Mucosal T-Cell Responses to HIV: Responding at the Front Lines

Abstract

Mucosal surfaces of the body serve as the major portal of entry for human immunodeficiency virus (HIV). These tissues also house a majority of the body’s lymphocytes, including the CD4+ T-cells that are the major cellular target for HIV infection. Mucosal surfaces are defended by innate and adaptive immune mechanisms, including secreted antibodies and CD8+ cytotoxic T-cells (CTL). CTL in mucosal lymphoid tissues may serve to limit viral replication, decreasing the host’s viral burden as well as reducing the likelihood of sexual transmission to a naive host. This review summarizes recent literature on HIV-specific T-cell responses in mucosal tissues, with an emphasis on the gastrointestinal tract.

Introduction

The genetic and immunologic correlates of protection from human immunodeficiency virus (HIV) infection and/or disease progression remain incompletely understood despite many years of intensive study. Most studies of HIV-specific immunity have focused primarily on peripheral blood, which is readily accessible for sampling. Nevertheless, the majority of the body’s T- and B-lymphocytes are housed in mucosal tissues lining the gastrointestinal tract [1]. These tissues also serve as a major site of HIV transmission, viral replication and CD4+ T-cell depletion. Because of the unique role of mucosal tissues in HIV transmission and pathogenesis, detailed studies of antiviral immune responses in these tissues can contribute important insights to our understanding of HIV pathogenesis. This review will provide an overview of recent literature on mucosal T-cell responses, with a major focus on CD8+ T-cells in HIV/SIV infection.

The gastrointestinal mucosa: inductive and effector sites

The gastrointestinal tract is thought of as the largest lymphoid organ in the body, with an estimated surface area 200 times that of skin [1]. The intestinal mucosa performs an important barrier function that permits peaceful coexistence with commensal flora while defending against potential pathogens [2]. Inductive sites for gut mucosal immunity include Peyer’s patches and lymphoid aggregates, where antigen is taken up and redirected towards priming of T and B-cell responses. Effector sites in the gut include the lamina propria and inter-epithelial lymphocytes (IEL) located at the basolateral surface of epithelial cells. Effector CD4+ T-cells are abundant in the lamina propria, while IEL are predominantly CD8+ T-cells [1, 3, 4].

Mucosal CD8+ T-cells differ from their counterparts in blood by increased expression of the IEL integrin CD103 (αEβ7), certain activation markers (CD69), chemokine receptors (CXCR4, CCR5), and a predominant effector memory phenotype (CD45RO+, CCR7−) [5–8]. Exposure to TGF-β in vitro can induce expression of CD103 and CD69 in PBMC [5], suggesting that locally-secreted cytokines may alter T-cell phenotype.

Recent studies in mice have led to a new appreciation for the role of mucosal dendritic cells in “imprinting” of newly primed lymphocytes to express trafficking molecules that will direct their eventual return to mucosal effector sites. Unlike most APCs in peripheral tissues, intestinal dendritic cells express enzymes critical for retinoic acid biosynthesis [9]. Retinoic acid induces the expression of integrin α4β7 on T and B-cells [10–13] (reviewed in [14]). Thus, lymphocytes primed in mucosal inductive sites receive an instructive signal directing them to express surface antigens that facilitate trafficking to mucosal effector sites. The ligand for α4β7 integrin, mucosal addressin cell adhesion molecule type 1 (MADCAM-1), is expressed on high endothelial venules in the gut mucosa. Engagement of MADCAM-1 by α4β7 facilitates T-cell extravasation from the bloodstream into mucosal tissues. Several chemokine receptors and their ligands also play a role in attracting lymphocytes to intestinal mucosa: CCR6/CCL20 [15], CCR9/TECK (CCL25) [16], and CCR10/MEC (CCL28) [17, 18]. Intriguingly, the activated α4β7 heterodimer can also bind HIV gp120 on infected CD4+ T-cells. This interaction then triggers activation of αLβ2 integrin (LFA-1) on the CD4+ T-cell surface, facilitating cell-cell interactions. Thus, activated α4β7 expressed by infected CD4 cells may increase cell-to-cell dissemination of HIV [19].

Acute HIV/SIV infection and the gut

As early as 1994, clinicians studying HIV-related diarrhea and wasting syndrome reported abnormalities in intestinal T-cell subsets in HIV-infected individuals [20, 21]. Studies of rhesus macaques infected with simian immunodeficiency virus (SIVmac) later revealed that acute infection results in rapid and profound CD4+ T-cell depletion, regardless of the route of transmission [22–24]. A recent resurgence of interest in this area has led to several detailed studies of acute HIV/SIV and the gastrointestinal tract, which have confirmed the rapid depletion of lamina propria CD4+ T-cells [24, 25]. This depletion may be mediated by direct infection [26], immune-mediated clearance of infected cells, bystander apoptosis [27], or a combination of mechanisms. During acute HIV/SIV infection, as lamina propria CD4+ T-cells are depleted, there is an expansion and/or influx of CD8+ T-cells [28] (Figure 1); however, these cells fail to clear infection or prevent widespread virus dissemination. This rapid depletion of gut CD4+ T-cells, and their slow reconstitution in patients on HAART, suggests that mucosal CD8+ T-cell responses may be inadequate, dysfunctional, and/or inhibited by inappropriate activity of regulatory T-cells (T_reg).

Figure 1. Mucosal CD8+ T-cells.

The figure shows an idealized rectal mucosa, with a single layer of columnar epithelium. HIV-1 may enter the mucosa through breaches in the epithelial layer (1); by binding to dendritic cell processes that extend through the epithelial layer to the lumen (2); via transcytosis across epithelial cells (3); or through a combination of mechanisms. CD8+ T-cells are present as intraepithelial cells (IEL) located between epithelial cells, and in the underlying layer as lamina propria lymphocytes (LPL). Also present in the lamina propria are CD4+ T-cells, macrophages, and antibody-producing plasma cells.

Are mucosal CD8+ T-cells “dysfunctional”?

Why are mucosal T-cells inefficient at clearing HIV/SIV infection? Despite expression of granzymes A and B, mucosal CD8+ T-cells express low levels of perforin as compared to CD8+ T-cells in PBMC [6, 29]; this is true of both healthy controls and individuals with chronic HIV infection and is therefore not directly linked to HIV. It is unclear whether this is due to the effects of the local ‘tolerogenic’ microenvironment; however, low perforin expression may restrict CTL cytotoxicity, limiting tissue damage and inflammation in the gut [6, 29]. Perforin expression is also typically low in CD8+ T-cells in other lymphoid sites such as lymph nodes and tonsil [30]. In contrast, perforin expression by gut T-cells is increased in inflammatory conditions such as Crohn’s disease [31], suggesting a link between high perforin expression and tissue damage.

Detailed kinetic studies of mucosal CD8+ T-cell function during acute HIV infection are lacking; however, in acute SIV infection, perforin is strongly expressed in gut by day 21 [29], consistent with the appearance of SIV-specific tetramer-binding CD8 T-cells in the colon at this time point [32]. Nevertheless, this perforin-positive CD8 T-cell response does not prevent the establishment and dissemination of SIV infection in the gut, and may be “too little and too late” [32]. In our studies of SIV infection, perforin expression (mRNA and protein) in the gut declined steadily after day 21 post-infection, and was indistinguishable from healthy control macaques by 180 days post-infection [29].

The PD-1/PDL-1 pathway may be important in regulating T-cell dysfunction during chronic viral infection [33–35]. In chronic HIV and SIV infection, mucosal T-cells appear to express higher levels of PD-1 than blood T-cells from the same individuals [36, 37]. This is consistent with the role of gut as a major site of HIV/SIV replication, but may also be related to the state of ‘partial activation’ typical of mucosal T-cells [5, 38].

Robust mucosal CD8+ T-cell responses in chronic HIV infection

Although the CD8+ T-cell response during acute infection is insufficient to eradicate infected cells, a robust and often polyfunctional CD8+ T-cell response continues to fight infection in the gut during the chronic phase. The persistence of virus in mucosal tissues throughout this phase of infection, including in patients on HAART [39], argues that the gut continues to be an important ‘battleground’ between the virus and the immune system.

Early studies of mucosal immunity to HIV and SIV relied upon ⁵¹Cr release assays to measure antigen-specific cytotoxic T cell activity [40–42]. These reports demonstrated that MHC class I restricted CTL were present in gastrointestinal mucosa of individuals with chronic HIV/SIV infection, but did not provide great detail about response specificity or breadth. In addition, these reports relied on mucosal cells that had been expanded and subjected to prolonged in vitro culture, so no conclusions could be drawn regarding the frequency of antigen-specific cells in vivo.

To address the frequency of HIV/SIV-specific CD8+ T-cells in mucosal tissues, several groups used MHC class I tetramer staining to quantify CD8+ T-cells specific for well-characterized immunodominant epitopes [7, 43]. These reports revealed that the frequency of Gag-specific CD8+ T-cells in both upper and lower GI tract during chronic infection was quite comparable to that observed in peripheral blood; at least, when immunodominant epitopes were studied. However, these studies looked only at tetramer binding and did not provide information on effector functions. Given that HIV/SIV-specific CD8+ T-cells in blood had been shown to be relatively ‘dysfunctional’ [44, 45], it was important to further explore the functions of mucosal CD8+ T-cells as well.

With the development of multiparameter cytokine flow cytometry (CFC), it became possible to assess the production of multiple cytokines, chemokines, and cytolytic granule constituents by antigen-specific T-cells in a single assay [46–48]. Importantly, CFC is performed on freshly isolated mucosal T-cells [49], without prior culture or in vitro expansion, so the results provide a reasonable estimate of the frequency and functionality of antigen-specific effector cells in a particular tissue. In a study of 28 patients with chronic HIV infection, we measured the production of IFN-γ, TNF-α, and release of the cytolytic granule constituent CD107 by CD8+ T-cells in blood and rectal mucosa in response to peptide stimulation [50]. In both blood and rectal mucosa, there was an immunodominant CD8+ IFN-γ response to Gag as compared to Pol and Env (P <0.01). In contrast, cytomegalovirus pp65 peptides elicited IFN-γ secretion strongly in blood but weakly in rectal CD8+ T cells (P = 0.015). Upon stimulation with HIV peptides, CD8+ T cells from both sites were capable of mounting complex responses, and in rectal mucosa CD107 release was frequently coupled with production of IFN-γ or TNF-α [50]. Thus, mucosal CD8+ T-cells can actively degranulate despite low levels of perforin expression. In a recent follow-up study, rectal HIV-specific CD8+ T-cells were also found to express MIP-1β and low levels of IL-2 [51, 52]. Expression of other chemokines or factors capable of inhibiting HIV replication by mucosal CD8+ T-cells has not been reported, nor has cell contact-dependent, non-cytolytic suppression been explored [53, 54]; these areas await further study. Taken together, these findings demonstrate that rectal CD8+ T cells are capable of robust and varied HIV-1-specific effector responses during chronic infection.

Mucosal T-cell responses are broad and largely shared with blood

An important question is the extent to which mucosal and blood T-cell populations overlap in terms of specificity and clonality. Musey and colleagues compared the TCR sequence and clonality of HIV-specific CTL that had been expanded from gastrointestinal mucosa, semen and cervix [55]. Their findings revealed that the majority of such clones were shared between mucosal sites and PBMC [55].

Due to the difficulty of obtaining large numbers of lymphocytes from mucosal biopsy tissue, comprehensive mapping of the fine specificity of mucosal responses has been challenging. However, relying on a polyclonal expansion approach, Ibarrondo and colleagues successfully mapped rectal HIV-specific CD8+ T cell responses to pools of overlapping peptides spanning the entire HIV genome [56]. These authors found a very similar pattern of responses in the two compartments, as well as a similar immunodominance hierarchy in terms of peptide pools; however, they did not map responses to specific epitopes. Subsequently, Lemongello et al. mapped responses to specific HIV peptides in Gag, Env and Nef [57]. This study reported a similar pattern of peptide recognition in both compartments, and determined that immunodominant peptide-specific responses were conserved between tissue sites [57]. Taken together, these studies reveal significant overlap in the specificity and clonality of CD8+ T cells isolated from mucosal tissues and blood of HIV-infected individuals.

Mucosal immunity in “long-term non-progressors” and “HIV controllers”

Crucial to our understanding of HIV pathogenesis is the study of individuals who remain healthy in the absence of antiretroviral therapy. Individuals who maintain normal CD4+ T-cell counts in the absence of therapy have been described as “long-term non-progressors”. Since many factors may contribute to limiting disease progression, investigators have increasingly defined HIV controllers based on plasma HIV RNA levels rather than on CD4 counts or survival time alone [58, 59]. Two distinct groups of HIV controllers have been described based on plasma viral load measurements: those who maintain HIV RNA below detectable limits (i.e., < 50 copies/mL), designated as “elite controllers”, and those with persistently detectable yet low plasma HIV RNA (i.e., 50–2,000 copies/mL), designated as “viremic controllers” [58]. Elite controllers are believed to represent fewer than 1% of the HIV infected population [58].

Although there have been isolated reports of strong mucosal CD8+ T-cell responses in LTNP and SIV/HIV controllers [50, 60], few studies have addressed these responses in detail. In one report, focused on gene expression analysis, differential expression of 34 genes associated with immune function was seen in jejunal biopsy tissue from 4 LTNP as compared to 11 patients with viral load >10,000 and progressive disease [61]. HIV RNA viral load in jejunal biopsies was low to undetectable in 3 of 3 LTNP, and mucosal Gag-specific CD8 T-cell responses were stronger in 3 LTNP than in 1 progressing individual [61].

Recently, we examined HIV Gag-specific CD8+ T-cell responses in blood and rectal mucosa of over 20 HIV controllers as compared to 14 non-controllers (VL >10,000 copies/vRNA/mL) and 10 antiretroviral treated individuals (VL <75 copies vRNA/mL) [62, 63]. Our hypothesis was that ‘polyfunctional’ T cells capable of producing multiple antiviral factors would be most abundant in mucosal tissues of HIV controllers. Using a 9-color flow cytometry panel measuring IFN-γ, TNF-α, IL-2, MIP-1β, and CD107, we found that mucosal responses were significantly stronger and more complex in controllers than in HAART-suppressed individuals (P=0.0004). Several controllers had unusually strong and complex mucosal CD8+ T-cell responses that were not mirrored in peripheral blood; these responses would have been overlooked if rectal mucosa had not been sampled (Figure 2). Furthermore, the frequency of 4-function HIV-specific CD8+ T-cells in rectal mucosa was significantly greater in controllers than non-controllers or patients on HAART (P<0.0001) [62, 63]. Many of the most robust mucosal responses were identified in subjects who had MHC class I alleles associated with non-progression (i.e., HLA-B57) [64–66]. The basis for HIV non-progression may be related to virologic, immunologic and/or genetic factors; nevertheless, these findings imply that mucosal T-cell responses play an important role in immune surveillance of gut mucosa.

Figure 2. Nine-color, 5-function analysis of HIV Gag-specific T-cell responses in rectal mucosa.

The figure summarizes data for one patient, an HIV controller whose Gag-specific CD8+ T-cell response is substantially stronger (in terms of percent responding cells) and more polyfunctional in rectal mucosa than in blood. Bar graphs show the magnitude of the Gag-specific response in each of 31 functional categories. The legend shows the combination of functions evaluated in each category. These combinations are color-coded based on the number of functions and correspond to wedges in the pie charts (5-function responses are black, 4-function red, 3-function orange, 2-function yellow, and single-function white). The rectal Gag-specific CD8+ T-cell response (white bars) of this HIV controller was heavily biased towards polyfunctional cells secreting CD107, MIP-1β, TNF-α and IFN-γ. In contrast, the Gag-specific response in PBMC was largely monofunctional. Data for over 20 HIV controllers, compared to non-controllers and patients on HAART, are reported elsewhere [62, 63].

The question also arises as to the role of mucosal CD8+ T-cell responses in “highly exposed, persistently seronegative” (HEPS) cohorts. HIV-specific CD8+ T-cells have been detected in cervicovaginal mucosa of highly-exposed sex workers [67], and SIV-specific CD8+ T-cells have been reported in the vaginal mucosa of exposed, uninfected rhesus macaques [68]. However, to our knowledge there have been no published reports documenting HIV-specific CD8+ T-cells in gastrointestinal mucosa of HEPS individuals, or of macaques repeatedly exposed to SIV via the rectal route [69].

Do T_reg modulate mucosal adaptive responses?

Given the role of immune activation in HIV/SIV pathogenesis, recent attention has focused on the balance between tolerogenic and pro-inflammatory T-cell responses in the intestinal mucosa. In the mouse, a tolerogenic cytokine environment dominated by retinoic acid and TGF-β induces differentiation of regulatory T-cells expressing the transcription factor Foxp3 [70, 71]. However, when both TGF-β and IL-6 are present, the cytokine balance favors induction of Th17-type responses [71, 72].

The role of T_reg in HIV/SIV pathogenesis remains controversial; these cells may reduce immune activation, but may also limit HIV-specific adaptive responses [73]. T_reg have been phenotypically identified in blood and mucosal tissues of HIV-infected humans [74, 75] and SIV-infected macaques [76, 77] by their expression of Foxp3, CTLA-4 and CD25. In acute SIV infection, T_reg are reportedly depleted from ileal mucosa [78], but expanded in lymph nodes [76]. In chronic SIV infection, FoxP3 and CTLA-4 mRNA are increased in mucosal tissues of macaques with high viremia, and FoxP3 mRNA is positively correlated with SIV RNA levels in tissues [79]. However, to date very few studies have addressed the presence of T_reg in mucosal tissues of HIV-infected humans [74, 75]. Assessing the functions of tissue T_reg in humans is complicated by the difficulty of obtaining sufficient numbers of cells from biopsy samples. Thus, it remains to be formally demonstrated that mucosal T_reg can actively limit HIV/SIV-specific CD8+ T-cell responses in tissues.

Conclusions

Given the recent failures of vaccine and microbicide trials, there is renewed interest in elucidating the role of mucosal immunity in HIV transmission and pathogenesis. The study of novel patient cohorts such as HIV controllers and exposed, seronegative individuals may shed light on this issue. Many HIV controllers have unusually robust, polyfunctional HIV-specific T-cell responses in mucosal tissues. Nevertheless, strong CD8+ T-cell responses, whether in mucosal tissues or blood, clearly do not account for all cases of HIV containment, suggesting that multiple mechanisms contribute to the controller phenotype. Additional studies will be required to address such issues as: (a) the role of T_reg in modulating mucosal T-cell responses; (b) the role of innate mucosal effector cells in antiviral immunity, and (c) fully characterizing the immunologic, virologic and genetic correlates of protection in HIV controllers.