Human Genomics of COVID-19 Pneumonia: Contributions of Rare and Common Variants

Abstract

SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) infection is silent or benign in most infected individuals, but causes hypoxemic COVID-19 pneumonia in about 10% of cases. We review studies of the human genetics of life-threatening COVID-19 pneumonia, focusing on both rare and common variants. Large-scale genome-wide association studies have identified more than 20 common loci robustly associated with COVID-19 pneumonia with modest effect sizes, some implicating genes expressed in the lungs or leukocytes. The most robust association, on chromosome 3, concerns a haplotype inherited from Neanderthals. Sequencing studies focusing on rare variants with a strong effect have been particularly successful, identifying inborn errors of type I interferon (IFN) immunity in 1-5% of unvaccinated patients with critical pneumonia, and their autoimmune phenocopy, autoantibodies against type I IFN, in another 15-20% of cases. Our growing understanding of the impact of human genetic variation on immunity to SARS-CoV-2 is enabling health systems to improve protection for individuals and populations.

Keywords: COVID-19 pneumonia; GWAS; SARS-CoV-2; inborn errors of immunity; type I interferons.

Figure 1.. Genetic and immunological determinants of COVID-19 pneumonia.

Odds ratios (OR) for the associations of auto-Abs against type I IFN, inborn errors of immunity (IEIs) and single genetic variants with the severity of COVID-19 pneumonia and resistance to SARS-CoV-2 infection are plotted according to risk factor frequency. For auto-Abs against type I IFN, the ORs and frequency were taken from (114). For X-linked TLR7 deficiency, the OR for the aggregated effect of biochemically proven loss of function (LOF) was taken from (74) and the cumulative frequency of biochemically proven LOF was taken from (76). For other IEIs, the OR for the aggregated effect of homozygous (autosomal recessive, AR) or heterozygous (autosomal dominant, AD) predicted LOF was taken from (74) and the corresponding cumulative frequencies were estimated from Gnomad v2.1.1. For single variants, the OR, assuming an additive model, and deleterious allele frequencies were taken from the most recent update of the COVID-19 Host Genetics Initiative GWAS study (43), except for the DOCK2 and APOE loci. The chromosomal region or closest gene is indicated. For the DOCK2 locus, the OR for the effect of the rs60200309-A variant on the severity of COVID-19 pneumonia under an additive model and allele frequency for rs60200309-A were taken from (156). For the APOE locus, the hazard ratio for the effect of APOE4 homozygosity as opposed to APOE3 homozygosity for COVID-19 mortality and the frequency of APOE4 homozygosity were taken from (157).

Figure 2.. Proportion of critical COVID-19 patients with IEIs and auto-Abs against type I IFN as a function of age.

Age-specific proportions of dominant and recessive IEIs among cases of critical COVID-19 pneumonia were estimated with Bayes theorem as a function of the probability of critical COVID-19 pneumonia for IEI carriers infected with SARS-CoV-2 (i.e. the penetrance). The age-specific frequency of IEIs in the general population and the age-specific critical infection rate were taken from (158). The age-specific frequency of IEIs in the general population was estimated from the frequency of IEIs at birth, assuming a non-specific mortality rate (i.e. not attributable to COVID-19) of 0 (panel A) or 1% (panel B) per year. Based on our previous findings (67, 159, 160), the frequency of IEIs at birth was set at 10⁻³ for dominant and 5 × 10⁻⁴ for recessive IEIs. We assumed a penetrance of 0.2 for dominant IEIs and 0.8 for recessive IEIs, consistent with the larger effect size estimated for recessive than for dominant IEIs. The age-specific proportions of patients with critical COVID-19 producing auto-Abs neutralizing low doses (100 pg/mL; gray line) or high doses (10 ng/mL; black line) of IFN-α and/or IFN-ω were taken from (114).

Figure 3.. Impaired type I IFN immunity underlies life-threatening COVID-19

Nearly 20% of patients with life-threatening COVID-19 pneumonia have impaired production of or response to type I interferons (IFNs), due to inborn errors of immunity (IEIs, 1 to 5%) or to blockade of type I IFN activity by pre-existing neutralizing autoantibodies (~15%). IEI of type I IFN immunity (genes shown in red) have been identified and lead to: (1) an impaired production of type I IFNs (left) in respiratory epithelial cells (RECs) and/or in blood plasmacytoid dendritic cells (pDCs), or (2) an impaired response to type I IFNs (right) in RECs, in response to SARS-CoV-2 infection. Genes for which X-linked or autosomal recessive defects have been identified are in italic. Genes for which common variants have been associated by GWAS with severe COVID-19 are circled in blue.