Simian Immunodeficiency Virus Infects Functionally Polarized Memory CD4 T Cells Equivalently In Vivo

Functional perturbations of memory CD4 T cells have been suggested to underlie important aspects of HIV disease progression. However, the mechanisms underlying these perturbations remain unclear. Using a nonhuman primate model of HIV, we show that SIV infects functionally defined populations of memory CD4 T cells equally in different anatomic sites. Thus, preferential infection by the virus is unlikely to cause functional perturbations.

ABSTRACT

Among the numerous immunological abnormalities observed in chronically human immunodeficiency virus (HIV)-infected individuals, perturbations in memory CD4 T cells are thought to contribute specifically to disease pathogenesis. Among these, functional imbalances in the frequencies of T regulatory cells (Tregs) and interleukin 17 (IL-17)/IL-22-producing Th cells (Th17/Th22) from mucosal sites and T follicular helper (Tfh) cells in lymph nodes are thought to facilitate specific aspects of disease pathogenesis. However, while preferential infection of Tfh cells is widely thought to create an important viral reservoir in an immunologically privileged site in vivo, whether immunological perturbations among memory CD4 T cell populations are attributable to their relative infectivity by the virus in vivo is unclear. Here we studied peripheral blood and lymphoid tissues from antiretroviral (ARV)-treated and ARV-naive Asian macaques and isolated functionally defined populations of memory CD4 T cells. We then assessed the degree to which these populations were infected by simian immunodeficiency virus (SIV) in vivo, to determine whether particular functionally identified populations of memory CD4 T cells were preferentially infected by the virus. We found that SIV did not preferentially infect Th17 cells, compared to Th1 cells, Th2 cells, or Tregs. Moreover, Th17 cells contributed proportionately to the total pool of infected cells. Taken together, our data suggest that, although Tfh cells are more prone to harbor viral DNA, other functionally polarized cells are equally infected by the virus in vivo and Th17 cells are not preferentially infected.

IMPORTANCE Functional perturbations of memory CD4 T cells have been suggested to underlie important aspects of HIV disease progression. However, the mechanisms underlying these perturbations remain unclear. Using a nonhuman primate model of HIV, we show that SIV infects functionally defined populations of memory CD4 T cells equally in different anatomic sites. Thus, preferential infection by the virus is unlikely to cause functional perturbations.

INTRODUCTION

Upon stimulation through the T cell receptor (TCR), with appropriate costimulatory and cytokine signals, naive CD4 T cells are capable of developing into distinct functional subsets of memory T cells. The differentiation of naive CD4 T cells toward a polarized effector phenotype depends on the environment in which the initial stimulation occurred. Immunological memory is then generally facilitated by differential expression of key transcription factors and microRNAs with epigenetic modifications (1). For example, stimulation of naive CD4 T cells in the presence of interleukin 12 (IL-12) generally leads to expression of the transcription factor T-bet and differentiation into the gamma interferon (IFN-γ)-producing Th1 lineage, whereas TCR stimulation in the presence of IL-4 and IL-13 promotes GATA3 expression and differentiation toward a Th2 lineage. Expression of these transcription factors during the differentiation of naive CD4 T cells is thought, largely, to imprint particular functional profiles into memory CD4 T cells, such that reexposure to the same antigen would result in appropriate effector functions being performed. Functional plasticity can exist in some circumstances, however, and CD4 T cells differentiated toward one particular functionality can become hyperfunctional or can fluctuate toward a different functional lineage (2–4). For example, IL-17- and IL-22-producing memory Th17 cells can, under some circumstances, produce the Th1 effector cytokine IFN-γ, with such coexpression being associated with autoimmunity (5).

In chronically human immunodeficiency virus (HIV)-infected individuals, key immunological pathologies exist that, in the vast majority of non-antiretroviral (ARV)-treated individuals, ultimately lead to susceptibility to opportunistic infections and neoplasia. While loss of CD4 T cells is a key immunological pathology that occurs in HIV-infected individuals who are ARV naive, other important phenomena occur and contribute to disease progression. CD4 T cells in mucosal tissues, such as the gastrointestinal (GI) tract, are preferentially depleted in HIV-infected humans and simian immunodeficiency virus (SIV)-infected Asian macaques (6–9). Moreover, the few CD4 T cells that remain within mucosal tissues are skewed in their production of IL-17, IL-22, and forkhead box protein P3 (FoxP3), exhibiting less polyfunctional profiles (10–13). The pathologies that occur within the GI tract damage its structural barrier, which results in the translocation of luminal microbial products into the lamina propria and systemic dissemination (14–16). These translocating microbial products lead to inflammation, which exacerbates HIV/SIV disease progression.

Even after long-term administration of ARVs, particularly if ARV treatment is initiated in the chronic phase of infection, CD4 T cell reconstitution in the GI tract is inefficient, Th17 cell perturbations remain, and residual microbial translocation and inflammation persist. Thus, there is great interest in understanding the potential causes of the preferential loss of Th17 cells in HIV-infected humans and SIV-infected Asian macaques (17). Th17 cells express CD4 and can express the HIV entry coreceptor C-C chemokine receptor type 5 (CCR5), and they are exquisitely sensitive to infection in vitro (18). Moreover, Th17 cells represent the first cells targeted by the virus after intravaginal infection of Asian macaques with SIV (19). Thus, preferential infection by HIV/SIV might explain the loss of these cells in vivo. However, numerous studies have demonstrated that IL-17-producing CD8 T cells (Tc17 cells) and innate lymphocytes type III (ILC3s) are also lost in HIV/SIV infection, despite lacking CD4 and CCR5 (20–22). Those studies suggested that loss of IL-17-producing cells in HIV/SIV-infected individuals might be attributable to the immunological milieu, specifically alterations to the landscape of antigen-presenting cells and to the levels of tryptophan-metabolizing enzymes such as indoleamine 2,3-dioxygenase, which produce metabolites that do not favor IL-17-producing lymphocyte differentiation and survival (20, 23). Thus, the degree to which infection of Th17 cells versus immunological milieu alterations contribute to the loss of IL-17-producing lymphocytes is unclear.

To understand whether immunodeficiency lentiviruses have an increased proclivity to infect particular functionally differentiated CD4 T cells, compared to others, here we used flow cytometric sorting of mitogenically stimulated lymphocytes to isolate memory CD4 T cells of similar differentiation status, based on their differential effector functions. We studied these cells in peripheral blood, spleen, and mesenteric lymph nodes (MLN) of SIV-infected Asian macaques, and we measured viral DNA to indicate the level to which these cells were infected by the virus in vivo. These data allow us to understand whether particular functionally defined populations of memory CD4 T cells are preferred targets for the virus in vivo and the degree to which each of these contributes to the total pool of infected cells.

RESULTS

Here we studied lymphocytes isolated from peripheral blood, spleen, and MLN of 16 SIVmac239-infected rhesus macaques, 10 SIVsmE543/E660-infected rhesus macaques, and 11 SIVmac239-infected pigtail macaques that had been treated with ARVs (animals with identifiers beginning with “PT”). All of the animals were progressively SIV infected and were sacrificed either during the chronic phase of infection or when they had progressed to fulminant simian AIDS (Table 1). We mitogenically stimulated lymphocytes and then used flow cytometry to isolate CD28⁺ CD95⁺ memory CD4⁺ T cells that expressed CCR6 and produced IL-17 (Th17 cells), that expressed CCR6 without production of IL-17 (Th17-like), that produced IFN-γ (Th1 cells), that expressed CCR4 (Th2-like), that expressed FoxP3 (T regulatory cells [Tregs], which overlap with CD25⁺ CD127^low CD4 T cells; data not shown [24]), or that expressed none of these (which we defined as “other”; a representative flow cytometric gating strategy is shown in Fig. 1). We specifically studied only CD28⁺ memory CD4 T cell subsets, because we found previously that cells that lose CD28 expression and are terminally differentiated harbor lower levels of viral DNA than do memory CD4 T cells that express CD28 in vivo (25). We then measured the frequency of each of the identified populations in peripheral blood (Fig. 2A), spleen (Fig. 2B), and MLN (Fig. 2C) of chronically SIV-infected animals that were ARV naive or ARV treated (Fig. 2D to F). In all anatomic sites, irrespective of ARV treatment, the population we defined as other, i.e., the cells that did not belong to one of the functional populations we defined, encompassed the largest frequencies of CD28⁺ memory CD4 T cells. The majority of the memory CD28⁺ CD4 T cells we defined as other expressed CXCR3 and were likely Th1-like cells (data not shown). Th17 cells, Th2-like cells, and Tregs represented a minority of CD28⁺ memory CD4 T cells. It is important to note that, given the limitation of our flow cytometer in having the capacity to sort only four populations simultaneously, we did not extensively study Tregs in the animals in our cohort.

TABLE 1.

Infection state of ARV-naive and ARV-treated Asian macaques used in this study^a

^aCD4 indicates the number of CD4 T cells per microliter of blood; viral load indicates the number of viral RNA copies per milliliter of plasma.

FIG 1.

Gating strategy used to sort functionally defined Th subsets. From live CD4⁺ CD28⁺ memory T cells, we sorted Tregs (Foxp3⁺ CD95⁺). From the Foxp3⁻ CD95⁺ population, we sorted Th17 cells (IL-17⁺ CCR6⁺) and Th17-like cells (IL-17⁻ CCR6⁺). Then, from the IL-17⁻ CCR6⁻ compartment, we sorted Th1 cells (IFN-γ⁺ CCR4⁻), Th2-like cells (IFN-γ⁻ CCR4⁺), and all other cells that did not express the phenotypic or functional properties we had defined. The CD28⁻ CD95⁺ memory T cell population was excluded, to avoid phenotypic differences that may greatly vary virus infectivity, due to a greater degree of differentiation.

FIG 2.

T helper subset population frequency as a percentage of CD28⁺ memory T cells for ARV-naive PBMC (A), ARV-naive spleen (B), ARV-naive MLN (C), ARV-treated PBMC (D), ARV-treated spleen (E), and ARV-treated MLN (F). Horizontal lines indicate the medians.

We next quantitatively measured the levels of viral DNA in each population of memory CD4 T cells in ARV-naive and ARV-treated animals. Consistent with our previous work, we found between 1 and 10 copies of viral DNA for every 100 CD28⁺ memory CD4 T cells in peripheral blood and lymphoid tissues of chronically infected macaques (26, 27). Tissues from macaques that had progressed to AIDS were capable of having viral DNA levels higher than 1 to 10 copies per 100 cells. Moreover, each population of CD28⁺ memory CD4 T cells contained viral DNA. Thus, all cell types displayed evidence of infection in vivo. However, no particular functionally defined subset of CD28⁺ memory CD4 T cells that we isolated was consistently more or less frequently infected by the virus, compared to any other population of cells. This was true in peripheral blood (Fig. 3A), spleen (Fig. 3B), and MLN (Fig. 3C) of animals that were ARV naive and those that were ARV treated (Fig. 3D to F). Moreover, similar to previous analyses of HIV-infected humans, we did not see that populations of memory CD4 T cells in lymphoid tissues were more prone to harbor viral DNA than were those isolated from peripheral blood (28). Importantly, while preferential loss of Th17 cells is not routinely observed in peripheral blood and peripheral lymphoid tissues, these cells are consistently observed to be preferentially lost from MLN (which respond to antigens from the GI tract) (10, 12, 17, 20, 21, 23, 29–31). We demonstrate that, irrespective of anatomic site and irrespective of how Th17 cells are identified based on expression patterns of CCR6 or IL-17, these cells are not preferentially infected, compared to any other population of similarly differentiated memory CD4 T cells.

FIG 3.

Infection frequency of T helper subsets in ARV-naive PBMC (A), ARV-naive spleen (B), ARV-naive MLN (C), ARV-treated PBMC (D), ARV-treated spleen (E), and ARV-treated MLN (F). Horizontal lines indicate the medians. vDNA, viral DNA. *, P < 0.05.

Finally, we compared the degree to which each population of cells contributed to the total pool of SIV-infected cells in vivo. For each subset, we multiplied the infection frequency by the respective population frequency to ascertain the contribution to the total pool of infected cells. We then divided the infected pool of each subset by the total infected pool to calculate each subset’s contribution to the total infected pool. We found in all anatomic sites, in both ARV-naive (Fig. 4A to C) and ARV-treated (Fig. 4D to F) animals, that the population we defined as other contributed the largest numbers of viral-DNA-positive cells to the total population of infected cells. Th17 cells and Th17-like cells each contributed less than 25% to the total population of SIV-infected cells, with infected Th17 and Th17-like cells being more prevalent in the peripheral blood and MLN than in the spleen. However, the contribution of each functional population to the total pool of infected cells differed across tissues, with Th17 and Th17-like cells contributing to the total population of infected cells more in the peripheral blood of ARV-naive macaques (Fig. 4).

FIG 4.

Proportional contribution of each T helper subset to the total infected pool for ARV-naive PBMC (A), ARV-naive spleen (B), ARV-naive MLN (C), ARV-treated PBMC (D), ARV-treated spleen (E), and ARV-treated MLN (F). Medians and ranges for each population are shown below the pie charts.

Given that the particular populations of cells seemed to be infected by the virus at similar levels, we next considered that the biggest predictor of how each population of cells contributed to the pool of infected cells would be the relative size of the population. Indeed, we found a significant positive correlation between the degree to which each population contributed to the pool of infected cells and the overall size of that population (Fig. 5). This was true in peripheral blood, spleen, and MLN of both ARV-naive (Fig. 5A to C) and ARV-treated (Fig. 5D to F) animals. Importantly, this was true if we considered only Th17 cells for our analysis (data not shown).

FIG 5.

Comparison of the size of each T helper subset population to its proportional contribution to the total infected pool for ARV-naive PBMC (A), ARV-naive spleen (B), ARV-naive MLN (C), ARV-treated PBMC (D), ARV-treated spleen (E), and ARV-treated MLN (F). Correlations were based on the Spearman rank correlation.

DISCUSSION

Here, we studied how the functional polarization of memory CD4 T cells might influence the proclivity of the cells to become infected by SIV in vivo. We studied peripheral blood, spleen, and MLN of SIV-infected animals that were ARV naive or ARV treated. We found that SIV infected all isolated subsets of memory CD4 T cells equally, suggesting that preferential loss of Th17 cells is not attributable to increased infection in vivo. These findings parallel a recent study with ARV-treated HIV-1-positive humans, in whom equal levels of viral DNA, on a per-cell basis, were observed in all functionally polarized cells examined (32).

Several previous studies found that T follicular helper (Tfh) cells are preferentially infected by HIV and SIV in vivo (33–38). The high levels of viral DNA within Tfh cells are thought to be attributable to two complementary factors, namely, their proximity to follicular resident dendritic cells, which trap replication-competent virus for a prolonged time (even after administration of ARVs) (38–40), and the inability of CD8 T cells to enter the lymphoid follicle effectively and to combat viral replication therein (41). Thus, the ability of the virus to target Tfh cells is predominantly influenced by the anatomic location of these cells. Our study aimed to understand whether CD4 T cell functionality, rather than the anatomic location, could influence how the virus targets cells in vivo. Considering that we studied peripheral blood, spleen, and lymph nodes, we did not include Tfh cells in our study because Tfh cells reside in follicles, their presence in peripheral blood is somewhat unclear, and we and others previously reported high levels of viral DNA within Tfh cells (34, 37, 38, 41). While we did not isolate Tfh cells, our analysis does allow direct comparisons of SIV infection of multiple populations of CD4 T cells isolated from the same animals but at different sites. Consistent with previous studies using samples from ARV-treated, HIV-infected individuals, we did not see higher levels of viral DNA within CD4 T cells isolated from lymphoid tissue, compared to those isolated from peripheral blood (28). These data might suggest that the circulation of memory CD4 T cells throughout peripheral blood and lymphatics leads to normalization of infection across these sites.

While Th17 and Th22 cells are present in systemic tissues and peripheral blood, they are present at higher frequencies within mucosal tissues (10). These T cells are important for antifungal and antibacterial immunity (42) and are also critical for the maintenance of epithelial cells (particularly within the GI tract) (20). Th17 and Th22 cells are routinely observed to be preferentially depleted from mucosal sites of HIV-infected humans and SIV-infected Asian macaques (reviewed in references 17 and 43) and represent some of the first cells infected after intravaginal infection (19). Moreover, CCR6⁺ CD4 T cells are exquisitely sensitive to infection in vitro, and naive CD4 T cells that are infected do not efficiently differentiate into Th17 cells (18, 44, 45). Thus, specific targeting of CCR6⁺ CD4 T cells (including CD4 T cells capable of producing IL-17 and IL-22) by HIV/SIV might explain their preferential loss. However, other IL-17- and IL-22-producing cells that do not express viral entry receptors, such as Tc17 cells and ILC3s, are also present at decreased frequencies in mucosal sites of HIV/SIV-infected individuals (20–22). Alterations to the landscape of antigen-presenting cells, with increased expression of the tryptophan-metabolizing enzyme indoleamine 2,3-dioxygenase, are also implicated in the loss of IL-17-producing cell types (20, 23), and low levels of IL-21 production in SIV-infected macaques can be important for the loss of IL-17-producing cells (29). Thus, there are several potential causes of Th17 cell loss that do not involve direct infection by the virus.

Our analysis of viral DNA does not distinguish between latent virus, preintegrated virus, replication-incompetent virus, and replication-competent virus. Rather, our quantitative PCR detects all of these, reflecting cells that were infected by the virus (46). Thus, our analysis allows us to understand whether particular populations of CD4 T cells are differentially infected by the virus in vivo. Considering that we show that CCR6⁺ CD4 T cells and IL-17-producing cells contain viral DNA at levels comparable to those in other populations of memory CD4 T cells, our data are consistent with immunological milieu factors contributing to the Th17 cell loss.

Moreover, the fact that we observe equivalent levels of infection across different functionally defined memory CD4 T cells suggests that therapeutic interventions aimed at decreasing viral replication, or purging latent virus in ARV-treated individuals, do not necessarily need to specifically consider particular transcription factors that define Th17 cells (RORγt), Th1 cells (T-bet), Th2 cells (GATA3), or Tregs (FoxP3). Each population of functionally defined memory CD4 T cells contributes to the total pool of infected cells, in direct proportion to the size of the respective population. However, it is important to note that the manner in which each functionally defined subset of memory CD4 T cells contributes to the reservoir in individuals treated with ARVs for prolonged periods is influenced by the life span of each population. Understanding how infection frequencies are combined with the life span of each memory CD4 T cell subset population would be an important consideration. Moreover, it would be important to understand how viral diversification occurs within individual populations of memory CD4 T cells. This would lead to an understanding of how, or whether, each memory CD4 T cell subset contributes to the low level of clonal viremia that is observed in ARV-treated, HIV-infected individuals (47). Importantly, given that we find equivalent levels of viral DNA in all populations of memory CD4 T cells, our data do suggest that therapeutic interventions aimed at targeting infected memory CD4 T cells would need to access all of these cells.

MATERIALS AND METHODS

Animals.

The NIAID Division of Intramural Research Animal Care and Use Program, as part of the NIH Intramural Research Program, approved all of the experimental procedures (protocol LVD 26). The program complies with all applicable provisions of the Animal Welfare Act and other federal statutes and regulations relating to animals.

Rhesus macaques and pigtail macaques of Indian origin were housed and cared for at the NIH Animal Center, under the supervision of the Association for the Assessment and Accreditation of Laboratory Animal Care (AAALAC)-accredited Division of Veterinary Resources and as recommended by the Office of Animal Care and Use nonhuman primate management plan. Husbandry and care met the standards set forth by the Animal Welfare Act, animal welfare regulations, and the National Cancer Institute (assurance A4149-01). The physical conditions of the animals were monitored daily. Animals were provided continuous access to water and were offered commercial monkey biscuits twice daily, as well as fresh produce, eggs, and bread products twice weekly and a foraging mix (consisting of raisins, nuts, and rice) thrice weekly. Enrichment to stimulate foraging and play activity was provided in the form of food puzzles, toys, cage furniture, and mirrors.

Euthanasia was initiated at endpoints and followed the protocols consistent with the American Veterinary Medical Association guidelines. Peripheral blood, spleen, and MLN were collected immediately after death, processed for cryopreservation, and stored in liquid nitrogen for subsequent use.

Plasma viral load assessment.

SIV viral RNA levels in plasma were determined by real-time reverse transcription-PCR using the ABI Prism 7900 sequence detection system (Applied Biosystems). Primer pairs for the assay corresponded to the forward nucleotides 1181 to 1208 and reverse nucleotides 1338 to 1317 of the SIVmac239 gag gene.

Sample processing.

Whole blood was centrifuged for plasma collection, and then peripheral blood mononuclear cells (PBMC) were isolated by standard density centrifugation and cryopreserved. Spleen and MLN were obtained at necropsy and processed into single-cell suspensions as described previously (26).

Flow cytometric analysis and cell sorting.

Cellular functionality was assessed after stimulation of single-cell suspensions for 6 h at 37°C in complete RPMI 1640 medium with phorbol myristate acetate (PMA) (2.5 ng/ml) and ionomycin (1 μg/ml), in the presence of brefeldin A (1 μg/ml). Four-way flow cytometric sorts were performed on stained cell suspensions using a BD FACSAria system equipped with FACS DiVa software (BD Biosciences). Cells were stained with Live/Dead aqua blue dye (Invitrogen, Carlsbad, CA) to assess viability. Antibodies against the following antigens were used for staining at predetermined concentrations: CD8 (clone RPA-T8) and IL-17α (eBio64DEC17) from eBioscience; IFN-γ (B27), tumor necrosis factor alpha (TNF-α) (Mab11), and CD3 (SP34-2) from BD; Foxp3 (206D), CD95 (DX2), CCR4 (L291H4), CCR6 (G034E3), and CD4 (OKT4) from Biolegend; and CD28 (CD28.2) from Beckman Coulter. CD4⁺ CD28⁺ memory T cell subsets were sorted as shown in Fig. 1.

Quantitative PCR.

Quantitative PCR was used to assess the viral infection frequencies of each sorted CD4 T cell subset, as described previously (26). Primers and probes targeting SIVmac239 or SIVsm were obtained from Biosearch Tech (Petaluma, CA).

Statistical analyses.

Statistical analyses of viral infection frequencies and correlations were performed using Prism v7.0 (GraphPad Software Inc.). The Wilcoxon t test (matched pairs, two way, nonparametric) was used to determine whether any Th subset population was significantly more or less infected by SIV. A Bonferroni multicomparison correction was used for determinations of significance. Analyses of the size of a subset population and its contribution to the total infected pool were performed using the Spearman correlation (nonparametric, using P values of <0.05).