Our limited understanding of the interaction between primate lentiviruses and the host immune system complicates the design of an effective HIV/AIDS vaccine. To identify immunological correlates of protection from SIV disease progression, we immunized two groups of five rhesus macaques (RMs) with either modified vaccinia Ankara (MVA) or MVAΔudg vectors that expressed SIVmac239 Gag and Tat. Both vectors raised a SIV-specific CD8+ T cell response, with a magnitude that was greater in mucosal tissues than in peripheral blood. After challenge with SIVmac239, all vaccinated RMs showed mucosal and systemic CD8+ T cell recall responses that appeared faster and were of greater magnitude than those in five unvaccinated control animals. All vaccinated RMs showed a ∼1-log lower peak and early set-point SIV viral load than the unvaccinated animals, and then, by 8 wk postchallenge, exhibited levels of viremia similar to the controls. We observed a significant direct correlation between the magnitude of postchallenge SIV-specific CD8+ T cell responses and SIV viral load. However, vaccinated RMs showed no protection from either systemic or mucosal CD4+ T cell depletion and no improved survival. The observation that vaccine-induced, SIV-specific CD8+ T cells that partially control SIVmac239 virus replication fail to protect from immunological or clinical progression of SIV infection underscores both the complexity of AIDS pathogenesis and the challenges of properly assessing the efficacy of candidate AIDS vaccines.
Human immunodeficiency virus causes ∼3 million deaths per year, with a rate of new infections that continues virtually unabated in developing countries. These facts emphasize the urgent need for an effective HIV/AIDS vaccine. However, development of an AIDS vaccine has proven difficult due to the adaptive nature and genetic diversity of a virus that typically does not induce any type of protective natural immunity (1, 2). Traditional vaccine approaches have been clearly unsuccessful, as live-attenuated lentivirus vaccines present unacceptably high risks, and the variability, plasticity, and glycosylation-mediated masking of the envelope glycoprotein have thwarted approaches based on the induction of broadly neutralizing Abs (3).
More recently, much emphasis has been placed on the potential efficacy of candidate HIV/AIDS vaccines based on the induction of virus-specific CTL responses, mainly mediated by CD8+ T cells (4). The importance of CD8+ T cells in the control of viral replication has been suggested by the following observations: 1) temporal association between emergence of HIV-specific CD8+ T cells and postpeak decline of viral load (5); 2) the presence of stronger and more functional CD8+ T cell responses in individuals with spontaneous (i.e., non-therapy induced) control of virus replication (6); 3) association between certain HLA class I haplotypes and rate of progression to AIDS (7); 4) the presence of CTL escape mutants in both HIV and SIV infections (8); 5) increased virus replication in SIV-infected macaques following CD8+ cell depletion (9). Based on these observations, intense efforts have been made to design candidate HIV/AIDS vaccines that elicit strong, broadly reactive, multifunctional, and long-lasting virus-specific CTLs (4). A number of vector platforms, including plasmid DNA, adenovirus (Ad),4 modified vaccinia Ankara (MVA), vescicular stomatitis virus, and various others have shown some protection from subsequent pathogenic challenge in rhesus macaques (RM; Refs. 10, 11, 12, 13, 14). However, these immunization protocols were mainly successful against the chimeric SHIV89.6P viral strain the relevance of which as a challenge model is questionable (15). In contrast, a much more limited protection has generally been observed against SIVmac239 or SIVmac251 (16, 17, 18, 19). Our limited knowledge of what features of vaccine-induced HIV/SIV-specific CD8+ T cell responses will confer actual protection from subsequent infection is arguably the main obstacle to the rational design of a CTL-based AIDS vaccine (1).
In this study, we examined the relationship between MVA vaccine- and/or SIVmac239 infection-elicited SIV-specific CD8+ T cell responses, SIV viral load, systemic and mucosal CD4+ T cell depletion, and survival in RMs. We studied a total of 15 macaques, organized into three groups of five animals each, that received immunizations with MVA-SIV, MVAΔudg-SIV, or were unvaccinated (control group). Recombinant MVA vectors that harbor a deletion of the essential poxvirus gene encoding uracil DNA glycosylase (udg) are defective for expression of late poxvirus genes during infection of noncomplementing cells and consequently elicit a restricted repertoire of vector-specific CD8+ T cells (D. A. Garber, unpublished observation). All animals were challenged with SIVmac239, and the SIV-specific CD8+ T cells were analyzed longitudinally pre- and postchallenge in peripheral blood, LNs, and MALTs that were sampled via rectal biopsies (RB) and bronchoalveolar lavage (BAL). Our principal findings are that immunization with MVA-SIV or MVAΔudg-SIV vectors: 1) elicited robust SIV Gag- and Tat-specific CD8+ T cell responses in the blood and high frequency peak SIV-specific CD8+ T cell responses in mucosal sites; 2) primed for robust anamnestic SIV-specific CD8+ T cell responses, following SIV challenge, both systemically and in mucosal compartments; 3) elicited T cell immunity that provided incomplete control of SIVmac239 viremia after challenge (∼1-log reduction for 8 wk); 4) did not confer protection against CD4+ T cell depletion from peripheral blood or mucosal compartments after SIV challenge; and 5) did not prolong the survival of SIVmac239-challenged macaques as compared with unvaccinated controls.
Materials and Methods
Fifteen healthy, SIV-uninfected, MaMu-A*01-positive Indian RMs were used in this study. All animals were housed at the Yerkes National Primate Research Center (Atlanta, GA) and maintained in accordance with National Institutes of Health guidelines. These studies were approved by the Emory University and University of Pennsylvania Institutional Animal Care and Use Committees.
MVA-SIV and MVAΔudg-SIV vaccines
Recombinant viruses MVA-SIV and MVAΔudg-SIV express identical SIV Gag and Tat Ags and were generated by homologous recombination of a bicistronic SIV gene expression cassette into deletion site III of either wild-type MVA or MVAΔudg (D. Garber, unpublished observation), respectively. This cassette directs expression, from independent modified H5 vaccinia promoters, SIVmac239 gag (nucleotides 1309–2841 of accession number M33262) and SIVmac239 tat as an in frame fusion among tat exon 1 (nucleotides 6558–6853 of M33262), tat exon 2 (nucleotides 9,062–9,158 of #M33262), and a 3′ (CAT)6 sequence encoding a hexahistidine tag.(http://www.ncbi.nlm.nih.gov/). Similar expression of SIV Gag and Tat Ags from recombinant viruses was confirmed by Western blot analyses of infected cell lysates (not shown). MVA vectors were propagated and titered on a udg-complementing cell line; virus stocks were purified via sucrose gradient centrifugation for use in immunization studies.
The homologous prime-boost-boost regimen used in this study consisted of three immunizations performed at day 0 (prime), after 6 wk (first boost), and then 6 wk after the first boost (second boost). Each immunization was composed of the MVA vector being injected i.m. (1 × 108 PFU) and intradermally (1 × 108 PFU) for a total vaccine dose of 2 × 108 PFU per macaque. One experimental group of five RMs received three immunizations with the parental MVA vaccine (MVA-SIV), while the other experimental group of five RMs received prime and boosts with MVAΔudg-SIV. Five additional animals were left unvaccinated before challenge.
One year after priming with the MVA-based vaccine, both experimental groups and an additional control group of five unvaccinated animals were challenged with SIVmac239 (10,000 fifty percent tissue culture-infective doses (TCID50) prepared from COS-1 cell supernatant after transfection with DNA from a pBR239 full-length SIVmac239 clone) provided by the laboratory of Louis Picker (Oregon Health Sciences University and Oregon National Primate Research Center, Beaverton, OR).
Tissue collection and processing
PBMCs were isolated by gradient centrifugation. Procedures for lymph node (LN) biopsy, rectal biopsies (RB), and BAL as well as isolation of lymphocytes from the obtained samples were performed as previously described (20).
Immunophenotyping and flow cytometry
Multicolor flow cytometric analysis was performed on mononuclear cells isolated from blood, LNs, and mucosal tissues (RB and BAL) according to standard procedures using human mAbs that were found to cross-react with RMs. Predetermined optimal concentrations of the following Abs were used: anti-CD4 PerCP-Cy5.5 (clone L200), anti-CD69 PE (clone L78), anti-CCR5 PE (clone 3A9), anti-CD123 PE (clone 9F5), anti-CD62L (clone SK11), anti-CD14 PE-Cy7 (clone M5E2), anti-CD8 PerCP-Cy5.5 (clone SK1), anti-CD20 PE-Cy5 (clone L27), anti-CD95 PE-Cy5 (clone DX2), anti-CD11c allophycocyanin (clone S-HCL-3), anti-HLA-DR allophycocyanin-Cy7 (clone L243), anti-CD25-allophycocyanin-Cy7 (clone MA251), anti-CD3 Alexa 700 (clone SP34-2), anti-CD8 Pacific Blue (clone RPA-T8), and anti-CD16 Pacific Blue (clone 3G8) all from BD Pharmingen; anti-CD127 PE (clone R34.34; Beckman Coulter); anti-CCR7 FITC (clone 150503; R&D Systems); and anti-CD28 PECy7 (clone 28.2; eBioscience). Flow cytometric acquisition and analysis of samples was performed on at least 100,000 events on a LSRII flow cytometer driven by the DiVa software package (BD Biosciences). Analysis of the acquired data was performed using FlowJo software (Tree Star).
MHC class I tetramers were prepared as previously described (21). The level of SIV-specific CD8+ T cells was assessed using soluble tetrameric Mamu-A*01 MHC class I tetramers specific for SIVmac239 immunodominant epitopes Gag181–189 CM9 (CTPYDINQM) and Tat28–35 SL8 (STPESANL). Tetramers were prepared and conjugated to streptavidin-allophycocyanin fluorophore (Molecular Probes) as previously described (21, 22). Lymphocytes isolated from blood and tissues were incubated with conjugated tetramer, along with surface Ab conjugates, and analyzed for tetramer and surface marker expression using an LSRII flow cytometer (BD Biosciences) equipped with FACS DiVa software.
Multifunctional assessment of SIV-specific T cell responses
The function of SIV-specific CD8+ T cells was assessed by flow cytometry after stimulation with peptide pools of 15-mers (overlapping by 11 aa) spanning the SIVmac239 Gag and Tat proteins (prechallenge pools) or Gag, Tat, Pol, and Nef proteins (postchallenge pools). Peptides were prepared from peptide stocks obtained from the National Institutes of Health AIDS Research and Reference Reagent Program, reconstituted in DMSO, and pooled. All peptides were used at a final concentration of 2 μg of each peptide per ml. Purified PBMCs were thawed, resuspended, and stimulated as previously described (6). Anti-CD107a FITC (BD Biosciences) was added at the start of all stimulation periods, as described previously (23). The mixture of Abs for surface staining included anti-CD4 PerCP Cy5-5 (BD Biosciences), anti-CD8 Qdot 605 (custom conjugated in the Betts laboratory with reagents purchased from Invitrogen, as previously described in Ref. 24), anti-CD14 Pacific Blue (BD Biosciences), anti-CD16 Pacific Blue (BD Biosciences), anti-CD20 Pacific Blue (eBioscience), anti-CD28 ECD (Beckman Coulter), and anti-CD95 PE Cy5 (BD Biosciences). The mixture for intracellular staining included anti-CD3 allophycocyanin Cy7 (BD Biosciences), anti-IFN-γ allophycocyanin (BD Biosciences), anti-IL-2 PE (BD Biosciences), and anti-TNF-a PE Cy7 (BD Biosciences).
Plasma viral load determination
Quantitative real-time RT-PCR assay to determine SIV viral load was performed as previously described (25).
Statistical analysis was performed using Microsoft Excel (Microsoft) and Prism Software (GraphPad). The area under the curve was calculated for the longitudinal immune responses of each animal and compared among MVA-SIV, MVAΔudg-SIV, and unvaccinated groups using ANOVA with Bonferroni multiple comparison adjustments. Measurements at each time point were compared between groups using repeated measures ANOVA with Bonferroni multiple comparison adjustments.
Experimental design: immunization and challenge
As shown in Fig. 1, 15 MaMuA01+ Indian RMs were included in this study. Ten received three homologous immunizations via split intradermal/i.m. routes, at 6-wk intervals, with MVA-SIV (n = 5) or MVAΔudg-SIV (n = 5) vectors that express identical SIVmac239 Gag and Tat Ags; five macaques were left unvaccinated. All animals were challenged i.v. at day 365 postimmunization with 10,000 TCID50 of SIVmac239. The rationale for the use of MamuA*01+ animals was to be able to directly assess SIV-specific CD8+ T cells in all examined tissues. The choice to immunize with SIVmac239 Gag- and Tat-expressing vectors and then to homologously challenge them with SIVmac239 reflects the adoption of a reductionist approach to the potential identification of correlates of immunological protection from disease progression. Fig. 1 also shows the time points postimmunization and postchallenge when the various analyzed tissues (blood, LNs, and MALT, via RB and BAL) were collected.
MVA-SIV and MVAΔudg-SIV induced strong levels of SIV-specific CD8+ T cells in peripheral blood and mucosal tissues
The SIV-specific T cell responses were enumerated in the 10 immunized RMs by tetramer staining for the immunodominant CD8+ T cell epitopes Tat-SL8 and Gag-CM9. As shown in Fig. 2,A, tetramer-positive CD8+ T cells were identifiable after the first immunization in both groups of immunized RMs, and their fraction of nonnaive (i.e., CD95+) CD8+ T cells increased significantly after the second vaccination. Gag- and Tat-specific CD8+ T cells were also detected by tetramer staining in LN-derived mononuclear cells (data not shown). As expected, virtually all tetramer-positive cells in blood and LNs were included within the memory or effector subsets and show variable levels of expression of the proliferation Ag Ki67 (data not shown). The kinetics of the vaccine-induced immune responses revealed a clear increase in the level of circulating SIV-specific CD8+ T cells after the second immunization (i.e., boosting effect) but not after the third immunization (Fig. 2 A). Consistent with previous studies (26, 27), the magnitude of Gag-specific responses was higher than Tat-specific throughout the follow-up period. Although it appears that the MVA-SIV vaccine vector induced slightly higher prechallenge Gag-specific responses, these differences were not significant.
Several recent studies have shown that pathogenic HIV/SIV infections are associated with a rapid, substantial, virus-induced depletion of memory CD4+ T cells from MALT (28, 29, 30, 31, 32). These studies led to the formulation of a model in which the MALT CD4+ T cell depletion plays a key role in the pathogenesis of AIDS (33). As such, a desirable property of a candidate AIDS vaccine would be to generate high numbers of HIV-specific memory CD8+ T cells in the mucosa that may limit this early loss of CD4+ T cells.
To determine the effect of the two tested immunization strategies (i.e., MVA-SIV and MVAΔudg-SIV) on mucosal immune responses, we quantified CD8+ T cell responses against the immunodominant Gag CM9 and Tat SL8 epitopes of SIV by tetramer staining in lymphocytes derived from RB and BAL. RB and BAL were conducted a total of eight times throughout the entire study. Of these eight time points, three were before challenge, i.e., 1 wk post-second and -third immunization (effector phase) and after 11.5 mo (memory phase), and five were postchallenge (Fig. 1). As shown in Fig. 2,B, we found that both MVA-SIV and MVAΔudg-SIV elicited a promptly detectable SIV-specific CD8+ T cell response in RB of the vaccinated animals that, at the 1-wk postimmunization time points (effector phase), was higher than those observed in blood or BAL (RB vs blood or BAL; ANOVA with Bonferroni multiple comparison test, p < 0.05). The levels of Gag-specific CD8+ T cells were 0.22–2.04% in BAL and 0.33–19.7% in RB, and the levels of Tat-specific CD8+ T cells were 0.10–0.77% in BAL and 0.50–15.9% in RB (Fig. 2,B). The highest levels of SIV-specific CD8+ T cells were found in the rectal mucosa, with the majority of these cells displaying a phenotype (CD28−CD95+) indicative of full differentiation to effector or effector-memory T cells (data not shown). As expected, the levels of SIV-specific T cell responses in the MALT were markedly decreased during the memory phase of the vaccine-induced immune response (day 337; Fig. 2 B). Collectively, these data indicate that expression of SIV Ags from MVA vectors (both udg+ and Δudg) effectively induces SIV-specific CD8+ T cells in the MALT.
MVA-SIV and MVAΔudg-SIV induced moderate levels of multifunctional SIV-specific T cell responses in the blood
To characterize the functional development of vaccine-induced SIV-specific T cell responses over time, we stimulated PBMCs from every RM with overlapping peptide pools of 15-mers from SIVmac239 Gag and Tat and assessed the production of IFN-γ, TNF-α, and IL-2, and the expression of CD107a using multiparametric flow cytometry. Both MVA-SIV- and MVAΔudg-SIV-immunized animals showed a detectable, albeit relatively low, number of mostly mono- and bifunctional SIV-specific CD8+ T cells (Fig. 3) and CD4+ T cells (not shown).
SIVmac239 challenge of vaccinated RMs results in more rapid generation and higher frequency of SIV-specific CD8+ T cells than in unvaccinated animals
One year after the first immunization with either MVA-SIV or MVAΔudg-SIV vaccines, all animals were challenged i.v. with 10,000 TCID50 of the CCR5-tropic SIVmac239 virus, which has been shown to be highly pathogenic in RMs (34). A control group of five unvaccinated RMs was infected simultaneously with the same virus stock and at the same dose. As expected, all 15 animals were successfully infected (see next paragraph and Fig. 5 A for more details).
The vaccinated animals exhibited greater postchallenge increase of circulating Gag-specific CD8+ T cells than the unvaccinated control animals (Fig. 4,A; day 14, ANOVA with Bonferroni multiple comparison test, p < 0.05). Gag-specific responses peaked and were significantly higher at day 10 in the vaccinated animals (ANOVA with Bonferroni multiple comparison test, p < 0.01), whereas the same responses in the unvaccinated animals did not peak until day 14. Whereas Tat-specific responses were generally higher in vaccinated RMs than in unvaccinated animals, only the MVA-SIV-vaccinated animals exhibited significantly higher peak responses (day 10 and 14; ANOVA with the Bonferroni multiple comparison test, p < 0.05). Similarly, the level of SIV-specific T cell responses was higher in the LNs of vaccinated RMs at their peaks (Fig. 4 B; ANOVA with the Bonferroni multiple comparison test, p < 0.05) except for Tat-specific responses elicited by the MVA-SIV which were indistinguishable from the unvaccinated controls.
This trend toward faster and greater SIV-specific CD8+ T cell responses in vaccinated animals was also observed in MALT-derived lymphocytes (Fig. 4, C and D). Specifically, we found that 1) MVA-SIV-vaccinated RMs have higher levels of both Gag- and Tat-specific CD8+ T cells in BAL-derived (but not in RB-derived) lymphocytes when compared with unvaccinated animals (day 14; ANOVA with the Bonferroni multiple comparison test, p < 0.001); and that 2) MVAΔudg-SIV-vaccinated RMs had higher levels of both Gag- and Tat-specific CD8+ T cells in both examined mucosal tissues (day 14; ANOVA with the Bonferroni multiple comparison test, p < 0.05). Postchallenge responses were largely monofunctional in both unvaccinated (Fig. 4,E) and vaccinated (Fig. 4 F) animals. Although the in vivo biological meaning of this latter finding is unclear, it is consistent with the recent observation that vaccine-elicited SHIV-specific CTL responses become indistinguishable from those of unvaccinated controls after challenge (35). Taken together, these data indicate that the current immunization strategy was effective in eliciting a rapid and robust generation of large numbers of SIV-specific CD8+ T cells in all examined tissues after challenge with a pathogenic virus.
MVA-immunized RMs show reduced levels of SIV viral load postchallenge
To directly assess the ability of vaccine-induced (or, in the control animals, the virus-induced) immune responses to control SIV replication, we longitudinally measured plasma viral load in all 15 animals by RT-PCR. As shown in Fig. 5,A, both immunized and control animals showed the typical dynamic of virus replication observed after experimental SIVmac239 infection, with peak viremia recorded at 10 days postchallenge and subsequent decline to a set-point level by wk 8. Importantly, the level of viremia was ∼1 log lower in vaccinated RMs (area under the curve of MVA-SIV or MVAΔudg-SIV vs unvaccinated between days 0 and 168 postchallenge; ANOVA with Bonferroni multiple comparison test, p < 0.05), and this difference persisted throughout most time points during the first 6 mo of follow-up (Fig. 5 B). However, viremia in the vaccinated groups converged with that of the unvaccinated controls at the 8-wk postinfection time point as well as later time points (i.e., after 6 mo or longer).
MVA-immunized RMs are not protected from systemic or mucosal CD4+ T cell depletion
To further evaluate the potential protective effect of our MVA-based vaccine on disease progression, we next examined the dynamics of circulating, LN, and mucosal CD4+ T cells in our cohort of RMs. As shown in Fig. 5, C and D, we found that the kinetics of CD4+ T cell decline after challenge was similar in vaccinated and unvaccinated animals. This result was the same when the effect of SIVmac239 challenge on CD4+ T cells was measured as absolute CD4+ T cell counts/mm3 (Fig. 5,C), fraction of CD3+CD4+ T cells out of the total T cell population (data not shown), or fraction from baseline (data not shown). Similarly, no differences were found between vaccinated and unvaccinated RMs in the fraction of LN-derived CD4+ T cells (Fig. 5,D). Given the importance of early CD4+ T cell dynamics in the MALT during HIV and SIV infection (28, 29, 30, 31, 32), we next measured longitudinally the fraction of CD3+CD4+ T cells in the RB and BAL of all infected animals. As shown in Fig. 5 D, both vaccinated and unvaccinated RMs manifested a rapid and strong decline of mucosal CD4+ T cells in both sampled sites. These results are consistent with previously reported findings (30, 32) and also indicate that the SIV-specific CD8+ T cells induced by the current immunization regimen failed to preserve these anatomic compartments from the same type of early CD4+ T cell depletion observed in unvaccinated animals.
Similar levels of T cell activation after SIV challenge in vaccinated and unvaccinated macaques
The observations that the level of immune activation is a strong correlate of disease progression in HIV-infected individuals (36, 37) and that natural SIV hosts do not progress to AIDS despite high viral loads in the presence of low immune activation (25, 38) support the hypothesis that immune activation plays a critical role in AIDS pathogenesis. In these experiments, we sought to assess the impact of previous vaccination on the postchallenge level of T cell activation by measuring longitudinally the expression of Ki67, a marker of recruitment to cell cycle, in CD4+ and CD8+ T cells. As shown in Fig. 5, E and F, we found that the overall level of T cell activation over the course of the study is largely similar between vaccinated and unvaccinated RMs, with the only exception of a significantly increased fraction of Ki67+CD4+ T cells at day 28 postchallenge in vaccinated animals (Mann-Whitney U test, p < 0.005). These results indicate that prior immunization to SIV Ags has no major impact on postchallenge immune activation.
No differences in overall survival between vaccinated macaques and controls
Ultimately, the efficacy of any immunization protocol aimed at protecting against a given infection must be measured in terms of survival after pathogenic challenge. In these experiments, the survival rate of each of the three groups of RMs (MVA-SIV vaccinated, MVAΔudg-SIV vaccinated, and controls) was similar (Fig. 6), thus indicating that the effects of the vaccine in 1) eliciting more rapid and greater magnitude SIV-specific CD8+ T cell responses upon challenge and 2) reducing viremia of ∼1 log in the early stages of infection were not sufficient to protect the animals from simian AIDS and death.
The magnitude of SIV-specific CD8+ T cell responses correlates with viral load but not with CD4+ T cell decline
To further investigate whether the SIV-specific CD8+ T cell responses (either pre- or postchallenge) exert some protective effect from SIV disease progression, we performed a series of correlation analyses between the level of these responses and the key virological and immunological markers of disease progression. In these analyses, all 15 macaques were included. We first assessed the relationship between SIV-specific T cell responses and virus replication and found a significant inverse correlation between peak viral load and the peak level of Gag-specific CD8+ T cells in blood, LNs, and RBs (Fig. 7 A). These correlations reflect the discreet, qualitative differences (magnitude of viral load and Gag-specific CD8+ T cells) between vaccinated and unvaccinated animals rather than a true continuum of protective effect, because the correlations did not hold when the unvaccinated animals were removed from the analysis (data not shown). No correlation was found between the peak level of SIV Tat-specific CD8+ T cells and either the peak viral load or set point (day 60 postchallenge) viral load (data not shown). This apparent lack of effect of Tat-specific CD8+ T cells on viral load could be due to the rapidity with which the Tat SL8 CTL epitope acquires escape mutations early during SIVmac239 infection of RMs (39). Collectively, this set of data suggests that vaccine- and/or infection-induced SIVgag-specific CD8+ T cell responses play a role in controlling virus replication in the infected animals.
We next examined a possible relationship between SIV-specific CD8+ T cell responses and preservation of the circulating or mucosal CD4+ T cell pool. However, we could not find any significant correlation of the peak level of Gag-specific CD8+ T cells in blood, LNs, BAL, or RBs and the depletion of CD4+ T cells in the corresponding anatomic site (measured as percentage of decline from baseline levels; Fig. 7 B). Similarly, no correlation was found when the prechallenge level of Gag-specific T cells was examined, or when the level of Tat-specific CD8+ T cells (either pre- or postchallenge) was considered (data not shown). In all, these results indicate a lack of any significant relationship between the magnitude of SIV-specific CD8+ T cell responses and the preservation of CD4+ T cell homeostasis.
Control of SIVmac239 viremia is associated with partial reconstitution of mucosal CD4+ T cells, but not with higher levels of SIV-specific CD8+ T cells
As shown in Fig. 8,A, 4 RMs (3 vaccinated and 1 unvaccinated) controlled plasma viremia to below 1000 copies/ml by the end of the study follow-up. These controller animals also displayed, during the chronic phase of infection, a substantial repopulation of mucosal CD4+ T cells as well as a better preservation of peripheral blood CD4+ T cell counts (Fig. 8,B). This control of SIV replication was not associated with either increased levels of Gag- and Tat-specific CD8+ T cells (Fig. 8,C), or with a more polyfunctional phenotype of these cells (Fig. 8 D). The CD8+ T cells of controller RMs showed a higher proportion of IL-2-producing SIV-specific CD8+ T cells. Taken together, these data do not support the hypothesis that, in this group of SIV-infected RMs, the achievement of a controller status was due to the presence of greater or more functional virus-specific CD8+ T cell responses.
Despite strong evidence that CTL may partially control HIV/SIV replication, the generation of an effective, CTL-based HIV/AIDS vaccine has proven to be elusive. This difficulty has been recently underscored by the negative results of the Merck STEP phase IIb test of concept trial of AIDS vaccination using an Ad5-based vector system (40). Although disappointing, this result emphasizes the importance of improving our understanding of the interaction between primate lentiviruses and the host immune system, a goal most likely to be obtained in preclinical AIDS vaccine studies in RMs.
In the current study, we challenged with SIVmac239 a total of 15 MaMuA01+ Indian RMs, of which 10 were previously immunized with either MVA-SIV or MVAΔudg-SIV vectors that express SIV Gag and Tat Ags. The aims of this study were to identify potential differences in the quantity, quality, or anatomic distributions of SIV-specific CD8+ T cells that were elicited by these two MVA vectors and to compare the SIV-specific T cell responses that were elicited after SIVmac239 challenge of both vaccinated and unvaccinated macaques, with the goal of further defining correlates of immune protection from SIV disease progression.
We elected to challenge our RMs with a high-dose highly pathogenic i.v. viral inoculum to test the protective effect of the vaccine-induced SIV-specific T cell responses in the context of a worst-case-type scenario. Although this approach sets the bar very high in terms of vaccine efficacy, it has the important advantage of avoiding the unrealistically promising result that is often obtained when less stringent challenges are used. We are also aware that the choice of using MaMuA01+ RMs does not reflect the MHC-disparate state of the human (or nonhuman primate) population. We reasoned, however, that in this experimental setting the possibility of promptly identifying SIV-specific CD8+ T cells by multiparametric flow cytometry outweighs the bias in the experimental system introduced by the presence of MaMuA*01 allele. Finally, due to the limited availability of mucosal tissues, we chose to use these samples to conduct immunophenotypic analysis with tetramer staining and thus to limit our functional T cell studies to PBLs.
In this study, we observed that both MVA-SIV and MVAΔudg-SIV raised a robust, although largely mono- or bifunctional SIV-specific T cell response in peripheral blood and, encouragingly, that our immunization protocols induced high levels of SIV-specific CD8+ T cells in mucosal tissues. As expected, CD8+ T cells responses postchallenge were stronger and more rapid (in peripheral blood, LNs, and MALT) in the vaccinated animals than in the unvaccinated controls. This latter finding indicates that our immunogens successfully elicited the type of immune memory capable of mounting an anamnestic T cell response after challenge (i.e., faster, greater magnitude, and more widespread) that should be key for the success of any vaccination regimen. Interestingly, the two vaccines induced similar anti-SIV responses, likely because the suppression of late gene expression through udg deletion had little effect on the immunogenicity of the SIV Ags. Taken together, the results of the prechallenge phase of our study appeared to indicate a satisfactory profile in terms of immunogenicity, which was then confirmed by the observation that vaccinated RMs showed a ∼1-log decrease in both peak and early set-point viral load when compared with unvaccinated animals. We emphasize that this level of protection, although certainly far from optimal, is comparable with the protection elicited in MamuA*01+ RMs by DNA-prime/Ad5-boost immunization regimens using SIV Gag as the sole SIV Ag (17) or by a DNA-prime/MVA-boost expressing a combination of CTL epitopes and full-length proteins (Tat, Rev, and Nef; Ref. 26). In contrast, in another study (27), a more persistent (>1 year) and robust suppression of SIV replication was observed in MamuA*01+ RMs immunized with a DNA-prime/Ad5-boost regimen expressing multiple SIV Gag Ags and then repeatedly challenged with low dose of SIVmac239.
In the current study, the partial control of virus replication observed in vaccinated RMs did not protect these animals from progressive CD4+ T cell depletion (in either blood or mucosal tissues) or AIDS-related death. In this regard, our results are not entirely consistent with those of a recently published study in which reduction of virus replication in RMs vaccinated with DNA prime Ad5 boost was found to be associated with preservation of mucosal memory CD4+ T cells (41). This discrepancy could be explained in several ways, including: 1) our immunogens were not designed to elicit anti-SIV Abs; 2) different mucosal tissues were sampled (i.e., jejunum vs rectum); 3) the magnitude of SIV-specific T cells in the rectal mucosa was lower in our study; 4) the postchallenge SIV-specific T cell responses that we measured were for the most part monofunctional; and 5) the level of mucosal SIV-specific CD8+ T cells declined substantially before challenge. Also, however, even in the study by Mattapallil (41) a significant degree of overlap was observed between vaccinees and controls in the level of mucosal CD4+ T cell depletion postchallenge, thus suggesting either significant interindividual variability and/or inherent limitations in the protective effect of the used vaccine. A discrepancy between control of virus replication and protection from mucosal CD4+ T cell depletion was also observed in a recent study in which RMs were vaccinated with a DNA/MVA prime/boost regimen and subsequently challenged with SIVmac251 (42). Ultimately, the fact that the vaccine-induced reduction in acute and early chronic viral load was limited (∼1 log) and transient (∼8 wk) may be the most logical explanation for the lack of protection from disease progression in the vaccinated animals included in the current study.
The observation that vaccine-induced cellular immune responses may reduce virus replication without protecting from disease progression is disappointing but not entirely surprising, especially when one considers the emerging complexity of the interaction between primate lentiviruses and the host immune system. For instance, it has now become more appreciated that although the level of virus replication predicts the rate of progression to AIDS in HIV-infected patients (43), this correlation is actually rather weak, given that virus replication is only a relatively ineffective predictor of the progression of immunodeficiency (44). Lack of relationship between SIV replication and disease progression is also a typical feature of natural SIV infection of sooty mangabeys and African green monkeys, in which the animals remain asymptomatic despite high levels of viremia (25, 45, 46).
In addition, although control of virus replication and preservation of systemic and mucosal CD4+ T cells are key determinants of survival during pathogenic HIV/SIV infections, the extent to which and the mechanisms by which these two phenomena are linked to each other are still unclear (47). Much emphasis has been recently placed on unconventional (i.e., non-directly virus-mediated) mechanisms of CD4+ T cell depletion during HIV infection of humans and SIV infection of macaques (48). Among these mechanisms, the apoptosis of bystander, uninfected CD4+ T cells may be prominent (49, 50). One intriguing possibility raised by our study is that the observed disconnection between control of viral load and disease progression is caused by the presence of different mechanisms of CD4+ T cell depletion in vaccinated vs unvaccinated animals. In this line of thinking, the CD4+ T cell depletion of unvaccinated RMs would be predominantly virus-induced whereas the depletion in vaccinated animals would be at least in part caused by CTL and/or involve higher levels of uninfected cells. Indeed, the higher levels of nonspecific CD4+ T cell activation seen in vaccinated animals (Fig. 5 E) could contribute significantly to depletion via activation-induced cell death, despite lower viral loads and, presumably, lower levels of virus-induced direct cytopathicity.
Finally, four of our SIVmac239-infected RMs appeared to control virus replication, and this control correlated neither with vaccination nor with the presence of levels of SIV-specific CD8+ T cells (Fig. 8). Such spontaneous control of virus replication is not infrequent among MamuA01+ RMs, although it is usually attributed to more effective responses of MamuA01-restricted CD8+ T cells (51).
For an AIDS vaccine to be effective, the immune responses generated by the vaccine need to be significantly better than those generated during natural infection. Although admittedly it is still unclear what are the exact correlates of this improved immunogenicity, protection from AIDS will likely require the induction of rapid and potent HIV-specific cellular immune responses in mucosal tissue that will limit early virus replication in the areas of virus transmission and early dissemination. The observation that our current vaccination protocol, despite its obvious immunogenicity at both the systemic and mucosal levels, did not provide the necessary protection from high-dose SIVmac239 challenge further emphasizes the tremendous difficulty of controlling retrovirus replication by the primate immune system. In this perspective, we believe that the presented results, although somewhat sobering, are important because they highlight the absolute need to define strong, robust, and reliable correlates of immune protection from HIV disease progression.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This research was supported by Integrated Preclinical and Clinical AIDS Vaccine Development Grant U19-AI 061728 (to J.D.A., S.I.S., D.A.G., M.B.F., and G.S.) and in part by the Virology and Drug Discovery Core of the Emory Center for AIDS Research (Grant P30 AI050409) and Grant P51-RR00165 to the Yerkes National Primate Research Center.
Abbreviations used in this paper: Ad, adenovirus; MVA, modified vaccinia Ankara; RM, rhesus macaque; BAL, bronchoalveolar lavage; TCID50, 50% tissue culture-infective dose; LN, lymph note; RB, rectal biopsy.