Effective Suppression of HIV-1 Replication by Cytotoxic T Lymphocytes Specific for Pol Epitopes in Conserved Mosaic Vaccine Immunogens

It is likely necessary for an effective AIDS vaccine to elicit CD8+ T cells with the ability to recognize circulating HIV-1 and suppress its replication. We recently developed novel bivalent mosaic T-cell vaccine immunogens composed of conserved regions of the Gag and Pol proteins matched to at least 80% globally circulating HIV-1 isolates. Nevertheless, it remains to be proven if vaccination with these immunogens can elicit T cells with the ability to suppress HIV-1 replication. It is well known that Gag-specific T cells can suppress HIV-1 replication more effectively than T cells specific for epitopes in other proteins. We recently identified 5 protective Gag epitopes in the vaccine immunogens. In this study, we identified T cells specific for 6 Pol epitopes present in the immunogens with strong abilities to suppress HIV-1 in vivo and in vitro. This study further encourages clinical testing of the conserved mosaic T-cell vaccine in HIV-1 prevention and cure.

H IV-1-specific cytotoxic T lymphocytes (CTLs) play an important role in suppression of HIV-1 replication (1)(2)(3)(4)(5)(6)(7)(8). Gag-specific CD8 ϩ T-cell responses were the most dominant among the nine proteins of HIV-1 in HIV-1-infected individuals (9). Studies of HIV-1-specific CTLs in chronic HIV-1 infection indeed showed associations of Gag-specific T-cell responses with good clinical outcomes (8,(10)(11)(12)(13), indicating that Gag-specific T cells have a stronger ability to suppress HIV-1 replication than responses specific for other proteins. This is explained by relative abundance of Gag protein and reduced viral fitness by escape mutations from Gag-specific CTLs (14)(15)(16)(17). On the other hand, a role for Pol-specific CD8 ϩ T cells in HIV-1 infection has been only partially analyzed. T-cell responses to Gag and Pol were frequently detected in natural HIV-1 infections (10,12,(18)(19)(20). A study in an HIV-1 clade C-infected African cohort showed that Pol-specific T-cell responses were not associated with a significant reduction in plasma viral load (pVL), whereas Gag-specific responses significantly correlated with lower pVL (10). In contrast, a study of clade B-infected treatment-naive Japanese individuals demonstrated that in addition to Gag-specific T cells, Pol-specific T cells also had a strong ability to suppress HIV-1 replication in vivo (20)(21)(22).
Although great efforts in T-cell vaccine development have been invested, no clinical trial has shown a definitive effect regarding prevention of HIV-1 infection (23,24). This is because the vaccine-elicited T cells may fail to recognize escape mutant viruses and/or the vaccines may fail to elicit strong T-cell immunity and suppress HIV-1 replication. To minimize escape and target HIV-1 "where it hurts," vaccines using conserved regions of HIV-1 proteins as immunogens have been proposed (25)(26)(27)(28). Ondondo et al. recently designed a second-generation conserved-region T-cell mosaic vaccine, tHIVconsvX, which consists of 2 Gag and 4 Pol protein regions functionally conserved across all M group viruses with high coverage of known protective epitopes and employs a bioinformatically designed bivalent mosaic to maximize the match of the vaccine potential T-cell epitopes to the global circulating HIV-1 isolates (29). Initial study of T cells recognizing the tHIVconsvX immunogens showed a significant correlation of both the total magnitude and breadth of the tHIVconsvX immunogen-specific T-cell responses to lower pVLs and higher CD4 ϩ T-cell counts (CD4 counts) in 120 treatment-naive HIV-1 clade B-infected patients in Japan (29). A following study demonstrated that CD8 ϩ T cells specific for five Gag epitopes in tHIVconsvX immunogens contribute to suppression of HIV-1 replication in vivo (30). However, it remains unknown whether CD8 ϩ T cells specific for the Pol region in the immunogen are equally effective.
In the present study, we clarified the role of CD8 ϩ T cells specific for the Pol regions in the tHIVconsvX immunogens in 200 HIV-1-infected Japanese individuals. We determined the fine specificities and HLA restriction of CD8 ϩ T cells specific for the Pol regions in the immunogens and further analyzed the correlation of these Pol epitopespecific T cells to clinical outcome as well as assessed their HIV-1 inhibition capacity in vitro. These results will inform and encourage clinical testing of the second-generation conserved-mosaic T-cell vaccines. and CD4 count in each pool between responders and nonresponders. The responders to P6, P8, and P9 peptides showed both significantly lower pVL and higher CD4 count than nonresponders (Fig. 3), suggesting a protective role against HIV-1 infection in vivo.
Mapping of the CD8 ؉ T-cell specificity to optimal Pol epitopes in the tHIVconsvX immunogens. We sought to map Pol epitopes included in P6, P8, and P9. We selected, respectively, 20, 16, and 17 individuals based on sufficient peripheral blood mononuclear cells (PBMCs) available for the determination of optimal epitopes. We found T-cell responses to 8 peptide pairs and one common single peptide in P6, 5 peptide pairs in P8, and 4 peptide pairs in P9 in at least one individual (Fig. 4A). These 15-mer peptides contained sequences of previously reported epitopes: 13 epitopes in P6, 4 epitopes in P8, and 3 epitopes in P9 (Fig. 4B). Upon inspection of the subjects' HLA molecules, most of the responders were found to have HLA alleles previously reported to restrict these optimal epitopes. However, all or some responders to 15-mer peptide pairs C256/257, C258/259, C300/301, C328/329, C346/347, C360/361, and C362/363 did not have the matching HLA alleles (Fig. 5A), suggesting that their CD8 ϩ T cells may recognize novel, previously unreported epitopes.
Pol epitope-specific CD8 ؉ T cells have strong abilities to suppress HIV-1 replication in vivo. As shown above, we mapped 20 reported and 5 novel HIV-1 Pol epitopes in HIV-1-infected Japanese individuals who responded to tHIVconsvX immunogenderived Pol peptides. To further investigate whether T cells specific for these epitopes have ability to inhibit HIV-1 replication in vivo, we selected 221 individuals whose PBMCs were available for this analysis and analyzed their T-cell responses to these epitope peptides in the IFN-␥ ELISPOT assay. Since a T-cell response to IK9/HLA-A11 was detected in only one patient (Table 1), we excluded this epitope from further statistical analysis. We analyzed differences in pVL or CD4 count between responders to each epitope and nonresponders. :06) had significantly lower pVLs and/or higher CD4 counts than nonresponders, suggesting that T cells specific for these 10 epitopes have the ability to suppress HIV-1 replication in vivo (Table 1).
We further analyzed the association of responses to these 10 epitopes with pVL and CD4 count in individuals having the epitopes' restricting HLA molecules. We found that  (Table 2). We confirmed a previously reported inhibition of HIV-1 in vitro and in vivo by T cells specific for the GI8/HLA-B*40:02 epitope (20,32). Thus, the present study showed that T cells specific for the 5 epitopes TI8, LI9, ER10, IL9, and GA9 could efficiently suppress HIV-1 replication in vivo.
Cross-recognition of IL9 and GA9 variants by specific T cells. We analyzed the sequences of the IL9-, ER10-, and GA9 Pol epitopes in our cohort of Japanese patients and found variations in IL9 and GA9 (Fig. 7A), while the sequence of ER10 was conserved in more than 80% of the individuals. We therefore assessed crossrecognition of variant epitopes IL9-4D and GA9-5I by IL9-and GA9-specific CD8 ϩ T cells, respectively. We found cross-recognition in both cases (Fig. 7B). Thus, T cells specific for IL9, GA9, and ER10 can recognize more than 80% of circulating HIV-1 in Japan. responses to the 6 epitopes and clinical parameters in the Japanese individuals. The breadth of the responses to the epitopes was correlated inversely with pVL and positively with CD4 count (Fig. 8A), suggesting that these Pol epitope-specific T cells play an important role in suppression of HIV-1 replication in this cohort. We recently demonstrated that the T cells specific for 5 Gag epitopes HR10/HLA-A*33:03, TL8/HLA-B*40:02, RI8/HLA-B*52:01, WV8/HLA-B*52:01, and AA9/HLA-A*02:06 in the tHIVconsvX immunogens had a strong ability to control HIV-1 in vivo (30). In order to predict an effect of T cells specific for these all protective epitopes on control of HIV-1, we analyzed correlations between T-cell responses to these 11 epitopes and clinical parameters. The breadth of the responses to the 11 epitopes showed highly significant correlations with both lower pVLs (P Ͻ 1 ϫ 10 Ϫ4 ; r ϭ Ϫ0.4909) and higher CD4 counts (P Ͻ 1 ϫ 10 Ϫ4 ; r ϭ 0.4542) in these Japanese individuals (Fig. 8B). Responders to both Pol and Gag epitopes had significantly lower pVLs and higher CD4 counts than responders to only Gag epitopes, though they had only a trend of lower pVLs and higher CD4 counts than with responders to only Pol epitopes (Fig. 8C). These results suggest additional protective effects of Gag-specific T-cell responses with Pol-specific ones on suppression of HIV-1 replication. Thus, the results indicate that all the strongly  protective epitopes we have identified in the tHIVconsvX immunogens play an important role in immunity against HIV-1 clade B infection in Japan.

DISCUSSION
Previous cross-sectional analysis of T-cell responses to 18-mer overlapping HIV-1 peptides spanning the entire HIV-1 proteome in approximately 600 clade C-infected Africans demonstrated that a breadth of T-cell responses to Gag peptides was inversely associated with pVL, but those to peptides in other proteins were not (10). In addition, analyses of T-cell responses in clade B-infected individuals at a small population level showed inverse correlation of Gag-specific T-cell responses with pVL but not to any other HIV-1 protein (12,35). Thus, it is well documented that Gag-specific T cells play a critical role in controlling HIV-1 in infected individuals. On the other hand, there is growing evidence that Pol-specific CD8 ϩ T cells have the ability to suppress HIV-1 replication. A previous study using a conserved immunogen showed that CD8 ϩ T cells specific for Pol peptides have a stronger ability to suppress HIV-1 replication in vitro than those specific for Gag, Env, and Vif peptides in healthy volunteers vaccinated with the first-generation conserved immunogen HIVconsv employing alternating clade consensus sequences (26). Moreover, we showed that both T-cell responses to Gag and Pol were significantly correlated with lower pVLs and higher CD4 counts in antiretroviral therapy (ART)-free HIV-1 clade B-infected Japanese individuals (20). In the present study, we demonstrated that six Pol epitope-specific CD8 ϩ T cells had strong associations with both lower pVLs and higher CD4 counts in treatment-naive HIV-1-infected Japanese individuals carrying HLA alleles restricting each epitope, suggesting that the 6 Pol epitope-specific T cells contribute to suppression of HIV-1 in vivo. In addition, we confirmed effective suppression of HIV-1 replication by T cells specific for the Pol ER10, IL9, and GA9 epitopes in vitro, while previous studies demonstrated that CTLs specific for TI8, LI9, and GI8 had strong abilities to suppress HIV-1 in vitro (32)(33)(34). Our recent study showed that T cells specific for five Gag epitopes in this vaccine immunogens effectively suppress HIV-1 replication (30). These findings together suggested that CD8 ϩ T cells specific for the six Pol and the five Gag epitopes can inhibit HIV-1 replication. It is our working hypothesis that when studying correlates of T-cell protection, the T-cell specificity is critically important. Any attempts to identify T-cell correlates using the full-length proteins is likely to yield a blurred picture. Much more granularity is required, and a clearer correlation can be achieved associating T-cell protection with responses to the common functionally conserved regions and/or protective epitopes.
The    (41), which is consistent with the present study in Japanese individuals (76%). Although the above-mentioned study did not analyze association of the T-cell responses with clinical outcome due to the low number of patients tested (41), we here clearly establish that the LI9-specific T-cell response was significantly associated with good clinical outcome in the Japanese individuals. In addition, our previous study demonstrated that T-cell clones specific for LI9 effectively suppress HIV-1 replication in vitro (33).
In Japan, the ER10 epitope was conserved among circulating viruses, whereas variations were found within IL9, GA9, LI9, GI8, and TI8. We here demonstrate that T-cell lines specific for IL-9 and GA9 cross-recognized the variant peptides IL9-4D and GA9-5I, respectively. In addition, previous studies showed that T-cell clones specific for GI8 and LI9 evenly recognized their mutant peptides (32,41), whereas TI8-specific HLA-B*52: 01-restricted T cells failed to recognize peptides mutated at the C terminus (34). We  show here that the TI8-specific T cells had the ability to suppress HIV-1 replication in vivo ( Table 2), suggesting that the ability of the T cells to suppress replication of wild-type virus may contribute to the suppression of HIV-1 replication at a population level.
In summary, we demonstrated strong abilities of the T cells specific for six conserved protective Pol epitopes present in the tHIVconsvX immunogens to control HIV-1 in vivo. Together with the five Gag epitopes identified in our previous study, broadly specific T-cell responses to 11 protective epitopes (5 in Gag and 6 in  (TI8, IL9, LI9, GA9, GI8, and ER10) epitope peptides were analyzed by using an IFN-␥ ELISPOT assay. The correlation of T-cell responses to protective 6 Pol epitopes (A) or 11 protective Gag and the Pol epitopes (B) with pVL and CD4 count was statistically analyzed using Spearman rank test. The pVL and CD4 count in responders to both Pol and Gag epitopes, to only Gag epitopes, and to only Pol epitopes as well as nonresponders were statistically analyzed by Mann-Whitney test (C). The value in each graph represents the median pVL and CD4 count. The averages of numbers of epitopes recognized by T cells are 1.17, 1.23, and 2.66 in only Pol, only Gag, and Gag and Pol groups, respectively. The P value between responders to only Pol epitopes and nonresponders was Ͻ0.0001 (pVL and CD4 count), while those between responders to only Gag epitopes and nonresponders was 0.032 (pVL) and 0.055 (CD4 count).
Pol) correlated strongly with low pVLs and high CD4 counts in HIV-1 clade B-infected Japanese individuals. Therefore, these findings indicate that the secondgeneration conserved mosaic tHIVconsvX immunogens contain a number of very useful protective epitopes. If these immunogens can induce high frequencies of these T cells in the right place in the right time with the strong T-cell functions, these vaccines have the potential contribute significantly to HIV-1 prevention and cure. The fact that the tHIVconsvX vaccine regions contain protective epitopes is a very encouraging first step.

MATERIALS AND METHODS
Subjects. All treatment-naive Japanese subjects chronically infected with HIV-1 subtype B were recruited from the National Center for Global Health and Medicine. PBMCs were separated from whole blood. HLA types of the individuals were determined by standard sequence-based genotyping. This study was approved by the ethics committees of Kumamoto University and the National Center for Global Health and Medicine. Informed consent was obtained from all individuals according to the Declaration of Helsinki.
Peptides. We generated seven pools containing pairs of 15-mer Pol peptides overlapped by 11 amino acids covering two mosaic regions in the tHIVconsvX immunogen (29). Ex vivo IFN-␥ ELISPOT assay. An ex vivo IFN-␥ enzyme-linked immunospot (ELISPOT) assay was performed as previously described (20). To standardize the number of spots to spot-forming units (SFU)/10 6 CD8 ϩ T cells, we measured a frequency of CD8 ϩ T cells among PBMCs using flow cytometry. Next 100,000 PBMCs from each individual were plated in each well in the ELISPOT plate that had been precoated with 5 g/ml of anti-IFN-␥ MAb 1-D1K (Mabtech, Stockholm, Sweden) at a concentration of 1 M (1.6 to 1.8 g/ml) of peptides. The plates were then incubated for 16 h at 37°C in 5% CO 2 , and then the cells were stained as previously described in detail (20). We calculated the number of CD8 ϩ T cells plated in each well containing 100,000 PBMCs by using the frequency of CD8 ϩ T cells among PBMCs and determined SFU/10 6 CD8 ϩ T cells in each well (20). The number of spots for each peptide-specific T cell response was finally calculated by subtracting the number of spots in wells without peptides. The mean ϩ 5 standard deviations (SD) of the SFU of samples (n ϭ 3) from 12 HIV-1-naive individuals for the peptide pool was 115 SFU/10 6 CD8 ϩ T cells (30). Therefore, we defined a positive IFN-␥ ELISPOT response as larger than 200 SFU/10 6 CD8 ϩ T cells to exclude false-positive results.
Establishment of T-cell lines specific for GA9, ER10, and IL-9 peptides using HLA/peptide tetramer complexes. To establish T-cell lines specific for the GA9, ER10, and IL9 epitopes, HLA-B*40: 06/GA9, HLA-A*33:03/ER10, and HLA-A*2402/IL9 tetrameric complexes (tetramers) were synthesized as previously described (51). PBMCs of HLA-B*40:06 ϩ KI-1268, HLA-A*33:03 ϩ KI-1427, and HLA-A*2402 ϩ KI-1105 individuals were stained with phycoerythrin (PE)-conjugated specific tetramers at a concentration of 100 nM at 37°C for 30 min. The cells were then washed twice with R10, followed by staining with FITC-conjugated anti-CD8 MAb and 7-aminoactinomycin D (7-AAD) at 4°C for 30 min. The CD8 ϩ T cells specific for the GA9, ER10, and IL9 epitopes were then sorted by FACS Aria (BD Bioscience, CA). The sorted cells were stimulated with 100 nM concentrations of the corresponding epitope peptides and cultured for 12 to 14 days to induce T-cell lines specific for each epitope. To confirm the purities of the specific T cells, the T-cell lines were analyzed by using the specific tetramers.
In vitro virus inhibition assay. The ability of HIV-1-specific CTLs to suppress HIV-1 replication in vitro was examined as previously described (33,52). CD4 ϩ T cells isolated from PBMCs of healthy donors carrying HLA-B*40:06, HLA-A*33:03, or HLA-A*24:02 were infected with HIV-1 NL4-3, and then the infected cells were cocultured with epitope-specific T-cell lines at effector/target (E:T) ratios of 1:1 and 0.1:1. On day 5 postinfection, the concentration of p24 Ag in the culture supernatant was measured by using an enzyme-linked immunosorbent assay.
Bulk sequence of autologous virus. Bulk sequencing of autologous plasma viral RNA from HIV-1-infected individuals was performed as described previously (53).
Statistical analyses. The two-tailed Mann-Whitney test was performed for comparison of two groups. Correlations between magnitudes and breadths of T-cell responses and pVL or CD4 count were statistically analyzed using Spearman rank test. P values of Ͻ0.05 were considered to be statistically significant.

ACKNOWLEDGMENTS
This research was supported by grants-in-aid (15fk0410019h0001, 16fk0410202h0002, and 17fk0410302h0003) for AIDS Research from AMED and by grants-in-aid (26293240, 17K10021) for scientific research from the Ministry of Education, Science, Sports, and Culture, Japan. C.Z. was supported by the China Scholarship Council (CSC) scholarship. T.H. is a Jenner Investigator.
We have no financial conflicts of interest except for T.H., who is a coinventor on the tHIVconsvX filings PCT/US2014/058422 and EP14846993.5.
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.