Association Between Stress and Bilateral Symmetrical Alopecia in Free-Ranging Formosan Macaques (Macaca Cyclopis) in A High-Population-Density Area

Since 2013, a high incidence of bilateral symmetrical alopecia has been observed in free-ranging Formosan macaques (Macaca cyclopis) in Mt. Longevity, Taiwan. We hypothesized that stress induces alopecia in this population. To verify our hypothesis, we evaluated the histopathological characteristics of skin biopsy and used a validated enzyme immunoassay (EIA) for fecal glucocorticoid metabolite (FGM) analysis, which act as an indicator of stress experienced by the individual. Follicular densities were lower (2.1–3.0 mm 2 ) in individuals with symmetrical alopecia than in those with normal hair conditions (4.7 mm 2 ). Furthermore, anagen to catagen/telogen ratios were lower in individuals with alopecia (0–1.4) than in those with normal hair (4.0). The histopathological characteristics of alopecia were similar to those of telogen e�uvium, which indicates that stress is one of the possible etiologies. On the basis of the analytical and biological validation of EIAs for FGM analysis, 11β-hydroxyetiocholanolone was considered suitable for monitoring adrenocortical activity in both sexes of Formosan macaques. The mean concentrations (standard error; sample size) of 11β-hydroxyetiocholanolone were 2.02 (0.17; n = 10) and 1.41 (0.10; n = 31) μg/g for individuals with and without alopecia, respectively. Furthermore, the results of logistic regression analysis show that 11β-hydroxyetiocholanolone (p = 0.012) concentration was positively associated with alopecia. Thus, stress was the most likely to trigger symmetrical alopecia in Formosan macaques in Mt. Longevity. Although stress can decrease the �tness of an individual, it should not impact the total population of Formosan macaque in Taiwan. Nonetheless, stress-induced immunosuppression might increase zoonosis risk due to frequent human–macaque contact in Mt. Longevity.


Introduction
Animal hair functions as an anatomical and physiological barrier for the animal against environmental hazards.Alopecia or hair loss is an indicator of severe health problems.Alopecia is commonly noted in captive macaques 1 .Factors that contribute to alopecia onset in nonhuman primates have been categorized as natural causes, such as aging, and seasonal changes, psychological factors 2 , nutritional de ciency 3 , hormonal imbalance 4 , immunological diseases 5 , and allergic reactions 6 .However, psychological factors were considered to lead to alopecia through the stimulation of abnormal hair-plucking behavior in macaques 7 .Recent studies have indicated that psychological stress is a direct cause of alopecia and is positively correlated with alopecia severity in rhesus monkeys 1,8 .Studies on mice have indicated that stress can induce telogen e uvium leading to alopecia through the possible mechanism of production of catagen-inducing hair growth inhibitors and hair-damaging proin ammatory cytokines 9 .
Stress is the physiological and psychological response of vertebrates to environmental stressors that lead to homeostasis imbalance 10 .Physiologically, the secretion of catecholamine and glucocorticoids (GCs) from the adrenal medulla and adrenal cortex, respectively, is the primary endocrine mechanism that mediates stress response 11 .Environmental stress condition facilitates a coping mechanism in animals through the immediate (within seconds) effect of catecholamine and a slow but long-lasting effect of GCs, such as cortisol or corticosterone, which ensure the survival of animals when exposed to stressors for a short term 12 .Nonetheless, prolonged exposure of animals to stressors results in a high concentration of GC secretion, which has been shown to result in decreased individual survival and tness through immunosuppression, muscle wasting, and reduced reproduction 10,12,13 .Therefore, GCs are commonly used to monitor chronic stress and identify possible stressors in both captive and free-ranging wild animals 14 .
After GCs are secreted from the adrenal cortex into the blood, they are metabolized by the liver and kidney and eliminated from the animal's body through urine or feces 15 .For measuring the amount of GCs secreted, collecting noninvasive samples, such as feces 16 , is recommended and is commonly adopted for avoiding the confounding increase in GCs during the restriction period for collecting blood samples from animals 17 .In addition, fecal samples can be collected repeatedly and easily without affecting the behavior and endocrine status of animals, which allows for the long-term and repeated monitoring of changes in GC concentrations in response to a speci c stimulus.Furthermore, GC metabolites accumulate in feces over a certain period, and their concentrations are not considerably affected by periodic uctuations or the pulsatility of GC secretion 17,18 .This property is particularly advantageous for studying wild animals 16,19 .
Although using fecal glucocorticoid metabolite (FGM) concentration to represent the stress status of an animal has its advantages, problems related to FGM analysis must be considered before the analysis.GCs are metabolized before elimination from an animal's body, and therefore, little or no native GCs may remain in feces or urine 20 .Furthermore, the primary route of the excretion of GC metabolites through urine or feces considerably varied among species 21,22 .Moreover, the metabolic process of GCs can substantially vary between species and sexes and even between phylogenetically closely related species, resulting in numerous different metabolites 18,20,23 .Overall, validating all the procedures of FGM analysis for each speci c species and sex is essential to accurately measuring FGMs and determining the stress status of individuals 15,18,22,24 .Usually, hormone concentrations are identi ed using immunoassays with hormone-or hormone-group-speci c antibodies to bind with the target hormone or hormone metabolites.Because native GCs are negligible or nonexistent in feces, selecting appropriate antibodies based on analytical and physiological (or biological) validation is crucial to obtaining meaningful results 14,25,26 .Since 2013, a high incidence of bilateral symmetrical alopecia was observed in free-ranging Formosan macaques (Macaca cyclopis) (Fig. 1) in Mt.Longevity (Fig. 2), Shoushan National Nature Park (SNNP), Taiwan.However, the cause of alopecia was unknown.We suspected that stress induced alopecia in Formosan macaques.We performed a skin biopsy of individuals with severe symmetrical alopecia for histopathological assessment to determine the possible etiology.In addition, to clarify the association between alopecia and GC secretion, we rst evaluated the validity of four enzyme immunoassays (EIAs) for monitoring adrenocortical activity through the quanti cation of FGMs in Formosan macaques.Furthermore, we used the validated EIAs to determine FGM concentrations of fecal samples collected from individuals with and without alopecia.We hypothesized that the level of GCs excreted was positively associated with bilateral symmetrical alopecia in Formosan macaques.

Ethics statement
The permission for trapping Formosan macaques in Mt.Longevity was obtained from the Forest Bureau (Permit no.: COA, Forestry Bureau, 1041701001).The procedures for trapping Formosan macaques in Mt.Longevity, administering anesthesia, and collecting samples were approved by Institutional Animal Care and Use Committee (IACUC) of National Pingtung University of Science and Technology (Approval no.: NPUST-104-021).All the procedures that involve live animals were performed in compliance with the IACUC regulations and Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines.No other permit was needed for collecting fecal samples.All details on sampling methods are provided in the following sections.

Study site
This study was conducted in Mt.Longevity, which is located in SNNP (latitude: 22.650 and longitude: 120.260) on the west side of Kaohsiung city (Fig. 2).A city and an ocean isolate Mt.Longevity from other natural habitats.The terrain of Mt.Longevity is characterized by uplifted coral reefs and limestone covered with natural forest.Formosan macaques are endemic to Mt. Longevity.The estimated population size of Formosan macaques ranged from 1200 to 1500 in the sampling area from 2012 to 2015 27 .The density of Formosan macaques in Mt.Longevity was high for a long period owing to food provision for a long term history 28 .The average density of Formosan macaques on Taiwan island was 0.72 troops/km 229 ; however, its population density in Mt.Longevity was 5.72 troops/km 230 .Food provision to macaques is prohibited, and a law was strongly enforced against food provision after the establishment of SNNP in 2011, which may have limited the food for macaques in Mt.Longevity.

Histopathological evaluation of alopecia
Free-ranging Formosan macaques with symmetrical alopecia were trapped for the histopathological evaluation of skin tissues with alopecia.We used steel-mesh box traps (Tomahawk Live Trap, LLC., Hazelhurst, WI, USA) modi ed for remote access, and a banana was used as a bait.
Individuals with severe symmetrical alopecia were trapped.The trapped macaques were anesthetized by veterinarians with a mixture of dexmedetomidine hydrochloride (25 µg/kg) and tiletamine HCl/zolazepam HCl (2 mg/kg).Skin tissues of the back area with alopecia of approximately 1 cm 2 were surgically collected and xed in 10% neutral buffered formalin for further histopathological evaluation.Vertical and transversal sections at the reticular dermis layer of skin biopsy were obtained and stained with hematoxylin and eosin stain.We examined the histopathological changes in skin sections to evaluate the possible etiology of alopecia.In addition, transversal skin sections were used for obtaining follicular counts and determining follicular stages, namely anagen, catagen, and telogen, based on Whiting 31 .We estimated follicular density per square millimeter by dividing the total count of follicles by the area covered in the image by using ImageJ software 32 .The follicles on the edge of the image were not counted.Moreover, we grouped catagen and telogen follicles together and estimated the anagen to catagen/telogen ratio 33 .

Stress manipulation of Formosan macaques in captivity
We adopted a biological validation procedure to induce GC secretion through restraint and anesthesia procedures 25,34,35 for routine health evaluation in captivity.We then assessed whether the increased GC secretion could be detected based on different EIA FGM quantities.In addition, we evaluated the capacity and speci city of different antibodies to quantify FGM abundance through integration with high-performance liquid chromatography (HPLC) analysis.Lastly, we compiled the results of the EIA and HPLC to evaluate the suitability of four EIAs in assessing adrenocortical activity in Formosan macaques.
For validating the EIAs, fresh fecal samples were collected for 5 consecutive days from one male and one female captive adult Formosan macaque housed individually in the Pingtung Rescue Center for Endangered Wild Animals (PTRC), National Pingtung University of Science and Technology.
The individual cages used for housing the macaques were 2 m high and 1 m wide and deep; the upper area of the cage had a platform for resting.
Pooled fecal samples of one day of each individual were collected by the PTRC staff before and after an anesthesia procedure for routine health evaluation of each individual.Anesthesia was administered to stimulate stress (physical challenge) in the macaques on the third day of sample collection.Collected feces were stored at −20°C immediately until analysis.The feces collected before the health check were used for baseline analysis, and the samples collected after the health check were used for evaluating the adrenal responses to stress.
Fecal sample collection from Formosan macaques in Mt.Longevity We collected fecal samples of Formosan macaques with and without alopecia by following troops with alopecia individuals in Mt.Longevity.When eld crews observed a macaque defecating, a fecal sample was collected and stored in a coolbox immediately.The feces were then stored at −20°C.We recorded the age, sex, and status of individuals with alopecia while collecting fecal samples.Age was recorded as juvenile, subadult, and adult according to the classi cation by Hsu and Lin 36 .We did not collect sample from infant in this study due to the limitation of availability.

Extraction of fecal hormone metabolites
Collected feces were transferred to 15-mL centrifuge tubes and covered with para lm.Holes were made on the para lm to allow evaporation.Lyophilization was then performed for 72 h.Dehydrated feces were ltered to exclude large particles, and the rest were ground to powder.The fecal powder of 50 mg was transferred to new 15-mL centrifuge tubes, and 3 mL of 80% methanol was then added.After vortexing for 10 min and centrifuging at 3000 rpm for 10 min, the supernatant was transferred to a 5-mL vial, labeled, and stored at −20°C until further analysis.

HPLC analysis
The samples collected at the PTRC from one male and one female macaque were used for reverse-phased high-performance liquid chromatography.To purify the fecal extractions, 1 mL of the fecal extract was mixed with 3 mL of sodium acetate buffer (0.2M, pH 4.2) and ltered using solid-phase extraction (SPE) cartridges (Sep-pak C18 cartridge 1 g, Waters, MA, USA).The SPE cartridge was activated by pumping 10 mL of 100% methanol and 10 mL of double-distilled water slowly with a syringe before addition of the extraction mixture.After the total volume of the extraction mixture was passed, 10 mL of double-distilled water was pumped using a syringe to wash the cartridge.The elution was discarded, and the cartridge was left to dry at room temperature for 3 h.After drying, 10 mL of methanol (100%) was added to elute the extraction from the cartridge.The elution was collected and evaporated to dryness through nitrogen sweeping and then reconstituted in 400 μL of 40% Acetonitrile (ACN) buffer for the HPLC process to separate the steroids.The HPLC settings include a ow speed of 0.4 mL/min and ACN:H 2 O ratio of 40:60 as mobile phase before injection of 100 μL of the puri ed extraction.Fractions (n = 80, 400 μL each) were collected every minute and were evaporated to dryness.An assay buffer (1 mL) was then added to each fraction and stored at −20℃ before the EIA analysis.Moreover, the standards of the four selected FGMs (cortisol, corticosterone, 11-oxoaetiocholanolone, and 11β-hydroxyetiocholanolone) were separated into fractions with HPLC to obtain elution positions for reference.

Enzyme immunoassays
Four competing EIAs, namely cortisol, corticosterone, 11-oxoaetiocholanolone, and 11β-hydroxyetiocholanolone, were used to detect targeted steroids and metabolites of HPLC elution and fecal extractions, as described by Heistermann, et al. 25 .The antibody, standard, and biotin-labeled steroid of each target used in this study were kindly provided by Dr. Palme from the Department of Biomedical Sciences, University of Veterinary Medicine, Vienna, Austria.The cross-reactivities of all antibodies applied in this study were described by Heistermann, et al. 25 .The EIAs were performed using 96-well microtiter plates (MTPs) coated with protein A (P-7837, Sigma-Aldrich, Vienna, Austria), according to procedures described by Palme and Möstl 37 .In brief, samples were diluted 20 times with an EIA assay buffer before reaction.Biotin-labeled steroids (100 μL) and diluted samples or standards (50 μL), as well as antibodies (100 μL), were added to each well and incubated overnight at 4°C.An enzyme solution (250 μL/well) containing streptavidin-horseradish peroxidase conjugate (RPN1231, GE Healthcare) was added to the plate after four washes with 0.02% Tween 20 (VE-V900548, Vetec) and then incubated with mild shaking in the dark at 4°C for 45 min.The plate was washed again before the addition of 250 μL of substrate solution containing tetramethylbenzidine (SI-T2885, Sigma-Aldrich) and incubated with mild shaking in the dark for another 45 min at 4°C.After incubation, 50 μL of 10% H 2 SO 4 was added to each well to stop the reaction, and the plate was read using a multimode microplate reader (Corona Electric Co., Ltd, Japan) within 30 min to measure absorbance at 450 and 620 nm as measuring and reference lters, respectively (Fig. S1).The differences in both readings were used for obtaining the standard curve formula, which was then used for the concentration calculation of each sample.Intra-and interassay coe cients of variance (CVs) of low and high concentration pool samples were calculated for quality control in EIA analysis.Only MTPs with less than 15% CV were considered valid results 38 .

Suitability of EIAs
The suitability of EIAs for analyzing the adrenal activity response of Formosan macaque to stress was assessed based on four criteria according to Heistermann, et al. 25 : (1) tremendous FGM response after physiological challenge, (2) substantial immunoreactivity in HPLC elution, (3) no indication of comeasurement of fecal androgen metabolites, and (4) low variation in baseline levels.

Data analysis
We used a multivariate logistic regression analysis to evaluate the relationship between alopecia, stress, and other possible covariates, such as age and sex.The analysis was conducted using the Generalized Liner Model (GLM) module in the computing environment R (R Development Core Team, 2010).The alopecia status during fecal sample collection was treated as the dependent variable.The fecal concentrations of 11oxoaetiocholanolone and 11β-hydroxyetiocholanolone were treated as explanatory variables.Furthermore, age and sex variables were treated as possible confounders for data analysis.We rst transformed all the explanatory variables to dummy variables and centered or standardized the numerical variables.Multicollinearity between the explanatory variables was evaluated using the variance in ation factor (VIF) 39 .The VIF threshold was set as 10 to avoid the problematic effect of multicollinearity on parameter estimations 40 .Variables were discarded from model construction if the VIF value was more than 10.The explanatory variables selected in the models were based on the Wald test with a p threshold value of 0.05 41,42 .We compared the model t based on the Akaike information criterion (AIC), with lower AIC values indicating a better model t to the dataset.

Histopathological evaluation of alopecia
We trapped three Formosan macaques of different ages, sexes, and rankings of severe symmetrical alopecia (Table 1).The skin sample of a roadkilled Formosan macaque in the study area with normal hair appearance was used to compare the histopathological changes in skin samples of symmetrical alopecia Formosan macaques.In the vertical sections of skin tissues of alopecia individuals, we did not observe the in ltration of in ammatory cells of skin tissues and pathogen infection, such as infection with fungi and ectoparasites.Therefore, alopecia due to infection and scarring was ruled out (Fig. 3).The transverse skin sections of individuals with symmetrical alopecia revealed lower follicular densities (2.1-3.0 mm 2 ) than in those of individuals with normal hair conditions (4.7 mm 2 ; Table 2).In addition, anagen to catagen/telogen ratios were lower in alopecia individuals (0-1.4)than in those with normal hair (4.0).A lower anagen to catagen/telogen ratio indicated a higher proportion of catagen and telogen follicles in individuals with symmetrical alopecia (Table 2; Fig. 4).

Validation of FGM analysis with HPLC and EIA
The characteristics of four EIAs and variation coe cients of intra-and interassay for each FGM EIA assay are presented in Table S1.The results of the four EIA analyses of all HPLC fractions collected from one male and one female are shown in Fig. 5.The 11-oxoetiocholanolone was mainly eluted at positions 18, 23, and 37-38 in both sexes.11β-Hydroxyetiocholanolone was mainly eluted at positions 20, 26, and 37-39 in both sexes.Both cortisol and corticosterone were presented in low amounts in all fractions, and they neither showed an obvious peak after EIA analysis nor coelution with authentic cortisol at position 14 and with authentic corticosterone at position 22 (Fig. 5).Most of the immunoassay reactions were located before position 40, which has been referred to as elution positions of GC metabolites in several studies 25,43,44 .

FGM baseline and response to biological challenge
The FGM concentrations in feces collected on the days before health checks in all individuals exhibited low variation at baseline, except for cortisol, and a substantial increase after health checks (Fig. 6), ranging from 1.8-to 6.9-fold above the baseline (Table 3).The peaks of each FGM appeared on the rst to third days after anesthesia procedure in all individuals and gradually returned to baseline.

Suitability of EIAs
The results of all EIAs indicated that the response of FGMs increased after anesthesia stimulation from 1.8-to 6.9-fold.However, only 11βhydroxyetiocholanolone and 11-oxoetiocholanolone showed measurable peaks in the HPLC pro les.Furthermore, coelution with androgen metabolites was not observed in HPLC fractions in all EIAs (Fig. 6).In addition, the baseline values of all tested individuals showed low variations, except cortisol.On the basis of the four criteria for the FGM EIA assay validation proposed by Heistermann, et al. 25 , we determined that cortisol and corticosterone were unsuitable EIAs for FGM analysis in Formosan macaque.The response of 11-oxoetiocholanolone after anesthesia stimulation was only 1.8-fold more than the baseline value in the male macaque.Therefore, we considered that 11β-hydroxyetiocholanolone was suitable for FGM analysis for monitoring adrenocortical activity in both sexes of Formosan macaques.However, 11-oxoetiocholanolone was suitable for use only for the female macaque (Table 3).

Analysis of FGMs from Formosan macaques with and without alopecia by using EIAs
We collected fresh fecal samples of identi ed individuals by following the movement of troops from August 2015 to August 2016.In total, 41 fecal samples were collected from 10 individuals with alopecia and 31 individuals without alopecia (Table 4).We selected the EIAs of 11oxoaetiocholanolone and 11β-hydroxyetiocholanolone for FGM assay in Formosan macaques based on EIA validation assays.Mean concentrations (standard error) of 11-oxoaetiocholanolone in individuals with and without alopecia were 1.47 (0.16) and 1.12 (0.07) μg/g, respectively (Table 3).Furthermore, mean concentrations (standard error) of 11β-hydroxyetiocholanolone in individuals with and without alopecia were 2.02 (0.17) and 1.41 (0.10) μg/g, respectively.
We adopted a logistic regression model to evaluate the association between alopecia and explanatory variables, including age, sex, concentrations of 11-oxoaetiocholanolone and 11β-hydroxyetiocholanolone.No variable was discard from the logistic regression model based on the VIF value.Results indicated that the FGM concentrations of 11-oxoaetiocholanolone (p = 0.04) and 11β-hydroxyetiocholanolone (p = 0.012) were signi cantly positively associated with alopecia when the variable was analyzed alone (Table 5).The AIC value of 11β-hydroxyetiocholanolone was lower (41.53)than that of 11-oxoaetiocholanolone (44.76;Table 5).Age and sex were not associated with alopecia (Table 5).In addition, we evaluated the full model, including the explanatory variables of 11-oxoetiocholanolone, 11β-hydroxyetiocholanolone, age, and sex, tted in the logistic regression.In the full model, 11β-hydroxyetiocholanolone was the only variable that showed a signi cant positive association with alopecia.The AIC value of the full model was 42.3, which was higher than that of the model with 11β-hydroxyetiocholanolone tted alone.The results indicated that 11β-hydroxyetiocholanolone better explained the mamcaque alopecia in our study.

Discussion
This study was conducted to identify the possible etiology of symmetrical alopecia in free-ranging Formosan macaques in Taiwan.To our knowledge, a massive bilateral symmetrical alopecia has never been described in a free-ranging population.Moreover, although bilateral symmetrical alopecia has been commonly observed in macaques in captivity, the etiology is still largly unclear 7,45 .For identifying the etiology, we rst evaluated the histopathological changes in skin tissues with symmetrical alopecia and observed that the histopathological characteristics of alopecia were similar to those of telogen e uvium 46,47,48 .The environmental stressor that induced telogen e uvium was suspected based on the observation of individuals with symmetrical alopecia.For assessing the relationship between symmetrical alopecia and stress, we validated the EIA of 11β-hydroxyetiocholanolone for Formosan macaque FGMs using noninvasive fecal samples.In addition, we found that the concentration of 11β-hydroxyetiocholanolone was positively associated with symmetrical alopecia.
Physiological or emotional stress is commonly identi ed as a cause of telogen e uvium in humans 49 .Other potential triggers identi ed in humans include febrile illness, severe trauma, postpartum hormonal changes, malnutrition, hypothyroidism, and iron de ciency 50 .Studies in mice have revealed that stress can induce profound hair growth inhibitory and hair-damaging proin ammatory effects, which generate a telogen e uvium pattern 9,51 .Telogen e uvium alopecia have been described in squirrel monkey and rhesus macaques 47,52 .However, both studies have concluded that stress was not the potential trigger based on the dominance hierarchy and blood cortisol concentration 47,52 .The assessment methods adopted in their studies might be incapable of determining the true status of the stress level in individuals because environmental stressors and restrictioninduced stress was not controlled.
The association between stress and alopecia has been demonstrated in different macaque species.A histopathological evaluation has rarely been conducted to identify the alopecia pattern 7,8,53,54 .However, many factors can contribute to alopecia in macaques 7 .Therefore, a skin biopsy of the alopecia site has been commonly recommended for histopathological examination and identi cation of the potential factor that triggers alopecia 7,46 .In this study, febrile illness, severe trauma, postpartum hormonal changes, malnutrition, and hypothyroidism were ruled out as potential causes of telogen e uvium based on the eld observation and serum chemical analysis of three trapped individuals (data not shown).
Stress was considered to be the primary factor inducing telogen e uvium in this study.Moreover, the logistic regression analysis showed that stress was associated with symmetrical alopecia in Formosan macaques in Mt.Longevity.The concentrations of both 11-oxoaetiocholanolone and 11β-hydroxyetiocholanolone were associated with symmetrical alopecia in the logistic regression analysis when each variable was tested alone.However, because 11-oxoaetiocholanolone was not considerably associated with alopecia in the multivariable analysis, and because 11βhydroxyetiocholanolone alone was the best t model compared with other models, 11β-hydroxyetiocholanolone was the most important explanatory variable and it implies that stress may be one of the primary factors of alopecia in the studied population.
FGM assay validation in a speci c species or different sexes of species before adopting the assay for measuring the FGM concentration was recommended and commonly followed as the standard procedure 55 .The validation of FGM analysis is described in different macaque species (Table S3).However, many studies have analyzed macaque FGMs by using assays without appropriate analytical and physiological (or biological) validation 55 .Analytical validation measures the precision, accuracy, sensitivity, and speci city, which might be in uenced by the cross-reaction of antibody to other substances 19 .Physiological validation is conducted through the pharmacological stimulation of hypothalamic-pituitary-adrenal (HPA) axis activity.The stimulation affects the amount of GC secretion into the blood, and then noninvasive samples (urine, saliva, or feces) that re ect the predicted changes of GCs in the blood are measured and validated 56 .The Adrenocorticotropic hormone (ACTH) challenge and dexamethasone suppression tests have been commonly adopted for physiological validation 55,56 .However, these methods have disadvantages, particularly in wild animals, such as unknown dosage for inducing the HPA axis activity and special permission required to conduct animal experimentation 55 .In this case, biological validation, which involves exposing animals to stressors such as restraint, anesthesia, or transportation, is recommended 15,18,19,56 .A proper FGM assay has never been validated and established for Formosan macaques.We evaluated the suitability of four EIAs for FGM assay in this study and determined that the 11β-hydroxyetiocholanolone EIA was suitable for Formosan macaques based on analytical and biological validation.In addition to Formosan macaque, the 11β-hydroxyetiocholanolone EIA is a validated FGM assay in various macaque species (Table S3).The validated 11β-hydroxyetiocholanolone EIA could be applied for the stress assessment of Formosan macaque in wild and captivity.
In this study, we identi ed that stress was the most likely factor to trigger symmetrical alopecia in a wild Formosan macaque population in Mt.
Longevity.The high density of Formosan macaque population in Mt.Longevity was suspected to be a stressor; however, it remained unclari ed.The population of Formosan macaque is stable in Taiwan.Although stress can decrease the tness of an individual, it should not impact the total population of Formosan macaque in Taiwan.Nonetheless, stress-induced immunosuppression might increase zoonosis risk in Mt.Longevity, where human-macaque contact is frequent 57 .

Declarations Tables
Due to technical limitations, table 1 is only available as a download in the Supplemental Files section.Formosan macaque with and without symmetrical alopecia as the dependent variable.

Table 2 .
Estimation of follicular density, stages, and anagen to catagen/telogen ratio for the skin biopsy of Formosan macaques with symmetrical alopecia in comparison with individuals with normal hair appearance.
a A/CT ratio denotes anagen to catagen/telogen ratio.

Table 3 .
25crease in FGMs (in folds) compared with baseline measured using four EIAs in response to anesthesia stimulation in both sexes.Numbers in bold indicate that the EIA was considered suitable for monitoring adrenocortical activity in a speci c sex of Formosan macaques.The suitability evaluation was based on the study by Heistermann, et al.25.

Table 4 .
Individual characteristics of fecal samples of Formosan macaques with and without alopecia in Mt.Longevity and the results of FGM concentration through EIA analysis.
1Dry weight of fecal sample2Mean and standard error of 11-oxoetiocholanolone3Mean and standard error of 11β-hydroxyetiocholanolone

Table 5 .
Explanatory variable statistics and AIC values of each variable tested alone and the full model in the logistic regression analysis by using