Neutrophil to lymphocyte ratio in captive olive baboons (Papio anubis): The effects of age, sex, rearing, stress, and pregnancy.

Subjects included a total of 387 captive olive baboons (59% female) housed at the Michale E. Keeling Center for Comparative Medicine and Research at The University of Texas MD Anderson Cancer Center in Bastrop, Texas. Baboons were housed in corrals or Primadomes ™ with inside access, or in indoor/outdoor runs, in two separate colonies on campus (see Figure 1 ). The first colony is the Specific Pathogen Free (SPF) colony (n=330) and the second is the conventional colony (non-virus-free) (n=57). Table 1 depicts the number of baboons across colony, sex, and rearing status. Across the entire sample, baboons ranged in age from 0 to 19 years old, including 265 juveniles (aged 0–4 years), 43 young adults (5–9 years), 52 older adults (10–14 years), and 25 geriatric individuals (15+ years). Baboons were housed in breeding groups consisting of one or two breeding males, 12–16 breeding females, and, in the SPF colony, their juvenile and infant offspring (0–3 years of age). Of the 387 baboons, 173 were mother-reared, 209 were nursery-reared, and five had an unknown rearing history. Nursery-reared individuals were defined as those baboons that were separated from the dam within 24 hours following birth and cared for by humans, raised in an incubator with access to human infant formula until they were put into small, same-age peer social groups until 2 years of age, when they were introduced to larger adult and sub-adult social groups. Mother-reared individuals were defined as baboons that were not separated from their dam for at least the first 6 months of life and were reared in their natal group during that time. Routine NLR values were taken from hematology records obtained during routine biannual physical exams conducted between 2017 and 2021. The sedation process included gathering the baboon social group into the indoor portion of the enclosure and subsequently funneling individuals through a mesh tunnel system one by one for a ketamine injection. Blood was taken when the animal was immobilized following injection. Only values from physical exams in which the animal was not injured, recovering from an injury, or pregnant (as dictated in the exam notes) were included as routine NLR values. To calculate NLR for each baboon, we divided percent values for neutrophils by percent values for lymphocytes, as we have done previously ( Neal Webb et al., 2020 ). Age of the baboon at the time at which NLR values were obtained was used as the age variable. All research and experimental protocols complied with those approved by the UTMDACC Institutional Animal Care and Use Committee, and complied with the legal requirements of the United States and the ethical guidelines put forth by AALAS, the Animal Welfare Act, and The Guide for the Care and Use of Laboratory Animals. The Keeling Center has been fully accredited continuously since 1979 by AAALAC. Between 2017 and 2018, baboons were transported from the University of Oklahoma Health Sciences Center (OUHSC) in Oklahoma to the Michale E. Keeling Center for Comparative Medicine and Research (KCCMR) in Bastrop, Texas. We calculated NLR from arrival/quarantine exams performed within 1 week of arrival to the KCCMR (referred to as “transport NLR” throughout). Transport NLR values were calculated for 156 baboons, including 126 females and 30 males. Transport NLRs were compared with routine NLRs (described above), which were obtained within 1 year of arrival during routine physical exams. We used clinical records to identify 81 exams at which female baboons were found to be pregnant. During routine physical exams, veterinarians conduct pregnancy ultrasounds and, in cases of pregnancy, measure the biparietal diameter (BPD) of the fetus. This measurement is then recorded in the animals’ clinical data. In our dataset, BPDs, which indicate length of gestation, ranged from 0 to 59.3 mm (60 mm is considered full term). We then calculated the NLR and age of the female from that physical exam. We also included control NLR values to compare to the gestational NLR values by finding the NLR value from the physical exam that was closest to the date of the gestational NLR. NLR values were not taken from every animal during every physical exam, so we used the NLR that was closest in date (either before or after) to the gestational NLR value date. Therefore, BPD, NLR at pregnancy, control NLR, and age at pregnancy were included as variables in the dataset. All pregnancies included in the dataset resulted in successful parturition. We used SPSS statistical software for all analyses (IBM, 2021). Data are available from the corresponding author upon reasonable request. Histograms and Q-Q plots showed that the data were positively skewed. Therefore, we used a Log10 transformation for all NLR outcome variables in our analyses (denoted with subscript “lg10”), and report mean and standard error of the mean for both the raw and log-transformed data. Visual inspection of residuals and QQ plots of residuals by fitted values for log-transformed data showed good homoscedasticity. Unless otherwise noted, analyses were limited to SPF baboons. Given that previous research has shown differences in lymphocyte counts between virus-free and non-virus-free baboons ( Oxford et al., 2015 ), we included colony (SPF vs conventional) as a predictor variable. We examined the effects of rearing, sex, and colony (conventional vs SPF) on NLR using a regression with sex (0=female, 1=male), rearing (0=mother-reared, 1= nursery-reared), and colony (0=SPF, 1=conventional) as predictors. We were unable to test for interactions given the limitations of our sample, including 1) confounded sex*age due to a low number of males over the age of 5 years; 2) confounded colony*sex*rearing due to a low number of mother-reared individuals as well as the low number of males in the conventional colony; and 3) confounded age*colony due to fact that the oldest animals are housed in the conventional colony. Therefore, we opted for the regression to examine each effect separately in the model, as it represents the individual effect (or unique contribution) of sex, colony, and rearing while controlling or adjusting for effects of the other variables ( Judd, McCellan, Ryan, 2011 ). For example, the coefficient for sex in the model represents the effect of sex while controlling for colony and rearing. Due to the low number of adult males in the colony, we utilized SPF females of all ages to determine whether NLR was predicted by age using a linear regression with NLR as the outcome variable and age as the predictor. Given the quadratic relationship between age and NLR in chimpanzees found by Neal Webb et al. (2020) , we then used curve estimation to further examine the linear and quadratic models of NLR. Using a subset of data for which we had transport NLRs (N=156, 78% female, 58% nursery-reared), we used a repeated-measures ANOVA with age group as the between-subjects factor to examine the difference between transport NLR values and NLR values taken during routine physical exams. Given that this was a within-subjects design, we used both conventional (n=46) and SPF (n=110) baboons in this analysis. We used the software program SOLAR (Sequential Oligogenic Linkage Analysis Routines) ( Almasy & Blangero, 1998 ) to estimate NLR heritability from the conventional and SPF baboon colonies at KCCMR using subjects for which we had pedigree information. Pedigree information consisted of known dam and sire for each subject (with unknown cases of paternity identified with paternity testing), going backwards three to six generations. The phenotypes examined included baseline NLR and transport NLR. For these analyses, we limited the sample to 156 subjects from whom we had baseline and transport values. Covariates included age, sex, rearing history, sex*age, sex*rearing history, and age*rearing history. To avoid convergence failure, rather than using the Log10 transformation, we transformed both the baseline NLR and transport NLR variables using the “inormalize” function in SOLAR. This is an inverse normalization used for heritability estimates that results in a more normally distributed variable that retains most of the original information relevant to heritability and linkage analysis ( Almasy & Blangero, 1998 ). To examine differences in NLR as a function of pregnancy, we examined differences in gestational NLR and control NLR values using a repeated-measures ANCOVA with age at pregnancy as a covariate. Given that this was a within-subjects analysis, we used both conventional (n=23) and SPF (n=59) baboons in this gestation analysis. Lastly, we examined cross-sectional differences in NLR as a function of trimester using a univariate ANCOVA with NLR at pregnancy as the dependent variable, trimester as the between-subjects factor, and age as a covariate. Because this was cross-sectional data, we used only SPF females in this analysis (N=61)

Histograms of routine NLR across the sample and transport NLR are shown in Figure 2 . As shown in the top panel, 95% of NLR values fell between 0.43 and 6.50 (mean = 2.37, standard deviation = 2.14, SEM=0.11, median = 1.63). Table 1 presents the mean NLR ± SEM values across sex and rearing status. Means (M) and standard errors of the mean (SEM) from the raw data as well as means and SEM from Log10 transformed data are reported in the results below. The regression examining whether NLR lg10 was predicted by sex, rearing, colony, and age was significant, F(4,377) = 11.59, p = 0.0001, R 2 adj = 0.10. As shown in Table 2 , the coefficients were significant for all effects included in the model. While adjusting for sex, rearing, and colony, age significantly predicted NLR lg10 : NLR lg10 was higher in older baboons ( Figure 3 ). While adjusting for age, rearing, and colony, sex significantly predicted NLR lg10 : males had lower NLR than females. While adjusting for sex, age, and colony, rearing significantly predicted NLR lg10 : mother-reared individuals had higher NLRs than nursery-reared baboons. Lastly, while adjusting for rearing, sex, and age, colony significantly predicted NLR lg10 : baboons in the SPF colony had significantly higher NLR than those housed in the conventional colony. The regression examining whether NLR lg10 was predicted by age in female baboons was significant, F(1,178) = 9.17, p=0.003, R 2 adj = 0.044, beta = 0.018, t(178) = 3.03, P =0.003. We further examined the effect of age on NLR lg10 using a curve estimation procedure. There was a significant linear and quadratic relationship between age and NLR lg10 [linear: F(1,327) = 23.97, p = 0.0001, R 2 adj =0.065; quadratic: F(2,326) = 12.83, p = 0.0001, R 2 adj =0.067]. As shown in Figure 4 , NLR lg10 tended to be highest in young adults (6–9 years of age), and lower in juveniles, older adults, and geriatric individuals. The repeated measures ANOVA examining differences between transport NLR lg10 and routine NLR lg10 showed that transport NLR lg10 was significantly higher (raw data: M = 3.79, SEM = 0.31; lg10: M = 0.49, SEM = 0.03) than routine NLR lg10 (raw data: M = 2.75, SEM = 0.28; lg10: M = .31, SEM = 0.03, F(1,152) = 23.89, p = 0.0001). The difference between transport and routine NLR lg10 was also significant across age groups F(3,152) = 6.35, p = 0.001). The difference in NLR lg10 was more pronounced in older animals, with older adult baboons (animals aged 10–15) showing the greatest difference, and juveniles (aged 4 years and younger) showing almost no difference ( Figure 5 ). We could not assess the effect of rearing due to the low number of mother-reared individuals in the adult age categories. We ran two additional analyses to eliminate the possibilities that 1) NLR lg10 was elevated as a function of being housed outdoors for the first time upon arrival to the new facility, and 2) NLR lg10 decreased as a function of housing at the new facility, rather than having increased as a function of the transfer itself. Using a subset of 46 baboons that had been housed outdoors prior to transfer, a paired-samples t-test showed the same pattern of elevated NLR lg10 at transfer (Raw data: M=4.36, SEM = 0.43; lg10: M=0.545, SEM = 0.043) compared to routine NLR lg10 taken approximately one year post-transfer (raw data: M = 2.26, SEM = 0.27; lg10: M = 0.219, SEM = 0.052), t(45) = −5.61, p < 0.001. Second, using a subset of 61 individuals for which we had pre-transfer NLR lg10 from the previous facility, a repeated-measures ANCOVA (with age as a covariate) showed that pre-transfer NLR lg10 (raw data: M = 2.57, SEM = 0.33; lg10: M = .29, SEM = 0.04) was significantly lower than transfer NLR lg10 (raw data: M = 4.5, SEM = 0.42; lg10: M = 0.55, SEM = 0.04), F(1,62) = 19.24, p < 0.001. Inverse normalized baseline NLRs in this study were not found to be significantly heritable (h 2 =.088, SE=.124, p=.204). However, inverse normalized transport NLRs did exhibit significant heritability (h 2 =.753, SE=.158, p<.001). For inormalized transport NLR, the interaction of sex*age was a significant covariate, and the proportion of variance accounted for by this covariate was .122. Regarding the direction of this sex*age effect, we found a significant positive correlation between age and transport NLR for both males (n=30, r =.627, =.0002) and females (n=126, r =.179, p=.045). Additional sex and rearing effects on mean NLR values are presented in Table 3 . Lastly, the repeated measures ANCOVA showed that gestational NLR lg10 was significantly higher (raw data: M = 4.61, SEM = 0.41; lg10: M = .55, SEM = 0.04) than control NLR lg10 (raw data: M = 2.74, SEM = 0.24; lg10: M = 0.33, SEM = 0.03, F(1,80) = 4.95, p = 0.03) while controlling for age ( Figure 6 ). Furthermore, there was no significant difference in NLR lg10 as a function of trimester (see Figure 6 : F(2,57) = 0.28, p = 0.76).

In the current study, we aimed to expand the characterization of non-human primate NLR by characterizing NLR in baboons. This is important given that baboons are increasingly used as research models in a variety of paradigms, ranging from Alzheimer’s disease and dementia, epilepsy, drug abuse, aging, pregnancy, vaccine development, and early life adversity, to name a few ( Chassen et al., 2020 ; Fagot et al., 2019 ; Lizarraga, Daadi, Roy-Choudhury, & Daadi, 2020 ; Moore, Zamarripa, & Weerts, 2023 ; Port et al., 2021 ; Szabó & Salinas, 2021 ), and NLR can be used as a predictive and prognostic tool in each of these models. We also aimed to determine whether patterns seen in previous studies with chimpanzees ( Neal Webb et al., 2020 ) and rhesus macaques ( Capitanio et al., 2023 ) generalize to or are replicated in baboons, and also extend understanding of NLR as a measure of stress (or change) using transport and gestational values. Mean baboon NLR across the sample (2.37 ± SD 2.14) was comparable to mean NLR found in rhesus macaques (2.24 ± SD 1.32; Capitanio et al., 2023 ), was lower than that found in chimpanzees (3.47 ± SD 1.86; Neal Webb et al., 2020 ), and was quite close to reference ranges reported in humans (range: 1.71–2.28; Azab, Camacho-Rivera, & Taioli, 2014 ). 95% of the NLR values in our sample fell between 0.43 and 6.50. As such, values below 0.43 and above 6.50 likely indicate an underlying pathological condition or stress. However, more research, including additional older, mother-reared individuals and more males, is needed to determine true reference intervals. Consistent with previous research in apes ( Neal Webb et al., 2020 ) and rhesus macaques ( Capitanio et al., 2023 ), we found that mother-reared individuals in the current study exhibited higher NLR values than nursery-reared individuals. Nursery-rearing is often associated with stress and used as a model for early-life adversity due to maternal separation and human-rearing, and can result in the development of abnormal behaviors in NHPs ( Sackett, Ruppenthal, & Elias, 2006 ). If nursery rearing is truly more stressful than mother rearing, we would have expected to find higher NLR values in nursery-reared individuals. Indeed, at other colonies, nursery-reared baboons showed delayed growth, increased stress-related behavior, poorer health, and shorter longevity compared to their mother-reared counterparts (Brent & Bode, 2002). There may be several explanations or factors that contribute to this finding. First, Capitanio and colleagues (2023) also found lower NLR values in nursery-reared infant macaques, and posit that one explanation may be that mother-reared individuals are exposed to more antigens in the outdoor environment (at least in their colony), which may contribute to higher immune responses, and thus, NLR ( Capitanio et al., 2023 ). Similarly, our nursery-reared animals are reared indoors for the first year of life, whereas mother-reared animals are reared outdoors; however, we did not explicitly analyze differences in NLR as a function of indoor vs. outdoor housing or antigen exposure (e.g., insects, other outdoor animals, fungi, etc.). Second, it is possible that immune function is indirectly affected through social mechanisms: mother-reared individuals may have increased probabilities of illness or infection due to their increased social contact given that they are housed in their natal group consisting of 30 or more individuals of mixed sex and age (more social interactions and grooming opportunities) compared to nursery-reared individuals ( Prall & Muehlenbein, 2014 ). Third, nursery-rearing in baboons may not be as stressful as conditions involved in mother-rearing. Perhaps there is a level of stress associated with the strict, matrilineal dominance hierarchy in large, complex social groups that is characteristic of a despotic species such as baboons, which affects immune-inflammatory responses ( Sapolsky, 2005 ). These dominance hierarchies are maintained through both aggression and psychological intimidation, which have impacts on physiological health, including increased cortisol levels, altered cardiovascular function, decreased gonadal hormone levels, and neurobiological changes ( Sapolsky, 2005 ). Indeed, there are reports of decreased lymphocytes following dominance-related social stressors in rodents, pigs, and chimpanzees ( Sapolsky, 2005 ). Lastly, nursery-rearing may result in a blunted or dysregulated immune-inflammatory response, similar to blunted cortisol responses that develop following chronic stress exposure in both humans ( Carpenter et al., 2009 ; Metz et al., 2020 ) and rhesus macaques ( Capitanio, Mendoza, Lerche, et al., 1998 ; Capitanio, Mendoza, Mason, et al., 2005 ). In this way, both very high and very low NLRs may indicate dysregulation. In humans, very low NLR (0.7) is also considered pathological (e.g., the highest mortality rate in septic shock occurred in patients with the lowest NLR values) ( Zahorec, 2021 ), and low NLR in rhesus macaques was associated with higher disease risk later in life ( Capitanio et al., 2023 ). More research is needed to examine which of these explanations, or other alternative explanations, may be most fitting and parsimonious with the data. Consistent with the relationship previously found in chimpanzees, baboon NLR in the current study also showed a quadratic relationship with age, such that NLR is highest in young adults (~7 years of age) and lowest in older adults and geriatric individuals. In baboons, this time frame may correspond to some growth processes associated with sexual maturation and the commencement of breeding for males and first births for females. This time of sexual maturation potentially corresponds to changes in sex steroids, which have been shown to impact lymphocyte proliferation ( Athreya, Pletcher, Zulian, Weiner, & Williams, 1993 ). For example, increases in estrogen levels can lead to reduced production of lymphocyte precursors in bone marrow ( Lang, 2004 ), which may contribute to overall higher NLR values. Additionally, Neal Webb et al. (2020) found that, 1) longitudinally, chimpanzee NLR was stable over a 10-year timespan, but 2) cross-sectionally, older individuals tended to have lower NLRs, and 3) individuals with higher NLRs died at younger ages. Together, this suggests that individuals with higher NLRs died at younger ages (potentially due to an underlying pathological condition), leaving a population of older individuals that had consistently low NLRs throughout life. It is possible that the quadratic relationship found in the current study in baboons reflects a similar longevity pattern, but longitudinal data are needed to determine whether baboon NLR is also stable over time, and whether individuals with higher NLR tend to die at younger ages. Baboon NLR within one week of arrival following transport was significantly higher than NLR taken during routine physical exams one-year post-transfer. Furthermore, this was strengthened by the additional result that pre-transfer NLRs (taken at the previous facility less than one year prior to transfer) were significantly lower than transfer NLRs and similar to post-transfer NLRs taken at the new facility, which may suggest a return to baseline following transfer. Lastly, we were able to eliminate the possibility that this result was an artifact of baboons being housed outdoors for the first time, since the subset of baboons that were outdoor housed prior to transport also showed elevated NLR following transport. This relationship between stress and NLR is consistent with previous research in humans showing elevated NLRs with higher levels of self-reported stress ( Demircan, Gozel, Kilinc, Ulu, & Atmaca, 2016 ). Transport involves a series of potentially taxing events, including removal from the social group, an extended period in a smaller enclosure on a transport truck, and arrival to a new, unfamiliar facility. This activates the sympathetic nervous system, which impacts processes associated with the immune system, thereby affecting NLR ( Kalelioglu & Karamustafalioglu, 2019 ). Indeed, lymphopenia (decreased lymphocyte counts) and neutrophilia (increased neutrophil counts), suggestive of a stress leukogram,, have been found to occur under conditions of acute psychological stress in humans ( Abramson & Melton, 2000 ; Benschop, Rodriguez-Feuerhahn, & Schedlowski, 1996 ; Iskandar, Griffeth, Sapra, Singh, & Giugale, 2011 ). Although we do not have data to investigate a return to baseline NLR in the period of time following arrival to the new facility, the fact that NLR values continued to be significantly higher within one week of arrival suggests some continued level of stress during that time period, perhaps related to acclimation in addition to transport. As such, this result seems to confirm NLR’s utility as a stress biomarker in captive baboons and highlights its use as a potential behavioral welfare indicator. We were interested in estimating the heritability (i.e., the proportion of variance that can be attributed to genetic factors) of both routine NLR values and transport NLR values. Contrary to what we hypothesized based on NLR research in humans ( Lin et al., 2016 ), we did not find routine NLR values to be heritable in this population of baboons. However, transport NLR values did exhibit significant heritability. Indeed, approximately 75% of the variance in transport NLR values can be attributed to variance in genetic factors, which suggests that stress responses have a strong genetic component in this population. We found the sex*age interaction to be a significant covariate for transport NLRs, whereby age was significantly correlated with transport NLR values for both males and females. However, it is worth noting the small sample of adult males in this study ( Table 1 ), which may be contributing to this observed sex*age effect. Lastly, we found that baboon NLR was elevated during pregnancy. This is perhaps unsurprising given that this result at least partly mirrors the pattern of NLR seen in human pregnancy ( Hai & Hu, 2020 ). However, our results are inconsistent with previous literature in humans showing that NLR changes across trimester (although this may be due to small sample sizes across trimester groups, given that the data showed an effect of pregnancy using the repeated-measures ANCOVA). Pregnancy is a physically stressful condition and involves physiological, including immunosuppressive, changes ( Chandra, Tripathi, Mishra, Amzarul, & Vaish, 2012 ). Given that baboons are used as models of certain conditions during pregnancy, this result provides some evidence demonstrating the utility of baboons in such models. From a clinical perspective, because NLR has been used to successfully predict certain pregnancy outcomes and has been identified as a risk factor for several pregnancy-related complications in humans, baboon NLRs that are higher than “normal” pregnancy-related NLRs may indicate potential complications, and identifying such complications is particularly important in breeding colonies. As such, future studies should examine NLR in relation to pregnancy outcomes and conditions (e.g., spontaneous abortion, pre-term delivery) in NHPs, which may have important clinical applications, particularly in breeding colonies. We would like to note that, although we termed NLR values taken during routine biannual physical exams “routine NLR,” it is possible that these NLR values reflect some level of stress. The sedation process involves some procedures associated with stress, including human handling and injections. Given that this process is the same for all baboons, the potential effect of stress on these NLR values is likely consistent across individuals, although individual variation in responses to stress likely exists. Although an NLR value taken following sedation (by far, the most common method of obtaining hematology) may reflect some stress, it is likely the closest approximation of “routine” that we have the current ability to obtain (outside of training for blood draws). Sample composition was a limiting factor in the current study. Specifically, there was a small number of mother-reared older adult and geriatric individuals in the sample (particularly males), which limited our ability to assess the interaction effects of rearing, age, and sex on NLR in one comprehensive analytical model. We anticipate having a larger sample of older, mother-reared individuals in the future as baboons continue to age within our breeding colony. Regardless, these data provide the first descriptive data of NLR in baboons, and suggest 1) there is something about mother rearing that seems to result in elevated NLR values, as our group has now found this in both chimpanzees and juvenile baboons, and this result has also been found in rhesus macaques ( Capitanio et al., 2023 ); 2) NLR seems to show the same quadratic relationship with age as has been found in chimpanzees ( Neal Webb et al., 2020 ); 3) NLR can be used as a measure of stress in baboons; 4) while routine NLRs do not appear to be heritable, transport NLR values exhibit significant heritability, perhaps indicating that this particular stress response has a strong genetic component; and 5) baboon NLR is elevated as a function of pregnancy, which is likely related to the normal immune-inflammatory changes characteristic of pregnancy, but provides additional evidence for the use of baboons as models for pregnancy-related complications in translational research. Additional studies are underway that further examine the heritability of NLR, intra-individual longitudinal changes in NLR, changes in NLR following other stressful events (e.g., repeated sedations), and relationships among NLR, clinical conditions (e.g., injury and illness), sociality, rank, and behavior. We believe that NLR has potential utility across clinical, welfare, basic research, and behavioral management settings. Our goal is to identify stressors and/or conditions that may contribute to elevated NLR in both the short term (i.e., stressors/conditions for which we may be able to intervene) and the long term (which may have implications for longevity and translational interventions aimed at increasing longevity). Currently, the evidence is building that NHP NLR can be used as an aging and longevity biomarker. Clinically, routine screening of NLR can assist veterinarians and staff to identify individuals with acutely or chronically “high” NLR. Further investigation into the potential causes of elevated NLR may lead to identification and evaluation of interventions aimed at resolving such issues. For example, individuals that exhibit chronically high NLR (i.e., individuals that exhibit higher NLR on average than other baboons on a long-term basis) may be put onto a welfare “watchlist”, as this may indicate issues with chronic physiological stress or inflammation. Additionally, identification of the reasons for chronically elevated NLR may lead to the development of interventions aimed at combating underlying physiologic stressors and/or “challenges” to longevity, which may have the potential to then be tested in translational research.