Elevated HLA-A expression impairs HIV control through inhibition of NKG2A-expressing cells

Abstract

The highly polymorphic human leukocyte antigen (HLA) locus encodes cell surface proteins that are critical for immunity. HLA-A expression levels vary in an allele-dependent manner, diversifying allele-specific effects beyond peptide-binding preference. Analysis of 9763 HIV-infected individuals from 21 cohorts shows that higher HLA-A levels confer poorer control of HIV. Elevated HLA-A expression provides enhanced levels of an HLA-A-derived signal peptide that specifically binds and determines expression levels of HLA-E, the ligand for the inhibitory NKG2A natural killer (NK) cell receptor. HLA-B haplotypes that favor NKG2A-mediated NK cell licensing (i.e., education) exacerbate the deleterious effect of high HLA-A on HIV control, consistent with NKG2A-mediated inhibition impairing NK cell clearance of HIV-infected targets. Therapeutic blockade of HLA-E:NKG2A interaction may yield benefit in HIV disease.

Fig. 1. Elevated HLA-A expression levels are associated with increased HIV viremia and reduced CD4⁺ T cell counts

(A) Data represent 2298 HIV-infected individuals from South Africa, Botswana, and Zambia, enrolled at 11 sites with cross-sectionally measured VLs. Each dot represents the average estimated expression level for a specific HLA-A allele by that allele’s reported effect on cross-sectional VL (11). A linear regression line is shown in blue with 95% confidence interval in gray. The size of each point is scaled by the number of contributing alleles; however, the correlation estimate is not weighted. (B) HIV viremia among 5818 HIV-infected adults and (C) CD4⁺ T cell counts among 2100 HIV-infected adults followed prospectively and grouped according to one-unit z-score change in HLA-A expression. VLs are plotted against time following seroconversion or date of enrollment (censored at ~5 years). In (B) and (C), lines are best fit (LOWESS lines) to unadjusted VL or CD4 counts.

Fig. 2. HLA-A expression and HLA-B −21M regulate HLA-E expression, resulting in biased licensing of NKG2A-expressing NK cells that are impaired in their killing of HIV-infected target cells

(A) HLA-E expression according to HLA-A expression and HLA-B −21M in 58 HIV-uninfected donors. Each dot represents HLA-E expression levels (expressed as median signal intensity on a linear scale), as determined by CyTOF (15), and imputed HLA-A expression (z-score) (R_pearson = 0.43; 95% CI 0.20–0.62; P = 5 × 10⁻⁴). (B) NKG2A⁺ NK cell licensing varies by HLA-A expression and HLA-B −21M. Peripheral blood mononuclear cells (PBMCs) from 10 HLA-B −21M/M and 10 HLA-B −21T/T donors were coincubated with Raji cells pretreated with mouse antibody (2.5 µg/ml) against human CD20 for 6 hours to probe NK cell licensing and education. Each point represents the proportion of IFN-γ⁺ NK cells from each individual that are NKG2A⁺/KIR⁻ (triangles) or KIR⁺/NKG2A⁻ (circles) as a function of HLA-A expression. Dotted and solid lines show best fit lines for NKG2A⁺ and KIR⁺ subsets, respectively. The association between NK cell responsiveness and HLA-A expression for NKG2A⁺ NK cells in HLA-B −21M/M donors was R_pearson = 0.69 (95% CI 0.10–0.92), P = 0.03; all other correlations were not significant. (C) PBMCs from 9 HLA-B −21M/M and 9 HLA-B −21T/T donors were cocultured for 6 hours with autologous T cell blasts that were left uninfected or were infected with HIV [vesicular stomatitis virus G glycoprotein (VSV-G) pseudotyped NL4-3] and stained for CD107A, a marker of NK cell degranulation (see fig. S3 for gating strategy). HLA-A expression was formally measured in these T cell blasts by quantitative polymerase chain reaction and is expressed relative to β2M expression levels. Plots show individual proportions of NK cells expressing CD107a among NKG2A⁺KIR⁻ and NKG2A⁻KIR⁻ subsets. A best fit line is shown for significantly correlated observations. Red and black lines and dots denote TT and MM donors, respectively. The association between NKG2A⁺KIR⁻ NK cell response to HIV-infected target cells, and HLA-A expression in HLA-B −21M/M donors was R_pearson = –0.77 (95% CI –0.21 to –0.95), P = 0.02; all other correlations were not significant.

Fig. 3. The effect of HLA-A expression on HIV VL is modified by HLA-B alleles encoding methionine at position −21 in the signal peptide

The magnitude of effect (slope) of HLA-A expression on HIV viral load is stronger among individuals with HLA-B −21 MM (VL from 428 individuals, black line, VL_effect-MM = 0.22 log₁₀ copies/ml, P = 1.5 × 10⁻²¹ adjusted for HLA-A, -B, and -C) compared with HLA-B TT (VL from 3071 individuals, red line, VL_effect-TT = 0.06 log₁₀ copies/ml, P = 1.8 × 10⁻⁹ adjusted for HLA-A, -B, and -C). Interaction P = 5.3 × 10⁻⁹. Gray shading represents 95% CI of the linear estimate.