Vaccine-induced cellular responses control simian immunodeficiency virus replication after heterologous challenge

Abstract

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.

FIG. 1.

Vaccination scheme and vectors. Eight rhesus macaques (blue outlines), expressing the MHC-I allele Mamu-A*02 (all were Mamu-A*01 negative, Mamu-B*08 negative, and Mamu-B*17 negative), were primed three times at 4-week intervals with DNA encoding each of the eight SIV proteins except Env. Animals were boosted at week 24 with six Ad5 vectors, encoding the same protein sequences as the DNA prime; one of the Ad5 vectors contained three of the SIV proteins in a single open reading frame (vif, vpr, and vpx). Thirty-seven weeks later, the vaccinees and eight naïve control animals which also expressed Mamu-A*02 and were negative for Mamu-A*01, Mamu-B*08, and Mamu-B*17 (in red) were challenged intrarectally with SIVsmE660 (800 TCID₅₀; 1.2 × 10⁷ copy eq). If animals were not productively infected (more than two consecutive positive viral loads) following a given challenge, they were challenged again 3 weeks later, up to a total of five low-dose challenges. If animals were not productively infected by five low-dose challenges, they were challenged with up to six high-dose challenges every 2 weeks with SIVsmE660 (4,000 TCID₅₀; 6 × 10⁷ vRNA copy eq). As with the previous dose, once an animal had two or more consecutive positive viral loads, it was considered to be infected and not challenged further.

FIG. 2.

Plasma virus concentrations after SIVsmE660 infection. Vaccinees (A) (solid blue lines) and controls (B) (dashed red lines) were infected after up to five low-dose mucosal challenges and up to six high-dose challenges (Table 2). The solid line on each graph indicates a viral load of 10⁶ copy eq/ml; the dashed line is the limit of detection for our viral load assay (30 copy eq/ml). Animals receiving high doses are indicated with an asterisk after the animal's identification number. Animal r02012 had extremely high viral loads and was sent to necropsy after 10 weeks of infection, as indicated by a symbol (†). (C) Means of the viral load for infected animals. This is the graphical representation of the data presented in Table 3. Chronic viral loads were determined for the range from 6 weeks to 20 weeks postinfection.

FIG. 3.

(A) Composite SIVsmE660 tree from vaccinated animals. Neighbor-joining tree includes inoculum sequences (black) which, together with numerous unique transmitted viruses, are distributed throughout the tree. Transmitted viruses are represented by discrete, low-diversity lineages which are color coded and labeled with animal designations and the number of transmitted/founder viruses. Sequences with enriched APOBEC-mediated G-to-A mutations are indicated by red asterisks. (B) Composite SIVsmE660 tree from control animals. The neighbor-joining tree includes inoculum sequences (black) which, together with numerous unique transmitted viruses, are distributed throughout the tree. Transmitted viruses are represented by discrete, low-diversity lineages which are color coded and labeled with animal designations and the number of transmitted/founder viruses. Sequences with enriched APOBEC-mediated G-to-A mutations are indicated by red asterisks.

FIG. 4.

Representative anamnestic vaccine-induced immune responses in the vaccinees. All vaccinees made CD8⁺ T-cell responses to several epitopes, including some restricted by the MHC-I molecule Mamu-A*02. In the interest of space, we are showing only two of the vaccinees in this figure; data from the remaining vaccinees are shown elsewhere (see Fig. S3 in the supplemental material). Whole-PBMC responses are indicated by blue bars, with responses postchallenge indicated in dark blue and responses observed immediately prior to challenge indicated in light blue. Responses observed in PBMCs depleted of CD8⁺ cells (red) are likely mediated by CD4⁺ T cells. CD8⁺ cell depletion was typically 99% complete (data not shown). Dark red bars indicate responses observed postchallenge, whereas light red bars indicate responses present immediately prior to the challenge. Green bars represent responses to minimal optimal peptides that bind to Mamu-A*02. As indicated in Table S1 in the supplemental material, some of these epitopes are conserved between SIVmac239 and SIVsmE660, whereas others have several substitutions, some of which could affect T-cell recognition. Again, light green bars indicate responses observed prior to the challenge, whereas dark green bars indicate anamnestic responses. Most anamnestic response analyses were performed at 14 to 15 days postinfection. For a couple of the animals (r00061, r02103), these assays were delayed to 21 days postinfection due to the very low viral loads observed.