Western-style diet in the presence of elevated circulating testosterone induces adipocyte hypertrophy without proinflammatory responses in rhesus macaques.

To assess the effects of T+WSD treatment on metabolic parameters, we determined body composition by DEXA and glucose tolerance by GTT. Animals treated with T+WSD for one year exhibited greater levels of circulating T, a significantly greater percentage of total body fat, and greater fasting insulin and glucose concentrations than control animals ( Table 1 ). T+WSD animals were also more insulin resistant than control animals, based on higher Homeostatic Model Assessment for Insulin Resistance (HOMA-IR) values, and more glucose intolerant, based on greater Area Under Curve (AUC) glucose values measured by GTT ( Table 1 ). We next assessed the effects of T+WSD on systemic pro-inflammatory mediators. Circulating levels of the proinflammatory mediators CRP and CCL5 were not significantly different between the control and T+WSD groups ( Table 1 ). Collectively, our data indicate that a one-year T+WSD treatment results in a significant increase in body fat and glucose intolerance, but has no significant effect on the levels of systemic proinflammatory markers in adult female rhesus macaques. To determine whether T+WSD treatment selectively modulates depot-specific morphological characteristics of WAT, we performed histological staining of WAT sections using picrosirius red and analyzed adipocyte size and tissue fibrosis. The mean areas of OM and subcutaneous (SC) adipocytes were significantly higher in T+WSD-treated animals compared to control animals ( Figure 1B and C ; Supplementary Table 1 ). Additionally, the fibrotic fractions in OM-WAT and SC-WAT were not significantly different between the groups ( Figure 1B and D ; Supplementary Table 1 ). These data show that T+WSD treatment of young adult female rhesus macaques led to the development of adipocyte hypertrophy without overt fibrosis. Furthermore, the mean area of OM adipocytes and body fat content were positively correlated with circulating T levels ( Figure 1E and F ). Collectively, these results are consistent with our previous studies showing similar metabolic (obesity and insulin resistance) and adipose-specific (adipocyte hypertrophy) phenotypes in female macaques exposed at puberty to the combination of T and WSD 9 – 12 . We next investigated whether T+WSD was associated with higher proinflammatory responses in tissue-resident macrophages using flow cytometry, which provided detailed insight into tissue-specific immune phenotypes. Cells residing in nonlymphoid tissues (OM-WAT, CL, OV, ENDO, and FBM) were stained with a myeloid-specific antibody panel. Our analysis revealed that T+WSD treatment had no significant effect on the frequencies of macrophages (CD45+HLADR+CD11b+) in the tissues examined. Additionally, we found no significant group differences in frequencies of proinflammatory M1-like macrophages (HLADR+CD11b+CD163-CD11c+) 22 – 24 , anti-inflammatory M2-like macrophages (HLADR+CD11b+CD163+CD11c-) 25 – 28 , or mixed M1/M2 macrophages (HLADR+CD11b+CD163+CD11c+) 24 , 29 in all tissues examined ( Figure 2 ). Collectively, our results indicate that macrophage polarization is not significantly affected by T+WSD treatment. To test whether T+WSD treatment impacts macrophage function, we measured cytokine responses of macrophages residing in OM-WAT and BM (including FBM and T12) to toll-like receptor (TLR) agonists in vitro. Cell mixtures were incubated with LPS (TLR4 agonist), Pam3CSK4 (TLR1/2 agonist), or HKLM (TLR2 agonist) in the presence of brefeldin A to enhance intracellular cytokine staining signals by blocking transport processes during cell activation. Cells were stained with antibodies to CD11b and HLA-DR to delineate monocytes/macrophages. Following surface staining, cells were permeabilized and stained with monoclonal antibodies to TNFα and IL-6 and analyzed by flow cytometry ( Figure 3A and B ). Because OM-WAT cells showed poor recovery following this procedure, we focused on the functional analysis of macrophages residing in FBM and T12 marrow. FBM and T12 macrophages showed variable TNF responses to different TLR agonists, while no significant group differences in TLR responses were observed ( Figure 3D and E ). Interestingly, IL-6 responses of BM macrophages to different TLR ligands were attenuated by T+WSD treatment, particularly in the FBM compartment ( Figure 3F and G ). These data suggest that T+WSD treatment does not enhance TLR-mediated proinflammatory responses, but may induce immune tolerance in BM macrophages. We next investigated whether T+WSD treatment is associated with higher proinflammatory responses in tissue-resident lymphoid cells. Cells residing in nonlymphoid and lymphoid tissues (OM-WAT, CL, OV, ENDO, FBM, and LNs) and PBMC were stained with a lymphoid-specific antibody panel. Our analysis of lymphoid subsets showed no significant group differences in frequencies of CD4+, CD8+, and CD4+CD8+ T cells in all tissues examined ( Supplementary Figures S1A – E ). We next determined whether T+WSD treatment affected tissue-resident Tregs, (CD3+CD4+CD25+FoxP3+) reported to play a role in the pathogenesis of obesity and insulin resistance 30 – 32 . However, we observed no differences in the frequencies of Tregs 33 in all tissues examined between control and T+WSD-treated animals ( Supplementary Figures S1A and F ). We next analyzed tissue-resident memory CD4+ and CD8+ T-cells. To discriminate between different memory T-cell subsets, CD8+ and CD4+ cells were co-stained with CD28 and CD95 antibodies to delineate CD28+CD95- naïve (Tn), CD28+CD95+ central memory (Tcm), and CD28- effector memory (Tem)/effector (Teff) T-cells ( Figures 4A and B ; Supplementary Figure S2A ). We found that CD28- T-cells did not express CD25 or the activation marker HLA-DR, suggesting that CD28- cells residing in OM-WAT are primarily composed of Tem in both control and treated groups ( Figures 4E and F ). T+WSD treatment resulted in a significant reduction in the frequencies of CD8+ Tem in OM-WAT, but not in other tissues or PBMCs ( Figure 4G ) and had no significant effect on frequencies of CD4+ memory subsets ( Supplementary Figures 2B and C ). We found no significant group differences in frequencies of proliferating Tems in the tissues examined ( Figures 4A , B , and H ). However, the frequencies of proliferating Tems in OM-WAT were negatively correlated with HOMA-IR ( Figure 4I ). To test whether T+WSD treatment accelerates T cell exhaustion, we analyzed the levels of Programmed Cell Death Ligand 1 (PD-1) in Tems. Tems from control and treated animals expressed very low levels of PD-1 ( Figures 4A and B , right panels). In contrast, T12 CD8+ Tems expressed low but detectable levels of PD-1 and low levels of Ki67 ( Figures 4C and D ), which validated our PD-1 staining in NHPs. These results suggest that T-cell exhaustion is not involved in the decline in CD8+ Tems in OM-WAT of T+WSD-treated animals. Notably, CD8+ Tem frequencies in OM-WAT were negatively correlated with HOMA-IR, fasting insulin, fasting glucose, glucose AUC, and % body fat ( Figure 4J and Table 2 ). In contrast, Tcm frequencies in OM-WAT were positively correlated with HOMA-IR ( Figure 4K ), while the frequencies of Tn in OM-WAT were not correlated with HOMA-IR ( Figure 4L ). Additionally, CD4+ Tem and M1 macrophage frequencies in OM-WAT were not correlated with any of the parameters examined ( Table 2 ).

This study was approved by the Oregon National Primate Research Center Institutional Animal Care and Use Committee and conforms to current Office of Laboratory Animal Welfare regulations as stipulated in assurance number A3304–01. Adult female rhesus macaques were pair-housed, with the cage size adjusted to animal weight according to the USDA Cage Size Guide, eighth Edition. Animals were maintained on either a monkey chow diet (n=5 Control; average age 6.04 ± 1.03) consisting of two daily meals of fiber-balanced monkey diet (15% calories from fat, 27% from protein and 59% from carbohydrates; no. 5052; Lab Diet, St. Louis, MO), supplemented with fruits and vegetables, or a WSD (n=9, average age 6.3 ± 0.57), containing 36% calories from fat, 18% from protein and 45% from carbohydrates (TAD Primate Diet 5LOP, 5A1F, Lab Diet), as previously described 10 . All WSD females also received testosterone implants as previously described 10 . Briefly, T-releasing capsules were prepared by packing Silastic tubing (0.34 cm i.d.; 0.64 cm o.d.; Dow Corning, Midland MI) with a mixture of cholesterol and T (Sigma-Aldrich, St. Louis, MO) at a ratio of 9:1. The capsules were implanted subcutaneously in the interscapular region. Weekly blood samples were assayed for serum T, and when levels dropped below 1.0 ng/ml, the capsules were replaced. A target range of ~1.5 ng/ml was maintained by changing the length of the T-releasing capsules from 1 to 5 cm over the study period. Androgen levels of T+WSD treated animals ( Table 1 ) are in the range for women with hyperandrogenemia/PCOS 18 – 20 . Two experimental groups, designated Control (n=5) and T + WSD (Treated; n=9), were used in this study. Both animal groups were maintained for 12 months prior to necropsy. Two months before necropsy, all animals underwent a controlled ovarian stimulation protocol to collect oocytes for analyses of the meiotic maturation, fertilization rate, and embryo development (data not shown, separate manuscript in preparation). Animals were allowed to undergo one menstrual cycle before tissues were collected for this study. Dual-energy X-ray absorptiometry (DEXA) analysis, intravenous glucose tolerance tests (GTTs), and determination of fasting glucose and insulin levels were performed before necropsy, as described 10 . For DEXA studies, monkeys were sedated with ketamine and positioned supine on the bed of a Hologic DEXA scanner (Discovery scanner, Hologic Inc., Bedford, MA, USA). Two to three scans were performed for each animal and averages were calculated for each measure. For glucose tolerance tests (GTTs), animals were fasted overnight, sedated with Telazol and a glucose bolus (50% dextrose solution) administered at a dose of 0.6 g/kg via the saphenous vein. Baseline blood samples were obtained prior to the glucose injection, and 1-ml blood samples were taken at 1, 3, 5, 10, 20, 40 and 60 min later by venipuncture. Glucose was measured immediately using a OneTouch Ultra Blood Glucose Monitor (LifeScan), and the remainder of the blood was placed in heparinized tubes on ice for insulin assay. Homeostatic Model Assessment of Insulin Resistance (HOMA-IR) was calculated as fasting serum insulin (μU/ml) × fasting plasma glucose (mg/dl)/405. Insulin was analyzed in the ONPRC Endocrine Technologies Core by immunoassay on a Roche cobas e411 chemiluminescence-based automated immunoassay platform (Roche Diagnostics, Indianapolis, IN). The assay range was 0.2–1000 μIU/ml. Intra-assay CV was 1.1%. Inter-assay CV for the insulin assay is 3.7%. Testosterone was analyzed by immunoassay on a Roche cobas e411 chemiluminescence-based automated immunoassay platform (Roche Diagnostics, Indianapolis, IN). The assay range was 0.025–15 ng/ml. Intra-assay CV was 2.2%. Inter-assay CV for the testosterone assay is 6.7%. Chemokine ligand 5 (CCL5) was analyzed by ELISA following the manufacturer’s instructions (R&D Systems, Minneapolis, MN). Samples were diluted 1:100 prior to analysis. The assay range was 31.3–2000 pg/ml; intra-assay CV was 3.5%. Inter-assay CV for the CCL5 ELISA is 5.7%. C-reactive peptide (CRP) was analyzed by ELISA following the manufacturer’s instructions (ALPCO Diagnostics, Salem, NH). Samples were diluted 1:100 prior to analysis. The assay range was 0.95–150 ng/ml. Intra-assay CV was 4.3%. Inter-assay CV for the CRP ELISA is 14.8%. WAT histology and image analysis were performed as described 11 . Briefly, 200 to 500-mg fragments of subcutaneous (SC)-WAT and OM-WAT were collected at necropsy and fixed in zinc formalin (Fisher Scientific, Hampton, NH, USA) at 4°C for 48 hours. Samples were transferred to 70% (v/v) ethanol for 5 days, embedded in Paraplast wax (Leica, Wetzlar, Germany), and 5-μm sections were prepared using a micron rotary microtome. Slides were stained with Picrosirius red Stain Kit (Abcam, Boston, MA, USA) according to the manufacturer’s instructions. Images representing tissue segments of approximately 5 to 10 mm in size were acquired using an Aperio AT2 System slide scanner (Leica Biosystems, Wetzlar, Germany) and saved as gigabyte-size TIFF files. The mean area of adipocytes and fibrotic fraction (the proportion of Picrosirius red-positive staining per WAT section) were measured using the thresholding function of Image J, resulting in foreground (Picrosirius red-positive) and background (Picrosirius red-negative) pixels within each section. Specific WAT parameters are shown in Supplementary Table 1 . All tissues were collected and cells isolated using the following protocol on the day of necropsy unless stated otherwise. Five grams of OM-WAT were collected at necropsy and transported to the lab in 25 mL ice-cold FACS buffer (DPBS (Thermofisher, Waltham, MA, USA) containing 0.5% BSA (Sigma-Aldrich, St. Louis, MO, USA) and 2 mM EDTA (Sigma-Aldrich)). Tissue was minced in a 6-well dish and transferred to 10 mL of freshly-made, filter-sterilized collagenase solution (120 mg collagenase type-II (Gibco, cat#17101–015), 1.4 g BSA (Sigma-Aldrich), DPBS to 40 mL, and 80 μL 1 M CaCl 2 ). Tissue was digested for 30–40 min, while vortexed every 5 min until tissue clumps were dissolved. Ice-cold FACS buffer (40 mL) was added to the digested tissue and the cell suspension was filtered through a 100-micron cell strainer. The tube was centrifuged at 1500 rpm (400 × g) for 10 min at 4 o C, and the supernatant was discarded. 5 mL red-blood-lysis (RBC) buffer was added to the cell pellet and incubated for 8 min at room temperature. The cell pellet was washed in ice-cold FACS buffer two more times by centrifugation at 1500 rpm for 10 min at 4 o C and resuspended in 1 mL FACS buffer on ice for cell counting. Cells were centrifuged and the supernatant was replaced with 1–2 mL ice-cold CryoStor CS10 (STEMCELL Technologies, Vancouver, BC, Canada). Cells were gently resuspended and cryopreserved in 2–4 aliquots at −80 o C for 24h and then stored in liquid N 2 until immune cell analyses. Femoral bone marrow (FBM) was removed by flashing the femur with 20 mL ice-cold FACS buffer. FBM was gently disrupted using a 25-ml syringe loaded with a blunt needle, and the cell suspension was filtered through a 100-μm cell strainer. The cell suspension was centrifuged for 10 min at 1500 rpm and the pellet was resuspended and incubated for 8 min at room temperature in the RBC buffer. Cells were washed and cryopreserved as described above. Vertebral BM was extracted from the T12 vertebra using a bone cutter and a ceramic mortar and pestle. BM cell suspensions were then filtered through a 100-micron cell strainer and processed as described for FBM. Reproductive tract and ovaries from rhesus macaques were obtained at necropsy, submerged in Hanks Balanced Salt solution and delivered to the lab within 15 minutes. For END isolation, the uterus was bisected along the longitudinal axis from the fundus to the cervix and then further divided into quarters. Full-thickness tissue sections (~5 mm) extending from the uterine lumen to the outer myometrium were biopsied from each quarter. OV were isolated from the reproductive tract, trimmed free of fat, and bisected. The CL was removed from the dominant ovary by use of curved dissecting scissors. END, CL and OV cell suspensions were then generated using the same collagenase digestion method described above and cryopreserved until immune cell analyses. LNs were collected into ice-cold FACS buffer, diced with scalpels, mashed through a 70-μm filter, and processed for cell isolation as described for WAT. Blood samples were collected in EDTA-treated tubes and peripheral blood mononuclear cells (PBMCs) were isolated by centrifugation at 2000 rpm (800 × g) for 30 min using a Lymphoprep gradient (STEMCELL Technologies) as described 21 . Cryopreserved cells were washed and analyzed by flow cytometry using monoclonal antibodies previously validated for specific binding to rhesus macaque antigens 17 and positive/negative cells were detected using lymphoid-specific and myeloid-specific gating strategies. For lymphoid staining, 5×10 5 -1×10 6 cells were incubated with a custom antibody panel that included anti-CD45 (D058–1283, Biolegend, San Diego, USA), anti-CD3 (clone SP34, BD Pharmingen, Franklin Lakes, NJ, USA), anti-CD4 (clone L200, BD Pharmingen), anti-CD8 (clone SK1, Biolegend), anti-HLA-DR (clone L243, Biolegend), anti-CD25 (clone 2A3, BD Pharmingen), anti-FoxP3 (clone PCH101, Thermofisher, Waltham, MA, USA), anti-CD28 (clone CD28.2, Biolegend), anti-Ki67 (clone B56, BD Biosciences), and fixable amine-reactive dye (Invitrogen, Waltham, MA, USA). For macrophage staining, 5×10 5 -1×10 6 cells were incubated with a custom antibody panel that included anti-CD45 (D058–1283, Biolegend), anti-CD123 (clone 6H6, Biolegend), anti-HLA-DR (clone L243, Biolegend), anti-CD11b (clone ICRF44, Biolegend), anti-CD163 (clone GHI/61, Biolegend), and anti-CD11c (clone 3.9, Biolegend) and fixable amine-reactive dye (Invitrogen). For lineage-negative staining, anti-CD3 (clone SP34, BD Pharmingen), anti-CD8 (clone SK1, Biolegend), and anti-CD20 (clone 2H7, Biolegend) were included in a separate panel. All cell aliquots were incubated with these aforementioned antibodies (at the manufacturer’s suggested dilution) for 30 minutes at 4 o C, then non-specific binding was reduced by washing twice with PBS containing 2% FBS before fluorescence detection. Fluorescence intensity was acquired using an LSRII flow cytometer (BD Biosciences). Data were then analyzed using FlowJo v10 (Tree Star, Ashland, OR, USA) and Prism v6 (GraphPad Software, San Diego, CA, USA). Functional responses in FBM and T12 macrophages were evaluated as described 21 . Specifically, 5 × 10 5 BM cells were stimulated with LPS (TLR4 agonist 1 μg/mL), Pam3CSK4 (TLR1/2 agonist, 1 μg/mL), or HKLM (TLR2 agonist, 108 particles/mL) for 8 h in the presence of 5 μg/mL brefeldin A (Sigma-Aldrich). TLR agonists were purchased from InvivoGen (San Diego, CA, USA). Cells were stained with antibodies to CD11b and HLA-DR to delineate monocytes/macrophages. Following surface staining, cells were permeabilized using the Foxp3/Transcription Factor Staining Buffer Set (eBioscience, cat# 00–5523-00) and stained with monoclonal antibodies TNFα, (clone MAb11, Invitrogen) and IL-6 (clone MQ2–6A3, BD Biosciences). Samples were washed twice with PBS containing 2% FBS and analyzed on an LSRII flow cytometer as described above. All statistical analyses were conducted in Excel and Prism 8 (GraphPad). Anatomical and metabolic parameters were compared using a two-tailed unpaired t-test. A correlation analysis was conducted by calculating the Pearson correlation coefficients and their associated p-values. Flow cytometry data analysis across multiple tissues and animal groups were conducted using 2-way ANOVA and a Bonferroni’s multiple comparisons test.

While the pathogenesis of PCOS is not fully understood, androgen excess, insulin resistance, adipocyte hypertrophy and low-grade inflammation can form a vicious cycle interconnecting metabolic and reproductive disarrangements in women with PCOS 4 , 6 , 8 . However, whether androgen excess and adipocyte hypertrophy alone can trigger a systemic and tissue-resident proinflammatory milieu remains to be defined. Here, we demonstrate that the combination of hyperandrogenemia and obesogenic WSD are not sufficient for inducing a proinflammatory phenotype after one year of treatment, suggesting that additional factors or a longer treatment period are required for immune cell activation in this experimental model. In the present study, we found that one year of exposure to hyperandrogenemia in the presence of WSD (T+WSD) induces obesity, glucose intolerance, insulin resistance, and adipocyte hypertrophy. Numerous studies in PCOS patients and in animal models of PCOS have shown that androgen excess is associated with metabolic abnormalities, including insulin resistance, obesity, and adipocyte hypertrophy 11 , 34 – 39 . The data presented here are also in line with our previous studies showing that T+WSD exposure for three years, starting at puberty, induced metabolic dysfunction and OM-WAT hypertrophy in female rhesus macaques 9 – 11 . Specifically, T+WSD treatment accelerated the development of WAT dysfunction through androgen-specific suppression of lipolysis and WSD-specific reduction in WAT capillary density 9 , 11 . In the same studies, exposure to T or WSD alone for three years, starting at puberty, did not induce adipocyte hypertrophy but did exert negative effects on some functional characteristics of WAT 9 , 11 . However, the effects of WSD or T alone on metabolic and WAT characteristics were not examined in the present study because our previous results demonstrated that the combination of T and WSD was required for the development of more severe metabolic effects. Here, we also report that T+WSD treatment was not associated with elevated levels of the pro-inflammatory mediators CRP and CCL5. In contrast, previous studies have shown that PCOS patients exhibit elevated markers of systemic inflammation, including CRP 40 , IL6, and TNFα 41 , 42 . CCL5 levels in the blood and visceral WAT have been shown to be elevated in letrozole-induced PCOS in mice and in PCOS patients 43 , 44 . Our previous study showed that T+WSD treatment or the consumption of WSD alone for three years was associated with elevated CRP levels in female rhesus macaques 10 , suggesting that longer treatment is necessary for inducing a systemic proinflammatory response in NHPs. One surprising finding of the present study was the lack of macrophage proinflammatory activation despite developing obesity, adipocyte hypertrophy, and insulin resistance in T+WSD-treated animals. Specifically, we found no evidence of macrophage tissue recruitment and proinflammatory polarization toward the classically activated M1 phenotype in all tissues examined, including OM-WAT, CL, OV, END, and BM. Furthermore, the frequencies of anti-inflammatory, alternatively activated M2 and mixed M1/M2 macrophages were not affected by T+WSD. Obesity has been shown to be associated with accumulation of proinflammatory macrophages in rodent 23 , 28 , 45 – 50 and human 22 , 24 , 29 , 51 WAT, based principally on immunohistochemistry. However, several subsequent studies using qPCR and immunohistochemistry as well as flow cytometry analyses similar to our study, showed that human obesity is not associated with proinflammatory M1 polarization in WAT 52 – 55 . In addition to our flow cytometry data showing that obesity in macaques is not associated with M1 polarization. Additionally, our functional data, using BM macrophages, further support the idea that T+WSD treatment is not associated with enhanced classical macrophage activation. Specifically, BM macrophages from the T+WSD and control groups produced similar amounts of TNFα in response to LPS and other TLR ligands. Interestingly, IL6 production by FBM macrophages in response to LPS (TLR4 agonist) and HKLM (TLR2 agonist) stimulation was significantly blunted by T+WSD treatment compared to control treatment, suggesting the potential development of immune tolerance to different TLR agonists in response to T+WSD. Further studies are needed to elucidate whether T+WSD and WSD and T alone modulates macrophage activation in other peripheral tissues. We also analyzed lymphoid immune cells residing in peripheral tissues and PBMCs and found that the frequencies of the major subsets of T cells, including Treg, were not affected by treatment. Treg cells play a role in the pathogenesis of obesity and insulin resistance and Treg frequencies are reduced in PCOS patients 56 , 57 . Recent studies have shown that obesity can alter the properties of WAT-resident T cells prior to activation of macrophages 58 – 64 . Specifically, effector CD8 + T-cells residing in WAT have been suggested to initiate a proinflammatory response leading to the accumulation of proinflammatory macrophages in WAT 65 , 66 . In contrast, in present study, short-term T+WSD treatment was not associated with proinflammatory T-cell activation in any of the tissues examined. Here, we also addressed the potential alteration of immune cell composition in reproductive organs. OM-WAT is a major human visceral fat depot associated with multiple abdominal organs, including the reproductive tract and lymphoid tissues. Therefore, changes in the local immune milieu in WAT may also affect the local immune composition of reproductive tissues. Immune cell composition is dynamic in reproductive tissues and is thought to play an important role in normal processes important for fertility 4 , 67 . For example, the primate CL undergoes significant changes in immune cell content during its development and regression at the end of the menstrual cycle 68 , 69 . Uterine immune cells are thought to play an important role in the process of embryo implantation 70 , 71 . There were no immunophenotypic changes detected in uterine and regional lymphoid tissues following T+WSD treatment. Collectively, our study indicates that short-term exposure to an obesogenic diet and mild hyperandrogenemia is not sufficient for inducing proinflammatory conditions. In the present study, we also studied the effect of T+WSD treatment on tissue-resident T-cells implicated in immunological memory formation. OM-WAT plays an active role in metabolism and regulation of immune responses to antigens 72 , 73 . Recent studies show that obesity and type 2 diabetes can alter CD8+ T-cell responses against viral infection and cancer 58 , 74 – 76 . The development of immunological dysfunction of WAT-resident T-cells has been reported in obese subjects 77 – 80 . However, the tissue-resident immune cell phenotypes of obese females with hyperandrogenemia remains poorly understood. We found that the frequencies of Tem were significantly reduced in OM-WAT of T+WSD-treated animals. Intriguingly, frequencies of WAT-resident Treg cells were inversely correlated with HOMA-IR, fasting insulin, fasting glucose, glucose AUC, and body fat, suggesting that metabolic dysfunction has a negative impact on the proliferating pool of Tem. In summary, our study demonstrates an intriguing connection between T+WSD-induced obesity and a decline in CD8+ Tem in OM-WAT, and warrants future studies addressing the role of T, insulin and other metabolic factors in regulating Tem differentiation and function in health and PCOS. Our study has several limitations. (1) The small sample size may have limited our power to detect smaller tissue-specific effects of T+WSD on immune cell subsets. (2) The higher number of animals recruited to the T+WSD group compared to the control group was dictated by the experimental design of this study. Specifically, additional T+WSD animals were initially generated to compensate for potential animal loss due to endometriosis, as our earlier studies demonstrated that long-term T+WSD treatment resulted in earlier onset and more severe types of endometriosis 81 . However, none of the animals treated with T+WSD for one year developed endometriosis, suggesting that longer treatment and/or animal age can influence disease progression. (3) The effects of short-term T+WSD treatment on reproductive phenotypes are not included here and will be reported in the separate study (manuscript in preparation). In our previous studies, long-term T+WSD treatment did not result in anovulatory cycles, but did result in ovarian and uterine dysfunction, leading to reduced fertility 12 , 13 . While long-term exposure to (experimentally induced) hyperandrogenemia and WSD increased the number of antral follicles 12 , it is clear that long-term and particularly short-term T+WSD regomens do not fully recapitulate all aspects of PCOS. (4) The present study did not examine the effects of T or WSD alone on immune cell characteristics and highlights the need to better discriminate the effects of WSD versus excess T on the immune system. (5) While the functional significance of WAT-resident Tems remains to be determined, it is possible that WAT forms a unique reservoir for pathogen-responsive Tems and that Tem responses to specific pathogens are modulated by obesity 82 .

Reproductive tract dysfunction and subfertility or infertility are associated with hyperandrogenemia (elevated testosterone, T) and metabolic abnormalities, including insulin resistance, obesity, and white adipose tissue (WAT) dysfunction. Hyperandrogenemia and metabolic dysfunction are hallmarks of polycystic ovary syndrome (PCOS), which can have varying degrees of ovulatory dysfunction and the development of arrested antral follicles that give rise to “cystic” appearing ovaries 1 – 3 . Adipocyte hypertrophy, induced by obesity and hyperandrogenemia, has been suggested to contribute to the systemic low-grade inflammation 4 – 7 thought to be one of the drivers leading to reproductive dysfunction in women with PCOS 8 . Therefore, the primary objective of this study was to focus on the metabolic and immunological effects of an obesogenic high-fat Western-style diet (WSD) and mild hyperandrogenemia (T+WSD) on the systemic and tissue-resident proinflammatory milieu using a nonhuman primate (NHP) model. We have previously demonstrated that chronic consumption of a WSD in combination with mild hyperandrogenemia induced greater metabolic and WAT dysfunction in female rhesus macaques than in either treatment alone 9 – 11 . In these earlier studies, elevated T and WSD consumption were initiated at puberty, resulting in an increase in body fat mass and insulin resistance following three years of treatment 10 , 11 . To address the effects of T+WSD in the absence of simultaneous peripubertal developmental endocrine effects, we generated a cohort of six-year-old adult female rhesus macaques exposed to a low-fat diet and no exogenous T (controls) versus those treated with T+WSD for one year ( Figure 1A ). Individual treatments (i.e., T or WSD alone) were not included in this study because we previously determined that combined treatment led to earlier and more severe metabolic 10 , 11 and reproductive imparements, including increased numbers of arrested antral follicles and reduced fertility 12 , 13 . We determined the proinflammatory responses of tissue-resident immune cells by flow cytometry and assessed systemic inflammation by measuring the levels of circulating proinflammatory mediators ( Figure 1A ). To test whether T+WSD disrupts the immune milieu that may affect normal egg implantation and other relevant reproductive processes, we conducted immune cell profiling during the mid-luteal phase, the interval during the menstrual cycle when the uterine endometrium is permissive for embryo implantation. Immune cells residing in omental (OM)-WAT, the corpus lutetium (CL), the contralateral ovary (OV) not containing the CL, endometrium (ENDO), lymph nodes (LNs), bone marrow (BM; a source of myeloid cells 14 , 15 ), and peripheral blood mononuclear cells (PBMCs) were characterized by flow cytometry, using NHP-specific antibodies 16 , 17 ( Figure 1A ). Our study demonstrates that T+WSD treatment of adult female rhesus macaques increases fat mass and induces insulin resistance and adipocyte hypertrophy, but has no significant effect on systemic or tissue-resident proinflammatory markers, suggesting that hyperandrogenemia and adipocyte hypertrophy alone are not sufficient to trigger a proinflammatory response.