Gag p27-Specific B- and T-Cell Responses in Simian Immunodeficiency Virus SIVagm-Infected African Green Monkeys

Nonpathogenic simian immunodeficiency virus SIVagm infection of African green monkeys (AGMs) is characterized by the absence of a robust antibody response against Gag p27. To determine if this is accompanied by a selective loss of T-cell responses to Gag p27, we studied CD4⁺ and CD8⁺ T-cell responses against Gag p27 and other SIVagm antigens in the peripheral blood and lymph nodes of acutely and chronically infected AGMs. Our data show that AGMs can mount a T-cell response against Gag p27, indicating that the absence of anti-p27 antibodies is not due to the absence of Gag p27-specific T cells.Simian immunodeficiency virus (SIV) infection in African green monkeys (AGM) is nonpathogenic, even though it is characterized by plasma viral load (PVL) levels similar to those found during acute and chronic pathogenic infection of humans with human immunodeficiency virus type 1 and macaques with SIVmac (14). This feature is shared with other African nonhuman primates, such as sooty mangabeys (SM) and mandrills (19, 20). SIV-infected AGMs also display high viral loads in the gastrointestinal mucosa (11), a transient decline of circulating CD4⁺ T cells during acute infection (13), and longer-lasting CD4⁺ T-cell depletion in the intestinal lamina propia (10). Concomitant with the peak viral load during acute infection, SIVagm-infected AGMs display transient increases of CD4⁺ and CD8⁺ T cells expressing activation, and proliferation markers, such as MHC-II DR and Ki-67 (4, 13), and anti-SIVagm antibodies (Ab) are induced with kinetics similar to those found in SIVmac infection (5). Interestingly, however, the Ab response against Gag p27 is weak, if present at all (1, 2, 12, 15, 17, 18). This observation is surprising since, in the context of human immunodeficiency virus type 1 and SIVmac infections, Ab responses to Gag p27 are usually quite strong. Weak or low reactivity to Gag p27 has also been observed in some other natural SIV infections (7, 8, 20) but not in all of them (21). We wondered whether such a selective lack of Ab reactivity in the SIV-infected AGM might be related to a lack of Gag p27-specific T cells. With this hypothesis in mind, we first confirmed and extended the studies of humoral responses against Gag p27 by characterizing the antigen-specific immunoglobulin G (IgG) responses and mid-point titers against total SIVagm antigens (SIVagm virions) and recombinant Gag p27 (rP27; SIVagm) in naturally and experimentally SIVagm-infected AGMs. Second, we searched for the presence of Gag p27-specific T-cell responses in SIVagm infection by analyzing the CD4⁺ and CD8⁺ T-cell responses specific for Gag p27 and other SIVagm proteins in blood and lymph nodes (LNs) of acutely and chronically infected animals.Humoral responses against SIV were analyzed in 50 wild-born AGMs (Chlorocebus sabaeus) and 17 rhesus macaques (RMs). The animals were housed at the Institut Pasteur in Dakar, Senegal, and the California National Primate Research Center, Davis, CA, respectively, according to institutional and national guidelines. RMs were either noninfected (n = 5) or intravenously infected with SIVmac251 (n = 12). AGMs were noninfected (n = 23), naturally infected (n = 17), or intravenously infected with wild-type SIVagm.sab92018 (n = 10) (5, 9). IgG titers against SIVagm.sab92018 virions or rP27 were determined by an enzyme-linked immunosorbent assay (ELISA) using monkey anti-IgG as secondary Ab (Fig. 1A and B). The virions had been purified by ultracentrifugation on an iodixanol cushion from cell-free supernatants of SIVagm.sab92018-infected SupT1 cells. The His-tagged rP27 was constructed using DNA from gut cells of an SIVagm.sab92018-infected AGM 96011 (11). A Gag p27 PCR product was subcloned into pET-14b, and the recombinant protein was produced in Escherichia coli BL21(DE3)(pLysS) and purified on nitrilotriacetic acid columns. SIV-infected macaques showed high IgG titers cross-reacting with both SIVagm virions (Fig. 1A and B, left panels) and rP27 (Fig. 1A and B, right panels). In contrast, only 2 out of 27 SIV-infected AGMs showed detectable IgG responses against rP27 (Fig. 1A and B, right panels), while 21 out of 27 displayed significant responses against SIVagm virions (Fig. 1A and B, left panels). Two AGMs out of 23 from the negative control group showed weak responses at the limit of detection against SIVagm and two against rP27, suggesting a natural response against SIVagm proteins, cross-reactivity with unknown pathogens, maternal Ab, or recent SIV infection. Of note, the titers against whole SIV in the infected monkeys were higher in macaques than in AGMs, which may be due to a lack of anti-p27 Ab in most AGMs.Open in a separate windowFIG. 1.Cross-sectional analysis of IgG Ab responses against SIVagm or Gag p27 in SIV-infected AGMs and RMs. (A and B) Cross-sectional analysis by ELISA. IgG Ab against SIVagm.sab₉₂₀₁₈ virions or recombinant p27-Gag antigens were determined in SIV-negative (Rh SIV−) and chronically SIVmac₂₅₁-infected (Rh SIV+) RMs and in SIV-negative and chronically SIVagm-infected AGMs that were either naturally (AGM Nat SIV+) or experimentally (AGM Exp SIV+) infected with SIVagm.sab₉₂₀₁₈. Ab titers were calculated for each animal by limited dilution of plasma on coated ELISA plates with 5 μg/ml of (p27 equivalent) virions (left) or 1 μg/ml of the monomeric recombinant protein (rP27) (right). IgG detection by ELISA displayed a high background for rP27, especially at the highest plasma concentration (e.g., 1/100 and 1/400 plasma dilution) in SIV-negative RMs and AGMs. To discriminate between positive responses and background, calculated dose-response curves were compared to theoretical sigmoid-dose response curves corresponding to the 95% confidence interval of SIV-negative animals. By convention, responses were considered background when sigmoid dose-response curves were graphically within the 95% confidence interval of SIV-negative animals and when the calculated negative log 50% effective concentration (EC₅₀) was lower than the top theoretical sigmoid dose-response curve from SIV-negative animals (corresponding to a threshold of negative log EC₅₀ of 2.8). (A) Results (optical density at 450 nm OD₄₅₀]) are represented for both virions (left) and rP27 (right) over plasma dilution (log₁₀) on a per animal basis (data points) and for each group (lines). Lines represent the sigmoid dose-response curves for each group (Prism 4; Graphpad). (B) Mid-point IgG titers were determined for each animal from individual sigmoid dose-response curves, and presented as the log₁₀ value from the reciprocal of the effective concentration that corresponds to 50% response between minimum and maximum OD₄₅₀ (negative log EC₅₀). Horizontal bars represent the median mid-point titer per each group. Mann-Whitney nonparametric tests were applied for statistical analysis (n.s., nonsignificant, with P values of >0.1) (C) Cross-sectional analysis of Ab against SIVagm proteins by Western blot analysis using denatured SIVagm.sab₉₂₀₁₈. For the positive controls on the left, we used sera from an SIVmac₂₅₁-infected macaque and a SIVagm.sab₉₂₀₁₈-infected AGM. Development times and reagents were identical for all Western blots. Mo, months of infection; y, years of infection; C−, negative control; C+, positive control.The study of IgGs by Western blot analysis using denatured SIVagm.sab92018 virions showed no or weak anti-Gag responses in SIV-infected AGMs, yet the anti-Env responses were often strong (Fig. ​(Fig.1C).1C). In contrast, SIV-infected macaques showed a dominant IgG cross-reactive response against the SIVagm Gag p27 protein. Even if responses in AGMs were detected more frequently with the Western blot analyses than with the ELISAs, these responses were different in magnitude and considerably weaker than those in macaques.To compare B- and T-cell responses over time, five simian T-cell leukemia virus-seronegative AGMs were infected with SIVagm.sab92018, and the animals were followed longitudinally during the acute and postacute phases of infection until day 90 postinfection (p.i.). Sequential blood samples were collected and biopsies of auxiliary and inguinal LNs were performed on day −5 and at three times p.i. (days 14, 43, and 62). PVL was measured by real-time PCR (5). Since we searched for Gag p27-specific responses, we also quantified Gag p27 antigen in the plasma (SIV p27 antigen assay; Coulter, Miami, FL). Viral RNA and p27 antigenemia peaks were observed between days 7 and 14 p.i. (Fig. 2A and B, respectively). The Gag p27 levels were variable among the animals but in a range similar to those reported previously in AGMs and macaques (3, 5). As has also been observed in SIVmac infection (except for rapid progressors), plasma Gag p27 levels fell below the detection level in the postacute phase (i.e., after day 28 p.i.) (Fig. ​(Fig.2B2B and data not shown). There were significant increases in circulating CD8⁺ DR⁺ T cells at days 7 and 14 p.i. and in CD8⁺ Ki-67⁺ T cells at days 14 and 28 p.i. (Fig. 2C and D, left panels). After day 28 p.i., the percentages were no longer statistically different from baseline levels. In LN cells (LNCs), the percentage of CD8⁺ Ki-67⁺ T cells rose from 3.1% ± 1.1% before infection to 6.1% ± 0.3% at day 62 p.i., but the difference was not statistically significant (Fig. ​(Fig.2D,2D, right panel). The levels of blood CD4⁺ DR⁺ Ki-67⁺, CD8⁺ DR⁺ Ki-67⁺, CD8⁺ Ki-67⁺ T cells, and LNC CD8⁺ Ki-67⁺ T cells were positively correlated with viremia (P values of 0.002 for DR⁺ cells and P values of <0.02 for Ki-67⁺ cells). Altogether, these results confirm previous data showing early, transient T-cell activation in the peripheral blood of SIVagm-infected AGMs (13).Open in a separate windowFIG. 2.Plasma viremia and T-cell activation in blood and LNs of five longitudinally followed SIVagm.sab₉₂₀₁₈-infected African green monkeys. (A) SIVagm.sab RNA copy numbers in plasma. (B) Plasma Gag p27 concentrations. (C) Percentages of MHC-II DR-positive CD4⁺ (•) and CD8⁺ (○) T cells within, respectively, total CD4⁺ and CD8⁺ T cells from PBMCs and LNCs. (D) Percentages of Ki-67⁺ CD4⁺ (•) and CD8⁺ (○) T cells within, respectively, total CD4⁺ and CD8⁺ T cells from PBMCs and LNCs. Results are shown as the mean ± the standard error of the mean. Asterisks indicate statistically significant differences compared to levels before infection (P < 0.05).We next looked for the presence of Ab responses against rP27 in these animals. No Ab were detected before infection. After infection, all five AGMs developed anti-SIVagm IgGs within 4 to 9 weeks p.i., with AGM 02001 showing the fastest response (Fig. ​(Fig.3A).3A). While the humoral responses against whole virions were significant (Fig. ​(Fig.3B),3B), the anti-rP27 responses were below the threshold for positivity (Fig. ​(Fig.3B),3B), with the exception of one animal (AGM 02001). The anti-rP27 response in this animal was only transient since it was no longer detectable at week 75 p.i., in contrast to the anti-SIV Ab that were sustained (Fig. ​(Fig.3B3B and data not shown).Open in a separate windowFIG. 3.Longitudinal analysis of IgG titers and T-cell proliferative responses against SIVagm and Gag p27 in five AGMs experimentally infected with SIVagm.sab₉₂₀₁₈. (A and B) Ab responses were analyzed by ELISA. (A) IgG dose-response curves against SIVagm (top) and rP27 (bottom) are shown over time (week −1 to week 24 p.i.). O.D.450, optical density at 450 nm. (B) Mid-point titers were calculated as described in the legend to Fig. ​Fig.1A.1A. Continuous lines correspond to median titers from all five animals. Red, anti-SIVagm IgGs; green, anti-p27 IgGs. (C) Proliferative responses of CD4⁺ and CD8⁺ T cells were assessed by flow cytometry using carboxy fluorescein succinimidyl ester staining (CFSE). CD4⁺ and CD8⁺ T-cell responses in PBMCs (left) and LNCs (right) after stimulation with peptide pools (Gag without P27, P27, and Tat) and Gag rP27 are shown for each animal. All data are reported after background subtraction. Results are presented in columns as the mean ± the standard error of the mean. Asterisks indicate statistically significant differences compared to individual values before infection (P < 0.05).We next searched for T-cell responses against Gag p27 compared to other SIVagm antigens in these animals. Gag p27 epitopes were presented in the following two ways: in the context of rP27 and as synthetic peptides. The peptide pools (comprised of overlapping 15-mers) spanned the following SIVagm proteins: Gag p27, Gag without p27, Env, and Tat. The amino acid sequences of the Gag and Env peptides corresponded to the autologous wild-type SIVagm.sab92018 sequence, and those of the Tat peptides corresponded to an SIVagm.sab consensus sequence. The latter was determined using Tat sequences of other SIVagm viruses from Senegal that are available in the databases (SIVagm.sab1c, SIVagm.sabD42, and SIVagm.sabD30). We measured T-cell responses by investigating the antigen-induced proliferation. T cells from blood (peripheral blood mononuclear cells PBMCs]) and LNs were analyzed. All assays were performed with fresh cells that were stimulated with 10 μg/ml of Gag rP27 and 5 μg/ml of peptides over a period of 4 days. Dead cells were gated out using 7-amino-actinomycin D, and dividing (CFSE^low) cells were analyzed after stimulation with medium alone, SIV antigens, or concanavalin A as a positive control. We detected significant Gag p27-specific proliferative responses for CD8⁺ T cells in PBMCs and for CD4⁺ and CD8⁺ T cells in LNCs (Fig. ​(Fig.3C).3C). The animal with the detectable anti-p27 Ab (AGM 02001) did not show stronger p27-specific T-cell responses than the other animals. Thus, all SIV-infected AGMs were able to mount a proliferative T-cell response against p27, while anti-p27 IgGs were lacking in four of the animals. However, the SIVagm-specific T-cell responses were detected at only a few time points p.i.We then analyzed the T-cell responses in the chronic phase of AGMs naturally and experimentally infected with SIVagm.sab92018. PVL, peripheral blood cell counts (CD4⁺ and CD8⁺ T cells; CD20⁺ B cells), and immune activation (Ki-67⁺ CD4⁺ and CD8⁺ T cells) were similar in naturally infected and in experimentally infected AGMs (Fig. ​(Fig.4A).4A). As expected, cell counts and immune activation levels were also not different from SIV-negative AGMs (Fig. ​(Fig.4A).4A). Again, we measured SIV-specific responses first by a proliferation assay (Fig. ​(Fig.4B).4B). One out of five animals tested had a proliferative SIV-specific CD4⁺ T-cell response (against Gag without p27, P27, rP27, Env GP120, and Tat), and two animals had a CD8⁺ T-cell response (against P27 in both animals and against Env GP120 and Tat in one). Two animals (one naturally infected and one experimentally infected with SIVagm.sab92018) did not show any detectable antigen-specific proliferative CD4⁺ or CD8⁺ T-cell response.Open in a separate windowFIG. 4.Immune parameters and SIVagm-specific proliferative and cytokine T-cell responses in chronically infected AGMs. (A) Cell counts (CD4⁺ and CD8⁺ T cells; B cells) and immune activation levels (percent of Ki-67⁺ in CD4⁺ and CD8⁺ T cells) in AGMs (n = 4) naturally infected with SIVagm (Nat SIV⁺) and AGMs (n = 6) experimentally infected with SIVagm.sab₉₂₀₁₈ (Exp SIV⁺) compared to uninfected AGMs (n = 10) (SIV⁻). PVL, if known, is indicated. Green, blue, and orange symbols correspond, respectively, to noninfected, naturally infected, and experimentally infected AGMs. (B) Proliferative response to SIVagm antigens in chronically infected AGMs (n = 5) compared to those in uninfected AGMs (n = 3). PBMCs were stimulated with the same antigens as those described in the legend to Fig. ​Fig.3.3. (C) Analysis of cytokine responses (gamma interferon IFN-γ] and tumor necrosis factor alpha TNF-α]) by SIVagm-specific T cells. ConA was used as a positive control. Representative results from a single animal are shown here. (D) Cumulative values of SIVagm-specific TNF-α and IFN-γ responses in chronically infected animals. The responses to SIVagm antigens were analyzed in peripheral blood specimens of 4 naturally and 5 experimentally infected AGMs as well as 10 uninfected AGMs. The data are reported after background subtraction corresponding to the subtraction of the frequency of positive events from the unstimulated samples to the frequency of positive events from the antigen-specific stimulation. Proliferative T-cell responses and cytokine T-cell responses in SIV-infected AGMs were defined as positive when higher than 3 standard deviations above the mean responses for uninfected animals. Freq, frequency; w/o, without.These results were extended to an analysis of SIV-specific T-cell cytokine responses, e.g., the production of IFN-γ and TNF-α in nine chronically infected compared to 10 noninfected AGMs (Fig. 4C and D). Fresh cells were stimulated for 8 h with the antigens described above. SIV-specific cytokine responses were detected in CD8⁺ but not in CD4⁺ T cells. Seven animals out of nine showed a response against at least one antigen. The two animals showing no response were among the four naturally infected animals tested. We therefore cannot exclude that the absence of response in these two animals is due to the presence of highly divergent viruses. However, a precise epitope mapping in SIVagm sequences would be necessary to confirm this. In those animals showing a SIVagm-specific cytokine T-cell response, the responses were directed against Gag p27 (four out of nine animals), other Gag proteins than p27 (two out of nine animals), and Env GP120 (four out of nine animals). In the experimentally infected animals, we might have underestimated the responses against Tat compared to Gag and Env antigens, since the Tat peptides corresponded to an SIVagm.sab consensus sequence and not to the autologous virus (SIVagm.sab92018). There was no correlation between the magnitude or breadth of SIV-specific T-cell responses and immune activation or PVL.Altogether, our study demonstrates that AGMs can mount T-cell proliferative and cytokine responses against Gag p27. The T-cell response was variable among the animals. In general, it appeared moderate, comparable to chronically SIV-infected RMs (9). Of note, T-cell responses were not consistently detected at all time points and not in all animals. We cannot exclude the possibility that we underestimated the magnitude of the cytokine responses. For instance, we did not costimulate the cells during the assays. However, cytokine responses were also variable in vervet AGMs, with a trend for reduced levels compared to those for RMs, even when more-sensitive assays were used (23). In SM, the responses were also reported to be not stronger than in RMs. This is in line with the lack of efficient control of viral replication in natural hosts (6, 22).In our study, we show that IgG responses against Gag p27 are either lacking, weak, or transient, while Ab against other SIVagm proteins are present. The mechanisms underlying this selective lack of Gag p27 Ab responses are unclear. It could be related to moderate and/or dysfunctional CD4⁺ T-cell responses and/or due to an unknown suppressive regulatory mechanism. SIV-specific T-cell cytokine responses were indeed principally found at the CD8⁺ T-cell level. This was also reported in SIVsm-infected SM (6, 22). Here, we also searched for SIVagm Gag p27-specific proliferative responses. Interestingly, they were detected for CD4⁺ T cells, indicating the presence of p27-specific CD4⁺ memory cells in AGMs. Moreover, AGMs can potentially mount a strong and sustained anti-Gag p27 humoral response, when appropriately immunized (D. Favre et al., unpublished data). This suggests that there is neither a central B-cell tolerance against p27 Gag protein in AGMs nor an inherent inability for CD4⁺ T cells to provide helper B-cell functions. The transient nature of anti-p27 Ab in one animal would be in favor of regulatory mechanisms, but that needs to be confirmed. Another explanation could be that AGMs are able to mount Ab responses against some p27 epitopes but not to those exposed by the native protein, which would explain why we and others detect more frequently humoral responses in Western blot analysis than in ELISAs (16).In conclusion, we characterized the IgG responses against SIVagm and confirmed a lower humoral response against p27 than in RMs. Moreover, our study reveals that cytokine and proliferative T-cell responses against SIVagm Gag p27 are detectable in AGMs. Thus, the reduced ability of the AGM to produce Ab against Gag p27 p.i. is not related to a lack of Gag p27-specific T cells.