Abstract
The clinical heterogeneity of COVID-19 is well known but its determinants are not fully understood, and the regulatory mechanisms connecting host genetic risk to the transcriptional signature of severe disease have seldom been addressed in one framework. Peripheral blood mononuclear cells (PBMCs) provide a convenient way to measure the systemic immune dysregulation seen in severe SARS-CoV-2 infection, yet most transcriptomic studies of this cell type stop at describing differential expression, rather than asking what drives it. In this study, a focused discovery cohort of 10 PBMC transcriptomes (five healthy controls and five severe COVID-19 patients) from the RNA-sequencing dataset GSE152418 processed through a HISAT2/featureCounts pipeline and differential expression was determined using DESeq2. The gene set was then tested against four largely complementary lines of evidence: Gene Ontology and KEGG pathway enrichment; a miRDB based microRNA target-network analysis; genome-wide ChIP-seq data for CTCF in CD14+ monocytes downloaded from ENCODE; and summary statistics from a published genome-wide association study of severe COVID-19. Thus, genes were prioritized on the basis of convergent support from multiple methods as opposed to any one analytical tool. Significant differential expression (padj < 0.01) was found in 1,459 genes. Pathway enrichment repeatedly highlighted two processes: disruption of the mitotic spindle checkpoint and impairment of mitochondrial oxidative phosphorylation. CTCF ChIP-seq identified 42,243 binding sites across the genome with about 40% in gene promoters, including direct occupancy at IL6, CXCL2, ACE2, TMPRSS2 and NFKBIA. A reanalysis of the genome-wide association study (GWAS) confirmed the previously reported risk loci at 3p21.31 and at the ABO locus on 9q34.2. When all four evidence layers were combined, three genes stood out, MMP8 (upregulated), and IL23R and CD209 (both downregulated), each supported by differential expression and CTCF occupancy, with convergent support from the microRNA target network (see Discussion). Altogether, we propose a model of how genetic susceptibility, the activity of transcription factors, and post-transcriptional control regulate the immunopathology of severe COVID-19. MMP8, IL23R and CD209 are candidates for future experimental work.