Differential escape patterns within the dominant HLA-B*57:03-restricted HIV Gag epitope reflect distinct clade-specific functional constraints

Abstract

HLA-B*57:01 and HLA-B*57:03, the most prevalent HLA-B*57 subtypes in Caucasian and African populations, respectively, are the HLA alleles most protective against HIV disease progression. Understanding the mechanisms underlying this immune control is of critical importance, yet they remain unclear. Unexplained differences are observed in the impact of the dominant cytotoxic T lymphocyte (CTL) response restricted by HLA-B*57:01 and HLA-B*57:03 in chronic infection on the Gag epitope KAFSPEVIPMF (KF11; Gag 162 to 172). We previously showed that the HLA-B*57:03-KF11 response is associated with a >1-log-lower viral setpoint in C clade virus infection and that this response selects escape mutants within the epitope. We first examined the relationship of KF11 responses in B clade virus-infected subjects with HLA-B*57:01 to immune control and observed that a detectable KF11 response was associated with a >1-log-higher viral load (P = 0.02). No evidence of HLA-B*57:01-KF11-associated selection pressure was identified in previous comprehensive analyses of >1,800 B clade virus-infected subjects. We then studied a B clade virus-infected cohort in Barbados, where HLA-B*57:03 is highly prevalent. In contrast to findings for B clade virus-infected subjects expressing HLA-B*57:01, we observed strong selection pressure driven by the HLA-B*57:03-KF11 response for the escape mutation S173T. This mutation reduces recognition of virus-infected cells by HLA-B*57:03-KF11 CTLs and is associated with a >1-log increase in viral load in HLA-B*57:03-positive subjects (P = 0.009). We demonstrate functional constraints imposed by HIV clade relating to the residue at Gag 173 that explain the differential clade-specific escape patterns in HLA-B*57:03 subjects. Further studies are needed to evaluate the role of the KF11 response in HLA-B*57:01-associated HIV disease protection.

Importance: HLA-B*57 is the HLA class I molecule that affords the greatest protection against disease progression in HIV infection. Understanding the key mechanism(s) underlying immunosuppression of HIV is of importance in guiding therapeutic and vaccine-related approaches to improve the levels of HIV control occurring in nature. Numerous mechanisms have been proposed to explain the HLA associations with differential HIV disease outcome, but no consensus exists. These studies focus on two subtypes of HLA-B*57 prevalent in Caucasian and African populations, HLA-B*57:01 and HLA-B*57:03, respectively. These alleles appear equally protective against HIV disease progression. The CTL epitopes presented are in many cases identical, and the dominant response in chronic infection in each case is to the Gag epitope KF11. However, there the similarity ends. This study sought to better understand the reasons for these differences and what they teach us about which immune responses contribute to immune control of HIV infection.

FIG 1

Median viral loads of KF11 responders versus nonresponders in B clade virus-infected individuals expressing HLA-B*57:01. PBMCs from B clade virus-infected, ART-naive individuals expressing HLA-B*57:01 were analyzed by IFN-γ ELISPOT assay for responses to the KF11 epitope. Viral loads of responders and nonresponders were compared. Mann-Whitney U tests were performed.

FIG 2

Median viral load and CD4 count of B*57:03-positive versus B*57:03-negative subjects in the Barbados cohort (B clade). B*57:03-positive subjects were compared to B*57:03-negative subjects for viral load (A) and CD4 count (B). Medians and 5th to 95th percentiles are shown. Mann-Whitney U tests were performed.

FIG 3

Effect of viral mutation S173T on epitope recognition of HIV-infected cells by KF11-specific CD8⁺ T cells. Ex vivo CD4⁺ T cells from B*57:03⁺ and B*57:03⁻ donors were infected with wild-type NL43 virus or NL43 virus harboring the S173T viral mutation. Infected CD4⁺ T cells were then cultured with KF11-specific CD8⁺ T cells (A) or IAW9-specific CD8⁺ T cells (B), and the level of CD8⁺ T-cell activation was monitored by expression of CD107 and Mip1β. Data from both experiments were standardized relative to percent recognition by wild-type virus (C). Experiments were performed in triplicate; means and standard deviations are shown. Student's t test was performed. *, P < 0.01; **, P < 0.001; ***, P < 0.0001. NS, not significant.

FIG 4

Median viral load and CD4 count of B*57:03-positive HIV-infected subjects with viral polymorphisms Ser-173 and Thr-173. Proviral DNA sequences from B*57:03-positive subjects from the Barbados cohort (B clade) were analyzed for the presence of the viral polymorphisms Ser-173 and Thr-173. Viral loads (A) and CD4 counts (B) were compared. Medians and 5th to 95th percentiles are shown. Mann-Whitney t tests were performed.

FIG 5

Viral replication capacity of NL43_GFP virus with multiple B*57:03-associated viral mutations. NL43_GFP virus was engineered to contain combinations of the viral mutations Thr-173, Gly-163, and Asn-165. MT4 cells were infected and monitored for GFP-positive cells over 14 days (A). The slope of the curve was calculated from the exponential growth phase using the LOGEST function and converted to natural logs (B). Experiments were performed in triplicate; mean and standard deviations are shown. Dunnett's multiple-comparison tests were performed. *, P < 0.01; **, P < 0.001; ***, P < 0.0001.

FIG 6

Frequency of KF11 mutations (A163G and S165N) and S173T in C clade virus Gag sequences (n = 1,899). HIV-1 p24 Gag sequences were analyzed for the presence of the KF11 mutations, A163G and S165N, in the presence of Thr-173 and Ser-173. B*57:03-positive subjects (A) and B*57-negative subjects (B) were analyzed. Fisher's exact tests were performed.