Simian Immunodeficiency Virus (SIV) Infection of a Rhesus Macaque Induces SIV-Specific CD8⁺ T Cells with a Defect in Effector Function That Is Reversible on Extended Interleukin-2 Incubation

Abstract

A vigorous expansion of antigen-specific CD8⁺ T cells lacking apparent effector function was observed in a rhesus macaque acutely infected with the simian immunodeficiency virus (SIV) strain SIVmac239. Antigen-specific CD8⁺ T cells were identified using antigenic-peptide class I major histocompatibility complex tetramers. As many as 8.3% of CD8⁺ cells recognized the Mamu-A*01-associated SIV epitope Gag_181–189 (CTPYDINQM); however, these cells demonstrated no effector function when presented with peptide-incubated targets, as measured by intracellular cytokine staining for gamma interferon (IFN-γ), interleukin-2 (IL-2) production, or direct cellular lysis. Similar results were observed with three other SIV peptide antigens. Nonresponsiveness did not correlate with apoptosis of the CD8⁺ cells, nor were cells from this macaque impaired in their ability to present peptide antigens. Associated with the nonresponsive state was a lack of IL-2 production and decreased IL-2 receptor expression. Exogenous IL-2 treatment for 1 week in the absence of antigenic stimulation restored antigen-specific responses and the quantitative correlation between tetramer recognition and antigen-responsive IFN-γ secretion. This case report suggests a regulatory mechanism that may impede the effector function of antigen-specific T cells during acute infection with SIV or human immunodeficiency virus in some cases. This mechanism may participate in the failure of the immune system to limit infection.

CD8⁺ T cells are thought to play an important role in the control of human immunodeficiency virus type 1 (HIV-1) (21, 24, 26) and simian immunodeficiency virus (SIV) (12, 28) infection. Induction of HIV-specific CD8⁺ T-lymphocyte responses is associated with the control of plasma viremia (23) and correlates with delayed disease progression in adults (3) and infants (26). The quantitation of antigen-specific CD8⁺ T cells in vivo has been greatly facilitated by the use of tetrameric complexes of antigenic peptide and MHC, which has brought about significant advances in the understanding of the dynamics of T-cell responses (2, 25). Current reviews of tetramer technology concur that there is a strong quantitative correlation between tetramer binding activity and functional measures of antigen-specific CD8⁺ T-cell activity (6, 19). However, several authors have recently reported examples of circulating antigen-specific CD8⁺ T cells that appear to be functionally inactive. For example, Zajac et al. have reported that in lymphocytic choriomeningitis virus-infected CD4-deficient mice, antigen-specific CD8⁺ T cells do not have the ability to secrete gamma interferon (IFN-γ) in response to antigen, unlike the analogous population in CD4-normal mice (34). Spiegel et al. have reported the persistence of high frequencies of HIV peptide-specific CD8⁺ cells, unable to secrete IFN-γ in response to antigenic peptide, in an HIV-1-infected individual with very low CD4⁺ T-cell counts (29). Lee et al. have shown the existence in a melanoma patient of melanoma-specific CD8⁺ cells that fail to exhibit cytolytic activity in vitro or to produce interleukin-2 (IL-2), IL-10, IFN-γ or tumor necrosis factor alpha in response to antigen (15). Finally, Lechner et al. recently showed that in acute hepatitis C virus infection, tetramer-positive cells are defective in IFN-γ production during the acute phase but recover over time. They and others have described these cells as “stunned lymphocytes” (14).

Here, we report a vigorous expansion of SIV-specific CD8⁺ T cells without effector function during an acute SIVmac239 infection. The nonresponsive cells did not produce cytokines or have cytolytic activity against a panel of SIV peptide antigens, despite having tetramer reactivity. However, incubation of the non-responsive cells with IL-2 in the absence of antigen restored the responses and the quantitative correlation between tetramer binding and cytokine secretion. The apparent nonresponsiveness of the antigen-specific CD8⁺ cells may reflect one mechanism by which immunodeficiency viruses escape the immune response.

Tetramer binding and IFN-γ staining.

Comparison of the frequencies of tetramer-positive cells and cells responding in an intracellular cytokine assay revealed nonresponsive SIV-specific CD8⁺ cells in an acutely infected but not a chronically infected macaque. We compared the frequency of antigen-specific CD8⁺ T cells (as measured by major histocompatibility complex [MHC]-antigenic peptide tetramers) to the frequency of antigen-reactive cells (as measured by an intracellular cytokine assay) in two infected Mamu-A*01-positive macaques (Mamu-A*01 typed by PCR with sequence-specific primers by the method of Knapp et al. [13]). One macaque, designated RPb4, was acutely infected with an intravenous challenge of highly pathogenic SIVmac239 (5). The other, designated RLc5, was chronically infected with SHIV-89.6, a nonpathogenic chimera of SIV and HIV (27). The two viruses have a common Gag antigen. For RPb4, peripheral blood mononuclear cells (PBMC), lymph node cells, and spleen cells were cryopreserved 30 to 60 days after infection (18). For RPb5, only cryopreserved PBMC were available. Thawed cells were stained with the Mamu-A*01 tetramer complex with the dominant SIV Gag epitope Gag_181–189 (CTPYDINQM). In the acutely infected macaque, RPb4, high levels of Gag_181–189-tetramer-specific CD8⁺ T cells were detected in each cellular compartment, with the highest frequency being found in lymph nodes and spleen (PBMC, 2.3% ± 0.5% [Fig. 1A]; lymph node cells, 8.3% ± 0.4%; spleen cells, 7.9% ± 0.8% [average of three determinations]). In the chronically infected macaque, RLc5, for which only PBMC were available, tetramer-positive cells were also detected (PBMC, 0.63% [two determinations] [Fig. 1A]).

FIG. 1.

(A and B) Tetramer binding does not correlate with IFN-γ secretion in the acutely infected macaque RPb4. (A) PBMC from macaque RPb4 acutely infected with SIVmac239 and macaque RLc5 chronically infected with SHIV89.6 were stained with the Mamu-A*01-Gag_181–189 tetramer. Values in the upper right quadrants represent the percentage of CD3⁺ CD8⁺ cells reactive with the tetramer. A total of 100,000 events were analyzed. (B) IFN-γ secretion by unstimulated, phorbol myristic acid (PMA)-ionomycin (mitogen)-treated, or Gag_181–189-incubated cells. (C and D) IL-2 treatment for 1 week recovers antigen responsiveness and the correlation with tetramer binding. (C) IL-2 treatment for 1 week in the absence of antigen restored IFN-γ secretion of RPb4 splenocytes in response to a panel of SIV epitopes. (D) T-cell binding to the corresponding tetramer in each case. The rightmost panel in panel D shows staining with an irrelevant tetramer, HLA-A2 with the antigenic Epstein-Barr virus peptide GLCLVAML. Values indicate the percentage of CD3⁺ CD8⁺ cells in the labeled quadrant. Data are representative of three experiments.

Intracellular cytokine assays were subsequently performed on cells from the two macaques based on protocols developed by Waldrop et al. (33). In the chronically infected animal RLc5, 0.56% of CD8⁺ T cells from PBMC produced IFN-γ in response to Gag_181–189 (Fig. 1B), concordant with the number of cells staining with the Gag_181–189 tetramer (0.63%). However, PBMC (Fig. 1B) and cells from the LN or spleen (data not shown) of RPb4 did not produce IFN-γ or IL-2 in response to Gag_181–189, despite the high frequency of tetramer-staining CD8⁺ cells (0.06% IFN-γ responsive, equivalent to background, compared to a tetramer-binding population of 2.4%).

The nonresponsiveness of antigen-specific T cells in the acutely infected macaque was also observed for three other epitopes for which tetramers were available. To determine whether the defect in peptide recognition in RPb4 was restricted to the Gag_181–189 epitope, we measured the IFN-γ response to other known Mamu-A*01-restricted epitopes including Env_234–242 (CAPPGYALL), Env_626–634 (TVPWPNETL), (7), and Tat_28–35 (STPESANL [D. Watkins, Wisconsin Regional Primate Center, personal communication]). In the acutely infected macaque, RPb4, no response to any of the epitopes could be measured, despite tetramer-staining cell frequencies ranging from 2.9 to 4.7% of CD8⁺ T cells (data not shown). A mitogen response was intact in both animals, since CD8⁺ T cells responded with IFN-γ production to stimulation with phorbol myristic acid and ionomycin (Fig. 1B). These results suggest that acute SIV infection with the pathogenic viral strain SIVmac239 induced a nonresponsive state in antigen-specific CD8⁺ T cells to subsequent SIV antigen stimulation. The possibility that the nonresponsive state of CD8⁺ T cells might extend to non-SIV antigens could not be examined because of the lack of defined peptide recall antigens in the macaque model system.

IL-2 recovery of the IFN-γ response.

Culture of the spleen cells of RPb4 with IL-2 (without antigen) for 1 week restored the responsiveness to peptide. It has been demonstrated that T-cell proliferation is triggered by the interaction of IL-2 and its receptor following T-cell activation (30). A lack of IL-2 during T-cell activation can result in anergy of the responding cells (17). In our studies, the CD8⁺ T-cell unresponsiveness in RPb4 was accompanied by undetectable levels of IL-2 production in response to Gag_181–189 (data not shown). This, together with a deficit of CD4⁺ cells in the nonresponsive animal (only 11% of total CD3⁺ cells [see below]), led us to postulate that low levels of IL-2 might be related to CD8⁺ cell nonresponsiveness. Spleen cells of RPb4 were cultured with recombinant human IL-2 (20 U/ml) for 7 days without antigenic stimulation. This culture restored the ability of CD8⁺ T cells to produce IFN-γ in response to subsequent antigenic stimulation by all peptides tested, although with a somewhat reduced mean IFN-γ expression (Fig. 1C). After recovery, the frequencies of IFN-γ-producing CD8⁺ T cells from the spleen were 2.4% against Gag_181–189, 1.6% against Env_234–242, 0.8% against Env_626–634, and 1.0% against Tat_28–35. The frequencies of tetramer-staining cells for the same antigens were in good concordance, ranging from 43 to 76% of responding cells (Fig. 1D). Incubation of RPb4 spleen cells with IL-2 for 6 h in the presence of Gag_181–189 peptide did not restore the response, indicating the requirement for the longer incubation. Addition of a monoclonal antibody to IL-2 during the IL-2 culture blocked the recovery of the IFN-γ response (data not shown). Thus, IL-2 treatment restored both the antigen-specific response and the correlation between responding cell frequency (IFN-γ assay) and antigen recognition frequency (tetramer assay).

Lytic activity.

Deficits in the lytic activity of RPb4 cells could be restored by a 1-week incubation in IL-2 in the absence of antigen. The primary cytotoxic activity of splenocytes and PBMC from macaque RPb4 and PBMC from RLc5 was tested by using freshly thawed cells as effectors in a standard 5-h chromium release assay (20). Mamu-A*01-transfected 721.221 cells (a cloned Epstein-Barr virus-transformed human B-cell line with homozygous deletions of the MHC class I loci [1]) were used as targets in the assay. No primary cytotoxic T-lymphocyte activity against Gag_181–189, Tat_28–35, Env_234–242, or Env_626–634 was found in either splenocytes (Fig. 2A panel i) or PBMC (Fig. 2A panel iii, Gag_181–189 only) from the acutely infected macaque, RPb4. A very low level of Gag_181–189-directed cytotoxicity was detected in the chronically infected macaque RLc5 (Fig. 2A panel iii), correlating with the lower frequency of Gag_181–189-specific CD8⁺ T cells in this animal. After 1 week of incubation with IL-2 in the absence of antigen, splenocytes from RPb4 (Fig. 2A panel ii) lysed all peptide-incubated targets tested, with the exception of the irrelevant control (Gag_2–10, GVRNSVLSG; lysis less than 6% [data not shown]). PBMC from both animals lysed Gag_181–189-incubated targets after a conventional in vitro boost with peptide plus IL-2 for 10 days (20 U of IL-2 per ml plus 5 μg of peptide per ml) (Fig. 2A panel iv).

FIG. 2.

(A) CTL activity of CD8⁺ T cells from SIV-infected macaques. Lymphocytes were tested for their ability to lyse peptide-incubated Mamu-A*01-positive 721.221 cells in a standard 5-h chromium release assay. Spleen cells from the acutely infected macaque, RPb4, were not treated (i) or were incubated for 1 week with IL-2 in the absence of antigen (ii). PBMC from RPb4 of RLc5 were not treated (iii) or were incubated for 1 week with IL-2 in the presence of antigen (iv) before the lytic assay. Data are representative of three experiments. (B) Defective induction of IL-2Rα (CD25). PBMC from RPb4 and RLc5 were gated for the expression of CD3, CD8, and Mamu-A*01-Gag_181–189 tetramer. These tetramer-positive cells were analyzed for CD25 expression before or after stimulation with antigenic Gag_181–189 peptide or were cultured in the presence of IL-2 for 1 week and then analyzed before or after Gag_181–189 stimulation. The histograms show the percentage of CD8⁺ tetramer-positive cells that are also CD25⁺ tetramer-positive. A total of 100,000 events were analyzed, and the frequency of tetramer-positive cells varied from 2.2 to 2.7% of the CD3⁺ CD8⁺ cells in RPb4 and from 0.21 to 0.61% in RLc5. The data are representative of two experiments. (C) Splenocytes from RPb4 can present peptide antigen. Cryopreserved splenocytes were thawed and cultured in the presence of Gag_181–189 peptide for 2 h and were washed and used as APC to present antigen to autologous spleen cells, previously treated with IL-2 for 1 week, in an ICC assay for IFN-γ. Data are expressed as the percentage of CD3⁺ CD8⁺ cells that are also IFN-γ⁺ CD69⁺ and are representative of three experiments. (D) Apoptosis of SIV-specific CD8⁺ T cells. Spleen cells from SIVmac239-infected RPb4 and PBMC from SHIV89.6-infected RLc5 were stimulated with Gag_181–189 for 6 h and stained for the expression of CD8, Gag_181–189 tetramer, and annexin V. Data are expressed as percentage of CD3⁺ CD8⁺ cells that are either tetramer-positive annexin V-negative or tetramer-positive annexin V-positive and are representative of three experiments.

Phenotypic analysis of T cells.

PBMC from RPb4 and the chronically infected macaque RLc5 were examined by flow cytometry for the expression of a number of phenotypic markers. A difference in the CD4⁺/CD8⁺ T-cell ratio (11:89 in RPb4 versus 50:50 in RLc5) was found, while CD3⁺ T cells remained in similar proportion to total PBMC (67.8% in RPb4 versus 68.3% in RLc5). This association of a relative CD4⁺ T-cell deficit in RPb4 with CD8⁺ T-cell nonresponsiveness is consistent with the observation of Zajac et al. of CD8⁺ T-cell nonresponsiveness to LCMV antigens in mice lacking CD4⁺ T cells (34). Further phenotypic analysis showed that Gag_181–189 tetramer-binding CD8⁺ T cells in the unresponsive macaque were positive for CD28⁺ (54% on CD8⁺ from RPb4 versus 87% on CD8⁺ from RLc5) but negative for CD45RA (12% on CD8⁺ from RPb4 versus 13% on CD8⁺ from RLc5). Although memory T-cell phenotypes have not been fully defined in macaques, the results suggest that the nonresponsive T cells are memory cells but not effectors, applying the subset definitions used in humans (11). The activation marker Mamu-DR (the homologue of HLA-DR) was expressed at low levels on tetramer-positive CD8⁺ T cells in both macaques (5% on RPb4 and 7.6% on RLc5).

IL-2Rα on unresponsive CD8⁺ T cells.

The high-affinity IL-2 receptor alpha (IL-2Rα) chain (CD25) is known to be important in regulating the T-cell response to antigen (22). Decreased expression of IL-2Rα has been found on CD8⁺ T cells from HIV-infected individuals (32). Therefore we examined the expression of CD25 on Gag_181–189 tetramer-positive CD8⁺ T cells. We found that CD25 was expressed at low levels on these T cells in both macaques (1.6 and 7.3% of tetramer-staining CD8⁺ T cells for acutely infected RPb4 and chronically infected RLc5, respectively) before in vitro antigenic stimulation (Fig. 2B). After 6 h of stimulation with Gag_181–189 peptide, 80% of the Gag_181–189 tetramer-positive CD8⁺ T cells from RLc5 expressed CD25 whereas only background levels of staining (6.4%) were observed in RPb4. However, after a 7-day culture of cells with 20 U of IL-2 per ml, 31 and 35% of antigen-specific CD8⁺ T cells expressed CD25 in RPb4 and in RLc5, respectively. Subsequent stimulation with Gag_181–189 peptide increased the expression of CD25 on antigen-specific CD8⁺ T cells to 88% in RPb4 and 72% in RLc5. Thus, the combination of IL-2 preculture and antigen exposure boosted the frequency of CD25-positive tetramer-reactive cells in the nonresponsive macaque to a similar level to that seen in the chronically infected macaque. A related finding was reported by Groux et al. (10), who found that that CD4⁺ T cells rendered anergic by culture with IL-10 failed to upregulate CD25 in response to antigen but that preculture with IL-2 restored CD25 expression in response to antigenic stimulation. On the other hand, failure to express CD25 in the presence of exogenous IL-2 correlated with a failure to reverse the anergic state. This parallel between the induction of CD25 and the recovery of the T-cell response suggests that defective CD25 expression may be involved in T-cell nonresponsiveness.

Antigen presentation.

Cells from the unresponsive macaque functioned normally to present antigen. HIV-1 has evolved a mechanism of immune evasion involving the downregulation of MHC class I expression mediated by the viral protein Nef (4, 9). Such a downregulation could affect the ability of cells to present exogenously added peptide antigens. To address the ability of unstimulated splenocytes from the acutely infected macaque RPb4 to present antigen, spleen cells were pulsed with Gag_181–189 peptide for 2 h and then washed extensively. They were incubated as antigen-presenting cells (APC) with autologous spleen cells previously cultured for 7 days with IL-2. The frequency of IFN-γ-producing CD8⁺ T cells was similar to that seen on direct addition of Gag_181–189 peptide (10 μg/ml) to the IL-2-cultured cells (4.6 and 3.3%, respectively [Fig. 2C]). Thus, APCs from the unresponsive macaque can present antigen, indicating that the CD8⁺ T-cell unresponsiveness seen in untreated cultures was not due to defective antigen presentation.

Apoptosis.

T cells from HIV-infected persons are highly prone to in vitro spontaneous and activation-induced apoptosis (8), and this tendency has been implicated in T-cell hyporesponsiveness in HIV disease (16). We therefore examined the role of apoptosis in the antigen responses of RLc5 and RPb4. Annexin V was used as a marker of apoptosis, and propidium iodide was used to differentiate apoptotic from necrotic cells (31). No significant necrosis (propidium iodide-positive annexin V-positive phenotype) was seen in any assay (data not shown). Cells from each macaque (freshly thawed and not pretreated with IL-2) were stained immediately for annexin V or were stimulated with Gag_181–189 peptide for 6 h, followed by staining and flow cytometric analysis. Untreated, total CD8⁺ cells in both macaques had a background frequency of apoptosis of about 10% (data not shown). Gag_181–189 tetramer-positive CD8⁺ cells in untreated samples were positive for the apoptotic marker at a higher frequency (35% for RPb4 and 28% for RLc5 [data not shown]), indicating that there was more apoptosis in the antigen-specific cell population than in the total CD8⁺ T-cell population in both animals. Incubation with Gag_181–189 peptide for 6 h increased the frequency of apoptotic tetramer-positive cells in both animals (58% for RPb4 [ratio of 3.6 tetramer-positive annexin V-positive cells to 3.2 tetramer-positive annexin V-negative cells] and 47% for RLc5 [0.14:0.18] [Fig. 2D]). However, it is clear that the slightly higher frequency of apoptotic tetramer-positive cells in RPb4 was not sufficient to fully account for the nonresponsive T-cell phenotype in that animal compared to RLc5. In other experiments where cells were thawed and treated with IL-2 alone for 1 week, we observed that the total lymphocyte counts changed by only ±15% between the initiation and termination of the cultures (data not shown). The frequency of Gag_181–189 tetramer-positive spleen CD8⁺ T cells, however, declined from 7.1% (data not shown) to 4.1% (Fig. 1D) over the same period, which suggests that apoptotic tetramer cells observed in the untreated sample were lost, as expected, during the week-long incubation.

In summary, acute SIVmac239 infection of a rhesus macaque resulted in high frequencies of SIV-specific CD8⁺ T cells without apparent effector function. Antigen responsiveness in the acutely infected macaque, as measured by IFN-γ production and cytotoxic T-lymphocyte activity was restored after a 1-week incubation with IL-2, as was the correlation between tetramer binding and IFN-γ production. A plausible interpretation of the data is that CD8⁺ T-cell unresponsiveness resulted from a lack of CD4⁺ T-cell help and IL-2 production, with a resulting defect in CD25 induction. The lack of CD8⁺ T-cell response that we observed in this acutely infected macaque may be a result of a specific mechanism used by SIV to escape immune surveillance or may reflect a generalized depletion of cofactors like IL-2 and CD4⁺ T-cell help necessary for the proper activation of mature, antigen-exposed CD8⁺ T-cell effectors. In either case, it will be important to understand the frequency and implications of the phenomenon in retroviral and potentially other viral infections to elucidate how to induce and and sustain beneficial virus-specific CD8⁺ T-cell effectors.