|Home | About | Journals | Submit | Contact Us | Français|
All human immunodeficiency virus (HIV) vaccine efficacy trials to date have ended in failure. Structural features of the Env glycoprotein and its enormous variability have frustrated efforts to induce broadly reactive neutralizing antibodies. To explore the extent to which vaccine-induced cellular immune responses, in the absence of neutralizing antibodies, can control replication of a heterologous, mucosal viral challenge, we vaccinated eight macaques with a DNA/Ad5 regimen expressing all of the proteins of SIVmac239 except Env. Vaccinees mounted high-frequency T-cell responses against 11 to 34 epitopes. We challenged the vaccinees and eight naïve animals with the heterologous biological isolate SIVsmE660, using a regimen intended to mimic typical HIV exposures resulting in infection. Viral loads in the vaccinees were significantly less at both the peak (1.9-log reduction; P < 0.03) and at the set point (2.6-log reduction; P < 0.006) than those in control naïve animals. Five of eight vaccinated macaques controlled acute peak viral replication to less than 80,000 viral RNA (vRNA) copy eq/ml and to less than 100 vRNA copy eq/ml in the chronic phase. Our results demonstrate that broad vaccine-induced cellular immune responses can effectively control replication of a pathogenic, heterologous AIDS virus, suggesting that T-cell-based vaccines may have greater potential than previously appreciated.
It has been impossible thus far for vaccines to engender broadly reactive neutralizing antibodies against human immunodeficiency virus (HIV) (12, 54). Investigators have therefore focused their attention on T-cell-based vaccines (9, 18, 26, 30, 34, 39, 48, 55). Previous preclinical studies in nonhuman primates have shown that vaccine-induced T-cell responses can partially control replication of homologous challenge viruses in the chronic phase (34, 56). Unfortunately, however, simian immunodeficiency virus (SIV) loads exceeded 1 million copies in almost every vaccinated animal during the acute phase. Given the high levels of viral replication observed in these vaccinated macaques, it is possible that such T-cell-based vaccines might not be able to reduce transmission during the acute phase of infection in humans. These high levels of replication during the acute phase likely resulted in the generation of diverse viral quasispecies, providing the substrate for immune selection and eventual escape. Furthermore, in these studies, vaccinated animals were challenged with viruses that were similar to the SIV sequences in the vaccine constructs. Given the diversity of HIV, human vaccinees will never be exposed to viruses with a comparable level of sequence similarity to the vaccine constructs.
An HIV-1 vaccine that induced T-cell responses exclusively has recently failed to show efficacy against the incidence of HIV infection and viremia in clinical testing. The STEP trial of a recombinant adenovirus 5 (Ad5)-vectored vaccine designed to induce HIV-specific T-cell responses in humans was widely seen as an important test of the T-cell vaccine concept (http://www.hvtn.org/media/pr/step111307.html) (11, 42). The lack of vaccine efficacy in the STEP trial has led some to conclude that T-cell-based vaccines may not be a viable approach to solving the AIDS epidemic (6, 49, 59). However, STEP trial vaccinees that became infected recognized a median of only five epitopes, mostly in the conserved proteins Gag and Pol. Given the sequence diversity of HIV (19), several of these vaccine-elicited T-cell responses may not have recognized epitopes in the infecting virus and, therefore, not constituted an adequate test of the T-cell vaccine concept.
We therefore sought to test whether high-frequency vaccine-induced T-cell responses against multiple T-cell epitopes in one of the simian AIDS viruses, SIVmac239, could effectively impact viral replication after a physiologically relevant heterologous mucosal challenge with SIVsmE660. The majority of virus challenges in macaques have been carried out with high doses of homologous viruses. We used a repeated low-dose mucosal challenge with a heterologous SIV strain. We also used a challenge dose intended to mimic HIV mucosal exposures that lead to infection. Here we show that vaccine-induced T-cell responses can reduce heterologous virus replication during both the acute and chronic phases after a relevant viral challenge.
The animals in this study were Indian rhesus macaques (Macaca mulatta) from the Wisconsin National Primate Research Center colony. They were typed for major histocompatibility complex class I (MHC-I) alleles Mamu-A*01, Mamu-A*02, Mamu-A*08, Mamu-A*11, Mamu-B*01, Mamu-B*03, Mamu-B*04, Mamu-B*08, Mamu-B*17, and Mamu-B*29 by sequence-specific PCR analysis (28, 36). Animals that were Mamu-A*02 positive were chosen for the study, but animals positive for Mamu-A*01, Mamu-B*08, or Mamu-B*17 were excluded. It has been observed that the presence of either the Mamu-B*08 or Mamu-B*17 allele alone is correlated with a reduction in plasma viremia (36, 60). The animals were cared for according to the regulations and guidelines of the University of Wisconsin Institutional Animal Care and Use Committee.
We synthesized genes coding for SIVmac239 Gag, Tat, Rev, Nef, Pol, Vif, Vpr, and Vpx based on codons frequently used in mammalian cells (51) and cloned these into the V1R vector. Five milligrams of plasmid DNA formulated with 7.5 mg of CRL1005 mixed with benzalkonium chloride was injected intramuscularly (i.m.) in a volume of 0.5 ml. Before the i.m. injection, the formulation was warmed slowly to room temperature from a frozen stock. DNA was injected at the following six different sites: the left and right triceps, the left and right quadriceps, and the left and right gastrocnemius. All plasmids were injected at a unique site, with the exception of Vif, Vpr, and Vpx. Although these were formulated on separate plasmids, all three injections were given at the same site. Sites to which each particular plasmid was administered were rotated every injection (including the Ad5 boost) to ensure that no site received a vector encoding the same protein twice. For instance, if the vector encoding Gag was injected into the right quadriceps for the first DNA prime, on the next prime, this vector would be injected into a different site, perhaps the left quadriceps, and then on the third prime into yet a different site, such as the right triceps, et cetera. In this way, the draining lymph nodes are exposed to different encoded proteins each time and are not reboosted with the same protein sequence, potentially inducing immunodominance. The same codon-optimized genes were cloned into an adenoviral vector based on Ad5 that had been rendered incompetent to replicate by deletions of the E1 viral gene and subsequently propagated in E1-expressing PER.C6 cells as previously described (51). Six different Ad5 vectors were created; five contained the single coding sequences for Gag, Tat, Rev, Nef, and Pol, while the sixth encoded Vif, Vpr, and Vpx, all very small proteins. Viral particles (1011) of Ad5 were delivered i.m. in a volume of 0.5 ml to the same six sites as used previously for DNA priming, again rotating the site to which each protein sequence was administered.
Thirty-seven weeks after vaccination was completed, the vaccinees and eight naïve control animals were challenged intrarectally with SIVsmE660 (800 50% tissue culture infective doses [TCID50]; 1.2 × 107 SIV RNA copy eq). We used a repeated mucosal challenge regimen (40) with SIVsmE660 to more closely mimic HIV transmission in humans. Animals were initially challenged with this dose every 3 weeks, up to five times. Animals were considered to be SIV positive after at least two subsequent positive viral load determinations and were no longer challenged after it was determined that they were positive. After five such challenges, animals that were still SIV negative were challenged intrarectally every other week with a dose of 4,000 TCID50. It took up to six additional challenges before all animals were infected.
Levels of circulating plasma virus were determined using a previously described quantitative reverse transcription-PCR assay (45). Virus concentrations were determined by interpolation onto a standard curve of in vitro-transcribed RNA standards in serial 10-fold dilutions using the LightCycler 2.0 (software version 4; Roche).
Fresh peripheral blood mononuclear cells (PBMCs) isolated from EDTA-anticoagulated blood were used in enzyme-linked immunospot (ELISPOT) assays for the detection of gamma interferon (IFN-γ)-secreting cells as previously described (35). As in previous publications, to determine positivity, the number of spots in duplicate wells (100,000 cells per well) was averaged, and the background was subtracted. This resultant number is required to be greater than five spots (50 spot-forming cells [SFC]/million PBMCs) and also greater than 2 standard deviations over the background. Positive results are multiplied by 10 to get SFCs/million cells. Some of the peptides used in these assays were obtained through the AIDS Research and Reference Reagent Program, the Division of AIDS, the National Institute of Allergy and Infectious Diseases, National Institutes of Health. Additionally, we examined responses by CD8-negative cells by depleting PBMCs of CD8+ cells using a CD8 MicroBead kit for nonhuman primates (Miltenyi Biotec) according to the manufacturer's instructions.
We monitored CD8+ and CD4+ subsets and memory populations by staining PBMCs with fluorescently labeled antibodies specific for CD3 Alexa 700 (BD Pharmingen, San Diego, CA), CD4 PerCP (Miltenyi, Auburn, CA), CD8 Pacific Blue (BD Pharmingen), CD95 fluorescein isothiocyanate (BD Pharmingen), CD28 phycoerythrin (Becton Dickinson, San Jose, CA), and beta7 integrin allophycocyanin (Becton Dickinson). In brief, 500,000 PBMCs were incubated with these antibodies for 30 min at room temperature. The samples were then washed twice, fixed with paraformaldehyde, and run on a BD-LSR-II flow cytometer (Becton Dickinson) using FACSDiva software (BD Biosciences). Data were analyzed using FlowJo software. Absolute counts were calculated by multiplying the frequency of CD3+/CD4+ T cells or CD3+/CD4+/CD95 plus CD28+ T cells (central memory) or CD3+/CD4+/CD95 plus CD28+ T cells (effector memory) within the lymphocyte gate by the lymphocyte counts per microliter of blood obtained from matching complete blood counts.
From each plasma specimen, approximately 20,000 viral RNA (vRNA) copies were extracted using the QIAamp vRNA mini kit (Qiagen, Valencia, CA). RNA was eluted and immediately subjected to cDNA synthesis. Reverse transcription of RNA to single-stranded cDNA was performed using SuperScript III reverse transcriptase according to the manufacturer's recommendations (Invitrogen Life Technologies, Carlsbad, CA). Briefly, each cDNA reaction mixture included 1× reverse transcription buffer, 0.5 mM of each deoxynucleoside triphosphate, 5 mM dithiothreitol, 2 U/ml RNaseOUT (RNase inhibitor), 10 U/ml of SuperScript III reverse transcriptase, and 0.25 mM antisense primer SIVsm/macEnvR1 (5′-TGTAATAAATCCCTTCCAGTCCCCCC-3′; nucleotides [nt] 9454 to 9479 in SIVmac239). The mixture was incubated at 50°C for 60 min, followed by an increase in temperature to 55°C for an additional 60 min. The reaction was then heat inactivated at 70°C for 15 min and treated with 2 U of RNase H at 37°C for 20 min. The newly synthesized cDNA was used immediately or frozen at −80°C.
cDNA was serially diluted and distributed among wells of replicate 96-well plates to identify a dilution where PCR-positive wells constituted less than 30% of the total number of reactions, as previously described (29, 47). At this dilution, most wells contain amplicons derived from a single cDNA molecule. This was confirmed in every positive well by direct sequencing of the amplicon and inspection of the sequence for mixed bases (double peaks), which would be indicative of priming from more than one original template or the introduction of PCR error in early cycles. Any sequence with mixed bases was excluded from further analysis. PCR amplification was carried out in the presence of 1× High Fidelity Platinum PCR buffer, 2 mM MgSO4, 0.2 mM of each deoxynucleoside triphosphate, 0.2 μM of each primer, and 0.025 U/μl Platinum Taq High Fidelity polymerase in a 20-μl reaction mixture (Invitrogen, Carlsbad, CA). First-round PCR primers included sense primer SIVsm/macEnvF1 (5′-CCTCCCCCTCCAGGACTAGC-3′; nt 6127 to 6146 in SIVmac239) and antisense primer SIVsm/macEnvR1 (5′-TGTAATAAATCCCTTCCAGTCCCCCC-3; nt 9454 to 9479 in SIVmac239), which generated a ~3.3-kb amplicon. PCR analysis was performed in MicroAmp 96-well reaction plates (Applied Biosystems, Foster City, CA) with the following PCR parameters: 1 cycle of 94°C for 2 min, 35 cycles of a denaturing step of 94°C for 15 s, an annealing step of 55°C for 30 s, and an extension step of 68°C for 4 min, followed by a final extension of 68°C for 10 min. Next, 2 μl from the first-round PCR product was added to a second-round PCR that included the sense primer SIVsmEnvF2 (5′-TATGATAGACATGGAGACACCCTTGAAGGAGC-3′; nt 6292 to 6323 in SIVmac239) or SIVmacEnvF2 (5′-TATAATAGACATGGAGACACCCTTGAGGGAGC-3′; nt 6292 to 6323 in SIVmac239) and antisense primer SIVsmEnvR2 (5′-ATGAGACATRTCTATTGCCAATTTGTA-3′; nt 9413 to 9439 in SIVmac239). The second-round PCR was carried out under the same conditions that were used for the first-round PCR but with a total of 45 cycles. Amplicons were inspected on precast 1% agarose E-Gel 96 (Invitrogen Life Technologies, Carlsbad, CA). All PCR procedures were carried out under PCR clean-room conditions using procedural safeguards against sample contamination, including prealiquoting of all reagents, use of dedicated equipment, and physical separation of sample processing from pre-PCR and post-PCR amplification steps.
Env gene amplicons were directly sequenced by cycle-sequencing using BigDye Terminator chemistry and protocols recommended by the manufacturer (Applied Biosystems, Foster City, CA). Sequencing reaction products were analyzed with an ABI 3730XL genetic analyzer (Applied Biosystems, Foster City, CA). Both DNA strands were sequenced by using partially overlapping fragments. Individual sequence fragments for each amplicon were assembled and edited using the Sequencher v4.7 program (Gene Codes, Ann Arbor, MI). Inspection of individual chromatograms allowed for the confirmation of amplicons derived from single versus multiple templates. The absence of mixed bases at each nucleotide position throughout the entire env gene was taken as evidence of SGA from a single vRNA/cDNA template. This quality control measure enabled us to exclude from the analysis amplicons that resulted from PCR-generated in vitro recombination events or Taq polymerase errors and to obtain multiple individual env sequences that proportionately represented those circulating in vivo.
All alignments and phylogenetic trees were made with ClustalW (52). Alignments are shown in Highlighter, which is a sequence analytical tool (http://www.hiv.lanl.gov) that displays the location and identity of nucleotide substitutions in a visually informative manner and allows tracing of common ancestry between sequences based on individual nucleotide polymorphisms.
Enrichment for mutations with APOBEC signatures was assessed by visual inspection of the sequences and confirmed using Hypermut 2.0 (http://www.hiv.lanl.gov).
Plasma viral concentrations were log transformed to reduce right-skewness and heteroscedasticity. Peak and average viral loads were computed for each animal from day 28 through 140, with mostly complete data present through day 112. P values were computed using two different statistical tests which compare controls to vaccinees. The Welch two-sample t test tests the null hypothesis of equal means against the two-sided alternative hypothesis, without assuming equal variance. The permutation test is a nonparametric distribution-free test of identical distributions across the two groups. Two-tailed P values under 0.05 are indicated by footnote d in Table Table33.
All 429 env sequences from ramp-up and peak viremia and 32 env sequences from the inoculum stock were deposited in GenBank (accession numbers FJ847845 to FJ847927).
We vaccinated eight Indian rhesus macaques with a DNA prime Ad5 boost regimen encoding all of the SIVmac239 proteins except Env (Fig. (Fig.1).1). We excluded any animals expressing the MHC-I allele Mamu-A*01, Mamu-B*08, or Mamu-B*17 that has been implicated in the control of viral replication (36). All of these eight vaccinees and an additional eight naïve control macaques expressed Mamu-A*02 so that we could follow CD8+ T-cell responses against a variety of well-defined CD8+ T-cell epitopes (37) (see Table S1 in the supplemental material). We measured vaccine-induced T-cell responses by performing whole-proteome ELISPOT assays with whole PBMCs at day 14 after the Ad5 boost. We used 83 pools each consisting of 10 15-mers (overlapping by 11 amino acids) spanning the nine SIV open reading frames. We subsequently mapped as many responses as we could (depending on blood availability) down to single 15-mer peptides in the ensuing weeks. Unfortunately, we were not able to distinguish between CD4+ and CD8+ T-cell responses during the vaccination stage. The vaccine induced high-frequency T-cell responses against 11 to 34 epitopes (Table (Table1)1) .
Interestingly, all vaccinated macaques made strong CD8+ T-cell responses against the Mamu-A*02-restricted epitope Env RY8(788-795) even though we did not have Env in the vaccine. Sequencing of all vaccine constructs revealed no env sequences, and Western blots were negative for antibody against Env (see Fig. S1 in the supplemental material). Since the envelope protein itself was not encoded by any of the vaccine vectors, we hypothesized that the Env RY8 epitope may have been derived from translation of an alternate reading frame of the Rev-encoding plasmid and Ad5 vector. To test this hypothesis directly, we transfected Mamu-A*02-expressing 721.221 cells with the Rev or the Gag plasmid or an Env-expressing plasmid not included in the vaccine. Twenty-four hours later, we tested whether these cells could present the RY8 epitope to a RY8-specific cytotoxic T-lymphocyte (CTL) line. The RY8-specific CTL line recognized the Rev- and Env-transfected cells but not the Gag-transfected cells, indicating that the Rev plasmid was the source of the RY8 epitope and likely the RY8-directed response in the vaccinated animals (data not shown). The sequence of the Rev plasmid does not contain an alternate methionine start codon upstream of the RY8 epitope. However, translation of the epitope could be due to ribosomal slippage or translation initiation at a non-Met start codon. In SIVsmE660, there are four substitutions in this Env RY8 (see Table S1 in the supplemental material), and vaccine-induced T-cell responses against this epitope were not expanded after challenge. In one animal, r02114, there was an additional response to an Env peptide in the N-terminal region of the protein, mapped to a 15-mer, NATIPLFCATKNRDT, Env(37-51). This response was sustained well into the vaccine phase, but was not expanded after the challenge. Thus, it is unlikely that these Env-specific CD8 T-cell responses affected the outcome after the challenge.
We then attempted to mimic clinically relevant HIV exposure in humans. To date, the majority of nonhuman primate vaccine trials have used challenge viruses homologous to the vaccine, and many of those have involved mucosal challenges that have been carried out using a high-dose inoculum of challenge virus to ensure that all control naïve animals become infected. In contrast, the majority of vaccinated humans will likely be exposed to multiple lower doses of a virus that is significantly different from the vaccine. Furthermore, this exposure will take place across a mucosal surface. Therefore, to model human exposure to HIV, we developed a repeated rectal challenge in macaques using a swarm virus (SIVsmE660 [10, 20, 25]). This virus is considerably different from the SIVmac239 viral sequences in the vaccine, and we used a dose that mimics the biology of HIV infection. HIV isolates within a single clade typically vary by as much as 10 to 20%, depending on the gene of reference (19). Vaccinated humans will likely encounter viruses that differ by at least this much from the vaccine strain. The sequence of SIVsmE660 differs from the SIVmac239-derived vaccine sequences by approximately 15% of its amino acids. The percent difference in amino acid sequences between SIVmac239 and SIVsmE660 is as follows: for Tat, 26%; Rev, 25%; Nef, 21.3%; Vif, 17%; Env, not in the vaccine; Vpr, 12%; Pol, 8.3%; Pol-Pro, 11.1%; Pol-RT, 6.2%; Pol-Int, 5.1%; Gag, 7.8%; Gag-p15, 5.2%; Gag-p27, 4.8%; and Gag-p18, 15%.
Recent reports indicate that in acutely HIV-infected individuals, only one or a few virus variants are involved in establishing the initial systemic infection (29). To recapitulate this scenario, we determined the titer of our pathogenic SIVsmE660 stock (45) using two different doses to define a challenge inoculum at which only one or a few viral variants are involved in establishing a disseminated systemic infection during the acute phase. A single rectal challenge with 4,000 TCID50 (233 μl of tissue culture fluid at 2.58 × 108 vRNA copy eq/ml, for a total of 6 × 107 vRNA copy eq) infected two of two macaques (see Fig. S2A in the supplemental material). A single rectal challenge with 1/10 of this initial dose, 400 TCID50 (23 μl of tissue culture fluid at 2.58 × 108 vRNA copy eq/ml for a total of 6 × 106 vRNA copy eq), infected only one of two macaques. After a second challenge at the lower dose, the second animal became infected. Two hundred seven full-length SIV env genes from plasma vRNA from the four animals were amplified and sequenced (median, 52 sequences per animal; range, 19 to 73). SGA (29) of plasma vRNA from acute-phase plasma revealed that a minimum of 3 to 10 viral variants had been transmitted to the two animals that received the single dose of 4,000 TCID50 (see Fig. S2B and C in the supplemental material). Productive infection by only one viral variant was evident in each of the two animals infected at the lower dose of 400 TCID50 (see Fig. S2D and E in the supplemental material). We made similar findings in a recent titration study of 18 Indian rhesus macaques infected intrarectally or intravenously with SIVsmE660 or SIVmac251 (29a). Thus, we elected to initially use a challenge dose of 800 TCID50 of the heterologous swarm virus SIVsmE660 to mimic mucosal HIV exposure in humans.
We began challenging our vaccinees (and eight naïve control animals) mucosally 37 weeks after the Ad5 boost with repeated doses (800 TCID50) of the heterologous swarm virus SIVsmE660 every 3 weeks for as many as five inoculations. At 33 weeks after the Ad5 boost, the vaccine-induced T-cell responses had diminished considerably, especially the CD4+ T-cell responses (see Table S2 in the supplemental material). Viral loads were assessed at days 7, 9, and 11 after each challenge, and if two of these were positive, the animal was not rechallenged. Five of eight vaccinees and six of eight naïve controls became infected after these challenges. Surprisingly, two of the vaccinees (r00061 and r02103) had peak viral loads of only 300 vRNA copy eq/ml (Fig. (Fig.2A).2A). Another vaccinee (r02114) had an acute phase peak of only 71,000 vRNA copy eq/ml. These three vaccinees have plasma viral loads that are either undetectable or less than 100 vRNA copy eq/ml in the chronic phase. Additionally, a fourth vaccinee (r01099) with a peak viremia of 456,000 vRNA copy eq/ml now has undetectable plasma viremia at 20 weeks postinfection. While none of the vaccinees had experienced a loss of either total or memory CD4+ T-cell counts after the challenge, three of the six infected control animals had reduced total memory CD4+ T-cell counts during the acute phase (see Fig. S3A and B in the supplemental material). However, the differences between these groups were not significant. We also analyzed CD4+ T cells via bronchoalveolar lavage. While the percentage of CD4+ T cells is higher in vaccinees than in control animals, the difference is not significant (data not shown). We were unable to obtain absolute counts of T cells in the bronchoalveolar lavage, however.
We then challenged our three remaining uninfected vaccinees, along with the two uninfected naïve control animals, with up to six repeated challenges of an increased dose of SIVsmE660 (4,000 TCID50). All five animals became infected. Remarkably, two of the three vaccinees (r97112 and r02089) controlled virus replication to less than 8,000 vRNA copies/ml during the acute phase (Fig. (Fig.2A).2A). Interestingly, these two animals were the ones that mounted the broadest and highest-frequency immune responses after vaccination (Table (Table11).
Even though there was no difference between the number of challenges required to cause infection in the vaccinees and controls (Table (Table2),2), the vaccinees exhibited significantly reduced viral replication in both the acute and chronic phases of infection. Peak plasma SIV RNA levels were 3.2 × 104 vRNA copy eq/ml for vaccinees versus 2.5 × 106 vRNA copy eq/ml for controls (P < 0.0263) (Fig. (Fig.2C;2C; Table Table3).3). At 16 weeks postinfection, vaccinated animals had a mean viral load of 7.9 vRNA copy eq/ml compared to 3.2 × 104 vRNA copy eq/ml (P < 0.0124) in control animals.
We used SGA-direct amplicon sequencing to identify and estimate the numbers of viral variants involved in establishing disseminated systemic acute infection in control and vaccinated animals. Two hundred twenty-two full-length SIV env genes from plasma vRNA from 10 animals (six controls and four vaccinees) 1 to 2 weeks after infection were amplified and sequenced (median, 22 sequences per animal; range, 10 to 45). Both control and vaccinated animals were infected by 1 to 4 viruses (median for controls, 3; median for vaccinees, 1.5) (Fig. (Fig.3;3; Table Table4).4). Although there was a trend toward lower numbers of transmitted variants in vaccinated animals than in control animals, this difference was not statistically significant. Interestingly, some animals productively infected by single viruses had a large proportion of viral sequences with APOBEC-mediated G-to-A hypermutation (Fig. (Fig.33 and see Fig. S2D and E in the supplemental material). Enrichment for G-to-A mutations was observed in titration animals (see Fig. S2D and E in the supplemental material), control animals (Fig. (Fig.3B),3B), and vaccinated animals (Fig. (Fig.3A3A).
All infected vaccinees experienced postchallenge expansions of vaccine-induced immune responses (Fig. (Fig.44 and see Fig. S4 in the supplemental material). Postchallenge, we measured the anamnestic vaccine-induced responses since we did not have a set of peptides that matched the sequences of SIVsmE660. Therefore, we examined only T-cell responses against regions of the virus that had previously been observed during the vaccination phase. As a result, we had fewer responses in this challenge phase to assess and we were able to carry out both whole-PBMC and CD8-depleted-PBMC IFN-γ ELISPOT assays using all of the reactive peptides seen 14 to 21 days after Ad5 vaccination. We assessed both CD4 and CD8 responses very early after the challenge, before CD4 responses diminished and prior to the onset of de novo challenge virus-specific responses. Seven of eight vaccinees showed evidence of anamnestic CD4+ and CD8+ T-cell responses expanded by the heterologous challenge virus (Fig. (Fig.44 and see Fig. S4B to F in the supplemental material). One vaccinee (r00061) mounted only a modest CD8+ T-cell response against the challenge virus, commensurate with its low viral load (Fig. (Fig.2A2A and see Fig. S4A in the supplemental material). Vaccine-induced cellular immune responses were, therefore, recalled effectively, with over 50% of the vaccine-induced responses expanded in the acute phase of infection (Tables (Tables55 and and66).
High-frequency CD4 responses were observed in five of the six vaccinees that successfully controlled replication of the heterologous challenge virus (Tables (Tables55 and and6).6). Gag was the main target of both of these CD4 responses and the anamnestic CD8 responses, perhaps due to its conservation between SIVmac239 and SIVsmE660. However, that explanation should also hold true for Pol, but anamnestic cellular responses against this conserved region were a fraction of those against Gag. Nef and Vif also served as CD8+ targets, but largely failed to induce anamnestic CD4 responses.
The level of containment of viral replication observed in the present study suggests that vaccine-induced T-cell responses might indeed be more effective against the AIDS virus than we had previously considered. We, and others, have already shown that SIV vaccines based solely on inducing cellular immune responses (i.e., no Env in the vaccine) can reduce both acute and chronic phase viral replication of homologous challenge viruses (34, 56). Unfortunately, however, viral replication exceeded 1 million copies/ml during the acute phase in these earlier studies using homologous viral challenges, suggesting that it might be difficult to control acute-phase viral replication with vaccine-induced T-cell responses alone. Encouragingly, our new results indicate that vaccine-induced T-cell responses alone can control replication of a heterologous virus during both the acute and the chronic phases of infection even after a heterologous challenge.
Given the low levels of viral loads in vaccinee r00061, it is possible that this animal was not productively infected (Fig. (Fig.2A).2A). Soon after the first challenge, we detected vRNA in plasma from this animal at days 9, 11, and 15. We also reamplified virus from frozen RNA extracted on these 3 days. Additionally, we observed a transient increase in 8 of 13 vaccine-induced CD8+ T-cell responses (see Fig. S4A in the supplemental material). However, two other independent laboratories were unable to amplify virus from shipped frozen samples from this animal. These two laboratories did amplify virus from a single time point (but not all time points) in shipped frozen plasma samples from vaccinee r02103, another animal with low acute-phase viral loads. This vaccinee was clearly infected given the expansion of 12 of 22 CD8+ and CD4+ vaccine-induced anamnestic T-cell responses at day 21 postinfection (Fig. (Fig.4A).4A). The Mamu-A*02/Gag GY9-specific T-cell response expanded from 102 SFC/106 PBMCs 1 month prior to challenge to 2,525 SFC/106 PBMCs 21 days after infection in this vaccinee (Fig. (Fig.4A).4A). We were also unable to isolate the virus from activated, CD8-depleted PBMCs from r00061 or r02103 on three different occasions. In contrast, we routinely cultured virus from PBMCs of two of the other vaccinees, r02114 and r01099, with higher acute-phase viral loads. Thus, we might have succeeded in productively infecting only seven of eight of our vaccinees. In contrast, we infected all of our naïve control animals.
Much of the attractiveness of vaccines that would elicit broadly reactive neutralizing antibodies relates to the ability of such antibodies to prevent infection or to limit viral replication during acute infection. This would diminish acute pathogenesis, reduce the generation of genetic diversity of the virus, and as an associated benefit, reduce the prospect of secondary transmission during acute infection, a time when higher levels of viremia are associated with increased transmission (21, 22, 44). The present results suggest that vaccine-induced cellular immunity, in the absence of neutralizing antibody, may also be able to achieve these same important objectives. More than half of the vaccinated animals had peak viral loads of less than 80,000 vRNA copy eq/ml. The achievement of this level of control of a heterologous challenge virus during the acute phase represents an encouraging new benchmark in the evaluation of prophylactic AIDS vaccines.
The mucosal challenge model employed in the present study recapitulates the results of human mucosal exposures leading to clinical HIV-1 infections, and the vaccine regimen used provided dramatic protective effects against this challenge that were greater than those seen in other studies. However, as there has been less experience with this repeated mucosal heterologous challenge model than some other challenge systems, we considered whether the protection we observed was a function of superior vaccine efficacy or potentially a result of a less-rigorous challenge than that used in other studies. Unlike clonal SIV challenge stocks (e.g., SIVmac239), challenge stocks of virus swarms like SIVsmE660 can vary in their pathogenicity depending on how they are propagated. Indeed, in recently published studies, five of eight naïve animals and four of six naïve animals controlled SIVsmE660 replication to undetectable levels (32, 61). However, these do not represent typical results for SIVsmE660 challenge, since only 3 of 29 Indian rhesus macaques controlled replication of SIVsmE660 to undetectable levels in many previous studies (1, 15, 16, 23, 24, 27, 43, 50, 57). We have previously used the same stock of SIVsmE660 used in the present study to intravenously challenge 10 naïve Indian rhesus macaques expressing a variety of MHC-I alleles, including six animals that expressed the “protective” alleles, Mamu-A*01, Mamu-B*08, and Mamu-B*17 (45) (see Fig. S5 in the supplemental material). All control animals became infected after one intravenous challenge with 100 TCID50 of this virus stock, and the mean peak plasma viremia during the acute phase was 5.1 × 106 vRNA copy eq/ml. All 10 of the naïve control animals had >100,000 vRNA copy eq/ml at 28 weeks postchallenge (see Fig. S5 in the supplemental material). Additionally, no “protective” effects were observed for any of the three MHC-I alleles after the SIVsmE660 challenge of these naïve control animals. Furthermore, in the current study, only 1 of 12 of our rectally challenged naïve macaques has controlled SIVsmE660 replication to undetectable levels (animal r96096) (see Fig. S2A in the supplemental material). Our stock of SIVsmE660, therefore, appears to be pathogenic and remarkably consistent from animal to animal after both intravenous and rectal challenge.
Despite the robust nature of SIVsmE660, it is formally possible that, like SHIV89.6P, SIVsmE660 may represent a less-than-stringent vaccine challenge virus, yielding misleadingly encouraging results in vaccine studies (5, 7, 17, 33). Vaccines designed to induce cellular immune responses only (i.e., using Gag/Pol and not Env) have had limited success at reducing acute-phase plasma viremia of SIVsmE660 below 1 × 106 vRNA copy eq/ml (16, 43, 50). Furthermore, vaccination using attenuated SIVmac239, our best current vaccine, has shown limited ability to control heterologous SIVsmE660 replication during the acute phase (1, 45, 57). These previous results are commensurate with our recent experiments using SIVmac239ΔNef. We vaccinated 10 Indian rhesus macaques with SIVmac239ΔNef and subsequently challenged them intravenously with SIVsmE660 using the stock of virus employed in the current study. Of these 10 SIVmac239ΔNef-vaccinated animals, only 4 showed some measure of control during the acute phase (see Fig. S6 in the supplemental material), and all 4 of these expressed the protective allele Mamu-B*08 or Mamu-B*17. Thus, the robust pathogenic stock of SIVsmE660 used in the current experiments appears to represent a rigorous challenge for vaccine studies.
Nonhuman primate challenge models, with pathogenic SIVs used for the challenge, have been criticized for being too stringent. Challenges using the pathogenic viruses SIVmac239 and SIVmac251 often result in plasma viral concentrations of greater than 500,000 vRNA copy eq/ml in the chronic phase (2-4, 8, 53). Our stock of SIVsmE660 given intravenously resulted in plasma viremia in excess of 100,000 vRNA copy eq/ml in the chronic phase (45). Here, we have used a novel mucosal challenge regimen with a heterologous virus swarm in order to mimic the results of typical mucosal HIV exposure resulting in infection in humans. We have succeeded in infecting macaques with 1 to 4 variants, replicating the results that have been observed in HIV-infected humans. Mean peak viral loads of 2.5 × 106 vRNA copy eq/ml and 80,000 vRNA copy eq/ml in the chronic phase were observed in our control (naïve) animals. It has been suggested that our intrarectal challenge method might also have resulted in a less-stringent challenge. However, when we compared the viral loads in this study with those used by Reynolds et al., in which SIVsmE660 from the same stock was given intravenously, there was no statistical difference between the viral loads (45) (see Fig. S7 in the supplemental material). These values are similar to average peak and chronic phase viral loads in humans (38, 46). Therefore, by several key measures we successfully mimicked HIV exposure in humans.
It will be important to calibrate the ability of other vaccine regimens to control the replication of SIVsmE660 after the challenge regimen we employed here. A benchmark of particular interest will be to formally assess whether the Merck Ad5 Gag, Pol, and Nef vaccine regimen, a vaccine approach that was not efficacious in the STEP study, can reduce viral replication after the type of mucosal heterologous SIVsmE660 challenge employed in the current study. Similarly, it will be critical to investigate whether SIVmac239ΔNef can control SIVsmE660 replication after a low-dose mucosal challenge in Indian rhesus macaques that do not express MHC-I alleles associated with the control of SIVmac239. Only moderate control of viral replication was observed in SIVmac239ΔNef-vaccinated monkeys after a high-dose intravenous challenge with this same stock of SIVsmE660 (45). Interestingly, animals expressing the MHC-I alleles Mamu-B*08 and Mamu-B*17 showed the best control of this heterologous challenge during acute infection. It is important to note that the majority of SIVmac239ΔNef-vaccinated monkeys have shown complete control of standard highly pathogenic homologous challenges with SIVmac239 or SIVmac251, administered either intravenously or mucosally (14, 48, 58).
To our knowledge, this is the first application of SGA-direct amplicon sequencing (29, 47) to the design and interpretation of an SIV vaccine trial. We combined a titration analysis of four naïve animals in this study, with nine additional naïve animals in another study (29a), to estimate a SIVsmE660 inoculum size that would productively infect animals with less than five viruses, recapitulating the results that characterize the majority of human mucosal HIV exposures resulting in clinical infection. We then confirmed in the present study in 10 vaccinated or control animals that the numbers of transmitted viruses leading to productive infection were indeed between one and four (Table (Table4).4). There were numerically lower numbers of transmitted viruses in vaccinated compared with control animals, but the difference was not statistically significant. An unexpected finding in the present study was a striking enrichment for G-to-A hypermutation observed in some (but not all) animals productively infected by a single virus. This was observed in vaccinated, control, and titration animals. While we have previously observed G-to-A hypermutation in HIV-1-infected humans (29) and in SIVsmE660- and SIVmac251-infected Indian rhesus macaques (29a), the extent of G-to-A hypermutation observed in some animals in the present study is unprecedented (e.g., animals 02103, 02114, and r92093) (Fig. (Fig.3A3A and see Fig. S2E in the supplemental material). With low-dose virus exposure, infection by single viruses with altered Vif function could lead to correspondingly high levels of APOBEC-mediated hypermutation, thereby affecting virus replication efficiency in naïve and vaccinated animals. Such a result might not be apparent in animals productively infected by multiple viruses if one or more of these exhibited wild-type Vif function and better overall replication fitness.
The breadth and frequency of the vaccine-induced T-cell responses achieved in the present study may have been critical for the enhanced control of replication of this heterologous, pathogenic challenge virus. No other vaccine regimens to date have achieved these frequencies or induced the breadth of T-cell responses observed in the current experiments (13, 31, 34, 41, 56). Whether the cellular immunity elicited by the replication-defective adenovirus HIV-1 vaccine in the STEP trial had comparable breadth and appropriate specificity such that one might have expected protection against the infecting strains in the reported cases is currently being investigated by using fine T-cell epitope mapping and viral sequencing. Those results will contribute to our understanding of the underlying reasons for the lack of vaccine efficacy in the STEP trial and help inform the next steps in HIV-1 vaccine research development.
D.I.W. and his laboratory are supported by the U.S. National Institutes of Health (grants R01 AI049120, R37 AI052056, R01 AI076114, R24 RR015371, R24 RR016038, and R21 AI077472; contract HHSN266200400088C). This publication was made possible in part by P51 RR000167 from the Wisconsin National Primate Research Center. This research was conducted at a facility constructed with support from the Research Facilities Improvement Program grants RR15459-01 and RR020141-01 and was also supported by grants from the International AIDS Vaccine Initiative. This work was supported in part by the CHAVI.
Thanks to Gary Schlei for database creation and graphic assistance.
The authors declare no competing interests.
Published ahead of print on 29 April 2009.
†Supplemental material for this article may be found at http://jvi.asm.org/.