The HLA-B -21 M/T dimorphism associates with disease severity in COVID-19

Abstract

Host genetics shape immune responses and influence severity of infectious diseases. The HLA-B -21 M/T dimorphism tunes the functionality of natural killer (NK) cells expressing the inhibitory receptor NKG2A. NKG2A⁺ NK cells have been reported to recognize SARS-CoV-2-infected cells, but it remains unclear whether the HLA-B -21 M/T dimorphism associates with COVID-19 severity. Here, we investigated the influence of the HLA-B -21 M/T dimorphism in a cohort of 230 unvaccinated patients hospitalized with COVID-19 and requiring respiratory support. We found that HLA-B -21 M/M genotypes were more prevalent in patients with moderate compared to severe COVID-19 (6.0% vs. 0.9%). Comparison of age- and sex-matched sub-groups revealed that patients with M/M genotypes required mechanical respiratory support less frequently (OR = 0.13, 95% CI = 0.01-0.76, P = 0.013). Furthermore, patients with M/M genotypes showed a coordinately shifted signature of clinical laboratory parameters, coinciding with elevated serum levels of the anti-viral cytokine IFN-γ. These findings demonstrate that HLA-B variants associate with COVID-19 severity and suggest that the robust functionality of NKG2A⁺ NK cells in patients carrying the M/M genotype may contribute to protection from severe disease.

Fig. 1. The HLA-B –21 M/M genotype is differentially distributed across COVID-19 severity.

A Distribution of M/M, M/T, and T/T genotypes in a reference (n = 8192) [6], all patients included in the study cohort (n = 230), patients with moderate disease (n = 116), and patients with severe disease (n = 114). Right bar graphs display M/M genotypes and fold enrichment/depletion compared to all patients. B Distribution of disease manifestation in patients with M/M (n = 8), M/T (n = 83), and T/T genotypes (n = 139). C Tabular summary of the study cohort (n = 230). D Tabular summary of a cohort previously published by Ellinghaus and colleagues [9] (n = 3193). Two-tailed Chi-square test (A, C, D).

Fig. 2. Patients with M/M genotypes require less mechanical respiratory support and display a signature of reduced severity in an age- and sex-matched sub-cohort.

A Schematic illustration of propensity score matching for age and sex. B Distribution of disease manifestations in genotype groups. C Disease courses of genotype groups. Distribution of patients D requiring intensive or non-intensive care, E requiring mechanical or non-mechanical respiratory support, and F with deceased or discharged outcome. PCA of clinical laboratory parameters. G Biplot showing individual patients (grey dots) and loading of variables (arrows). H Contribution of parameters to loading to PC1. I PC1 scores of discharged and deceased patients. Bars indicate mean and error bars SEM. J Score plot of patients stratified for genotype group and outcome. K PC1 scores of M/M and T/X genotype groups. Bars indicate mean and error bars SEM. L Serum levels of selected cytokines determined by proximity extension assay. Dots represent individual patients and lines indicate mean. Numbers below graphs denote number of included samples. M Graphical summary of findings. n = 8 for M/M, n = 80 for T/X (B–F), n = 6 for M/M, n = 56 for T/X (G–K), and n = 4 for M/M, n = 27 for T/X (L). One-tailed chi-squared test (B, D–F) and one-tailed t-test (L).