest Dr. Halpern-Felsher is the Founder and Executive Director of the Tobacco Prevention Toolkit. She is also a paid expert scientist in some e-cigarette litigation and an unpaid scientific advisor and expert witness regarding some tobacco-related policies. Ms. Liu and Dr. Mathur Gaiha have no conflicts of interest to disclose. Financial Disclosure All authors have indicated they have no financial relationships relevant to this article to disclose. Disclaimer The content is solely the responsibility of the authors and does not necessarily represent the official views of the Taube Research Faculty Scholar Endowment or TRDRP. The 2-μm plasmid encoded protein Raf1 regulates both stability and copy number of the plasmid by blocking the formation of the Rep1–Rep2 repressor complex Abstract The 2-μm plasmid of the budding yeast Saccharomyces cerevisiae achieves a high chromosome-like stability with the help of four plasmid-encoded (Rep1, Rep2, Raf1 and Flp) and several host-encoded proteins. Rep1 and Rep2 and the DNA locus STB form the partitioning system ensuring equal segregation of the plasmid. The Flp recombinase and its target sites FRTs form the amplification system which is responsible for the steady state plasmid copy number. In this work we show that the absence of Raf1 can affect both the plasmid stability and the steady sate copy number. We also show that the Rep proteins do bind to the promoter regions of the 2-μm encoded genes, as predicted by earlier models and Raf1 indeed blocks the formation of the Rep1–Rep2 repressor complex not by blocking the transcription of the REP1 and REP2 genes but by physically associating with the Rep proteins and negating their interactions. This explains the role of Raf1 in both the partitioning and the amplification systems as the Rep1–Rep2 complex is believed to modulate both these systems. Based on this study, we have provided, from a systems biology perspective, a model for the mechanism of the 2-μm plasmid maintenance. INTRODUCTION The high copy number 2-m plasmid of the Saccharomyces strains (1,2) is a classic example of selfish DNA elements present in the nucleus without jeopardizing the fitness of its host (reviewed in (3,4)). Even though it confers no obvious advantage to its host, it is among the most stable plasmids with a loss rate of about 10 −5 per cell per generation (5)(6)(7)(8). Owing to its high mitotic stability and copy number, components of the 2-m plasmid are being used successfully for heterologous high copy gene expression in yeast. This near chromosome-like stability is due to two plasmid borne systems, namely the partitioning and the amplification systems comprising of the four proteins (Rep1, Rep2, Raf1 and Flp), a partitioning locus STB and two recombination sites FRTs (9)(10)(11)(12)(13). These proteins and DNA loci are involved in the faithful segregation and the maintenance of steady state copy number of the plasmids. Rep1 and Rep2 proteins along with the cis-acting locus STB form the partitioning system that helps in the equal segregation of the plasmids between the mother and the daughter. If the copy number drops due to missegregation, the amplification system is activated till the steady state copy number is restored. Once the steady state is reached, the amplification system is presumably switched off. The amplification system comprises of the plasmid encoded recombinase Flp and its target sites FRTs. The amplification of the copy number is achieved as the Flp-mediated recombination switches between the theta mode and a rolling circle mode of plasmid replication, as proposed by Futcher (14). A relatively less studied but an important member among the plasmid encoded genes is RAF1. The partitioning and the amplification systems together ensure the stable maintenance of the 2-m plasmid at a high copy number in the cell (6,15,16). However, as the proteins involved in the partitioning and amplification systems are

not mutually exclusive, perturbations in one are likely to affect the other. The communication between the partitioning and amplification systems is likely mediated by the Rep1 and Rep2 proteins. The roles of the Rep proteins in the utilization of several host factors required for equal plasmid partitioning are well documented (3,(17)(18)(19)(20)(21)(22)(23)(24)(25). In addition, Rep1 and Rep2 form a bipartite repressor that regulates plasmid gene expression, in particular, to prevent inappropriate or unregulated plasmid amplification (15,26,27). 7168 Nucleic Acids Research, 2017, Vol. 45, No. 12 Whereas FLP, REP1 and RAF1 genes are subject to downregulation by the Rep1-Rep2 repressor, REP2 appears to be constitutively expressed (27). However, there is no direct evidence for the interaction of the Rep proteins with the plasmid genes presumed to be under their control. Genetic evidence implicates Raf1 in triggering a prompt amplification in response to low plasmid copy number (PCN) by antagonizing the repressor and thus enhancing FLP expression (15). It is not clear whether Raf1 acts by disrupting the formation of the heteromeric repressor, or by blocking the action of the mature repressor at the FLP promoter. The role of Raf1 is not limited to PCN maintenance, but also extends to partitioning (28). Raf1 has been shown to associate with STB (29), and its absence results in increased plasmid missegregation (28). However, the mechanisms by which Raf1 promotes plasmid stability are not understood. In this report, we have analyzed the interrelationships between Raf1 and the Rep1-Rep2 repressor to better understand the interplay between the 2-m plasmid partitioning and amplification systems. We provide direct evidence for the interactions of Rep1, Rep2 and Raf1 with promoters of the 2-m plasmid genes. Results from yeast two hybrid, competitive yeast two hybrid and bimolecular fluorescence complementation (BiFC) assays demonstrate that Raf1 physically interacts with Rep1 and Rep2. Raf1-Rep1 and Raf1-Rep2 interactions are independent of Rep2 and Rep1, respectively. These interactions disrupt Rep1-Rep2 repressor formation, and provide a molecular explanation for Raf1 being a Rep1-Rep2 antagonist. We integrate these protein interactions into a model in which a repressor amplified negative feedback loop sets up cell cycle-dependent oscillations of the Rep1-Rep2 repressor, and possibly of the Flp, Rep1 and Raf1 proteins as well. The rhythmic change in the repressor levels proposed by the model is typical of several biochemical oscillators (30). The model is consistent with the experimental observation that the Rep protein levels increase from a minimum at the G1 phase, peaks at the S/G2 stage and then decreases till the end of the mitosis (11). To explain the role of Raf1 in copy number amplification, it has been hypothesized that Raf1 promotes transcriptional activation of Flp expression by antagonizing the Rep1-Rep2 repressor complex (16,29). Raf1 may either disrupt formation of the heteromeric repressor or may block its role as a negative regulator of Flp gene expression. Raf1 has also been implicated to play a role in the partitioning system to promote plasmid segregation, although the mechanistic details are not well understood (28). In this report we sought to test the above hypothesis to unveil the role of Raf1 in further detail. The communication between the partitioning and amplification systems is likely mediated by the Rep1-Rep2 repressor complex. It generates sufficient scientific curiosity to address how faithful plasmid partitioning and maintenance of a steady state PCN are simultaneously achieved through the Rep1-Rep2-mediated cross-talk between the partitioning and the amplification systems. Functions of the Rep proteins through utilization of the host factors in equal segregation of the plasmids are well documented (3,(17)(18)(19)(20)(21)(22)(23)(24)(25). In a recent study, Prajapati et. al, demonstrated that along with the microtubules and Kip1 motor (23,31), the microtubuleassociated proteins Bik1 and Bim1 play a significant role in the faithful segregation of the 2-m plasmid (32), adding yet another class of host factors to the working model of plasmid segregation. To link Rep1-Rep2 complex with Flpmediated amplification functions, several studies have proposed that the complex forms a bipartite repressor and attenuates Flp activity by repressing its transcription. A feedback repression by the Rep1-Rep2 complex also attenuates the expression of REP1 and RAF1 (15,16,27). Raf1, on the other hand, has been proposed to antagonize Rep1-Rep2 repressor due to its ability to act as a positive regulator of Flp which has been inferred by observing an increased level of FLP transcripts when Raf1 is over-expressed although no effect on the transcript level of REP2 was reported (27). Although it has been demonstrated that the disruption of RAF1 leads to the missegregation of the plasmid (28) and Raf1 has the ability to associate with STB (29), the mechanism by which Raf1 promotes the stability is not yet understood. In this report cells bearing the 2-m plasmids devoid of RAF1 gene were used to demonstrate that Raf1 can influence both the equal partitioning and the copy number of the plasmids. All the interactions pertaining to the transcriptional control of the plasmid encoded genes leading to faithful maintenance have been predicted primarily through indirect evidences based on which a model of plasmid maintenance has been proposed where Rep1-Rep2 complex acts as a repressor of transcription and Raf1 as an anti-repressor (16,27). However, there has been no direct evidence demonstrating the interaction of either the Rep proteins or Raf1 to the promoters to modulate transcription. Moreover, there was no evidence to demonstrate the interaction of Raf1 directly with Rep1-Rep2 complex to attenuate the repressor activity (16,27). In this study we have provided direct evidence to validate the model by demonstrating the interactions of Rep1, Rep2 and Raf1 to the promoters of the 2-m encoded proteins. With the help of yeast two hybrid assay, competitive yeast two hybrid assay and BiFC assay we have also demonstrated that Raf1 physically interacts with both the Rep proteins independently and while doing so it indeed blocks the formation of the Rep1-Rep2 repressor complex. The 2-m plasmid has the capability to revert to a steady state optimum for its stable maintenance in its host. It has been demonstrated that the expression of the Rep proteins is not constitutive and goes through a cyclic change during the cell growth. The cellular concentration of the Rep proteins increases from a minimum at the G1 phase, reaches a maxima at the S/G2 stage and then decreases till the end of the mitosis (11). This rhythmic change in the concentration is suggestive of a system-level characteristic typical of many biochemical oscillators (33). Toward the end, we have proposed that the interactions among the 2-m plasmid encoded proteins form a repressor amplified negative feedback system that maintains a stably oscillating level of the Rep1-Rep2 complex and possibly Flp, Rep1 and Raf1 proteins which is crucial for faithful propagation of the plasmids at a steady state copy number within the host. Reagents, plasmids and yeast strains Reagents are listed in the Supplementary Data. Plasmids are listed and described in Supplementary Table S1. Yeast strains are listed and described in Supplementary Table S2. Bi-molecular fluorescence complementation (BiFC) assay Earlier studies have provided extensive cell biological evidence for the localization and dynamics of Rep1 and Rep2 (11), but the complex itself has never been visualized. A system based on BiFC (34) (see the Supplementary Data for BiFC tagging and Supplementary Figure S5) was developed to visualize the complex and its dynamics. BiFC was confirmed by detecting the fluorescence in the presence of 2% galactose. Cells were grown overnight in SC-Raffinose and were subcultured in fresh SC-Raffinose till the OD 600 reached 0.5-0.8. Dextrose or galactose was then added to a final concentration of 2%, and samples were harvested at different time intervals. The cells were washed with and resuspended in 0.1M phosphate buffer before imaging. Imaging was done with the YFP filter (Zeiss filter set 46 (excitation: BP 500/20; beamsplitter: FT 515; emission: BP 535/30)). Other methods The other methods are described in detail in the Supplementary Data. Raf1 is required for plasmid stability Earlier studies with Raf1 demonstrated its role in the plasmid stability (28) and its ability to bind to the plasmid partitioning locus, STB (29), however, the mechanism through which Raf1 affects plasmid segregation is not well understood. The role that Raf1 plays in the maintenance of plasmid could be through a combined effect of this protein both on the partitioning and the amplification systems. This is because Raf1 also affects the expression of FLP (and hence the PCN (16)) and therefore deletion of RAF1 is not only expected to affect the partitioning but also the amplification system (16,27). It is instructive to note that the stability of the 2-m plasmid is affected by the gene dosages of REP1 and REP2 (10). Therefore, a change in the PCN can alter the relative stoichiometry of Rep1 and Rep2 which, in turn, affects the localization and function of the Rep proteins at STB as well as the function of the Rep1-Rep2 complex as a transcriptional repressor (26). Consequently, the plasmid missegregation is a concerted effect of both the loss of partitioning and the variation in the PCN. Unlike earlier experiment (28), in this study the effect of Raf1 on the 2-m plasmid stability has been visualized using a fluorescently labeled 2-m derived plasmid, pSV1 as described earlier (25,35). pSV1, harboring lac operator array appeared as green foci within the yeast cells expressing lac repressor fused to GFP ( Figure 1A). Budded cells with separated DAPI (post anaphase) were analyzed for the distribution of the plasmid foci. Equal number of foci found in the mother and daughter nuclei of a cell is counted as 'equal segregation' whereas unequal distribution of foci is counted as 'missegregation' or 'no segregation'. Since pSV1 harbors only the partitioning locus STB, it segregates equally in the [Cir + ] cells where Rep proteins are provided in trans from the endogenous 2-m plasmid whereas it missegregates in the [Cir 0 ] cells (36), due to absence of any native plasmid and hence the Rep proteins. To find if Raf1 acts as an independent component of the partitioning system or if it acts through the Flp recombinase, the segregation of pSV1 was visualized and the extent of missegregation or 'no segregation' (unequal distribution of plasmid foci) was measured in wild-type [Cir + ], [Cir 0 ] and raf1Δ strains. The extent of missegregation was found to be much higher in the raf1Δ strain as compared to wild-type [Cir + ] ( Figure 1B). As expected [Cir 0 ] cells showed highest 'no-segregation' population type (all the plasmid foci segregated to the mother, Figure 1B green bar). Further validation of the plasmid segregation analysis was performed using plate-based plasmid stability assay as described in Supplementary Data using another 2-m derived plasmid, YEpLac181 harboring an auxotrophic marker (LEU2) ( Figure 1A). Plasmid stability was assayed by calculating the plasmid bearing cells following a growth in the non-selective media. Consistent with the cell biological assay the instability (i%) of YEpLac181 was found to be minimum for the [Cir + ] wild-type strain but the instability increased gradually and consistently for raf1Δ, and [Cir 0 ] strains ( Figure 1B). To analyse if Raf1 functions independent of its role in Flp expression, the plasmid instability was also measured for flpΔ and raf1ΔflpΔ double deletion strains. Comparison of the average instability of the plasmid in raf1Δ with raf1ΔflpΔ strain revealed a slight difference (paired t-test P = 0.14, 95% confidence interval) while comparison of flpΔ and raf1ΔflpΔ strains revealed a significant difference (paired t-test P = 0.04, 95% confidence interval) in the average instability ( Figure 1C). These results are consistent with the earlier study (28) and suggest that Raf1 might have a role in promoting equal segregation of the plasmids independent of its effect on Flp and hence PCN. Surprisingly, Raf1 when over-expressed from a GAL10 promoter also showed an increased plasmid loss rate much higher than the [Cir + ] wild-type strain ( Figure 1C). This increased instability was not due to the growth medium (SC-galactose) since there was no difference in the instability (paired t-test P = 0.299, 95% confidence interval) when raf1Δ was grown in either YPD or SC-galactose. Raf1 influences plasmid amplification and hence the copy number A direct consequence of FLP activation function of Raf1, as proposed in the previous study (15), suggests that there should be an increase in the PCN when RAF1 is deleted. To measure this perturbation, we