Male and female C57BL/6 mice were obtained from The Jackson Laboratory (Stock 664) or the National Institute on Aging. T cell receptor delta chain knockout mice (TCRδ KO, B6.129P2-Tcrdtm1Mom/J, Stock 2120) were obtained from The Jackson Laboratory and bred in-house. Different age groups of mice were utilized ranging from 2 to 24 months; age is specified for each experiment in figures and/or legends. Mice were housed in pressurized intraventilated cages and maintained in an environment under controlled temperature (21°C–23°C), humidity (30%–70%), and lighting (14 h/10 h, light/dark) with free access to drinking water and chow (Teklad Global No. 2918). All procedures were approved by the Institutional Animal Care and Use Committee at the University of Kentucky and performed in accord with the National Institutes of Health guidelines for ethical animal treatment.

Isochronic parabiotic pairs were constructed using age- and sex-matched WT and TCRδ KO male and female mice (male to male or female to female). Young males were cohoused at weaning. Aged males were acclimated for pairing using an in-house developed clear plexiglass barrier for 2 weeks (Supplementary Figure S1). After removal of the barrier, pairs were cohoused for an additional 2 weeks before surgery to assure compatibility. Female mice were cohoused for at least 2 weeks prior to surgery without the barrier. Surgeries were performed following the protocol described by Kamran et al. (2013). Briefly, C57BL/6 and TCRδ KO mice were anesthetized with isoflurane, 4%–5%, provided by a precision vaporizer followed by application of ophthalmic ointment, meloxicam (5 mg/kg s. c), buprenorphine SR-LAB (Zoopharm, 1 mg/kg s.c.), and enrofloxacin (10 mg/kg, s.c.). The appropriate side of each mouse was clipped of fur and the mice transferred to a heated surgical platform with two separate side-by-side nose cones, with isoflurane maintained at 1.5%–2%. With the mice positioned in lateral recumbency, back-to-back, the exposed side of each mouse was aseptically prepped using betadine and isopropyl alcohol, and an incision was made through the skin extending from 5 mm above the elbow to 5 mm below the knee joint. The skin adjoining the incision was gently separated from underlying tissue to create a 5 mm flap surrounding the incision and the exposed joints were joined using 3-0 Prolene suture. The skin along the ventral and dorsal sides of the incision was then approximated and joined with 5–0 Vicryl suture in a continuous pattern. Subsequently, the mice were transferred to a recovery cage, administered 1 mL warm sterile physiological saline (s.c.), and monitored until recovered from anesthesia and were ambulatory. Additional doses of meloxicam were provided at 24- and 48-h post-op and remaining exposed sutures were removed at 14 days. Parabiotic pairs were euthanized at 4 weeks to harvest blood and tissues for flow cytometry analysis.

Mice were deeply anesthetized by isoflurane inhalation (5%), laparotomy performed, and blood collected from the inferior vena cava (IVC) by syringe needle with 10% volume of 0.1 M sodium citrate. Subsequently, the IVC was cut, and the entire vasculature was perfused with 30 mL physiological saline through the cardiac ventricles to eliminate circulating cells. For parabiotic pairs, perfusion was initiated from the heart of one mouse and the IVC of the other mouse was cut to visualize combined blood flow. For flow cytometry, fresh tissues including visceral adipose tissue (VAT) from the gonadal fat pads, spleen, liver (a portion of left lateral lobe) and lymph nodes (inguinal and brachial) were dissected and kept on ice until processing.

VAT, spleen, and blood were processed as previously described (Bruno et al., 2022). Liver and lymph nodes were pressed through a 100-μm cell strainer with a 5 mL syringe plunger, washed with 20 mL digestion buffer (0.5% BSA in HBSS with Ca²⁺Mg²⁺) and centrifuged at 500 × g for 10 min. RBC lysis was performed for liver (BioLegend 420302), followed by passing through a 70-μm strainer. After a second round of centrifugation, cells were resuspended in digestion buffer for subsequent counting, and processing for flow cytometry.

Single-cell suspensions in 250 µL of Dulbecco’s Phosphate Buffered Saline (DPBS, Gibco 14190-144) were stained with Fixable Viability Dye eFluor 450 (eBioscience 65-0843-14) according to the manufacturer’s protocol, and Fc receptor blocking was performed using TruStain FcX (Biolegend 156604) for 10 min on ice in 250 µL of DPBS containing 1 mM EDTA, 25 mM HEPES, 1% FBS. Cells were further incubated for 30 min at 4°C in the dark with respective antibodies for cell surface staining (Supplementary Table S1). For cell surface analyses only, stained cells were fixed with 4% paraformaldehyde (Biolegend 420801) for 20 min. For intracellular staining, SVF cells were permeabilized following surface staining according to standard protocol (Invitrogen 00-5523-00), and intracellular staining was performed using antibodies for Ki-67 and isotype control. Cells were analyzed on a FACSymphony A3 Cell Analyzer (BD, San Jose, CA, United States). Analysis of flow cytometry data was performed using the FlowJo data analysis software (FlowJo, LLC, Ashland, OR, United States).

Stock solution of 5-ethynyl-2′-deoxyuridine (EdU) at 40 mg/mL was prepared by dissolving 10 mg in 250 µL of DMSO which was stored at −20°C. Each mouse was injected with 1 mg (Chen et al., 2019) dosage using 25 µL of stock solution and 375 µL PBS to make 400 µL of total injection volume. Mice were given intraperitoneal injections once daily, 24 h apart, for 3 consecutive days and sacrificed on day four. Body weights were measured before and after 3 injections and did not change significantly. VAT was harvested and processed to obtain single-cell suspensions which were stained with Fixable Viability Dye eFluor 450 and antibodies against cell surface markers (Supplementary Table S1), followed by EdU staining according to manufacturer’s guidelines using the Click-iT Plus EdU Alexa Fluor 594 flow cytometry assay kit (Invitrogen C10646).

VAT and lymph nodes were collected from mice and processed to obtain single-cell suspensions. Cells were resuspended in DPBS containing 1 mM EDTA, 25 mM HEPES, 1% FBS, and Fc receptor blocking was performed followed by surface staining. Cells were then stained with CellEvent™ Caspase-3/7 Green Flow Cytometry Assay kit according to manufacturer’s protocol (Invitrogen C10740). In order to induce apoptosis, cells (∼1 × 10⁶) were incubated with 2 µM ABT737 (Selleck chem S1002) and 1.6 µM Mcl-1 Inhibitor II (Sigma-Aldrich 5.08053) or 0.2% DMSO (vehicle) for 3 h at 37°C with 5% CO₂ similar to the protocol described by Tan et al. (2019). Subsequently, cells were washed with DPBS containing 1 mM EDTA, 25 mM HEPES, 1% FBS and Fc receptor blocking, cell surface staining, and staining for caspase 3/7 performed as described above.

Continuous bivariate associations were analyzed using simple linear regression. Categorical explanatory variables with a continuous response were analyzed using one-way ANOVA with or without repeated measures as appropriate. Post-hoc pairwise comparisons were done using Tukey’s Honest Significant Difference for independent samples and paired t-tests for repeated measures. Normality with constant variance was assumed in all analyses. No evidence of violations of normality were detected.

We previously reported an age-associated increase in γδ T cells in aged (19–24 months) compared to young (4–5 months) mice (Bruno et al., 2022); however, at what age this becomes apparent was not determined. To identify the trend of γδ T cell accumulation in VAT over the lifespan, γδ T cells were evaluated in seven age groups of mice (2 months, 4 months, 8 months, 12 months, 16 months, 20 months, 24 months). CD3⁺ T cells were quantified among lymphocytes and γδ T cells were distinguished by their TCR expression; conventional T cells (T_conv, i.e., αβ T cells) were evaluated in parallel (Figure 1A). The entire flow cytometry gating scheme is shown in Supplementary Figures S2A. Among CD3⁺ lymphocytes, the percentage of γδ T cells progressively increased in VAT with aging (Figure 1B, R ² = 0.4684, p < 0.0001), while the percentage of T_conv cells reciprocally declined with age (Figure 1C, R ² = 0.4335, p < 0.0001). The total number of γδ T cells showed a significant increase at middle age (12–16 months) which further increased at old age (20–24 months) (Figure 1D). The number of T_conv cells likewise significantly increased at middle age (12–16 months); however, no further significant increase at old age was observed (Figure 1E). To understand if adiposity plays a role in the age-related accumulation of T cells in VAT, the number of cells was adjusted for fat mass (see Supplementary Figures S3A, B for body weight and fat mass data). In the resulting data, a significant increase in γδ T cells at old age remained (Figure 1F), indicating that age-related γδ T cell accumulation is adiposity-independent. In contrast, statistical significance for the increase was lost for T_conv cells (Figure 1G), which points to adiposity as a major determinant for T_conv accumulation in VAT over the lifespan. Furthermore, the significant increase in γδ T cells at middle age was also lost after fat mass adjustment, indicating a contribution of mid-age weight gain to γδ T cell abundance. In comparison, aging minimally influenced abundance of γδ T cells and T_conv cells in peripheral lymph nodes (Supplementary Figures S4A–D). Collectively, these data indicate that γδ T cells progressively accumulate in VAT over the lifespan and that the late age-associated increase is independent of adiposity and unique to this tissue-resident population of cells.

Previous studies using parabiosis of CD45.1 and CD45.2 congenic mice have reported VAT γδ T cells in young mice to be a tissue-resident population (Kohlgruber et al., 2018; Goldberg et al., 2020). However, since adipose tissue is chronically inflamed in aging, promoting the secretion of inflammatory cytokines and chemokines, we tested the hypothesis that the inflamed microenvironment of aged VAT provides recruitment signals for peripheral γδ T cells. Isochronic parabiotic pairs of wild-type (WT) and TCRδ KO mice at both young and old age were used to evaluate peripheral γδ T cell migration into VAT (Figure 2A). γδ T cells in blood showed comparable chimerism between WT and TCRδ KO mice at both ages (Figure 2B, young: 55.4% ± 3.8% in WT vs. 44.6% ± 3.8% in KO; aged: 64.1% ± 8.6% in WT vs. 36% ± 8.6% in KO), confirming distribution of circulating cells and validating our model system. In the young pairs, minimal recirculation of γδ T cells was observed into the VAT of the TCRδ KO mouse from its WT pair (Figure 2C, 6.6% ± 6% in young TCRδ KO), confirming previous reports that VAT γδ T cells are tissue-resident (Kohlgruber et al., 2018; Goldberg et al., 2020). Similar data were observed in the aged pairs (2.75% ± 2.3% γδ T cells in the aged TCRδ KO mouse, p = 0.1239 vs. young KO), indicating that the aged microenvironment does not enhance γδ T cell accumulation. This is likely to be unique to VAT as we observed recirculation in liver (Figure 2D) and spleen (Figure 2E). Therefore, while there is continuous low-level migration of γδ T cells into VAT (<10%), which could contribute to accumulation over the lifespan, recruitment of these cells is not enhanced by aging.

Since de novo generation of γδ T cells exists only in the thymus and peripheral recruitment to the VAT is minimal and not affected by aging, we evaluated whether proliferation of tissue-resident γδ T cells is enhanced in the aged. Two different approaches were utilized: in vivo EdU incorporation (Figure 3A) and intracellular staining for the proliferation marker Ki67 (flow cytometry gating schemes are shown in Supplementary Figure S2). The percentage of EdU⁺ γδ T cells significantly declined from young to aged (Figure 3B, 23% ± 12% in young vs. 11% ± 2% in aged, p = 0.0367), indicating that the proportion of proliferating to non-proliferating γδ T cells decreases by aging. The total number of EdU⁺ γδ T cells per gram of VAT, however, showed a significant 6.3-fold increase in the aged (Figure 3C, p = 0.0008). Note that mass of VAT was not significantly different between young and aged mice in this study (data not shown). T_conv cells in VAT showed similar results (Supplementary Figures S5A, B).

We next sought to validate these results utilizing the proliferation marker Ki67, in doing so we added a middle-aged group. The percentage of Ki67⁺ γδ T cells progressively declined from young to old age (Figure 3D, 25% on average in young, 18% in middle age, and 12% in old age, p < 0.001 for young vs. aged comparison). The total number of Ki67⁺ γδ T cells per gram of VAT, however, showed a significant 4.4-fold increase in the aged group only (Figure 3E, p < 0.0001: young vs. aged, p = 0.0025: middle-aged vs. aged). In T_conv cells, the percentage of Ki67⁺ cells declined from young to middle age without further change at old age; however, the total number of Ki67⁺ T_conv cells per gram of VAT showed a significant 5.6-fold increase in only the aged group without any significant difference between young and middle-aged (Supplementary Figures S5C, D). In peripheral lymph nodes, the percentage of Ki67⁺ γδ T cells remained unchanged over the lifespan, while there was a slight increase in total number in the aged (Supplementary Figures S6A, C). T_conv cells in peripheral lymph nodes, on the other hand, showed a slightly higher proportion of Ki67⁺ cells in the aged but no change in total number (Supplementary Figures S6B, D).

Collectively, these results indicate that although the proportion of proliferating to non-proliferating γδ T cells decreases with aging, the total number of proliferating cells within the population increases in old age likely on account of the population itself being expanded over the lifespan. Further, both γδ T cells and T_conv cells in VAT show similar results with respect to the effect of age on proliferation, while cells in the peripheral lymph nodes were minimally affected.

To understand whether aging influences VAT-resident γδ T cell turnover, we analyzed apoptosis via caspase 3/7 activity (FACS gating scheme shown in Supplementary Figures S7). The proportion of apoptotic γδ T cells significantly declined at middle age compared to young without any further change in the aged (Figure 4A, 25% ± 8% in young, 13 %± 10% in middle-aged, 19% ± 7% in aged, p = 0.0363 young vs. middle-aged, p = 0.3685 middle-aged vs. aged). A concomitant increase in the proportion of live cells was observed starting at middle age (Figure 4B, p = 0.0117 young vs. middle-aged, p = 0.2993 middle-aged vs. aged), suggesting that γδ T cells are protected from apoptosis from middle age onwards. However, the absolute number of apoptotic and live γδ T cells per gram of VAT was significantly higher in the aged (Figures 4C, D, p < 0.001 young vs. aged and p < 0.001 middle-aged vs. aged).

To address the role of anti-apoptotic proteins in VAT-resident γδ T cell survival, apoptosis was induced by treating SVF cells isolated from adipose tissues with a combination of Bcl-2 family inhibitors: ABT737 (inhibitor of Bcl-2, Bcl-xl, and Bcl-w) and Mcl-1 inhibitor II (specific to Mcl-1). Apoptosis via caspase 3/7 activity in γδ T cells was compared between control (treated with DMSO) and inhibitor-treated cells (Figure 4E) by FACS with cell surface staining for γδ T cells. In the presence of inhibitors, apoptosis was significantly induced in young γδ T cells (Figure 4F, 18% ± 4% to 45% ± 8%, p = 0.0008). In contrast, apoptosis was not significantly induced in cells from middle-aged mice (8% ± 8% to 16% ± 12%, p = 0.1062). γδ T cells from aged mice were responsive to inhibitor treatment with apoptosis induced from 20% ± 5% to 36% ± 10% (p = 0.0011), although induction of apoptosis appeared mildly blunted compared to young. These data suggest that γδ T cells at middle age may be protected from apoptosis via a mechanism unrelated to the targeted Bcl2 family of anti-apoptotic proteins.

VAT-resident T_conv cells likewise showed a significant decline in the proportion of apoptotic cells from young to middle age, suggesting protection from apoptosis similar to γδ T cells (Supplementary Figures S8A, B). The absolute number of apoptotic and live T_conv cells after adjusting for fat mass also showed similar results to γδ T cells with numbers increasing only in the old age group (Supplementary Figures S8C, D). However, in the presence of Bcl2 family inhibitors, induction of apoptosis was significant in all groups (Supplementary Figures S6E), indicating a unique pro-survival mechanism for γδ T cells.

Contrarily, apoptosis was not reduced in γδ T cells or T_conv cells from the peripheral lymph nodes at middle age or old age. Instead a significant increase in the percent of apoptotic γδ T cells was observed from middle age to old age (Supplementary Figures S9A, B and Supplementary Figures 10A, B). The total number of apoptotic or live γδ T cells in lymph nodes did not significantly change over the lifespan (Supplementary Figures S9C, D); however, T_conv cells showed significant changes by aging (Supplementary Figures S10C, D). Apoptosis of lymph node γδ T cells and T_conv cells was enhanced in all age groups in the presence of Bcl2-family inhibitors although the level of apoptotic cells was already high at baseline in these cells (Supplementary Figures S9E, 10E).

Collectively, our data indicate that T cells (both γδ T and Tconv) are uniquely protected from apoptosis in VAT, beginning at middle age. Further, the presence of a distinctive mechanism for VAT-resident γδ T cells exists which is independent of the traditional Bcl2-family of anti-apoptotic proteins.