Background:  Crucial protection is conferred by the CD8⁺ve T-cell response. Certain HLA alleles are associated with slower disease progression (eg, HLA B27 and B57). The mechanism behind this differential HLA Class I effect is elusive. We studied the relationship between epitope sequence variation and HLA, to see how escape might contribute to HLA advantage.

Methods:  The Swiss–Spanish Intermittent Therapy Trial (SSITT; n = 133) was a European study of structured treatment interruptions. A second cohort of African patients (n = 516) was recruited in southern Africa. From the SSITT cohort, HLA Class I immune responses were measured by gamma-interferon ELISspot using optimal peptides, tested repeatedly over a mean of 14 months. Immune responses for the African patients were tested using a panel of 410 overlapping peptides that spanned the HIV genome. HIV-1 gag, pol, and nef genes were sequenced from DNA. The difference in proportions of polymorphic epitopes in the HLA-matched and -unmatched individuals was scored as a measure of HLA-driven variation. The significance of sites within epitopes that varied from the consensus peptide sequence was tested using odds ratios and Fisher’s Exact Test. CodeML was used to identify sites with dn/ds ratios indicative of positive selection pressure.

Results:  Studying sequence variation in 54 epitopes in the European cohort and 70 epitopes in the African cohort, we found that those epitopes that were restricted by protective HLA Class I molecules (B27, B57, B51 in European patients, and B5703, B5801, and B8101 in Africans) were those that were most polymorphic in HLA-matched individuals. Of the 10 most variable epitopes for each cohort, 7 were restricted by protective HLA Class I alleles. A “quality” score for each epitope in the European cohort reflected sequence variation and immune responses. Of the 10 most highly scoring epitopes, 9 were restricted by protective HLA Class I alleles. There was a significant inverse correlation between the prevalence of mutations within restricted epitopes and the HLA-associated relative hazard of disease progression (p = 0.028; R² = 0.34).

Conclusions:  We conclude that protective HLA Class I alleles drive out significant variation in the epitopes they target, but in conjunction with persisting immune responses to the equivalent wild-type peptides. We hypothesise that escape from these beneficial HLA molecules confers a fitness cost that contributes to their clinical advantage.