Therapeutic vaccine-mediated Gag-specific CD8+ T-cell induction under anti-retroviral therapy augments anti-virus efficacy of CD8+ cells in simian immunodeficiency virus-infected macaques

Anti-retroviral therapy (ART) can inhibit HIV proliferation but not achieve virus eradication from HIV-infected individuals. Under ART-based HIV control, virus-specific CD8+ T-cell responses are often reduced. Here, we investigated the impact of therapeutic vaccination inducing virus-specific CD8+ T-cell responses under ART on viral control in a macaque AIDS model. Twelve rhesus macaques received ART from week 12 to 32 after simian immunodeficiency virus (SIV) infection. Six of them were vaccinated with Sendai virus vectors expressing SIV Gag and Vif at weeks 26 and 32, and Gag/Vif-specific CD8+ T-cell responses were enhanced and became predominant. All macaques controlled viremia during ART but showed viremia rebound after ART cessation. Analysis of in vitro CD8+ cell ability to suppress replication of autologous lymphocytes-derived SIVs found augmentation of anti-SIV efficacy of CD8+ cells after vaccination. In the vaccinated animals, the anti-SIV efficacy of CD8+ cells at week 34 was correlated positively with Gag-specific CD8+ T-cell frequencies and inversely with rebound viral loads at week 34. These results indicate that Gag-specific CD8+ T-cell induction by therapeutic vaccination can augment anti-virus efficacy of CD8+ cells, which may be insufficient for functional cure but contribute to more stable viral control under ART.

Scientific RepoRtS | (2020) 10:11394 | https://doi.org/10.1038/s41598-020-68267-w www.nature.com/scientificreports/ may contribute to more complete shutdown of viral replication under ART, although therapeutic vaccine trials under ART have failed to achieve functional HIV cure, sustained viral control after ART cessation [17][18][19][20] . Optimization of vaccine immunogens is important for induction of effective CD8 + T-cell responses. Cumulative studies have indicated strong anti-HIV efficacy of Gag-specific CD8 + T cells 14,[21][22][23][24][25][26] . Furthermore, Vif-specific CD8 + T cells have recently been indicated to exert strong suppressive pressure against HIV replication 27,28 . However, the impact of therapeutic vaccination inducing these CD8 + T cells under ART have not fully been determined. We have previously developed Sendai virus (SeV) vectors expressing SIV/HIV antigens and have shown the potential of these vectors to induce viral antigen-specific T-cell responses in macaques and humans [29][30][31] . In the present study, SIVmac239-infected rhesus macaques were immunized with SeV vectors expressing SIV-mac239 Gag and Vif antigens under ART. This immunization enhanced Gag/Vif-specific CD8 + T-cell responses. Our analysis indicated association of Gag-specific CD8 + T-cell responses with anti-SIV efficacy of CD8 + cells post-vaccination.

Results
Analysis of plasma viral loads. Twelve Burmese rhesus macaques consisting of six animals sharing the MHC-I haplotype 90-010-Ie (E), four sharing 89-075-Iw (W) and two sharing 91-010-Is (S) were used in the present study as shown in Table 1. These macaques were intravenously inoculated with SIVmac239 and received ART from week 12 to 32 post-infection. All the macaques showed persistent viremia after the SIV infection and reduced plasma viral loads after ART initiation at week 12 (Fig. 1). These macaques were divided into two groups, Group N (n = 6) receiving no vaccination and Group V (n = 6) receiving vaccination at weeks 26 and 32 post-infection. The Group V macaques consisting of three E-positive, two W-positive and one S-positive were intranasally immunized with Gag-and Vif-expressing SeV vectors at weeks 26 and 32 post-infection. Two of the Group N (NE3 and NW5) and all the six Group V animals were intravenously administered with polyclonal anti-SIV immunoglobulin G (anti-SIV IgG). After ART cessation at week 32, all macaques showed reappearance of plasma viremia. No clear difference was observed in viral loads post-ART between two anti-SIV IgG-treated and four untreated macaques in Group N. While two Group V macaques VE2 and VW4 exhibited relatively lower viral loads post-ART, no significant difference in viral loads post-ART was observed between Groups N and V.
Analysis of antigen-specific CD8 + T-cell responses. We examined CD8 + T-cell responses specific for SIV individual antigens in these macaques before ART initiation, during ART, and after ART cessation by detection of antigen-specific interferon-γ (IFN-γ) induction (Fig. 2). Before the ART initiation at week 12, macaques possessing the MHC-I haplotype E showed predominant induction of Nef-, Tat/Rev-and Env-specific CD8 + T-cell responses, whereas those possessing the haplotypes W/S predominantly induced Gag/Vif-specific CD8 + T-cell responses. At week 26 during ART, these antigen-specific CD8 + T-cell responses were reduced as expected. All the vaccinated Group V macaques showed induction and/or enhancement of Gag/Vif-specific CD8 + T-cell responses at week 27 post-infection, one week after the first SeV-Gag/Vif vaccination. After the second SeV-Gag/ Vif vaccination and the ART cessation at week 32, SIV antigen-specific CD8 + T-cell responses were enhanced. Comparison revealed significantly higher Gag-and Vif-specific CD8 + T-cell responses in Group V than in Group N at weeks 27 and 34, whereas no significant difference was observed before vaccination (at week 26) (Fig. 3a,b). There was no significant difference in CD8 + T-cell responses targeting Nef that was not included in the vaccine antigens at week 26, 27 or 34 between Groups N and V (Fig. 3c).
The unvaccinated Group N animals showed similar patterns of immunodominance pre-ART and post-ART; i.e., Nef-and Env-specific CD8 + T-cell responses were predominant post-ART in the E-positive unvaccinated  Analysis of in vitro anti-SIV efficacy of CD8 + cells. We performed in vitro viral suppression assay to examine the ability of CD8 + cells pre-ART, during ART and post-ART to suppress replication of SIVs derived from autologous peripheral blood mononuclear cells (PBMCs) pre-ART (referred to as pre-ART SIVs). The ability of CD8 + cells to suppress replication of SIV derived from pre-ART PBMCs (anti-pre-ART SIV CD8 + cell efficacy) was examined by using pre-ART CD8cells obtained at week 10 post-infection from individual animals as the target cells. CD8 + cells derived from PBMCs at weeks 10, 27 (1 week after the first vaccination in Group V) and 34 were prepared as effector cells pre-ART, under ART and post-ART, respectively, and reductions in SIV production from the target cells by coculture with autologous CD8 + cells pre-ART, under ART and post-ART were assessed, respectively (Fig. 4a). The ability of CD8 + cells under ART to suppress replication of pre-ART www.nature.com/scientificreports/ SIVs was marginal and/or reduced compared to CD8 + cells pre-ART in Group N but enhanced in all the Group V except for one animal (VE1). The ratio of anti-pre-ART SIV efficacy of CD8 + cells under ART to pre-ART was significantly higher in Group V than Group N (P = 0.0173 by Mann-Whitney U test) (Fig. 4b), implying that the therapeutic SeV-Gag/Vif vaccination can augment in vitro anti-SIV efficacy of CD8 + cells. The ability of CD8 + cells post-ART to suppress replication of pre-ART SIVs was much higher than that under ART in all the Group N animals, and there was no significant difference in the ratio of anti-pre-ART SIV efficacy of CD8 + cells post-ART to pre-ART between Groups N and V. Next, correlation analyses were performed between in vitro anti-pre-ART SIV efficacy of CD8 + cells and antigen-specific CD8 + T-cell responses in the Group V animals. Anti-SIV efficacy of CD8 + cells pre-ART (at week 10) showed no correlation with Gag-specific nor Vif-specific CD8 + T-cell responses at week 10 post-infection ( Then, the ability of CD8 + cells to suppress replication of SIVs derived from autologous PBMCs post-ART (referred to as post-ART SIV) was examined by using post-ART CD8cells obtained at week 34 post-infection from individual Group V animals as the target cells (Fig. 6a). The ability of CD8 + cells under ART to suppress replication of post-ART SIVs (anti-post-ART SIV efficacy of CD8 + cells under ART) showed no significant correlation with Gag-specific nor Vif-specific CD8 + T-cell responses at week 27 post-infection (Fig. 6b). Instead, anti-post-ART SIV efficacy of CD8 + cells post-ART exhibited a significant correlation with Gag-specific CD8 + T-cell responses (R = 0.8857, P = 0.0333) but not with Vif-specific CD8 + T-cell responses at week 34 post-infection (2 weeks after ART cessation) (Fig. 6b).
There were some discrepancies between anti-pre-ART SIV and anti-post-ART SIV efficacy of CD8 + cells as expected. Analysis found some difference in viral gag and vif sequences between individual pre-ART and post-ART SIVs obtained from the vaccinated macaque PBMCs for in vitro viral suppression assay (Fig. 7). Additional non-synonymous mutations in post-ART SIV gag were detected in macaques VW5 (GagM457V) and VS6 (GagK28R). Especially in the latter macaque VS6, Gag 20-34 peptide-specific CD8 + T-cell responses were detected, and CD8 + cells under ART (at week 27) showed much lower anti-post-ART SIV efficacy compared to anti-pre-ART SIV efficacy, suggesting that the GagK28R mutation may be responsible for the difference between anti-pre-ART SIV and anti-post-ART SIV efficacy of CD8 + cells.

Correlation analyses between in vitro anti-SIV efficacy of CD8 + cells and plasma viral load in
the vaccinated macaques. Finally, we examined correlation of in vitro anti-SIV CD8 + cell efficacy with plasma viral loads at weeks 34 and 36 (2 and 4 weeks after ART cessation) in Group V animals. There was no significant correlation between these plasma viral loads after ART and anti-pre-ART SIV efficacy of CD8 + cells under ART (Fig. 8). In contrast, anti-post-ART SIV efficacy of CD8 + cells post-ART was inversely correlated with plasma viral loads at weeks 34 (R = − 0.8857, P = 0.0333 by Spearman test) and 36 (R = − 0.9429, P = 0.0167), respectively (Fig. 8). Anti-post-ART SIV efficacy of CD8 + cells under ART exhibited a non-significant trend of association with lower viral load at week 34 (R = − 0.7714, P = 0.1028) but a significant inverse correlation with plasma viral load at week 36 (R = − 0.8857, P = 0.0333) (Fig. 8).

Discussion
Prophylactic vaccines which are administered to naive (unvaccinated) individuals can dominantly induce vaccine antigen-specific CD8 + T-cell responses. In contrast, therapeutic vaccines are supposed to be administered to individuals already primed by HIV infection. Some HIV-infected individuals have MHC/HLA (human leukocyte antigen) class I genotypes associated with predominant Gag/Vif-specific CD8 + T-cell response but others have those with predominant CD8 + T-cell responses targeting non-Gag/Vif antigens. It is important to define the patterns of CD8 + T-cell responses after therapeutic vaccination in the latter cases as well as the former. In the present study, we examined the impact of therapeutic SeV-Gag/Vif vaccination under ART on SIV antigen-specific CD8 + T-cell responses in SIVmac239-infected rhesus macaques. In the three Group V macaques possessing MHC-I haplotype W/S associated with predominant Gag/Vif-specific CD8 + T-cell responses, Gag/Vif-specific CD8 + T-cell responses were enhanced after SeV-Gag/Vif vaccination and remained predominant after ART cessation (at week 34). In the remaining three Group V macaques possessing MHC-I haplotype E associated with predominant non-Gag/Vif antigen-specific CD8 + T-cell responses, Gag/Vif-specific CD8 + T-cell responses were enhanced   Early diagnosis of HIV infection is not easy, and most HIV-infected individuals start ART after the viremia setpoint. We then began ART in macaques at week 12 post-infection, after the viremia setpoint with induction of SIV-specific T-cell responses, in the present study. The first therapeutic vaccination was performed 3 months after the ART initiation, when SIV-specific CD8 + T-cell responses were largely reduced. This protocol is considered to be good for the analysis of changes in immunodominance by therapeutic vaccination described above. However, CD8 + T-cell escape SIV variants are expected to appear during the 12 weeks before the ART initiation. We thus examined efficacy of CD8 + cells to suppress replication of autologous PBMC-derived SIVs as described in the  32,33 . However, our experiments did not exhibit significant effect on viremia rebound after ART. One possible important difference between these previous studies and the present study may be the time point of ART initiation. ART was initiated in a week post-infection in the previous studies but at week 12 in the present study. It is speculated that CD8 + T-cell escape SIV variants appear during the 12 weeks post-infection 14,34 , which can disturb the efficacy of wild-type antigen-specific CD8 + T-cell responses induced by vaccination against viral replication. Indeed, several gag/vif mutations were detected before the ART initiation at week 12. Therefore, induction of CD8 + T-cell responses targeting Gag, the most conserved HIV/SIV antigen, by therapeutic vaccination could have greater impact on viral control as indicated by our results. In addition, SIV-specific T-cell responses induced before ART initiation post-infection can affect dominant/subdominant T-cell responses induced by therapeutic vaccination as described above. These results would provide insights into the development of therapeutic vaccination under ART initiated after the viremia setpoint. www.nature.com/scientificreports/ In summary, the present study showed that therapeutic SeV-Gag/Vif vaccination under ART initiated after the viremia setpoint can enhance subdominant as well as predominant Gag/Vif-specific CD8 + T-cell responses in SIV-infected rhesus macaques. Our results indicate that induction of Gag-specific CD8 + T-cell responses by therapeutic vaccination can augment anti-virus efficacy of CD8 + cells, which may be insufficient for functional cure but result in more stable viral control under ART.

Animal experiments. Animal experiments were carried out in the Institute for Frontier Life and Medical
Sciences, Kyoto University after approval by the Committee on the Ethics of Animal Experiments in Kyoto University (permission number: R10-05 and R11-05) under the guidelines for animal experiments in accordance with the Guidelines for Proper Conduct of Animal Experiments established by Science Council of Japan (https ://www.scj.go.jp/ja/info/kohyo /pdf/kohyo -20-k16-2e.pdf). The experiments were in accordance with Weatherall report for the use of non-human primates in research (https ://royal socie ty.org/topic s-polic y/publi catio ns/2006/ weath erall -repor t/). Blood collection, virus inoculation, vaccination, and antibody administration were performed under ketamine anesthesia.
In vitro viral suppression assay. We examined the ability of CD8 + cells to suppress replication of SIVs derived from autologous CD8 − cells. PBMCs were separated into CD8 + and CD8 − cells by using Macs CD8 MicroBeads (Miltenyi Biotec). CD8 − cells selected from PBMCs at weeks 10 (before ART initiation) and 34 (after ART cessation) were used for targets pre-ART and post-ART, respectively. These target cells (10 5   . Non-synonymous gag and vif mutations in the vaccinated macaques. Viral gag and vif cDNAs were amplified from pre-ART and post-ART CD8cells-derived SIVs obtained at in vitro viral suppression assay and subjected to sequence analyses. Viral gag and vif cDNAs were also amplified from plasma-derived RNAs at weeks 12 and 34 and subjected to sequence analyses. + indicates that the non-synonymous mutation resulting in amino acid substitution was detected predominantly or equivalently with the wild-type sequence. (+) indicates that the wild-type sequence was predominant while the non-synonymous mutation was subdominantly detected.
Scientific RepoRtS | (2020) 10:11394 | https://doi.org/10.1038/s41598-020-68267-w www.nature.com/scientificreports/ immunosorbent assay (ELISA) using SIV p27 Antigen Capture Assay (Advanced Bioscience Laboratories). If the non-CD8 culture (the culture without CD8 + cells) had lower than 100 ng/ml of p27, the culture supernatants on day 8 (VE2 pre-ART, VE3 post-ART, VW4 post-ART, and VS6 pre-ART) or 10 (VE2 post-ART) were used for the analysis. Because SIV replication kinetics are considered to be different among the cultures of target cells derived from individual SIV-infected macaque PBMCs, it is difficult to exactly adjust effector-target ratios in all the in vitro viral suppression assay by using limited available samples. Then, as described above, the viral suppression was analyzed under the condition that the individual negative control target cultures without CD8 + cells were adjusted to produce more than 100 ng/ml of p27 not on day 4 but on day 6 or later. The ratios of p27 concentrations in the supernatants from the coculture to those without CD8 + cells are shown as fold of reduction (in vitro anti-SIV efficacy) 42 .
Analysis of viral gag and vif sequences. Viral RNAs were extracted using the high pure viral RNA kit (Roche Diagnostics) from SIVs pre-ART and post-ART, the culture supernatants of CD8cells derived from PBMCs at weeks 10 and 34 post-infection in individual animals. Viral RNAs were also extracted from plasma samples at weeks 12 and 34. Viral gag and vif cDNAs were amplified from these RNAs by reverse transcription and nested PCR using the PrimeScript one-step RT-PCR kit version 2 (Takara) and KOD-Plus version 2 (Toyobo), and subjected to direct sequencing by using dye terminator chemistry and an automated DNA sequencer (Applied Biosystems) as described before 37 .
Statistical analysis. Statistical analysis was performed using Prism software version 6.0d (GraphPad Software Inc.) with significance levels set at a P value of less than 0.05. Specific CD8 + T-cell frequencies, in vitro anti-SIV efficacy and plasma viral loads were log transformed. Comparison was performed by Mann-Whitney U test. Correlation analysis was performed by Spearman test.