AIDS Progression Is Associated with the Emergence of IL-17–Producing Cells Early After Simian Immunodeficiency Virus Infection

Immunopathology mediated by inappropriate or poorly controlled effector T cell responses has typically been viewed in the context of the Th cells (Th1 and Th2) paradigm. More recently, a CD4⁺ T cell subset characterized by the production of IL-17 and crucially involved in certain autoimmune, allergic, and inflammatory diseases was identified (1–4). Invariant NK T cells (NKT), which have a role in antitumor immune responses and antiviral immunity (5), have also been recently reported to produce IL-17 in the context of inflammatory diseases (6, 7).

TGF-β and inflammatory cytokines together induce the development of Th17 cells from CD4⁺ T cells in mice and humans (8–10). In the immune system, TGF-β affects multiple cell lineages by either promoting or opposing their differentiation, survival, and proliferation (11). At present, TGF-β is mainly viewed as an immune suppressive cytokine because TGF-β is a critical factor for regulatory T cells, and its loss is associated with a fatal lymphoproliferative disease (12). TGF-β regulates the components of adaptive immunity, such as T cells, as well as innate immunity, such as NK cells (13). In contrast, an inflammatory cytokine environment inhibits the generation of regulatory T cells and instead leads to the differentiation of Th17 cells. Inflammatory cytokines are strongly induced in cells of the innate immune system following engagement of specific pattern-recognition receptors such as TLRs and C-type lectin receptors. Thus, in addition to IL-6, TNF-α and IL-1 have been proposed to have an additional role in the amplification of Th17 responses (14).

The elevation of TGF-β (15–17) plus an inflammatory environment may suggest the possible induction of Th17 population in this context. However, recent reports indicated a lower frequency of Th17 CD4⁺ T cells at mucosal and systemic sites during HIV infection and SIV infection (18–21). Given the role of IL-17 in controlling commensal bacteria, it has been proposed that depletion of Th17 may participate in the disruption of the mucosal barrier. However, the absence of Th17 CD4⁺ during the acute phase is not associated with an increase in LPS translocation (22), which suggests the possibility of additional IL-17⁺–expressing cells that control intestinal flora translocation compensating for the defect in Th17 CD4⁺ T cells. However, whether the loss of IL-17–expressing cells is also true in intact secondary lymphoid tissues and whether other cells, such as NKT expressing IL-17, emerge during SIV infection is presently unknown.

Twenty-four rhesus macaques (RMs; Macaca mulatta) were inoculated i.v. with the pathogenic SIVmac251 strain (10 50% animal infectious doses). All the animals were challenged with the same batch of virus, titrated in vivo in RMs, and stored in liquid nitrogen. Six African green monkeys (AGMs) of sabaeus species were experimentally infected with 300 50% tissue culture-infective dose of SIVagm.sab92018 strain. As previously described (23), this virus derives from a naturally infected AGM and has never been cultured in vitro. Animals were demonstrated as being seronegative for simian T leukemia virus type-1), SRV-1 (type D retrovirus), herpes B viruses, and SIVmac. Animals were housed and cared for in compliance with existing French regulations (Institut Pasteur, Paris, France). The organs were collected and frozen in isopentane cooled in liquid nitrogen. Organs were cut into 4-μm sections on a cryostat, and the sections were stored at −80°C until use (16).

RNA was extracted from plasma of SIV-infected monkeys using the TRI Reagent BD Kit (Molecular Research Center, Cincinnati, OH). Real-time quantitative RT-PCR was used to determine viral loads as previously described (24). Productively infected cells were assessed in lymph nodes (LNs) by in situ hybridization using a [³⁵S]-labeled RNA probe derived from the SIVmac nef gene, and the frequency of SIV infected cells was measured by limiting dilution PCR as described (24).

Tissues were prepared for immunohistochemistry as described (16). Briefly, cryostat sections (4 to 5 μm) were fixed in acetone for 10 min and stored at −20°C until use. For staining, tissue sections were fixed in 2% paraformaldehyde, permeabilized with Triton 0.05%, then incubated with mAbs against IL-17 (eBioscience, San Diego, CA), and TGF-β (9016, R&D Systems, Minneapolis, MN), for 2 h, followed by biotinylated goat anti-mouse Ab and streptavidin peroxidase complex (Vectastain ABC kit, Vector Laboratories, Burlingame, CA). Analyses were performed on four different slides, in a blind manner, by two investigators using a Nikon-FXA microscope (Nikon, Melville, NY) and the Optilab Pro 2.5 software (Graftek Imaging, Austin, TX). The numbers of cells counted were divided by the surface of the entire tissue section. The results were expressed as cell numbers/2 mm² sections. Semiquantitative analyses were also determined by measuring the level of staining (numbers of pixels of the entire tissue section) divided by the total surface. The results were expressed as a number of positive cells.

Blood from monkeys was collected in sterile EDTA-treated vacuum tubes and immediately centrifuged at 400 g for 15 min at 4°C to avoid cytokine synthesis in vitro. IL-6, TNF-α, and IL-1β were detected simultaneously in plasma by using the human inflammatory cytokine cytometric bead array kit (BD Pharmingen, San Diego, CA), which has been validated for cytokine measurement in RMs and AGMs. The cytometric bead array working range was 20–5000 pg/ml for each cytokine. IL-18 in plasma was measured by using an ELISA kit (MBL Biomedical, Clinisciences), and stimulation of cells from healthy monkeys demonstrated no difference in IL-18 expression between RM and AGM. We also used ELISA to measure type I IFN (PBL interferon source, Clinisciences) validated for both species.

PBMCs from monkeys were isolated from peripheral blood by density gradient centrifugation LymphoPrep (PAA Laboratories, Pasching, Austria) and cultured in RPMI 1640 supplemented with 10% FCS, 1% glutamine, 1% pyruvate, and 1% antibiotics. Up to 500,000 PBMCs were cultured overnight. Cells were incubated in the presence of brefeldin A (5 μg/ml for the last 12 h) before being stained for FITC–anti-CD4 and PerCP–anti-CD3 (BD Biosciences, San Jose, CA), then washed twice in PBS, and further incubated with PE–anti-IFN-γ or PE–anti-TNF-α and APC–anti-IL-17A. Cells were also stained with PE-CD8 (RPA-T8), PE-CD56 (MY31), PE-HLA-DR (G46-6), FITC-CD161 (DX12), PE-CD25 (M-A251), PE-CD27 (M-T271), and PE-CD95 (DX2) from BD Biosciences; PE-CXCR6 (56811) and PE-CCR6 (53103) from R&D Systems; PE-CD127 (R34.34) and FITC–anti-Vα24 (clone C15) from Beckman Coulter (Fullerton, CA); and PE–anti-human invariant NKT (clone 6B11, BD Pharmingen), PerCP–anti-CD4, and APC–anti-IL-17A. We also used anti-human FITC–anti-CD45RA (clone 2H4, Beckman Coulter, Fullerton, CA) and PE–anti-CD62L (clone SK11, BD Biosciences). Samples were analyzed on an FACScalibur cytometer with CellQuest software (CellQuest, Tampa, FL).

Human PBMCs were isolated from peripheral blood by density gradient centrifugation on LymphoPrep (PAA Laboratories). Up to 500,000 PBMCs were stimulated with Con A (1 μg/ml) and then incubated with either TGF-β alone (5 ng/ml), IL-18 alone (50 ng/ml), or TGF-β plus IL-18 for 5 d at 37°C with 5% CO₂ in RPMI 1640 with 5% of FCS. At day 6, cells were activated with PMA (100 ng/ml) plus ionomycin (300 ng/ml) for 5 h in the presence of brefeldin A (5 μg/ml). This experiment was conducted using human cells and human cytokines to preclude the possibility that these cytokines might not have cross-species reactivity.

Data are reported as means ± SEM. Comparisons were based on ANOVA and Tukey’s post hoc test, using Prism 3.0 software (GraphPad, San Diego, CA). Student t test and Mann-Whitney U test (Prism software, GraphPad) were also used to determine whether differences were significant if p < 0.05. Correlations were identified by means of the Spearman rank correlation coefficient (ρ). Differences were considered significant if p < 0.05.

By immunochemistry analyses, we clearly observed IL-17–expressing cells in axillary (Ax) and mesenteric (Mes) LNs of RMs infected with the pathogenic SIVmac251 strain (SIV⁺ RM) (Fig. 1A). The quantification of IL-17⁺–producing cells from RMs, killed at different time points postinfection (n = 3 RMs, at each time point), revealed a significant increase at day 14 (Fig. 1A, 1E) concomitantly with the peak of viral replication in these organs (Fig. 1D). These cells were mostly located within the paracortical T cell zone. TGF-β, assessed in the same tissue sections, progressively increased and peaked at day 14 (Fig. 1B, 1F). Thereafter, at day 60, IL-17– as well as TGF-β–expressing cells decreased both in Ax and Mes LNs but remained higher than preinfection. By analyzing a group of 12 monkeys (24, 25) at day 14, we found a positive correlation between the numbers of IL-17–expressing cells and the numbers of TGF-β–expressing cells in Ax LNs of SIV⁺ RM, whereas no correlation was observed with the extent of viral replication in LNs (Fig. 1H). Although SIV infection of RMs leads to progressive CD4⁺ T cell depletion and AIDS, SIVagm infection of AGM is nonpathogenic despite levels of plasma viral load similar to those observed in SIVmac-infected RMs (23, 26, 27). In this study, in stark contrast to RMs, we did not detect an increase in IL-17– and TGF-β–positive cells in peripheral LNs from six SIV-infected AGMs (Fig. 1C, 1E, 1F). A role for IL-17 in the recruitment of neutrophils in inflamed tissues has been proposed (2). In this study, we found, despite the absence of IL-17, the presence of neutrophils in LNs of SIV-infected AGMs at day 14, similar to that observed in Ax LN of SIV-infected RMs (Fig. 1G), suggesting the possible role of additional cytokines (i.e., IL-8) in the recruitment of neutrophils in nonpathogenic monkeys.

These results demonstrate the differential expansion of IL-17–expressing cells early postinfection in the pathogenic primate model of SIV infection in RMs.

We recently demonstrated during the chronic phase of SIV⁺ RM that TGF-β expression is significantly increased in noncontrollers versus controllers (16). We thus quantified the expression of IL-17–expressing cells in Ax and Mes LNs of 10 SIV⁺ RMs (5 were noncontrollers as previously described (16). Whereas no IL-17–expressing cells were found in a healthy monkey (#93750), we observed increased levels of IL-17–expressing cells in Ax LNs from SIV⁺ monkeys; these levels were higher in noncontrollers (as shown in monkeys #94746 and #272) than in controllers (as shown in monkey #94422) (Fig. 2A, 2B). In Mes LNs, the levels of IL-17–expressing cells in SIV⁺ RMs were also higher compared with healthy monkeys (Fig. 2A, 2B), a result that is consistent with the expression of TGF-β (16). We also studied IL-17 expression in GALTs discriminating small and large intestine (ileum and colon, respectively). We found that SIV infection induces increased numbers of IL-17⁺ cells in the ileum that is higher in noncontrollers (monkeys #94746 and #272) than in controllers (monkey #94422) (Fig. 2C, 2D) and higher than that observed in healthy macaques (#93750). Moreover, whereas we detected large numbers of IL-17⁺ cells in the colon of healthy RMs (#93750), which may reflect the presence of commensal bacteria (28), we observed a progressive depletion of these cells in animals progressing toward AIDS (Fig. 2C, 2D). In Indian RMs, which are more rapid progressors than Chinese RMs, it has also been shown that there are differential dynamics between large and small intestines (19). Thus, in large intestine, the depletion is higher.

Altogether, our tissue analyses demonstrate greater expression of IL-17 cells associated with the progression to AIDS, except in the colon.

In addition to classical Th17 cells, NKT have been recently reported to produce IL-17 (6, 7, 29). Therefore, we next investigated the phenotype of cells that ex vivo spontaneously produced IL-17 in SIV⁺ RM. We also stimulated the cells with PMA/ionomycin that measure the pool of IL-17–expressing cells. Our data revealed that ex vivo cells (CD4^low) from SIV⁺ macaques spontaneously produced IL-17 compared with cells from healthy macaques. In contrast, in vitro stimulation revealed that healthy macaques have a pool of effector IL-17⁺CD4⁺ cells higher than SIV⁺ RMs, consistent with recent published data (18–20). Most importantly, in vitro stimulation induced the loss of this subset, which produced spontaneous IL-17 (data not shown). Therefore, in studying the spontaneous production of this cytokine, we found in PBMCs of SIV⁺ RMs an IL-17–producing population that is higher in noncontrollers than in controllers at day 60 (Fig. 3A). As shown in Fig. 3A and 3C, the IL-17–producing subset expressed low levels of CD4 and CD3 molecules. These cells were negative for CD8 and CD56, a marker used to detect macrophages in primates. We then examined whether this subset exhibits phenotypic characteristics of NKT. We found that IL-17⁺–expressing cells from peripheral blood of noncontrollers are NKT⁺Vα24⁺ and also express CD161⁺ and CCR6⁺ molecules (Fig. 3C), which have been reported to characterize this type of cell, although other cell types express these markers (30, 31). We demonstrated that IL-17⁺CD4^low cells spontaneously produced TNF-α and IFN-γ (Fig. 3B), consistent with the fact that freshly isolated NKT from histologically tumor-bearing human tissues, were more potent producers of TNF-α and IFN-γ (32). We also analyzed the expression of CD62L and CD45RA markers. Instead of being either CD62L^bright or CD62L^neg and either CD45RA^bright or CD45RA^neg cells like CD4⁺ or CD4⁻ T cells from peripheral blood of noncontrollers (Fig. 3D), this population, as previously described (33), is mostly CD45RA^int and CD62L^int (Fig. 3D). Gating on IL-17⁺NKT⁺ compared with CD4⁺ T cells from peripheral blood, we found that these cells were negative for CD127 and CD25; half of them expressed HLA-DR and CXCR6 but they expressed CD27 (TNF receptor) at a low level and a higher level of CD95 (death receptor) (Fig. 4). We also analyzed the phenotype of freshly isolated IL-17⁺ cells in Ax LNs and the spleen of noncontrollers, and we found that IL-17–expressing cells were NKT⁺, CD161⁺, CCR6⁺, and CD8⁻ but, unlike peripheral blood, were all positive for HLA-DR (Fig. 5).

During the acute phase, we found that the percentage of the IL-17⁺NKT⁺ subset increased both in peripheral blood and in Ax LNs of SIV⁺ RMs (n = 10); in contrast, the percentage of IL-17⁺CD4⁺ cells decreased both in the blood and peripheral LNs (Fig. 6A), consistent with other recent reports (18–20). We also examined in the different organs the fraction of CD8, NKT, and CD4 cells expressing IL-17⁺ at the chronic phase. Six RMs were analyzed at necropsy. The levels of the IL-17⁺NKT⁺ subset were higher in SIV⁺ RMs than in healthy monkeys and greater in Mes LNs and the spleen than in the Ax LNs (Fig. 6B). These results are consistent with our recent report showing the same profile of TGF-β expression in these tissues (16). Our data also revealed that in splenic and Mes LN tissues, no changes in IL-17⁺CD8⁺ cells were observed, whereas the fraction of IL-17⁺CD4⁺ T cells increased due to the fact that the percentage of CD4⁺ T cells in LNs decreased in SIV-infected RMs compared with healthy macaques (Fig. 6B).

Altogether, these data demonstrate that AIDS progression is associated with the emergence of IL-17⁺NKT⁺ early after SIV infection, and these cells persist in animals progressing more rapidly to AIDS.

Because an inflammatory environment concomitantly with exogenous TGF-β has been reported to induce the differentiation of IL-17–expressing cells (14) and IL-18 is essential for NKT (34), we next investigated the presence of proinflammatory cytokines during primary SIV infection. In addition to the pyrogenic cytokines IL-1β, TNF-α, and IL-6, we also assessed the expression of IL-18, a proinflammatory cytokine and member of the IL-1 “superfamily” that is unable to induce fever (35), as well as a type I IFN, that has been reported to be increased during SIV infection and favors NKT (36, 37). In SIV⁺ RMs, we found that type I IFN as well as IL-18 plasma levels increased postinfection, reaching a peak at day 11 and day 14, respectively, and then decreased (Fig. 7A) (n = 12); we also observed a moderate increase in IL-6 levels, whereas IL-1β and TNFα were undetectable. This peak of IL-18 in the sera is associated with increased expression of IL-18 in the LN tissues (Ax and Mes LNs) (Fig. 7B) and is consistent with the absence of fever in SIV-infected RMs during the acute phase. Most interestingly, despite similar levels of viral replication, IL-18 and type I IFN were only slightly increased in SIV-infected AGMs.

Thus, we found cytokine profiles that were distinct in pathogenic compared with nonpathogenic SIV infection.

Because IL-18 is essential for NKT (34) and because it has been proposed that IL-6 is dispensable for IL-17⁺NKT (29), we then determined the involvement of IL-18 in the expansion and in the induction of IL-17⁺NKT in SIV⁺ RMs. Thus, we investigated whether IL-18 concomitantly with TGF-β might induce in vitro the differentiation of IL-17⁺NKT. PBMCs from human healthy donors were used to test the effect of these cytokines to avoid any possible defect in cross-reactivity in monkeys. Cells were first activated with Con A and then incubated in the absence or presence of TGF-β and IL-18. At day 6, the cells were restimulated with PMA plus ionomycin (Fig. 8A, 8B). We found that TGF-β alone increased the proportion of the cells that differentiated into IL-17⁺NKT as well as IL-17⁺CD4⁺ cells. Given the role of stimulated APCs in the release of inflammatory cytokines, it is possible that the initial Con A stimulation also contributed to the in vitro differentiation of IL-17⁺ cells. IL-18 alone also increased the percentage of IL-17⁺NKT, whereas it had no effect on the percentage of IL-17⁺CD4⁺ cells. The combination of both TGF-β and IL-18 enhanced the percentage of both IL-17⁺CD4⁺ and IL-17⁺NKT as compared with each cytokine alone. To determine whether this increase was reflected by the expansion and/or differentiation of these subsets, we evaluated the numbers of these cells in culture. Each cytokine, TGF-β and IL-18 alone, reduced the numbers of IL-17⁻NKT⁺, leading to a relative increase in the percentage of IL-17–expressing NKT (Fig. 8C); in contrast, TGF-β together with IL-18 induced the expansion of these IL-17⁺NKT. Whereas TGF-β alone induced the expansion of IL-17⁺CD4⁺, the addition of IL-18 in culture with TGF-β enhanced their expansion. Thus, increased IL-18 expression concomitantly with elevated TGF-β in SIV⁺ RMs may contribute to the induction of the IL-17⁺NKT subset.

Recently, several groups have shown a decline of Th17 CD4⁺ T cells during HIV and SIV infection and proposed that the absence of this subset may participate in the disruption of mucosal barrier integrity (18–21). In this study, our results demonstrated the early expansion of IL-17–expressing cells in SIV⁺ RMs that is correlated with TGF-β expression. Thus, an innate IL-17 production by NKT is rapid and precedes the adaptive Th17 response. The emergence of this IL-17⁺NKT⁺ population in SIV-infected RMs, therefore, could compensate for the defect in Th17 CD4⁺ T cells (19, 21), preventing microbial translocation (no LPS was detected during the acute phase (20) and the occurrence of a wasting syndrome early postinfection. Moreover, in chronically SIV-infected monkeys (19) as well as in HIV-infected Africans (38), however, microbial translocation was not correlated with immune activation and viral replication. Thus, microbial translocation might be a symptom of the defect of IL-17 populations and not a direct cause of HIV-1 disease. The observation that SIV-infected AGMs that do not progress to disease have no expansion of IL-17–expressing cells is consistent with the results from another group (20) that found to be unchanged numbers of IL-17 cells in peripheral blood in the nonpathogenic model of Sooty mangabey. The possible difference in the previous reports could be the origin of the monkeys used, given that we used RMs of Chinese origin instead of RMs of Indian origin, the latter being more susceptible to SIV infection (26), and we used freshly isolated cells instead of frozen samples to perform these assays. Another point is our observation that ex vivo activation induced the loss of this subset, suggesting the hypothesis that cells might be more prone to die (activation-induced cell death) because they express CD95 and CD27. Finally, to our knowledge, no other groups have assessed the expression of IL-17 in the tissues, and therefore this is the first report of this observation. Therefore, although we cannot exclude the possibility that other cells in the tissues express IL-17, our data strongly suggest that most of the freshly isolated cells that produced IL-17 were NKT.

It is now clear that IL-17 is associated with a number of human autoimmune disorders by inducing excessive inflammation. IL-17 exerts a potent effect against fibroblasts, epithelial cells, and endothelial cells and promotes the infiltration of inflammatory T cells and neutrophils (39, 40). Thus, the early emergence of this novel population expressing IL-17 could contribute to local tissue damage and favor viral dissemination. Despite this side effect, there is also considerable evidence, in mice, that IL-17 is important in host responses to infection by Gram-negative bacteria, specifically Klebsiella, Pseudomonas, Escherichia coli, Salmonella, and Bordetella species (41, 42). Whereas we found higher levels of IL-17–expressing cells in the ileum, our data demonstrated the absence of IL-17–expressing cells in the colon of noncontrollers SIV⁺ RMs during the chronic phase. This could contribute to the absence of bacterial control, consistent with the clinical observation that animals progressing toward AIDS develop a wasting syndrome characterized by a loss of weight and frequent diarrhea. These results may be related to a distinct dynamic of IL-17⁺ cells in the small and large intestine of SIV-infected monkeys or a redistribution of IL-17⁺ cells from the colon to the ileum.

TGF-β and inflammatory cytokines together induce the development of IL-17–expressing CD4⁺ T cells (8–10) as well as NKT (6, 7). Our results demonstrate for the first time a role for IL-18 in the genesis of IL-17⁺ cells consistent with its role in the emergence of NKT (34). Caspase-1, a member of the cysteine protease family involved in apoptosis, is involved in the maturation and secretion of IL-18 (35). Therefore, the absence of IL-18 in SIV-infected AGMs could reflect the absence of apoptosis that we and others have described in this nonpathogenic primate model, compared with pathogenic primate models, during both the acute (26) and the chronic phase (24, 43–46). Thus, a vicious cycle may be generated by the occurrence of apoptotic cells in the tissues of SIV⁺ RMs, favoring not only the expression of IL-18, but also, through their phagocytosis by macrophages and dendritic cells, the production of TGF-β and the absence of TNF-α and IL-1β (47), favoring the emergence of IL-17⁺NKT⁺ over IL-17⁺CD4⁺ T cells.

However, how is the expansion of NKT early after SIV infection generated? Viral danger signals through the type I IFN pathway trigger NKT activation by myeloid dendritic cells (48, 49). DNA and RNA viruses induce type I IFN through TLR7 and TLR9, and, in particular, HIV has been shown to induce type I IFN (50). Recently, several groups (36, 37) have shown increased levels of type I IFN early after SIV infection of RMs, whereas in nonpathogenic SIVagm infection, these levels are relatively low (23). Thus, the dynamics of IL-17⁺NKT is consistent with the expression of type I IFN during the acute phase observed in SIV-infected RMs.

In conclusion, we demonstrated the emergence of IL-17–producing cells early after SIV infection. The levels of these cells, and their subsequent persistence in peripheral blood and in a large number of tissues, except colon, were associated with progression to AIDS. Moreover, our results show for the first time the expansion of the IL-17⁺NKT⁺ subset in response to microbial infections. Our results strongly suggest the role of IL-18, in addition to TGF-β, in the differentiation of this novel subpopulation.

We thank J.M. Panaud (Institut Pasteur) for assistance with microphotographs.

Disclosures The authors have no financial conflicts of interest.