Role of Escape Mutant-Specific T Cells in Suppression of HIV-1 Replication and Coevolution with HIV-1

Escape mutant-specific CD8+ T cells were elicited in some individuals infected with escape mutants, but it is still unknown whether these CD8+ T cells can suppress HIV-1 replication. We clarified that Gag280V mutation were selected by HLA-B*52:01-restricted CD8+ T cells specific for the GagRI8 protective epitope, whereas the Gag280V virus could frequently elicit GagRI8-6V mutant-specific CD8+ T cells. GagRI8-6V mutant-specific T cells had a strong ability to suppress the replication of the Gag280V mutant virus both in vitro and in vivo. In addition, these T cells contributed to the selection of wild-type virus in HLA-B*52:01+ Japanese individuals. We for the first time demonstrated that escape mutant-specific CD8+ T cells can suppress HIV-1 replication and play an important role in the coevolution with HIV-1. Thus, the present study highlighted an important role of escape mutant-specific T cells in the control of HIV-1 and coevolution with HIV-1.

cells, is proposed to eradicate latent HIV-1 reservoirs from antiretroviral therapy (ART)treated individuals (10,11). A previous study in a nonhuman primate model of simian immunodeficiency virus showed that mosaic vaccines in combination with an immune modulator Toll-like receptor 7 (TLR7) agonist improved virologic control and delayed viral rebound following discontinuation of antiretroviral therapy and that the breadth of cellular immune responses correlated inversely with set point viral loads and correlated directly with time to viral rebound (12), suggesting that effective cellular immunity is required in kick-and-kill treatment. A recent clinical trial of a therapeutic vaccine in 26 ART-suppressed HIV-infected individuals who had started with ART during an acute infection demonstrated that the mosaic vaccine induced high levels of polyfunctional CD4 ϩ T cells and CD8 ϩ T cells, as well as Env-specific antibodies, but the effect of this vaccine to delay viral rebound following discontinuation of antiretroviral therapy was small compared to that of placebo controls (13).
HIV-1-specifc T cells select HIV-1 escape mutants affecting T cell recognition (14)(15)(16)(17). Therefore, the existence of escape mutations in reservoir viruses and circulating viruses is a critical barrier for the eradication of latent HIV-1 reservoirs and prevention of HIV-1 infections. These escape mutant viruses can elicit mutant-specific T cells in some cases (18)(19)(20)(21). A recent study showed that the transmission of human leukocyte antigen (HLA)-adapted mutations affects the clinical outcome in the acute phase of an HIV-1 infection (22). T cell responses to epitopes including HLA-adapted mutations are frequently detected in HIV-1 chronic infections (23), whereas they are rarely found in acute infections (22). Although some HLA-adapted mutations are known to be escape ones, it remains unknown whether T cells specific for epitopes having HLA-adapted or escape mutations can effectively suppress HIV-1 replication in chronic infections. Previous studies demonstrated that escape mutant-specific T cells fail to suppress replication of the mutant virus in vitro (19,24,25).
In the present study, we investigated the mechanisms for the selection and accumulation of escape mutations at Gag280 in HIV-1 subtype B-infected Japanese individuals and for elicitation of escape mutant-specific T cells. Furthermore, we investi-as only 0.08% of them in August 2013 (Fig. 2B). We established RI8-6T-specific T cell lines from peripheral blood mononuclear cells (PBMCs) at these time points by sorting for RI8-6T-specific T cells and analyzed the ability of these T cell lines to recognize target cells prepulsed with RI8-6V peptide and those infected with Gag280-6V virus. Two RI8-6T-specific T cell lines recognized 721.221-B*52:01 cells prepulsed with RI8-6T peptide and those infected with Gag280-6T virus, whereas they failed to recognize those prepulsed with RI8-6V peptide and those infected with Gag280-6V virus ( Fig. 2C and D). In addition, a viral suppression assay showed that these T cell lines strongly suppressed the replication of the Gag280T virus but not that of the Gag280-6V one (Fig.  2E). Taken together, these results support the idea that the Gag280V mutation could be selected by RI8-6T-specific T cells. However, it remains unclear as to why the Gag280V mutant did not accumulate in the subtype B-infected HLA-B*52:01 ϩ individuals.  3D). On the other hand, 2 RI8-6T-specific T cell clones, 2F and 8F, effectively recognized 721.221-B*52:01 cells prepulsed with RI8-6T peptide, though the latter one had a weak ability to cross-recognize those prepulsed with RI8-6V peptide at a high concentration (Fig. 3C). Both clones effectively recognized 721.221-B*52:01 cells infected with the Gag280T virus, whereas clone 2F and clone 8F failed to recognize and weakly recognized, respectively, those infected with the Gag280V virus (Fig. 3D). Analysis using the B*52:01 tetramers demonstrated that clone 2F and clones 6C and 11B were RI8-6Tspecific and RI8-6V-specific T cells, respectively, and that clone 8F was cross-reactive T cells that strongly bound to the RI8-6T tetramer but weakly to the RI8-6V one (Fig. 3E). Thus, 3 types of RI8-specific T cells (RI8-6V-specific, RI8-6T-specific, and cross-reactive T cells) were elicited in KI-855. Further analyses using viral suppression assays demonstrated that the RI8-6V-specific T cell clone effectively suppressed the replication of the Gag280V virus but not that of the Gag280T one and that the RI8-6T-specific T cell clone suppressed the replication of the Gag280-6T virus but not that of the Gag280-6V one (Fig. 3F). The cross-reactive T cell clone revealed a strong ability to suppress the replication of both viruses, though the viral suppression ability for the Gag280-6V virus was weaker than that for the Gag280-6T one (Fig. 3F). These results indicate that T cells having a strong ability to suppress the replication of the Gag280V virus were elicited in HLA-B*5201 ϩ individuals infected with the Gag280V virus. KI-855 revealed a reversion of Gag280V to Gag280T after the elicitation of RI8-6V-specific and cross-reactive T cells. This finding supports the idea that RI8-6V-specific T cells and/or cross-reactive T cells selected the wild-type virus.
We next analyzed RI8-6V-specific and/or cross-reactive T cells in all 12 HLA-B*5201 ϩ individuals infected with the Gag280V virus. RI8-6V-specific T cells were detected in 10 of these individuals, though RI8-6T-specific T cells were also found in 4 of them (Fig.  4A). The analysis of PBMCs from 5 individuals, performed by using specific tetramers, confirmed the existence of RI8-6V-specific T cells in these 5 individuals (Fig. 4B). RI8-6V-specific T cell lines established from 4 individuals demonstrated a strong ability to suppress replication of Gag280V mutant virus, though those from KI-917 exhibited a strong ability to suppress the replication of both viruses (Fig. 4C). These results demonstrated that RI8-6V-specific T cells and/or cross-reactive T cells were frequently

Contribution of RI8-6T and RI8-6V-specific CD8 ؉ T cells to control of HIV-1 in subtype B infection.
Next, we analyzed the effect of Gag280 mutations on the clinical outcome in subtype B-infected HLA-B*52:01 ϩ Japanese individuals. The individuals infected with the Gag280T virus had significantly higher CD4 counts than those with Gag280S/A virus, whereas the Gag280V-infected individuals showed a trend for a higher CD4 count than the Gag280S/A-infected ones (Fig. 5A). These results suggest that RI8-6T/6V-specific T cells may have suppressed the replication of HIV-1 in these individuals. We therefore investigated the association of T cell responses to RI8-6T/6V with the clinical outcome. Responders to RI8-6T or 6V peptide showed significantly higher CD4 counts and trends toward a lower plasma viral load (pVL) than nonresponders (Fig. 5B), indicating that both RI8-6T-specific and RI8-6V-specific T cells contributed to the suppression of HIV-1 replication in subtype B-infected HLA-B*52:01 ϩ Japanese individuals.

DISCUSSION
A previous study on HLA-associated HIV-1 polymorphisms in HIV-1 subtype B-infected Japanese individuals showed that Gag280S and Gag280A are HLA-B*52:01associated mutations, whereas Gag280V is not (46). This finding suggested that Gag280S and Gag280A, but not Gag280V, are escape mutations selected by HLA-B*52: 01-restricted RI8-specific T cells. However, the present study clearly demonstrated that HLA-B*52:01-restricted RI8-specific T cells failed to recognize cells infected with Gag280V, Gag280A, or Gag280S mutant virus, indicating that these mutations were escape ones. Gag280V had not accumulated in the HLA-B*52:01 ϩ individuals, whereas this mutation was found more frequently than the Gag280A or Gag280S mutation in Japanese individuals. These findings together suggested the presence of a mechanism responsible for the absence of accumulation of the Gag280V mutation in the HLA-B*52: 01 ϩ individuals. Our hypothesis is that RI8-6V-specific T cells were elicited in HLA-B*52: 01 ϩ individuals infected with the Gag280V mutant virus and that these T cells selected the wild-type virus. Indeed, we demonstrated that RI8-6V-specific T cells were detected in most of the HLA-B*52:01 ϩ individuals infected with Gag280V mutant virus and that these T cells had a strong ability to suppress replication of the Gag280V mutant virus.
The results of an HLA class I stabilization assay showed that the binding affinity of RI8-6S peptide for HLA-B*52:01 molecules was much weaker than that of the RI8-6T peptide but that the affinity of the RI8-6V peptide was identical to that of the RI8-6T. These findings together suggest that position 6 is a critical site for both the peptide binding to HLA-B*52:01 and TCR recognition, though this position is not an anchor residue (47). The affinity of the RI8-6S peptide was much weaker than that of the RI8-6T or the RI8-6V one, suggesting that the RI8-6S epitope peptide could not be presented in the cells infected with the Gag280S mutant virus. On the other hand, the Gag280A mutation weakly affected the peptide binding affinity, suggesting this mutation may have affected TCR recognition. RI8-6T-specific T cells failed to recognize the RI8-6V epitope, whereas RI8-6V-specific T cells were elicited in the individuals infected with Gag280V mutant virus. These findings suggest that there were 2 T cell repertoires for RI8 in HLA-B*52:01 ϩ individuals, one having high-affinity TCRs for RI8-6T and the other for RI8-6V.
HLA-B*52:01 is protective allele in the subtype B and C infections (31,33,40), whereas Gag RI8 is one of protective T cell epitopes restricted by HLA-B*52:01 (6). RI8-6T-specific T cells failed to recognize the cells infected with Gag280S/A mutant viruses, and T cells specific for RI8-6A/6S mutant epitopes were not elicited in the individuals infected with these viruses (Fig. 1H). These findings suggest that the accumulation of Gag280S/A mutations would critically affect suppression of HIV-1 replication by these specific T cells in vivo. Indeed, HLA-B*52:01 ϩ Japanese individuals infected with Gag280S/A mutant viruses had significantly lower CD4 counts than those infected with the wild-type virus. In contrast, RI8-6V-specific T cells, which were frequently elicited in Gag280V virus-infected HLA-B*52:01 ϩ individuals, had a strong ability to suppress replication of Gag280V mutant viruses in vitro. Indeed, our analysis showed that no significant difference in CD4 count was found between individuals infected with Gag280T virus and those with the Gag280V one, suggesting that the Gag280V mutation did not affect the control of HIV-1. Since the accumulation of Gag280S/A mutations was found in only 20% of the HLA-B*52:01 ϩ individuals, GagRI8 is still a protective T cell epitope in them.
Three of 4 HLA-B*52:01-restricted epitopes are conserved among circulating HIV-1 subtype B viruses (6), and T cells specific for these epitopes have a strong ability to suppress HIV-1 replication in vivo (6,44). These epitopes may be targets for prophylactic T cell vaccines and a cure for HIV-1. The wild-type sequence of RI8 is found in only 60% of Japanese individuals infected with the subtype B virus, suggesting that this epitope may not be useful for a T cell vaccine and AIDS cure. However, the Gag280V mutant virus could elicit RI8-6V mutant virus-specific T cells in individuals infected with this mutant virus, and these T cells could suppress replication of the mutant virus. Since approximately 80% of circulating viruses have Gag280T/V, chimeric antigens (Ags) containing both RI8-6T and RI8-6V epitopes could be useful for a vaccine and cure of AIDS. Thus, the present study showed that a T cell epitope including an escape mutation could be target for a T cell vaccine and AIDS cure. However, since it is still unknown whether other escape mutant epitopes also could elicit specific T cells that could effectively suppress HIV-1 mutant viruses, further studies on T cell recognition for escape HIV-1 mutants are required for generation of chimeric vaccine antigens that should contribute to the development of a prophylactic T cell vaccine and AIDS cure.
In the present study, we demonstrated a mechanism for the accumulation of different Gag280 mutations in subtype B-infected Japanese and for coevolution of HIV-1 with HIV-1-specific T cells as well as the important role of mutant specific T cells in the suppression of HIV-1 replication in vivo (Fig. 6). The results of the present study strongly impact our understanding of the role of mutant epitope-specific T cells in the control of HIV-1 and imply their potential usefulness for a prophylactic AIDS vaccine and AIDS cure.  (47,48). These cells were maintained in RPMI 1640 medium (Invitrogen) containing 5% fetal calf serum (FCS; R5) and 0.15 mg/ml of hygromycin B or 0.2 mg/ml of neomycin.
Sequence of autologous virus. Determination of the epitope sequence for RI8 was performed as previously described (46). The RI8 sequence data from 390 chronically HIV-1 subtype B-infected treatment-naive Japanese individuals were analyzed after excluding individuals having a mixture of amino acid sequences at Gag280 from previously analyzed ones (46) and adding new data from 16 individuals.
IFN-␥ ELISPOT assay. Ex vivo gamma interferon (IFN-␥) ELISPOT assays were performed as previously described (6,49). The number of spots for a T cell response to each peptide was finally calculated by subtracting the number of spots in wells without peptides. The mean ϩ 3 standard deviations (SD) of the spot number of samples from 13 HIV-1 naive individuals for the peptides was 162 spots/10 6 CD8 ϩ T cells (6,49). Therefore, we defined Ͼ200 spots/10 6 CD8 ϩ T cells as positive responses.
Generation of epitope-specific T cell clones or lines. PBMCs were stained with PE-or APCconjugated tetramers, FITC-conjugated anti-CD3 (Dako, Glostrup, Denmark), Pacific blue-conjugated anti-CD8 MAb (BD Biosciences), and 7-AAD (BD Pharmingen), after which CD3 ϩ CD8 ϩ 7AAD Ϫ tetramer ϩ T cells were sorted in U-bottomed 96-well microtiter plates (1 cell/well for T cell clones and 100 to 500 cells/well for T cell lines) by using a FACSAria (BD Biosciences). The sorted cells were stimulated with the corresponding epitope peptide and cultured as previously described (51). After 2 to 3 weeks in culture, epitope-specific CD8 ϩ T cells were used in functional assays after their purity had been confirmed by flow cytometry analysis using tetramers.
Intracellular cytokine staining (ICS) assay. 721.221 cells prepulsed with HIV-1 epitope peptide or 721.221 cells infected with HIV-1 were cocultured with T cell clones or lines in a 96-well plate for 2 h at 37°C. Brefeldin A (10 g/ml) was then added, and the cells were incubated further for 4 h at 37°C. The cells were then fixed with 4% paraformaldehyde and incubated in permeabilization buffer (0.1% saponin-10% fetal bovine serum [FBS]-phosphate-buffered saline [PBS]), after which they were stained with APC-conjugated anti-CD8 MAb (Dako, Denmark) followed by FITC-conjugated anti-IFN-␥ MAb (BD Biosciences). The percentage of IFN-␥-producing cells among the CD8 ϩ T cell population was determined by use of the FACSCanto II.
HLA stabilization assay. The affinity of peptide binding to HLA-B*52:01 was examined by using RMA-S-B*52:01 cells as previously described (52,53). Briefly, these RMA-S transfectant cells were cultured at 26°C for 16 h, then pulsed with peptides at 26°C for 1 h, and subsequently incubated at 37°C for 3 h. Staining of cell surface HLA molecules was performed by using anti-HLA class I ␣3 domain MAb TP25.99 (54) and FITC-conjugated sheep anti-mouse IgG (Jackson ImmunoResearch). The fluorescence intensity was measured with the FACSCanto II. HIV-1 replication suppression assay. The ability of epitope-specific CD8 ϩ T cells to suppress HIV-1 replication was measured as described previously (24,55). CD4 ϩ T cells isolated from HLA-matched healthy donor PBMCs were infected with HIV-1 virus and then cocultured with epitope-specific T cells at effector-to-target cell (E:T) ratios of 1:1, 0.1:1, and 0:1. When RI8-6V-specific bulk T cells were used as effector T cells (Fig. 4C), the number of effector T cells was calculated as a total number of T cells ϫ percent RI8-6V tetramer ϩ T cells. On day 5 postinfection, the concentration of p24 antigen in the culture supernatant was measured by using an enzyme-linked immunosorbent assay (ELISA) kit (HIV-1 p24 Ag ELISA kit; ZeptoMetrix). The percent suppression was calculated as follows: (concentration of p24 without CTLs -concentration of p24 with CTLs)/concentration of p24 without CTLs ϫ 100.
Statistical analysis. The frequency of the mutation between HLA ϩ and HLA Ϫ individuals was statistically analyzed by using Fisher's exact test. Groups were compared by performing the unpaired t test or two-tailed Mann-Whitney U test. P values of Ͻ0.05 were considered significant