Viral escape mutations do not account for non-protection from SIVmac239 challenge in RhCMV/SIV vaccinated rhesus macaques

Abstract

Simian immunodeficiency virus (SIV) vaccines based upon 68-1 Rhesus Cytomegalovirus (RhCMV) vectors show remarkable protection against pathogenic SIVmac239 challenge. Across multiple independent rhesus macaque (RM) challenge studies, nearly 60% of vaccinated RM show early, complete arrest of SIVmac239 replication after effective challenge, whereas the remainder show progressive infection similar to controls. Here, we performed viral sequencing to determine whether the failure to control viral replication in non-protected RMs is associated with the acquisition of viral escape mutations. While low level viral mutations accumulated in all animals by 28 days-post-challenge, which is after the establishment of viral control in protected animals, the dominant circulating virus in virtually all unprotected RMs was nearly identical to the challenge stock, and there was no difference in mutation patterns between this cohort and unvaccinated controls. These data definitively demonstrate that viral mutation does not explain lack of viral control in RMs not protected by RhCMV/SIV vaccination. We further demonstrate that during chronic infection RhCMV/SIV vaccinated RMs do not acquire escape mutation in epitopes targeted by RhCMV/SIV, but instead display mutation in canonical MHC-Ia epitopes similar to unvaccinated RMs. This suggests that after the initial failure of viral control, unconventional T cell responses induced by 68-1 RhCMV/SIV vaccination do not exert strong selective pressure on systemically replicating SIV.

Keywords: CMV vaccine vector; HIV/SIV; SIV; viral escape; viral sequence analysis.

Figure 1

Summary of SIVmac239 viral evolution at 28-35 days post infection. Deep sequencing was performed on plasma virus sampled at 28-35 DPI from four cohorts, identifying the position and frequency of nucleotide mutations. (A) The plot displays all detected SIVmac239 mutations in the labeled vaccine cohorts. Each dot represents a single mutation detected in one RM, and the y-axis indicates the frequency of that mutation in the sample. Mutations are colored according to functional category (see legend). (B) The graph summarizes the number of dominant mutations (detected in >50% of reads) per animal, grouped by functional category. (C) The plot displays every position where a dominant variant was detected in at least one animal, displaying the RMs per cohort with each mutation. Most mutations are private, detected in a single animal. The only shared non-synonymous mutation was Env V67M, detected in three unvaccinated RMs. (D) The plot displays the total nucleotide distance from the challenge sequence, defined as the sum of the frequencies of all mutations detected in that sample. Collectively, these data demonstrate that across all vaccine groups, except for known sub-optimal nucleotide A9110G, the dominant circulating virus has nearly identical coding potential to the challenge stock, and there are no significant differences in levels of mutation between cohorts. NS, non-synonymous; Synon, synonymous.

Figure 2

Summary of SIVmac239 viral evolution at 70-84 days post infection. Deep sequencing was performed on plasma virus sampled at 70-84 days post-infection from all cohorts, identifying the position and frequency of nucleotide mutations. (A) The plot displays all detected SIVmac239 mutations in the labeled vaccine cohorts. Each dot represents a single mutation detected in one RM, and the y-axis indicates the frequency of that mutation in the sample. Mutations are colored according to functional category (see legend). (B) The graph summarizes the number of dominant (detected in >50% of reads) mutations per animal, grouped by category. These data show that all cohorts are accumulating additional dominant mutations, although there are no significant differences between cohorts. (C) The plot displays the 17 dominant mutations detected in multiple RMs (out of 58 total dominant mutations), for the purpose of identifying patterns of common selection. While three positions were enriched in every group (Env V67M, Nef E93E, and the sub-optimal nucleotide A9110G), the pattern of mutation was otherwise diverse, and there were not any variants enriched in a specific vaccine cohort. (D) The plot displays the total nucleotide distance from the challenge sequence, defined as the sum of the frequencies of all mutations detected in that sample. NS, non-synonymous; Synon, synonymous.

Figure 3

Nucleotide mutation within MHC-E and MHC-II restricted supertopes. Supertopes are a unique property of 68-1 RhCMV vaccines, with different RhCMV versions eliciting different patterns of responses. (A) Each boxplot displays the sum of nucleotide mutation per RM within the indicated MHC-E restricted supertope. Plots are colored based on whether the vaccine is expected to elicit a T cell response against that supertope, which is determined by the vector backbone and whether the epitope is encoded by the vector. (B) Analogous to (A), each boxplot displays the sum of nucleotide mutation per RM within the indicated MHC-II restricted supertope. Collectively, these data demonstrate that there is relatively little viral mutation within most supertopes of 68-1 RhCMV vaccinated RMs. While there are examples specific RMs with high frequency mutations overlapping a particular supertope, there is no significant difference between RMs that do or do not have a vaccine-elicited response against the supertope, suggesting other factors are driving this selection.

Figure 4

Nucleotide mutation within conventional MHC-Ia restricted supertopes. Each boxplot displays the sum of nucleotide mutation per RM within the indicated immunodominant MHC-Ia restricted epitope. While none of the vaccines elicit responses against canonical MHC-Ia epitopes, de novo responses will be primed by the challenge itself. RMs within each group are separated based on whether the RM expresses the restricting MHC-Ia allele for that epitope. Collectively, these data demonstrate that viral escape consistent with MHC-Ia restricted CTL pressure occurs in both RhCMV/SIV vaccinated and unvaccinated RMs. Asterisks after the plot title indicate epitopes with a statistically significant difference in mutation rates between RMs that do or not express the restricting MHC-Ia allele, using a pairwise Wilcoxon rank sum test (Holm-corrected), with adjusted p-value threshold <0.05.