NEW YORK – Numerous research groups are trying to solve the mystery of why some individuals infected with SARS-CoV-2 show no or mild symptoms while others need to be hospitalized or put on a ventilator.
Studying the genetics of the human host is one way to get a grip on these puzzling effects, and multiple teams have been gathering data from large human cohorts to look for associations between genetic variants and COVID-19 susceptibility or severity. During a plenary session at the American Society of Human Genetics 2020 virtual meeting on Wednesday, three researchers presented results from their respective studies.
The COVID-19 Host Genetics Initiative, an open collaboration between scientists around the globe formed earlier this year, has been looking for genetic determinants of disease susceptibility and severity and has more than 200 studies registered on its website so far.
Andrea Ganna, a researcher at the University of Helsinki's Institute for Molecular Medicine Finland, presented results from a genome-wide association study meta-analysis on behalf of the initiative that included data from 36 studies in 16 different countries, a total of 31,000 COVID-19 cases and 1.7 million controls. The data came from a variety of biobanks, COVID-19 clinical studies, and studies from direct-to-consumer companies Ancestry, 23andMe, and Helix.
For their study, Ganna and his team looked for genetic associations with hospitalized cases, of which they had 8,600, and with infections in general, including self-reported ones, a total of 31,000 cases. Altogether, they found nine significant loci on chromosomes 3, 6, 9, 12, 19, and 21.
The strongest signal came from the 3p21.31 region on chromosome 3 that had previously been implicated in COVID-19 in a study published this summer in the New England Journal of Medicine. The researchers also saw the ABO blood group locus on chromosome 9 that was reported in the same study.
They further looked for links between the chromosome 3 region and phenotypic markers they had available for hospitalized patients and discovered significant associations with two markers of tissue damage: lactate dehydrogenase and creatine kinase.
They also found hints for biological functions in some of the other loci they identified. For example, the signal on chromosome 6 included a variant that has been associated with lung cancer and is more common in East Asian than other populations. The locus on chromosome 12 included a gene cluster that encodes antiviral restriction enzyme activators, as well as variants associated with protection from chronic lymphocyte leukemia. Further, the signal on chromosome 19 contained a variant previously linked to protection from autoimmune disease, and another one associated with interstitial lung disease. Finally, the chromosome 21 locus had a region associated with inflammatory response.
Gita Pathak, a researcher at Yale School of Medicine, presented additional results on behalf of the COVID-19 Host Genetics Initiative. She and her team analyzed GWAS data from 3,200 hospitalized COVID-19 patients and 900,000 controls, combining it with gene expression and splicing data from the Genotype-Tissue Expression (GTEx) project.
They conducted a multi-tissue transcriptome-wide association study, which can be used to predict genetically regulated gene expression, and identified six risk genes for COVID-19 hospitalization. In particular, they recapitulated the chromosome 3 region found in the previous GWAS studies, including genes involved in cytokine regulation, and identified IFNAR2, a gene encoding a type-1 interferon receptor that is part of the innate antiviral response, on chromosome 21 as a risk gene. Next, they tested for allele-specific expression of the six genes and found this to be the case for the gene SLC6A20.
To see whether alternative splicing also played a role, they then performed a splice-TWAS and found 11 significant splice sites in four genes, recapitulating the prior TWAS associations.
To better understand the clinical relevance of the regions they identified, they conducted a phenome-wide association study using 1,700 clinical traits available for 70,000 individuals in the Vanderbilt Biobank. They found associations between IFNAR2 and bundle branch block, an abnormal electric impulse of the heart, as well as with migraine. Both heart arrhythmia and migraines, Pathak noted, have been observed in COVID-19 patients. A similar association study with lab test results revealed links with blood markers of immune response.
In a third study, researchers at Regeneron looked for genetic associations using two sets of data: imputed genotyping data from about 880,000 individuals, among them 13,000 COVID-19-positive cases, and exome sequencing data from almost 600,000 individuals, including 2,400 COVID-19 cases. The data came from a variety of sources, including the UK Biobank, Geisinger Health System and Regeneron, Ancestry, the Penn Medicine Biobank, and FinnGen.
Jack Kosmicki, a researcher at the Regeneron Genetics Center, explained that the goal was to find genetic associations with three phenotypes: infection, hospitalization, and severe disease as defined by the need for oxygen, ventilation, or death.
In total, they identified four significant loci, three of them associated with hospitalization and two with infection. These included the previously published chromosome 3 locus, which was associated with all three phenotypes, in particular hospitalization risk, and three other loci on chromosome 9, chromosome 19, and chromosome 21. The ABO blood group locus on chromosome 9 only appeared to be linked with infection, not hospitalization risk, he noted.
Their exome data yielded no associations, likely because the data contained too few cases and was underpowered, Kosmicki explained. The researchers were also unable to fine-map their GWAS loci with the exome data, as no genes in those regions showed significant burdens of rare variants.
Likewise, the exome data was unable to replicate the findings of two previous exome sequencing studies, published in Science and in the Journal of the American Medical Association, that found rare mutations affecting the immune response in severely ill COVID-19 patients. However, the researchers did find an association of the chromosome 21 locus with IFNAR2, similar to the Yale group.