Selective escape from CD8+ T-cell responses represents a major driving force of human immunodeficiency virus type 1 (HIV-1) sequence diversity and reveals constraints on HIV-1 evolution

Abstract

The sequence diversity of human immunodeficiency virus type 1 (HIV-1) represents a major obstacle to the development of an effective vaccine, yet the forces impacting the evolution of this pathogen remain unclear. To address this issue we assessed the relationship between genome-wide viral evolution and adaptive CD8+ T-cell responses in four clade B virus-infected patients studied longitudinally for as long as 5 years after acute infection. Of the 98 amino acid mutations identified in nonenvelope antigens, 53% were associated with detectable CD8+ T-cell responses, indicative of positive selective immune pressures. An additional 18% of amino acid mutations represented substitutions toward common clade B consensus sequence residues, nine of which were strongly associated with HLA class I alleles not expressed by the subjects and thus indicative of reversions of transmitted CD8 escape mutations. Thus, nearly two-thirds of all mutations were attributable to CD8+ T-cell selective pressures. A closer examination of CD8 escape mutations in additional persons with chronic disease indicated that not only did immune pressures frequently result in selection of identical amino acid substitutions in mutating epitopes, but mutating residues also correlated with highly polymorphic sites in both clade B and C viruses. These data indicate a dominant role for cellular immune selective pressures in driving both individual and global HIV-1 evolution. The stereotypic nature of acquired mutations provides support for biochemical constraints limiting HIV-1 evolution and for the impact of CD8 escape mutations on viral fitness.

FIG. 1.

Clinical course of HIV-1 infection in four study subjects. Viral loads and CD4⁺ T-cell counts are shown for four subjects identified during acute HIV-1 infection and followed longitudinally. Arrows indicate time points from which viral sequences were derived. Shaded regions, time under treatment with highly active antiretroviral therapy.

FIG. 2.

Sites of sequence variation and associated CD8 T-cell responses. Longitudinal whole-genome sequencing, excluding Env, was conducted on four subjects. (A) Mutations accumulating in subject AC-38 over 4 years of follow-up (○). (B) Mutations developing in each subject over 0.3 to 5.3 years of follow-up (○). •, mutations associated with detectable IFN-γ ELISPOT CD8⁺ T-cell responses. Numbers corresponding to mutations associated with reversions are boxed in black. (C) Summary of mutations associated with detectable CD8⁺ T-cell responses in either defined or newly identified epitopes or with reversions. Mutations strongly associated with HLA alleles at the population level are indicated separately.

FIG. 3.

Escape mutations in Gag and Pol occur at highly variable residues in the population. The degrees of variability of Gag and Pol residues were assessed by plotting normalized Shannon entropy scores. The residue observed to arise during sequence evolution in a subject is shown below each epitope. Solid bars, escape mutations corresponding to the most polymorphic residue in each epitope; hatched bars, escaping residues not corresponding to the most polymorphic residue in each epitope. CD8-targeted regions include (A) B57 epitopes and (B) published epitopes (Pub), epitopes in study subjects (with subject designation given above the bar graph, e.g., AC-38), and partially mapped epitopes in study subjects (UNMAPPED).

FIG. 4.

Shared polymorphic residues in Gag and Pol in clades B and C. The degrees of variability of clade C sequences are plotted against clade B Gag and Pol epitopes shown in Fig. 3. Black bars, clade B sequences; gray bars, clade C sequences. Asterisks indicate highly polymorphic residues shown to mutate in Fig. 3 which are shared across clade B and clade C. (A) B57 epitopes; (B) published epitopes (Pub), optimal epitopes in study subjects (with subject designation given above the bar graph, e.g., AC-38), and partially mapped epitopes in study subjects (unmapped).