A Novel Immunogen Selectively Eliciting CD8⁺ T Cells but Not CD4⁺ T Cells Targeting Immunodeficiency Virus Antigens

Induction of effective CD8⁺ T-cell responses is an important HIV vaccine strategy. Several promising vaccine delivery tools have been developed, and immunogen optimization is now crucial for effective T-cell induction. Conventional immunogens have been designed to induce virus-specific CD4⁺ T cells as well as CD8⁺ T cells, but induction of virus-specific CD4⁺ T cells that are preferential targets for HIV infection could enhance acute HIV proliferation. Here, we designed a novel immunogen to induce HIV-specific CD8⁺ T cells without HIV-specific CD4⁺ T-cell induction but with non-HIV antigen-specific CD4⁺ T-cell help. Our analysis in a macaque AIDS model showed that our immunogen can efficiently elicit effective CD8⁺ T but not CD4⁺ T cells targeting viral antigens, resulting in no enhancement of acute viral replication after virus exposure. This immunogen design, also applicable for other currently developed immunogens, could be a promising method for selective induction of effective anti-HIV CD8⁺ T-cell responses.

ABSTRACT

Optimization of immunogen is crucial for induction of effective T-cell responses in the development of a human immunodeficiency virus (HIV) vaccine. Conventional T-cell-based vaccines have been designed to induce virus-specific CD4⁺ T as well as CD8⁺ T cells. However, it has been indicated that induction of HIV-specific CD4⁺ T cells, preferential targets for HIV infection, by vaccination may be detrimental and accelerate viral replication after HIV exposure. In the present study, we present a novel immunogen to selectively induce CD8⁺ T cells but not CD4⁺ T cells targeting viral antigens. The immunogen, CaV11, was constructed by tandem connection of overlapping 11-mer peptides spanning simian immunodeficiency virus (SIV) Gag capsid (CA) and Vif. Prime-boost immunization with DNA and Sendai virus (SeV) vectors expressing CaV11 efficiently induced Gag/Vif-specific CD8⁺ T-cell responses with inefficient Gag/Vif-specific CD4⁺ T-cell induction in rhesus macaques (n = 6). None of the macaques exhibited the enhancement of acute viral replication after an intravenous high-dose SIV challenge, which was observed in those immunized with DNA and SeV expressing the whole Gag protein in our previous study. Set point viral control postinfection was associated with SeV-specific CD4⁺ T-cell responses postimmunization, suggesting contribution of SeV-specific helper responses to effective Gag/Vif-specific CD8⁺ T-cell induction by vaccination. This immunogen design could be a promising method for selective induction of effective anti-HIV CD8⁺ T-cell responses.

IMPORTANCE Induction of effective CD8⁺ T-cell responses is an important HIV vaccine strategy. Several promising vaccine delivery tools have been developed, and immunogen optimization is now crucial for effective T-cell induction. Conventional immunogens have been designed to induce virus-specific CD4⁺ T cells as well as CD8⁺ T cells, but induction of virus-specific CD4⁺ T cells that are preferential targets for HIV infection could enhance acute HIV proliferation. Here, we designed a novel immunogen to induce HIV-specific CD8⁺ T cells without HIV-specific CD4⁺ T-cell induction but with non-HIV antigen-specific CD4⁺ T-cell help. Our analysis in a macaque AIDS model showed that our immunogen can efficiently elicit effective CD8⁺ T but not CD4⁺ T cells targeting viral antigens, resulting in no enhancement of acute viral replication after virus exposure. This immunogen design, also applicable for other currently developed immunogens, could be a promising method for selective induction of effective anti-HIV CD8⁺ T-cell responses.

INTRODUCTION

In the estimation by UNAIDS (1), more than 37 million people are living with human immunodeficiency virus (HIV), and approximately 1.7 million people are newly infected per year worldwide. Several approaches, including the Care Cascade and preexposure prophylaxis (PrEP), may have contributed to the reduction in the number of new HIV infections, but an effective vaccine is essential for global control of the pandemic (2). In the estimation of a current report (3), globally, 49 million incident cases under status quo interventions are predicted from 2015 to 2035, but achieving the HIV/AIDS 95-95-95 targets of the UNAIDS program would reduce the total by 25 million. Introduction in 2020 of a 50% efficacy vaccine gradually scaled up to 70% coverage could result in an additional reduction of 6.3 million, suggesting a considerable impact of even partially effective vaccines on global HIV control.

Induction of effective antibody and/or T-cell responses is a principal vaccine strategy. At present, it may be difficult to obtain an HIV vaccine with 100% efficacy, but a combination of antibody-inducing and T-cell-inducing vaccines even with partial efficacy could be a promising strategy for improving vaccine efficacy. We have been working mainly on the development of T-cell-inducing HIV vaccines. CD8⁺ T-cell responses are crucial for the control of HIV replication (4–6). Viral vectors are considered to be a promising vaccine delivery tool for inducing CD8⁺ T-cell responses. Several vaccine attempts using viral vectors to induce T-cell responses have shown partial but significant protective efficacy against simian immunodeficiency virus (SIV) challenge in rhesus macaques (7–11). In particular, heterologous prime-boost regimens such as DNA-prime/viral vector-boost have been shown to efficiently induce antigen-specific T-cell responses (12). We have developed a vaccine regimen using Sendai virus (SeV) vectors and shown its potential to efficiently induce antigen-specific T-cell responses, resulting in the control of an SIVmac239 challenge in rhesus macaques (7, 13). An HIV vaccine clinical trial has indicated that the SeV vector can be a delivery tool candidate for heterologous prime-boost HIV vaccine regimens for efficient T-cell induction (14).

Optimization of vaccine antigens as well as delivery tools is crucial for effective T-cell induction. Recent studies have implicated CD8⁺ T-cell responses targeting viral Gag and possibly Vif antigens in the control of HIV/SIV replication (15–20). Furthermore, several promising immunogens, including HIVconsv, consisting of multiple conserved regions of the HIV proteome, and HTI, consisting of 16 HIV Gag/Pol/Vif/Nef-derived regions that were relatively conserved and predominantly targeted by individuals with reduced viral loads, have been designed (21–24).

On the other hand, virus-specific CD4⁺ T cells can be preferential targets for HIV infection (25) although helper CD4⁺ T-cell responses are essential for effective CD8⁺ T-cell responses (26–29). Our previous study (30) has indicated that virus-specific CD107a⁻ CD4⁺ T-cell induction by vaccination could be detrimental and accelerate viral replication after HIV/SIV exposure, supporting a rationale for vaccine design inducing HIV-specific CD8⁺ T-cell responses without HIV-specific CD4⁺ T-cell induction but with non-HIV antigen-specific CD4⁺ T-cell help (31). In the present study, based on this concept, we constructed a novel immunogen, CaV11, consisting of tandemly connected overlapping 11-mer peptides (TC11) spanning simian immunodeficiency virus (SIV) Gag capsid (CA) and Vif amino acid sequences. This CaV11 antigen is expected to selectively elicit Gag/Vif-specific CD8⁺ T cells with inefficient Gag/Vif-specific CD4⁺ T-cell induction because the ideal length of CD4⁺ T-cell epitopes is longer than 11 mers, whereas CD8⁺ T-cell epitopes are 8 to 11 mers (32). Our analysis in rhesus macaques revealed that vaccination with CaV11-expressing SeV vectors can selectively and efficiently elicit CD8⁺ T cells but not CD4⁺ T cells targeting viral Gag and Vif antigens.

RESULTS

Gag- and Vif-specific T-cell responses after CaV11 vaccination in an SIV-infected macaque.

Eight kinds of CaV11 immunogens, CaV11A, CaV11B, CaV11C, CaV11D, CaV11E, CaV11F, CaV11G, and CaV11H, were constructed (Fig. 1 and 2). Each CaV11 consists of 3-mer-overlapping 11-mer peptides connected in tandem and spanning SIVmac239 CA and Vif amino acid sequences, and every overlapping 11-mer peptide derived from CA and Vif was included in one of these eight CaV11 immunogens.

FIG 1.

CaV11 immunogen consisting of overlapping 11-mer peptides connected in tandem and spanning SIV Gag CA and Vif amino acid sequences. (A) Schema of a representative CaV11 immunogen, CaV11A. (B) SIV Gag and Vif regions that individual CaV11 (A to H) immunogens cover. aa, amino acid.

FIG 2.

Amino acid sequences of CaV11 immunogens. Start codons (M) and alanine (A) linkers are underlined.

First, a rhesus macaque possessing protective major histocompatibility complex class I (MHC-I) haplotype 90-120-Ia and controlling SIVmac239 replication for more than 5 years, which was described in our previous study (33, 34), was vaccinated with CaV11-expressing F-deleted SeV [F(−)SeV-CaV11] vectors. This macaque received F(−)SeV-CaV11A and F(−)SeV-CaV11F vectors at week 281 postinfection and, 1 week later, F(−)SeV-CaV11B and F(−)SeV-CaV11H (Fig. 3A). Analyses of MHC-I haplotype 90-120-Ia-associated CD8⁺ T-cell responses by tetramer staining and by detection of specific gamma interferon (IFN-γ) induction found enhancement of Gag_206–216-, Gag_241–249-, and Vif_114–124-specific, but not Nef_193–203-specific, CD8⁺ T-cell responses after the vaccination (Fig. 3B, C, and D). Further investigation confirmed that the SeV-CaV11 vaccination expands CD8⁺ T cells but not CD4⁺ T cells targeting viral Gag and Vif antigens (Fig. 3E). Nef-specific CD8⁺ T and CD4⁺ T cells were not enhanced, while SeV-specific CD8⁺ T and CD4⁺ T cells were induced (Fig. 3E).

FIG 3.

Efficient induction of CD8⁺ T cells but not CD4⁺ T cells targeting SIV Gag and Vif by CaV11 vaccination in an SIV controller. (A) Experimental protocol. A macaque possessing protective MHC-I haplotype 90-120-Ia that controlled SIV replication received CaV11A- and CaV11F-expressing F(−)SeV vectors at week 281 postinfection and, 1 week later, CaV11B- and CaV11H-expressing F(−)SeV vectors. (B) Representative gating schema for detection of Gag_241–249-specific CD8⁺ T cells by tetramer at week 282 postinfection by flow cytometric analysis. (C) Changes in frequencies of Gag_241–249-Mamu-A1*065:01 and Nef_193–203-Mamu-A1*065:01 tetramer-positive cells after the first vaccination at week 281 postinfection. (D) Changes in frequencies of Gag_206–216, Gag_241–249, Vif_114–124, and Nef_193–203 epitope-specific CD8⁺ T cells detected by specific IFN-γ induction after the first vaccination. (E) Changes in frequencies of Gag-, Vif-, Nef-, and SeV-specific CD4⁺ and CD8⁺ T cells detected by specific IFN-γ induction after the first vaccination. n.d., not determined.

Gag- and Vif-specific T-cell responses after CaV11 vaccination in naive macaques.

Second, six naive rhesus macaques were vaccinated with CaV11-expressing DNAs and SeV vectors (Fig. 4A). Macaques possessing protective MHC-I haplotype 90-120-Ia or 90-010-Id were not included, and four of the six animals shared MHC-I haplotype 90-010-Ie, which is associated with typical disease progression (18, 35, 36). No significant changes in peripheral percent CD4⁺ and percent CD8⁺ T cells were observed after vaccination (Fig. 5). Gag/Vif-specific T-cell responses were undetectable or marginal after the DNA vaccination. After the SeV-CaV11 vaccination, all six animals showed efficient Gag- and/or Vif-specific CD8⁺ T-cell responses; efficient Gag-specific CD8⁺ T-cell responses were induced in all of them (Fig. 4B and C and Fig. 6). In contrast, Gag/Vif-specific CD4⁺ T-cell induction was inefficient; Gag-specific CD4⁺ T-cell responses were undetectable in all six animals. Comparison between Gag-specific CD8⁺ T-cell and CD4⁺ T-cell frequencies showed that CD8⁺ T cells are dominantly induced by SeV-CaV11 vaccination (Fig. 4D). These results indicate that the CaV11 immunogen can efficiently induce CD8⁺ T cells but not CD4⁺ T cells targeting viral Gag and Vif antigens.

FIG 4.

Efficient induction of CD8⁺ T cells but not CD4⁺ T cells targeting SIV Gag and Vif by CaV11 vaccination in rhesus macaques. (A) Experimental protocol in six naive rhesus macaques. (B) Representative gating schema for detection of IFN-γ induction after stimulation with overlapping Gag peptide pools in macaque 2 at week 19 postvaccination. (C) Frequencies of CD4⁺ and CD8⁺ T cells targeting Gag and Vif, as indicated, 1 week after the last vaccination (at week 19 after the first vaccination). (D) Comparison of Gag-specific CD4⁺ and CD8⁺ T-cell frequencies 1 week after the last vaccination in CaV11-expressing SeV-immunized macaques. Gag-specific CD4⁺ T-cell responses were poor and significantly lower than Gag-specific CD8⁺ T-cell responses (P = 0.0313 by Wilcoxon signed-rank test).

FIG 5.

Percentages of peripheral CD4⁺ T and CD8⁺ T cells in macaques postvaccination and postinfection. Percentages of peripheral CD4⁺ T cells (upper panels) and CD8⁺ T cells (lower panels) in macaques postvaccination and postinfection are shown.

FIG 6.

Gag/Vif-specific CD8⁺ T-cell responses postvaccination and postinfection. Frequencies of CD8⁺ T cells targeting Gag and Vif, as indicated, at weeks 8, 13, and 19 post-initial vaccination and at weeks 2, 6, and 12 postinfection.

Virological and immunological analyses after an intravenous high-dose SIVmac239 challenge in CaV11-vaccinated macaques.

To examine whether the CaV11 immunogen can mitigate the detrimental effect of conventional vaccine-induced CD4⁺ T cells on vaccine efficacy postexposure, we analyzed acute viral infection after intravenous high-dose SIV challenge for comparison with our previous results (30). Six SeV-CaV11-vaccinated macaques were intravenously challenged with a high dose (1,000 50% tissue culture infective doses [TCID₅₀]) of SIVmac239 6 weeks after the last vaccination. Two (controllers, 2 and 6) of the six animals controlled SIV replication at the set point, while the remaining four (noncontrollers, 1, 3, 4, and 5) failed to control viremia (Fig. 7A). We compared viral loads at week 1 postinfection in these 6 macaques with those in 15 unvaccinated and 10 Gag-vaccinated macaques in our previous study (30) (Fig. 7B). Six of the 15 unvaccinated macaques and 5 of the 10 macaques immunized with DNA and SeV expressing Gag had MHC-I haplotype 90-010-Ie, while none of them had protective MHC-I haplotype 90-120-Ia or 90-010-Id. Seven Gag-vaccinated macaques that failed to control set point viremia (noncontrollers) showed significantly higher viral loads at week 1 than unvaccinated animals (P = 0.0017 by one-way analysis of variance [ANOVA] and Dunn’s multiple-comparison test). In contrast, no clear increase in viral loads at week 1 was observed in CaV11-vaccinated SIV noncontrollers (Fig. 7B).

FIG 7.

Plasma viral loads after intravenous high-dose SIVmac239 challenge. (A) Changes in plasma viral loads after SIV challenge. Viral loads (SIV gag RNA copies/milliliter of plasma) were determined as described previously (7). The lower limit of detection is approximately 4 × 10² copies/ml. Viral loads in the six CaV11-immunized macaques are shown in red. Viral loads in 15 unvaccinated macaques in our previous study (30) are represented by black dotted lines as historical controls. (B) Comparison of plasma viral loads at week 1 postinfection in the 6 CaV11-vaccinated macaques (CaV11) with those in 15 unvaccinated (unvac) macaques and 10 Gag-vaccinated (Gag) macaques including three SIV controllers (closed diamonds) in our previous study (30). Multiple comparison among unvaccinated (n = 15), Gag-vaccinated noncontrollers (open diamonds; n = 7), and CaV11-vaccinated noncontrollers (open symbols; n = 4) showed a significant difference between unvaccinated and Gag-vaccinated noncontrollers (P = 0.0017 by one-way ANOVA and Dunn’s multiple-comparison test).

Analysis of SIV-specific induction of CD107a, IFN-γ, macrophage inflammatory protein-1β (MIP-1β), tumor necrosis factor-α (TNF-α), and interleukin-2 (IL-2) in CD4⁺ T cells detected SIV-specific CD4⁺ T-cell responses at week 1 postinfection in all of the CaV11-vaccinated macaques (Fig. 8A). In our previous study (30), SIV-specific CD107a⁻ CD4⁺ T cells induced by Gag vaccination were depleted following SIVmac239 infection, and none of the seven Gag-vaccinated noncontrollers had detectable SIV-specific IFN-γ⁺, MIP-1β⁺, or TNF-α⁺ CD4⁺ T cells at week 1. In contrast, all of the CaV11-vaccinated macaques, including noncontrollers, showed detectable SIV-specific IFN-γ⁺, MIP-1β⁺, or TNF-α⁺ CD4⁺ T cells at week 1 in the present study. A comparison demonstrated significantly higher SIV-specific IFN-γ⁺, MIP-1β⁺, and TNF-α⁺ CD4⁺ T-cell responses at week 1 postinfection in these CaV11-vaccinated macaques than in the Gag-vaccinated as well as in the unvaccinated macaques (Fig. 8B). These results suggest mitigation of depletion of de novo-induced SIV-specific CD4⁺ T cells postinfection due to no enhancement of acute SIV replication in CaV11-vaccinated macaques.

FIG 8.

SIV-specific CD4⁺ T-cell responses at week 1 postinfection. (A) SIV-specific CD107a, IFN-γ, MIP-1β, TNF-α, and IL-2 responses in CD4⁺ T cells at week 1 postinfection in CaV11-vaccinated macaques. (B) Comparison of SIV-specific CD4⁺ T-cell responses at week 1 postinfection in CaV11-vaccinated macaques (CaV11) with those in unvaccinated (unvac) and Gag-vaccinated (Gag) macaques in our previous study (30).

Comparison of T-cell responses between vaccinated SIV controllers and noncontrollers.

Finally, we compared T-cell responses 1 week after the last vaccination (at week 19 postvaccination) between CaV11-vaccinated SIV controllers (macaques 2 and 6) and noncontrollers (animals 1, 3, 4, and 5) to examine immune correlates with SIV control. No significant difference in Gag/Vif-specific CD8⁺ T-cell responses postvaccination was observed between the controllers and the noncontrollers (Fig. 4C). However, a significant inverse correlation was observed between Vif-specific CD8⁺ T-cell responses at week 19 postvaccination and plasma viral loads at week 1 postinfection (Fig. 9A). This inverse relationship was not observed between the Vif-specific CD8⁺ T-cell responses and viral loads later. Viral genome sequences (Fig. 9B) imply that this may be due to rapid selection of viral Vif-specific CD8⁺ T-cell escape mutations. The two controllers induced relatively high Gag-specific CD8⁺ T-cell responses at week 19 postvaccination, but no significant association was observed between the Gag-specific CD8⁺ T-cell responses and viral loads. Interestingly, the two SIV controllers induced higher SeV-specific CD4⁺ T-cell responses 1 week after the last vaccination (at week 19 postvaccination) (Fig. 10A), and a significant inverse correlation between postvaccination SeV-specific CD4⁺ T-cell frequencies and set point viral loads was observed (Fig. 10B). Macaques (animals 2, 3, and 6) showing relatively high SeV-specific CD4⁺ T-cell responses induced relatively high Gag-specific CD8⁺ T-cell responses, but no significant correlation was observed between SeV-specific CD4⁺ and Gag- or Vif-specific CD8⁺ T-cell frequencies. However, postvaccination CD8⁺ cells derived from the two SIV controllers showed higher anti-SIV efficacy in vitro (Fig. 10C). These results suggest that helper CD4⁺ T-cell responses targeting vector SeV antigens contribute to induction of effective Gag/Vif-specific CD8⁺ T cells by CaV11 vaccination.

FIG 9.

Correlation analyses between postvaccination T-cell responses and plasma viral loads. (A) Correlation analyses between CD8⁺ T-cell responses targeting Gag or Vif, as indicated, 1 week after the last CaV11-expressing F(−)SeV vector vaccination (at week 19 post-initial vaccination) and plasma viral loads at week 1 and month 6 postinfection were performed (Spearman’s test). There was a significant inverse correlation between Vif-specific CD8⁺ T-cell responses postvaccination and viral loads at week 1 postinfection (P = 0.0278; r = −0.8986). (B) Viral gag and vif mutations in the early phase of SIV infection. We examined sequences of viral Gag- and Vif-encoding cDNAs amplified from plasma RNAs obtained from CaV11-vaccinated macaques at months 1.5, 3, and 6 postinfection. Amino acid (aa) substitutions derived from dominant nonsynonymous mutations are shown. Residues that are not covered by CaV11 immunogens are shaded. Asterisks indicate substitutions by multiple amino acids.

FIG 10.

SeV-specific T-cell responses and in vitro anti-SIV efficacy of CD8⁺ cells postvaccination. (A) SeV-specific CD4⁺ and CD8⁺ T-cell responses 1 week after the last vaccination (at week 19 postvaccination). (B) Correlation analysis between SeV-specific CD4⁺ T-cell responses 1 week after the last vaccination and plasma viral loads at month 6 postinfection. There was a significant inverse correlation (P = 0.0167; r = −0.9429, by Spearman’s test). (C) Anti-SIV efficacy in vitro of CD8⁺ cells 1 week after the last vaccination. CD8⁻ target (T) cells 2 days after SIVmac239 infection were cultured alone (no CD8) or cocultured with autologous CD8⁺ effector (E) cells obtained from PBMCs at week 19 postvaccination (E/T = 1:1). SIV Gag p27 concentrations in the culture supernatants at day 8 postinfection are shown in the left panel. Mean values of duplicate experiments are shown except for macaques 4 and 5, whose values are from a single experiment because of the limitation of sample availability. The ratios of p27 concentrations in the supernatants from the coculture (E/T = 1:1) to those without CD8⁺ cells (no CD8) are shown as fold reductions in viral replication (in vitro anti-SIV efficacy) in the right panel.

DISCUSSION

Virus-specific CD4⁺ T-cell responses are crucial for induction of effective CD8⁺ T-cell responses against virus infection (26–29). Current vaccine strategies inducing CD8⁺ T-cell responses are accompanied by CD4⁺ T-cell induction. Vaccine-induced CD4⁺ T cells, however, can be preferential targets for HIV/SIV infection (25). Indeed, our recent study found that vaccine-elicited SIV-specific CD107a⁻ CD4⁺ T cells are depleted in the acute phase of infection after SIV challenge, indicating that virus-specific CD107a⁻ CD4⁺ T-cell induction by vaccination could accelerate viral replication after HIV/SIV exposure (30). This has been supported by a current report indicating the detrimental effect of vaccine-induced CD4⁺ T cells on HIV vaccine efficacy (37). These studies support the rationale for vaccine design inducing HIV-specific CD8⁺ T-cell responses without HIV-specific CD4⁺ T-cell induction but with non-HIV antigen-specific CD4⁺ T-cell help. In the present study, based on this concept, we constructed the CaV11 immunogen, TC11, spanning SIV Gag CA and Vif amino acid sequences. Our results in rhesus macaques confirmed that the CaV11 vaccination can selectively and efficiently elicit CD8⁺ T cells but not CD4⁺ T cells targeting viral Gag and Vif antigens.

We showed efficient induction of Gag/Vif-specific CD8⁺ T-cell responses by our vaccine regimen using DNAs and F(−)SeV vectors expressing immunogens CaV11A to CaV11H. This might be due to efficient processing of CD8⁺ T-cell target epitopes from overlapping 11 mers. We used eight CaV11 immunogens, CaV11A, CaV11B, CaV11C, CaV11D, CaV11E, CaV11F, CaV11G, and CaV11H. Individual CD8⁺ T-cell epitopes are not included in all of them. For example, an 11-mer epitope is included in only one of the CaV11 immunogens. This may contribute to induction of subdominant as well as dominant CD8⁺ T-cell responses, resulting in broader CD8⁺ T-cell induction.

The CaV11 immunogen was designed to mitigate the detrimental effect of HIV-specific CD4⁺ T-cell induction on vaccine efficacy in the acute phase postexposure. Therefore, to examine whether CaV11 can really mitigate the detrimental effect, the present study analyzed acute viral infection after intravenous high-dose SIV challenge for comparison with our previous results (30). In the previous study of vaccination with SeV vectors expressing the whole Gag antigen, Gag-vaccinated SIV noncontrollers showed higher viral loads at week 1 postinfection and depletion of vaccine-induced CD4⁺ T cells following SIV infection; SIV-specific CD4⁺ T cells were almost undetectable at week 1 postinfection. This implies that even naive-derived de novo SIV-specific CD4⁺ T cells induced postinfection were killed because of enhancement of acute SIV replication. In contrast, CaV11-immunized macaques, including noncontrollers, showed detectable de novo SIV-specific CD4⁺ T cells without a clear increase in acute viral loads at week 1 postinfection in the present study (Fig. 7B and 8). It is inferred that naive-derived SIV-specific CD4⁺ T cells induced postinfection were not completely killed and detectable, suggesting no enhancement of SIV replication without a preferential SIV target, vaccine-induced SIV-specific CD4⁺ T cells, at virus exposure. These results suggest that the acute depletion of virus-specific CD4⁺ T cells and acceleration of viral replication post-SIV exposure that are observed in Gag-vaccinated macaques can be mitigated in CaV11-vaccinated macaques although no significant enhancement of viral control was observed in the CaV11-vaccinated macaques compared to levels in the Gag-vaccinated macaques after intravenous high-dose SIV challenge.

It is speculated that induction of HIV/SIV-specific CD4⁺ T cells may more largely accelerate acute viral replication and/or viral acquisition after intrarectal low-dose HIV/SIV exposure than that after intravenous challenge. Indeed, a current report has indicated that HIV vaccine-induced TH1-biased CD4⁺ T cells in the intestinal and genital mucosa can diminish beneficial effects of protective antibodies and CD8⁺ T cells (37). Thus, the CaV11 immunogen that can mitigate the detrimental effect of vaccine-induced CD4⁺ T cells may exhibit protective efficacy against intrarectal low-dose SIV challenge, which needs to be addressed in the next study.

Remarkably, the two CaV11-vaccinated SIV controllers showed higher SeV-specific CD4⁺ T-cell responses and higher in vitro anti-SIV efficacy of CD8⁺ cells postvaccination (Fig. 10). A significant inverse correlation between SeV-specific CD4⁺ T-cell responses postvaccination and set point viral loads was shown. These results suggest that SeV-specific helper CD4⁺ T-cell responses may contribute to efficient induction of effective Gag/Vif-specific CD8⁺ T-cell responses without Gag/Vif-specific CD4⁺ T-cell induction by the CaV11 vaccine regimen using CaV11-expressing F(−)SeV vectors. In addition to SeV-specific CD4⁺ T cells, CD4⁺ T cells targeting CaV11-derived neoepitopes encompassing two 11 mers connected with an alanine spacer could be induced by one of the CaV11 immunogens, CaV11A to CaV11H. These CaV11 neoepitope-specific CD4⁺ T cells can also exert a helper function at vaccination but are not SIV specific and, thus, are not considered to be preferential targets for SIV infection.

Our TC11 design is applicable to other prophylactic and therapeutic HIV vaccine candidates (21–23), such as HIVconsv and HTI, to induce effective HIV-specific CD8⁺ T-cell responses. This immunogen can also be introduced into other viral vectors, and a combination of several vectors may lead to induction of durable effective CD8⁺ T-cell responses. Development of a method for induction of effective anti-Env antibodies without Env-specific CD4⁺ T-cell induction may be a next issue. An immunogen selectively inducing HIV-specific CD8⁺ T-cell responses without induction of HIV-specific CD4⁺ T-cell responses would be advantageous not only for a prophylactic HIV vaccine but also for a therapeutic vaccine toward functional HIV cure.

In summary, to mitigate the detrimental effect of vaccine-induced CD4⁺ T cells on vaccine efficacy, we developed a novel TC11 immunogen, CaV11, for selective induction of CD8⁺ T cells, but not CD4⁺ T cells, targeting SIV Gag CA and Vif. This study showed that DNA-prime/SeV-boost vaccination expressing our immunogen can efficiently elicit effective CD8⁺ T cells, but not CD4⁺ T cells, targeting viral Gag CA and Vif antigens. Further analysis suggests a contribution of SeV-specific CD4⁺ T-cell helper responses for effective Gag/Vif-specific CD8⁺ T-cell induction by vaccination. This immunogen design of tandemly connecting overlapping 11-mer peptides, which is also applicable for other vaccine antigens, could be a promising method for selective induction of effective anti-HIV CD8⁺ T-cell responses.

MATERIALS AND METHODS

Immunogen design.

Eight CaV11 immunogens were prepared and administered separately to reduce possible recombination frequency. Amino acid sequences of the CaV11 immunogens, CaV11A, CaV11B, CaV11C, CaV11D, CaV11E, CaV11F, CaV11G, and CaV11H, are described in Fig. 2. As shown in Fig. 1, CaV11A consists of 3-mer overlapping 11-mer peptides spanning Gag_135–393 and Vif_1–195 (32 Gag-derived 11 mers and 24 Vif-derived 11 mers) that are connected in tandem with an alanine (A) spacer. CaV11B, CaV11C, CaV11D, CaV11E, CaV11F, CaV11G, and CaV11H span Gag_136–394/Vif_2–196, Gag_137–395/Vif_3–197, Gag_138–396/Vif_4–198, Gag_139–397/Vif_5–199, Gag_140–398/Vif_6–200, Gag_141–399/Vif_7–201, and Gag_142–400/Vif_8–202, respectively. Thus, these CaV11 immunogens in total include 1-mer overlapping 11-mer peptides spanning Gag_135–400 and Vif_1–202 amino acid sequences. Individual cDNAs encoding these CaV11 immunogens were synthesized (Eurofins Genomics) and introduced into pcDNA3.1 vector plasmid (Invitrogen) to obtain DNAs expressing CaV11A, CaV11B, CaV11C, CaV11D, CaV11E, CaV11F, CaV11G, and CaV11H for vaccination. F-deleted [F(−)] SeV vectors expressing these individual CaV11 immunogens were also constructed (38).

Animal experiments.

Animal experiments were carried out in the Tsukuba Primate Research Center, National Institutes of Biomedical Innovation, Health and Nutrition (NIBIOHN), with the help of the Corporation for Production and Research of Laboratory Primates after approval by the Committee on the Ethics of Animal Experiments in NIBIOHN (permission numbers DS27-17 and DS28-18) in accordance with the Guidelines for Proper Conduct of Animal Experiments established by the Science Council of Japan (39). The experiments were in accordance with recommendations of the Weatherall report on the use of nonhuman primates in research (40). Blood collection, vaccination, and virus inoculation were performed under ketamine anesthesia. A rhesus macaque possessing protective MHC-I haplotype 90-120-Ia and controlling SIVmac239 replication for more than 5 years, which was described in our previous studies (33, 34), received 6 × 10⁹ cell infectious units (CIU) of F(−)SeV-CaV11A and F(−)SeV-CaV11F vectors and, 1 week later, 6 × 10⁹ CIU of F(−)SeV-CaV11B and F(−)SeV-CaV11H. In the vaccine/challenge experiment, six Burmese rhesus macaques received four CaV11-expressing DNA vaccinations at weeks 0, 1, 3, and 4, followed by four CaV11-expressing SeV vector vaccinations at weeks 6, 7, 12, and 18, as shown in Fig. 4A. Macaques possessing protective MHC-I haplotype 90-120-Ia or 90-010-Id were not included, and four of the six animals shared MHC-I haplotype 90-010-Ie, which is associated with typical disease progression (18, 35, 36). At each DNA vaccination, animals intramuscularly received four of eight kinds of CaV11-expressing DNAs (5 mg each) in the right upper limb, left upper limb, right lower limb, and left lower limb. Gag/Vif-specific T-cell responses after DNA vaccination (before SeV vector boost) were expected to be undetectable or marginal based on our previous experience (7). At each SeV vector vaccination, animals received two kinds of CaV11-expressing F(−)SeV vectors (1 × 10⁹ CIU each) intranasally and two intramuscularly. These animals were intravenously challenged with 1,000 TCID₅₀ of SIVmac239 6 weeks after the last vaccination. In addition, we used data obtained in our previous study using 15 unvaccinated and 10 Gag-vaccinated macaques (30). Six of these 15 unvaccinated macaques and 5 of the 10 macaques immunized with DNA and SeV-expressing Gag had MHC-I haplotype 90-010-Ie, while macaques possessing protective MHC-I haplotype 90-120-Ia and 90-010-Id were excluded.

Analysis of antigen-specific T-cell responses.

In tetramer staining, peripheral blood mononuclear cells (PBMCs) were stained with phycoerythrin (PE)-conjugated Gag_241–249-Mamu-A1*065:01 tetramer and allophycocyanin (APC)-conjugated Nef_193–203-Mamu-A1*065:01 tetramer (MBL), followed by staining with fluorescein isothiocyanate (FITC)-conjugated anti-human CD4, peridinin chlorophyll protein (PerCP)-conjugated anti-human CD8, and APC-Cy7-conjugated anti-human CD3 monoclonal antibodies (BD).

For analysis of SIV antigen-specific T-cell responses by detection of IFN-γ induction after specific stimulation as described previously (34, 41), autologous herpesvirus papio-immortalized B-lymphoblastoid cell lines (B-LCLs) were pulsed with epitope peptides (Gag_206–216, Gag_241–249, Vif_114–124, and Nef_193–203) or peptide pools (at a final concentration of 1 to 2 μM for each peptide) derived from panels of overlapping peptides spanning SIVmac239 Gag, Vif, and Nef amino acid sequences (Sigma-Aldrich). For analysis of SeV-specific T-cell responses, autologous B-LCLs were infected with SeV. PBMCs were cocultured with these B-LCLs in the presence of GolgiStop (monensin; BD) for 6 h. Intracellular IFN-γ staining was performed with a Cytofix/Cytoperm kit (BD) and FITC-conjugated anti-human CD4, PerCP-conjugated anti-human CD8, APC-Cy7-conjugated anti-human CD3, and PE-conjugated anti-human IFN-γ (BioLegend) monoclonal antibodies. Antigen-specific T-cell frequencies were calculated by subtracting nonspecific IFN-γ⁺ T-cell frequencies from those after antigen-specific stimulation. Antigen-specific T-cell frequencies lower than 0.04% of CD4⁺ or CD8⁺ T cells were considered negative.

For analysis of SIV-specific polyfunctional T-cell responses, we examined SIV-specific induction of CD107a, IFN-γ, MIP-1β, TNF-α, and IL-2 in CD4⁺ or CD8⁺ T cells as described previously (30). PBMCs were prestimulated with 5 μg/ml immobilized anti-human CD28 (BioLegend) and 5 μg/ml immobilized anti-human CD49d (BioLegend) for 12 h, followed by 6-h coculture in the presence of Brilliant Violet 785-conjugated anti-human CD107a (BioLegend) with B-LCLs infected with vesicular stomatitis virus G protein (VSV-G)-pseudotyped env- or nef-deleted SIV (SIVGP1) (30). Monensin and brefeldin A (Sigma-Aldrich) were added to the culture 1 h after the start of coculture. Immunostaining was performed using a Fix & Perm fixation and permeabilization kit (Invitrogen) and the following monoclonal antibodies: APC-Cy7-conjugated anti-nonhuman primate CD3 (BD), Brilliant Violet 605-conjugated anti-human CD4 (BioLegend), Alexa Fluor 700-conjugated anti-human CD8 (BD), PE-Cy7-conjugated anti-human IFN-γ (eBioscience), Pacific Blue-conjugated anti-human TNF-α (BioLegend), PerCP-Cy5.5-conjugated anti-human IL-2 (BioLegend), and PE-conjugated anti-human MIP-1β (BD). Dead cells were stained using a Live/Dead fixable dead-cell stain kit (Invitrogen). Frequencies of SIV-specific T-cell subsets were calculated by subtracting frequencies after nonspecific stimulation from those after SIV-specific stimulation.

Analysis of in vitro anti-SIV efficacy of CD8⁺ cells.

To evaluate in vitro anti-SIV efficacy of CD8⁺ cells postvaccination, we examined SIVmac239 replication on CD8-depleted PBMCs in the presence of CD8⁺ cells positively selected from PBMCs, as described previously (42). In brief, PBMCs were separated into CD8⁺ and CD8⁻ cells by using magnetically activated cell sorting (MACS) CD8 MicroBeads (Miltenyi Biotec). For preparing target cells, prevaccination PBMC-derived CD8⁻ cells were cultured in the presence of 2 μg/ml phytohemagglutinin-L and 20 IU/ml recombinant human IL-2 (Roche Diagnostics) and infected with SIVmac239 at a multiplicity of infection (MOI) of 1:10³ TCID₅₀/cell. The CD8⁻ target (T) cells 2 days after SIVmac239 infection were cultured alone (no CD8) or cocultured with autologous CD8⁺ effector (E) cells obtained from PBMCs at week 19 postvaccination (1 week after the last vaccination) at an effector/target (E/T) ratio of 1:1. SIV Gag p27 concentrations in the culture supernatants at day 8 postinfection were measured by a p27 antigen capture assay (Advanced BioScience Laboratories). Reduction in viral production by addition of CD8⁺ cells was shown as fold reduction in the concentration of p27 compared with that in the supernatant from virus-infected CD8⁻ cell culture without CD8⁺ cells.

Statistical analysis.

Statistical analyses were performed using Prism software (GraphPad Software, Inc.) with significance set at a P value of <0.05. Comparisons were performed by Wilcoxon signed-rank test, and multiple comparisons were performed by one-way ANOVA and Dunn’s multiple-comparison test. Correlation analyses were performed by Spearman’s test.