Mucosal and Systemic γδ⁺ T Cells Associated with Control of Simian Immunodeficiency Virus Infection

As a major component of the mucosal immune system, γδ T cells may play a crucial role in early events during HIV transmission. Mucosal surfaces are the main entry point for HIV, and γδ T cells represent a key constituent of GALT (1). Compared with αβ T cells, γδ T cells have few available V gene segments in the TCR, with three main Vδ segments and seven functional Vγ segments (2). Most studies have addressed two γδ T cell subsets, Vδ1 and Vδ2. Vδ1⁺ T cells are mainly found in epithelial tissues of the intestine (2) and are efficiently triggered by host ligands, including stress-induced self-Ags, glycolipids presented by CD1c (3), and MIC A/B molecules (4). In contrast, Vγ9Vδ2 T cells, the Vδ2 subset, are mainly found in peripheral blood and are activated by nonpeptidic molecules, such as phosphoantigens, produced by many microbial pathogens and stressed cells (5, 6). Activated γδ T cells exhibit a multiplicity of effector functions, including direct killing of infected cells, Ab-dependent cellular cytotoxicity, and production of cytokines such as IFN-γ (7) and IL-17 (8, 9). γδ T cells have Ag-presenting and regulatory functions and through cross-talk enhance the cytotoxic effector function of NK cells (10). Activated γδ T cells can acquire B cell helper activity and thus might modify adaptive immunity by regulating Ab responses (11, 12). Interaction of γδ T cells with dendritic cells also impacts responses of both cell types (13).

γδ T cells in HIV infection have received only limited attention. HIV affects several lymphocyte subsets and impairs both acquired and innate immunity. Both Vδ1 and Vδ2 T cells are altered during HIV infection. Disease progression is associated with depletion of Vδ2 T cells and expansion of Vδ1 T cells, leading to an inversion of the normal Vδ2/Vδ1 ratio in peripheral blood and mucosal tissues (6, 14–18). This alteration may result from infected cells accumulating phosphoantigens, leading to a brief period of activation and rapid expansion of Vδ2 T cells that subsequently decline and become dysfunctional by an unknown mechanism (18–20). In contrast, microbial translocation across the gut epithelium may induce expansion of peripheral Vδ1 T cells (8). The HIV/SIV-mediated changes in γδ T cells appear to be part of a strategy for evading antiviral immunity and establishing persistent infection with chronic disease.

In addition to the blood and intestinal tissue, γδ T cells are also present in uterine tissue (21) and the endocervix (17). In the female reproductive tract (FRT), the upper tract (fallopian tubes, uterus, and endocervix) is lined by a single layer of columnar epithelial cells joined by tight junctions. The lower tract (ectocervix and vagina) lining is composed of stratified squamous epithelium that, unlike the upper reproductive tract, relies primarily on the presence of multiple layers to provide a protective barrier against the entry of organisms such as HIV/SIV (22, 23). Recently, HIV-1 transmission was shown to occur in both the upper and lower FRT (24). This compartment exhibits different immune microenvironments that influence HIV infection. Increased activation and factors important for immune responses have been reported in the ectocervix/endocervix, whereas the endometrium has shown high expression of factors that support HIV infectivity and favor HIV replication (25). Vδ1 is the predominant γδ T cell subset in the endocervix of uninfected women, but its frequency decreases in HIV-positive women, along with that of Vδ2 T cells in both the endocervix and blood (17). However, no functionality of γδ T cells in the FRT has been assessed, and in general innate and adaptive immune responses in the FRT to HIV/SIV infection have not been completely defined because of complex regulation by female sex hormones and the degree of compartmentalization (22, 23, 26–29). Elucidating immunological mechanisms operative in the FRT and determining their impact on HIV transmission and control will be important for developing better strategies for prevention and treatment, as ∼50% of HIV infections worldwide are in women (30).

Most of the studies characterizing lymphocytes in mucosal tissue, including the FRT and rectum, have focused on αβ T cells, dendritic cells, B cells, and NK cells (31–34). Although the impact of SIV infection on peripheral and mucosal γδ T cells has been reported in some studies (8, 16), γδ T cells from the upper and lower FRT have been poorly explored in nonhuman primate models. In this study, we focus on the distribution and functional properties of Vδ1 and Vδ2 T cells in peripheral blood of naive and SIV-infected rhesus macaques and in mucosal tissues: rectal, ectocervix, endocervix, and vagina, of chronically SIV-infected macaques to define the relationship of these cells to disease progression. In view of previously observed sex differences in HIV pathogenesis (35, 36) and SIV vaccine outcome (37), we also explored potential phenotypic and functional differences of peripheral and mucosal γδ T cells between males and females.

Freshly isolated PBMCs from 11 naive and viably frozen PBMC from 14 (4 females, 10 males) SIV_mac251-infected Indian rhesus macaques (Macaca mulatta) were used for the assays. Eleven of the infected macaques had been previously vaccinated with various vectored SIV vaccines, followed by boosts with SIV envelope protein, whereas three had served as unvaccinated controls (Supplemental Table I) (37–39). Blood was collected between 14 and 76 wk postinfection, and viral loads ranged from <50 to 1 × 10⁷ SIV RNA copies/ml plasma (geometric mean of <3.1 × 10⁴). Additionally, vaginal, ectocervical, endocervical, and/or rectal tissues were collected at necropsy (40–52 wk postinfection) from 16 (10 females and 6 males) SIV_mac251-infected macaques, part of a previous vaccine study (37). Their viral loads ranged from <50 to 7.3 × 10⁶ SIV RNA copies/ml plasma (geometric mean of <1.3 × 10⁴). Although the 14 blood and 16 mucosal tissues were collected from 26 different SIV-infected animals, no differences were observed in the chronic viral load between both groups (data not shown).

Peripheral blood samples were collected into EDTA-treated collection tubes. PBMCs were obtained by centrifugation on Ficoll-Paque PLUS gradients according to the product insert (GE Healthcare, Piscataway, NJ). Cells were washed with PBS and resuspended in R10 medium (RPMI 1640 containing 10% FBS, 2 mM l-glutamine [Invitrogen], and antibiotics). Rectal pinches (20 per animal) and pinch biopsies of the vagina, ectocervix, and endocervix (40 per tissue) were obtained from each macaque following euthanasia and rinsed with prewarmed R10. The pinches were minced in 5 ml 40 μg/ml Liberase (Roche) solution using a scalpel, transferred to a 50-ml tube, and brought up to 10 ml with Liberase solution. Tissues were digested for 25 min (rectal tissue) or 45 min (FRT tissue) at 37°C with pulse vortexing every 5 min. The dissociated cells and tissue fragments were passed five times through a blunt end cannula using a 20-ml syringe. Cell suspensions were finally passed through a 70-μm cell strainer and washed with 40 ml R10. Cell pellets were resuspended in R10, and cells were counted and distributed for ex vivo phenotypic and/or in vitro functional analyses by FACS.

γδ T cell subsets were identified using a fixable aqua blue dead cell stain (Life Technologies) and a combination of Abs, including CD3-allophycocyanin-Cy7 (SP34-2), pan γδ TCR-PE (B1), and CD4-Pac blue (L200) (all from BD Biosciences); Vδ2-FITC (15D; Thermofisher); and CD8-A×700 (RPA-T8; eBioscience). CD3⁺ T cells were divided into Vδ1⁺ and Vδ2⁺ populations, as described by Harris et al. (8) and as illustrated for a mucosal sample (Supplemental Fig. 1), and were further subdivided by their CD4 and CD8 expression patterns. NKG2 receptor expression was assessed using combinations of anti–NKG2A-allophycocyanin (Z99; Beckman Coulter) and anti–NKG2D-PE-Cy7 (1D11; BioLegend). For homing receptor expression and activation markers, the following Abs were used: α4β7-allophycocyanin (NIH NHP Reagent Resource) and CCR7-PE-Cy7 (3D12 (BD Biosciences) and CD69-Pac blue (FN50; BioLegend). For representative staining, see Supplemental Fig. 2. For intracellular staining, cells were fixed and permeabilized using a Perm/Fix solution (BD Biosciences) prior to incubation with IFN-γ PE-Cy7 (4S.B3; BD Biosciences), TNF-α PerCpCy5.5 (Mab11; BioLegend), CD107-PE-Cy5 (H4A3; BD Biosciences), and granzyme B (Grz B) allophycocyanin (GB12; Invitrogen). At least 50,000 CD3⁺ T cell events were acquired on a LSRII (BD Biosciences) and analyzed using FlowJo software version 9.8.5 (Tree Star).

Peripheral and mucosal lymphocytes were mitogenically stimulated with PMA (50 ng/ml) and ionomycin (250 ng/ml). Mononuclear cells were incubated with monensin (Golgi Stop; BD Biosciences), according to the manufacturer’s instructions, and added at the start of the incubation. After 6 h, cells were washed and stained for γδ T cell subsets and cytokines and cytotoxicity markers, as described above. Values reported are after subtraction of nonstimulated control values. For ex vivo analysis, after 6 h resting without stimulation, each γδ T cell subset was further interrogated for the expression of NKG2 receptors, α4β7 (GALT homing), CCR7 (lymph node [LN] homing), and CD69 (activation). In some cases, a limited number of mucosal cells was obtained, and only some analyses could be performed.

The Wilcoxon rank-sum analysis was used for the comparison of phenotypic and functional data between naive and SIV-infected macaques. The Wilcoxon signed-rank test was used to test for differences in paired samples within groups. The Spearman rank correlation test was used to assess the relationships of γδ T cell phenotype and function with viral loads. Figures display means with or without SEM or medians with or without interquartile ranges. All p values are two sided. Corrections for multiple comparisons have been addressed as follows: in panels with three or four sets of values to compare, the p values shown are not corrected, but marginal p values >0.025 were considered nonsignificant. In panels with eight or more sets of values, the p values shown have been corrected by the Hochberg method for the number of unpaired or paired comparisons in the panel. Statistical analysis was performed using GraphPad Prism V6.01 (GraphPad Prism Software, La Jolla, CA) and SAS/STAT software version 9.3 (SAS Institute, Cary, NC).

In view of the increasing interest in defining protective immune mechanisms against HIV/SIV infection in both peripheral and mucosal compartments, we investigated γδ T cells in SIV-infected and noninfected rhesus macaques, initially determining the distribution of various subpopulations. Unlike humans, the peripheral blood of naive and SIV-infected macaques exhibited a higher frequency of Vδ1 T cells than Vδ2 T cells (p = 0.0010 and 0.0002, respectively; Fig. 1A) in agreement with Wang et al. (40). Within the FRT, Vδ1 and Vδ2 T cell frequencies were comparable across all three compartments (Fig. 1B). In rectal tissue of all macaques, Vδ1 cells were more prevalent than Vδ2 cells (p = 0.029; Fig. 1C); no differences were seen when macaques were analyzed by sex (data not shown). The Vδ1/Vδ2 ratio was seen to be similar in peripheral blood of naive and SIV-infected macaques (Fig. 1D). Unfortunately, no mucosal tissue was available from naive macaques for this study. Examination of FRT tissue from SIV-infected animals showed a nonsignificant upward trend of the Vδ1/Vδ2 ratio in the endocervix compared with the other two tissues (Fig. 1E). The Vδ1/Vδ2 ratio in rectal tissue showed no differences between males and females (Fig. 1F).

Expression of CD4 and CD8 was next explored in both Vδ1 and Vδ2 T cells. Analysis of peripheral blood revealed that both Vδ1 and Vδ2 γδ T cells in naive and SIV-positive macaques are mainly CD8⁺ and exhibit significant differences in prevalence compared with each of the other three CD4/CD8 subsets, with the exception of the marginally nonsignificant differences between the CD8⁺ and CD4⁻CD8⁻ subsets for both Vδ1 and Vδ2 populations in naive macaques (Fig. 1G, 1H). Although no alteration in the Vδ1/Vδ2 ratio was observed, we found that in SIV-infected macaques the frequency of Vδ1 T cells expressing CD4⁺ and CD4⁻CD8⁻ decreased, but CD8⁺ and CD4⁺CD8⁺ subsets increased compared with naive animals (p < 0.02 for all four comparisons; Fig. 1G). In the rectal mucosa, the majority of Vδ1 cells were CD8⁺ or CD4⁻CD8⁻ (Fig. 1I), whereas rectal Vδ2 T cells predominantly expressed CD8⁺ compared with the CD4⁻CD8⁻ population (p = 0.0002; Fig. 1J). In the different FRT compartments, statistically significant differences in the CD8⁺ and CD4⁻CD8⁻ subpopulations were not observed in either γδ T cell subset (Fig. 1I, 1J). Frequencies of Vδ1 and Vδ2 CD8⁺ and CD4⁻CD8⁻ T cell subsets were similar across the mucosal tissues, and few CD4⁺ and CD4⁺CD8⁺ cells were observed (Fig. 1I, 1J). No differences were found between females and males in the different rectal Vδ1 and Vδ2 CD4 and CD8 T cell subsets (data not shown), so results of both sexes were plotted together.

To elucidate potential trafficking of γδ T cells, we next investigated peripheral blood γδ T cells for expression of the homing markers: α4β7 (GALT) and CCR7 (LN). We also explored activation status by examining the expression of CD69, the inhibitory NK receptor NKG2A, and the activating NK receptor NKG2D. γδ T cell subsets, Vδ1 and Vδ2, from both naive and SIV-infected macaques, exhibited similar expression levels of α4β7 (Fig. 2A), whereas CCR7 expression on Vδ2 cells was lower in SIV-infected macaques compared with naive (p = 0.0072; Fig. 2B). Vδ2 also exhibited lower expression of CCR7 than Vδ1 T cells (p = 0.0017). However, levels of the activation marker, CD69, in Vδ1 (p = 0.0003) and Vδ2 (p = 0.0042) cells were higher in SIV-infected compared with uninfected macaques (Fig. 2C). Additionally, the Vδ1 subset exhibited a more highly activated profile than Vδ2 T cells regardless of SIV infection (p = 0.0020 and 0.0067 for naive and SIV⁺ cells, respectively; Fig. 2C). Vδ1 cells of SIV-infected macaques had increased expression of the activating receptor NKG2D (p < 0.0001; Fig. 2D). For the Vδ2 T cell subsets, no differences were observed in expression of NKG2D or NKG2A receptors between infected and uninfected rhesus macaques (Fig. 2D, 2E). For both naive and SIV-infected macaques, peripheral Vδ2 cells compared with Vδ1 cells showed higher expression of NKG2A (p = 0.0010 and 0.0001, respectively; Fig. 2E). However, higher expression of NKG2D was seen on Vδ2 compared with Vδ1 T cells from naive macaques (p = 0.0010; Fig. 2D), suggesting a different regulatory mechanism in infected animals.

We next examined homing markers on peripheral blood CD4 and CD8 γδ T cell subsets to assess potential trafficking to GALT and LN. Vδ1 CD4⁺ T cells of SIV-infected macaques showed a lower expression of α4β7 (p = 0.0056; Fig. 2F) and a similar although nonsignificant trend for CCR7 (Fig. 2G) than naive macaques. Also, Vδ2 CD8⁺ and CD4⁻CD8⁻ T cells from SIV-infected macaques showed marginally decreased expression of CCR7 compared with naive animals (p = 0.06 for both; Fig. 2I). Regardless of SIV infection, α4β7 is mainly expressed by CD4⁺CD8⁺ and CD8⁺ subpopulations in Vδ1 and Vδ2 T cells, respectively (Fig. 2F, 2H). In contrast, CCR7 expression was high in CD4⁺ Vδ1 cells from naive macaques, whereas expression was high in the CD4⁺CD8⁺ subset from SIV-positive macaques (Fig. 2G). Overall, however, fewer Vδ2 T cells expressed CCR7 in both naive and SIV⁺ animals (Fig. 2I). Because only a limited number of cells was obtained from mucosal biopsies, homing, NK receptors, and activation markers could not be assessed in those tissues.

During HIV/SIV infection, peripheral blood γδ T cell functions are dysregulated (6–8, 18). Therefore, the functionality of peripheral and mucosal Vδ1 and Vδ2 subsets from naive and SIV-infected macaques was of interest. Lymphocytes from peripheral blood and mucosal tissues were left unstimulated or were mitogenically stimulated with PMA/ionomycin for 6 h and then evaluated for cytokine production. Unstimulated, resting peripheral blood Vδ1 T cells from SIV-infected macaques exhibited higher expression of both CD107 (p = 0.0014; Fig. 3A) and Grz B (p = 0.0012; Fig. 3B) compared with naive macaques. However, no difference in IFN-γ or TNF-α expression between infected and uninfected macaques was observed (Fig. 3C, 3D). When the response of Vδ1 and Vδ2 T cells to PMA/ionomycin stimulation was evaluated, the ability of both subsets of γδ T cells to express CD107, Grz B, and TNF-α was comparable to naive macaques (Fig. 3E, 3F, 3H). However, in SIV-infected macaques, the ability of Vδ2 T cells to secrete IFN-γ (p = 0.0003; Fig. 3G) was lower compared with naive macaques.

Due to a low cell yield from mucosal biopsies, Vδ1 and Vδ2 T cells of SIV-infected macaques were only analyzed following PMA/ionomycin stimulation. No differences in CD107 expression levels were seen between Vδ1 and Vδ2 T cells of the rectal or FRT compartments (Fig. 3I). However, endocervical Vδ2 T cells exhibited greater IFN-γ secretion compared with vaginal Vδ2 T cells (p = 0.016; Fig. 3J) and also higher levels of IFN-γ expression compared with endocervical Vδ1 T cells (p = 0.031; Fig. 3J). This might suggest that γδ T cells from the endocervix exhibit a higher inflammatory profile compared with cells from the other compartments within the FRT. No difference was found between females and males in expression of CD107 or IFN-γ by rectal Vδ1 or Vδ2 T cells after stimulation (data not shown), so the results of both sexes were plotted together.

We also evaluated the response of peripheral CD4 and CD8 Vδ1 and Vδ2 T cell subpopulations to PMA/ionomycin stimulation (Fig. 4A–D). In view of the large number of tests, few of the differences observed remain significant after correction for multiple comparisons. We have called those that exhibited a p value <0.05 but were not significant after correction “tentative” (see legend to Fig. 4), indicating that they are of interest, but require additional confirmation. Thus, no differences of note are seen among the various CD4/CD8 subpopulations of Vδ1⁺IFN-γ⁺ (Fig. 4A) or Vδ2⁺TNF-α⁺ T cells (Fig. 4D) of naive and SIV⁺ macaques. However, following stimulation, circulating Vδ1 CD8⁺ and CD4⁻CD8⁻ T cells of infected macaques produced higher levels of TNF-α (tentatively significant; Fig. 4B) compared with naive, whereas CD8⁺ and CD4⁻CD8⁻ Vδ2 T cells produced lower levels of IFN-γ (p = 0.019 and p = 0.0021, respectively; Fig. 4C) than naive macaques. No differences in CD107 and Grz B expression in SIV-infected or naive macaques were seen in the various subpopulations of γδ T cells expressing CD4 or CD8 (data not shown).

In mucosal tissue from SIV-infected macaques, Vδ1 CD8⁺ T cells from both ectocervix and endocervix showed higher expression of CD107 compared with vaginal γδ T cells (tentatively significant; Fig. 4E). No differences were observed in the expression of CD107 by Vδ2 T cells across the different mucosal tissues analyzed (Fig. 4F). Endocervical Vδ1 CD8⁺ T cells also exhibited higher IFN-γ production compared with vaginal cells (tentatively significant; Fig. 4G), and they were also significantly elevated above the endocervical CD4⁻CD8⁻ cells (tentatively significant; Fig. 4G). For Vδ2 T cells, both CD8⁺ and CD4⁻CD8⁻ endocervical subpopulations exhibited higher frequencies of IFN-γ⁺ cells (p = 0.023 and tentatively significant, respectively; Fig. 4H) compared with vaginal cells.

Sex-related differences have been documented for various immunological parameters in mucosal tissues (41) as well as for their impact on AIDS disease progression (18, 35, 36). However, a potential sex bias in γδ T cell distribution and functionality has not been explored in HIV/SIV infection. In this work, we interrogated γδ T cells in rectal tissue from SIV-infected female and male macaques for the expression of IFN-γ and the CD107 degranulation marker. No difference was seen between sexes in IFN-γ production after stimulation with PMA/ionomycin (Fig. 5A). Although chronically SIV-infected males maintained higher levels of Vδ1 CD8⁺ T cells expressing CD107 compared with females, suggesting that viremia control in rectal tissue of males might be associated with cytolytic activity of Vδ1 T cells, this difference was only tentatively significant due to the number of tests (Fig. 5B).

Natural virus suppressors that control virus replication maintain levels of Vδ2 T cells equivalent to those of healthy controls (42–44). We explored whether the control of viral load in chronically SIV-infected macaques was associated with better preservation of γδ T cells in peripheral blood and/or in mucosal tissue. A significant negative correlation was observed between levels of circulating Vδ1⁺ CD4⁺ T cells and chronic viremia (p = 0.0018; Fig. 6A) and a marginally significant negative correlation between Vδ2⁺ CD8⁺ T cells and chronic viremia (p = 0.045; Fig. 6A). Also, significant negative correlations were seen between Vδ1 CD4⁺ IFN-γ⁺ and Vδ2 CD8⁺ IFN-γ⁺ T cells and chronic viremia (p = 0.012 and p = 0.031, respectively; Fig. 6B, 6C). Circulating NKG2D⁺ Vδ2 T cells were also negatively correlated with chronic viral loads (p = 0.020; Fig. 6D). No significant correlations between γδ T cells in rectal tissue and viremia of either males or females were observed (data not shown). In contrast, in mucosal tissue of the FRT, total endocervical Vδ2 T cells correlated negatively with chronic viremia (p = 0.046; Fig. 6E), as did CD8⁺ Vδ2 T cells (p = 0.0032; Fig. 6F).

Although effects of HIV/SIV infection on peripheral blood γδ T cells have been reported, few studies have examined the more populous mucosal γδ T cells during HIV/SIV infection, and, to our knowledge, none have investigated the effects of SIV infection on γδ T cells in the different regions of the FRT of rhesus macaques. Our interest in this area was stimulated by the known sex bias in HIV pathogenesis (35, 36) and our recent observation of a sex bias in vaccine-induced protective efficacy (37). Sex differences in viral pathogenesis are associated with differences in immune responses (45). The observed sex bias in vaccine-induced protection was correlated with viral-specific mucosal IgA and mucosal memory B cells (37). These findings suggested that differences might also occur between males and females in γδ T cell populations at the mucosal sites of SIV infection and replication. Therefore, in this study, we investigated effects of SIV infection on phenotypic and functional characteristics of γδ T cell populations in the rectum and the FRT in comparison with those in peripheral blood.

To obtain a comprehensive picture of the γδ T cell populations in our rhesus macaques, we initially characterized the cells in the peripheral blood of naive and SIV-infected animals. Unlike humans, macaques exhibit a predominance of Vδ1 T cells in peripheral blood (40). A further expansion of peripheral Vδ1 T cells in SIV infection has been attributed to microbial translocation across the mucosal epithelium (8). Although in this study we did not observe a change in the Vδ1/Vδ2 ratio in blood of SIV-positive macaques compared with naive animals (Fig. 1D), we did see changes in Vδ1⁺ CD4⁺ and CD8⁺ T cell subsets (Fig. 1G). The decreases in CD4⁺ and CD4⁻CD8⁻ γδ T cells of SIV-infected macaques might reflect CD4 depletion occurring by a direct or indirect mechanism. Concomitantly, an increase in the Vδ1 CD8⁺ and CD4⁺CD8⁺ subsets suggests the expansion of cytotoxic cells in response to infection. Although we observed no depletion of Vδ2 T cells in SIV-infected macaques in comparison with the naive animals, this population of cells has been reported to be lost and/or to become dysfunctional early following HIV/SIV infection (6–8, 16–18). Alteration in γδ T cell subsets has been widely shown during HIV infection (5, 17, 18, 44). In contrast, only a single report concerning Vδ1/Vδ2 ratio alterations in SIV-infected rhesus macaques has appeared (8), showing a marginally significant difference in blood between naive and SIV-infected macaques. Unfortunately, we were unable to obtain mucosal tissues from naive macaques to determine whether we could confirm a similar increase in Vδ1/Vδ2 ratio in gut tissue following SIV infection, as seen previously (8). Future longitudinal studies following SIV infection of rhesus macaques might help to explain the lack of alteration of the Vδ1/Vδ2 ratio seen in our study. Analysis of the microbiome might elucidate the finding as well, as the altered ratio observed previously was attributed to microbial translocation.

The trafficking of γδ T cells between peripheral blood and secondary lymphoid organs has not been explored to any extent. In this work, we investigated the impact of chronic SIV infection on the homing profile of peripheral γδ T cells to gut and LN. No alteration in the frequency of total α4β7⁺ Vδ1 or Vδ2 T cells was observed (Fig. 2A), whereas the frequency of CCR7⁺ Vδ2 T cells was decreased during SIV infection (Fig. 2B), suggesting diminished active trafficking of Vδ2 T cells to LN. The proportions of Vδ1⁺α4β7⁺ were decreased in the blood of infected macaques (Fig. 2F, along with a downward trend inVδ1⁺ CCR7⁺ CD4⁺ T cells, Fig. 2G). We speculate that a decline in Vδ2⁺ T cells trafficking to the LN might influence B cell responses. Human peripheral blood Vδ2⁺ T cells have been reported to help B cells secrete Ab (46), and mouse splenic γδ T cells have been shown to modulate preimmune B cell function (47). Moreover, a recent finding in mice has shown that, in the absence of αβ T cells, γδ T cells are localized in close proximity to B cells within germinal centers (48). Whether the Vδ2 T cell subset in macaques can perform functions similar to those of LN T follicular helper cells that are lost or dysregulated during SIV infection (49) will require further investigation.

The cytolytic function of γδ T cells is tightly regulated by receptors such as NKG2A, NKG2D, and NKG2C (50, 51). HIV/SIV infection activates γδ T cells, as shown in this work by the increase in CD69⁺ γδ T cells (Fig. 2C), consistent with previous reports (52). Activated γδ T cells also express the NKG2C receptor that triggers cytotoxic function in Vδ1 T cells (53). Unfortunately, we lacked an Ab reactive with NKG2C for macaque cells, but observed an increased frequency in SIV-infected animals of γδ T cells expressing NKG2D (Fig. 2D), another activating receptor. Potent in vitro antiviral activity of Vδ1 T cells from HIV-infected patients has been described involving engagement of the NKp30 receptor and the combined effect of NKp30-induced CC chemokines: CCL3, CCL4, and CCL15 (54). Cytotoxicity of Vδ1 T cells against multiple myeloma cells has been reported to be mediated in part by the TCR, also involving NKG2D (55). In this work, support for a role of Vδ1 cells as cytotoxic effectors includes expression of CD107a and Grz B (Fig. 3A, 3B). Vδ2 cells were also shown to express IFN-γ (Fig. 3G). However, activated Vδ1 cells may also function as regulatory T cells, as they have been reported to have strong suppressive activity (56–58). Clearly, this subpopulation of γδ T cells requires further in-depth investigation to elucidate which role it plays in HIV/SIV infection and whether it would be a useful therapeutic target.

Our main goal was to explore the distribution and functionality of γδ T cells in mucosal tissues as the primary site of HIV entry. Mucosal immune responses to HIV/SIV are likely to be critical for providing protection against infection and disease progression (37, 59–61). Unfortunately, we did not have access to mucosal tissues of naive macaques, but were only able to examine mucosal samples (rectal and FRT biopsies) of SIV-infected animals. Nevertheless, we were able to acquire a generalized overview of mucosal γδ T cells in the SIV-infected rhesus macaque. As previously reported in nonhuman primates (8), the intestinal mucosa of SIV-positive macaques exhibits predominantly Vδ1 T cells. We confirmed that in this work, showing that Vδ1/Vδ2 ratios in rectal tissue (Fig. 1F) were somewhat elevated compared with the three tissues of the FRT (Fig. 1E). In both rectal tissue and the FRT, Vδ1 and Vδ2 CD8⁺ and CD4⁻CD8⁻ were the main subsets of γδ T cells. CD4⁻CD8⁻ γδ T cells are particularly important during pregnancy, as they exhibit regulatory function by secreting TGF-β and IL-10 to support tolerance and avoid maternal rejection of the fetus, thereby maintaining the pregnancy (62). This illustrates the importance of defining the roles of different CD4 and CD8 γδ T cell subsets in different tissues. In this work, we saw elevated CD8⁺CD107⁺ and IFN-γ⁺ Vδ1 cells in the ectocervix and endocervix of the FRT (Fig. 4E, 4G), suggesting cytotoxic function at sites of SIV infection. IFN-γ⁺ Vδ2 cells also appeared to be elevated in the endocervix compared with the vagina (Fig. 4H), perhaps enhancing recruitment of other immune cells and mediating effector functions (7, 63). Overall, γδ T cells in the endocervix exhibited a high inflammatory profile compared with the other two compartments within the FRT. An examination of rectal γδ T cells revealed no differences between males and females. Both sexes showed similar Vδ1/Vδ2 ratios (Fig. 1F). Therefore, these cell populations did not seem responsible for the sex bias we observed previously in which female macaques exhibited delayed SIV acquisition following repeated low-dose rectal challenges (37). We noticed, however, that males tended to have higher frequencies of rectal CD107⁺ Vδ1 T cells (Fig. 5B), suggesting that they might be able to better control local viremia, again highlighting the potential cytotoxic function of this cell population. This potential sex difference in mucosal γδ T cells in SIV infection requires confirmation. In general, several subpopulations of both Vδ1 and Vδ2 cells in peripheral blood were correlated with decreased chronic viremia (Fig. 6A–D), strengthening the presumed protective role of these cells. However, in the FRT, it was the Vδ2 subpopulation in the endocervix that was significantly correlated with control of plasma viremia (Fig. 6E, 6F).

The potential role of peripheral γδ T cells in viremia control was previously reported in natural virus suppressors (42). However, to our knowledge, this is the first report concerning a protective role for Vδ2 T cells of the FRT. Results of clinical studies have supported a protective role of peripheral Vδ2 T cells. HIV suppressors maintain equivalent levels of Vδ2 T cells compared with healthy controls (42–44). Although we also observed in this work in SIV-infected macaques significant correlations of Vδ1 subpopulations with decreased viremia (Fig. 6A, 6B), no in vivo protective role of these cells has been demonstrated in HIV disease. However, the frequency of peripheral Vδ1 CD4⁺ T cells was positively correlated with the CD4 T cell count in HIV-infected individuals (44).

To our knowledge, this is the first study of phenotypic and functional characteristics of γδ T cells in the FRT that has revealed a potentially protective role of Vδ2 T cells in controlling viremia. Further studies examining γδ T cells in the FRT early in SIV infection and in naive animals are needed to determine whether these cells contribute to protection against acquisition (by control of infectious foci) and/or to control of acute viremia. Although we did not track menstrual cycle phases in this study, it will be important to include this parameter in future investigations to determine whether the complex regulation of the FRT by female sex hormones impacts γδ T cell frequencies and function. Additionally, the contribution of peripheral γδ T cells, although a small population of cells, should not be overlooked in overall control of chronic viremia. Efforts to strengthen the innate γδ T cell immune response against HIV/SIV may have important implications for both treatment strategies and prevention of transmission through mucosal surfaces.

We thank the veterinarians and staff at the National Institutes of Health Animal Facility for expert care of the macaques and implementation of the research protocols; Kathy McKinnon and Sophia Brown for flow cytometric support; and Ranajit Pal and Maria Grazia Ferrari (Advanced BioScience Laboratories) for quantification of SIV RNA viral loads.

The authors have no financial conflicts of interest.