High-resolution epigenome analysis in nasal samples derived from children with respiratory viral infections reveals striking changes upon SARS-CoV-2 infection

Abstract

Background: DNA methylation patterns of the human genome can be modified by environmental stimuli and provide dense information on gene regulatory circuitries. We studied genome-wide DNA methylation in nasal samples from infants (<6 months) applying whole-genome bisulfite sequencing (WGBS) to characterize epigenome response to 10 different respiratory viral infections including SARS-CoV-2.

Results: We identified virus-specific differentially methylated regions (vDMR) with human metapneumovirus (hMPV) and SARS-CoV-2 followed by Influenza B (Flu B) causing the weakest vs. strongest epigenome response with 496 vs. 78541 and 14361 vDMR, respectively. We found a strong replication rate of FluB (52%) and SARS-CoV-2 (42%) vDMR in independent samples indicating robust epigenome perturbation upon infection. Among the FluB and SARS-CoV-2 vDMRs, around 70% were hypomethylated and significantly enriched among epithelial cell-specific regulatory elements whereas the hypermethylated vDMRs for these viruses mapped more frequently to immune cell regulatory elements, especially those of the myeloid lineage. The hypermethylated vDMRs were also enriched among genes and genetic loci in monocyte activation pathways and monocyte count. Finally, we perform single-cell RNA-sequencing characterization of nasal mucosa in response to these two viruses to functionally analyze the epigenome perturbations. Which supports the trends we identified in methylation data and highlights and important role for monocytes.

Conclusions: All together, we find evidence indicating genetic predisposition to innate immune response upon a respiratory viral infection. Our genome-wide monitoring of infant viral response provides first catalogue of associated host regulatory elements. Assessing epigenetic variation in individual patients may reveal evidence for viral triggers of childhood disease.

Keywords: DNA methylation; GWAS; SARS-CoV-2; infant; influenza; viral infection.

Figure 1:

Hierarchical clustering of top 50% variable CpGs in nasal samples derived from ten respiratory viral infections.

Figure 2:

Respiratory viruses elicit differential magnitude and direction of host methylation response. A) Discovery rate of vDMR. hMPV, N=496; RV, N=665; Adeno, N=943; Corona (Oc43), N=1281; EVD68, N=3129; FluA, N=3439; PIV3, N=4788; RSV, N=12557; FluB, N=14361; and SARS-CoV-2, N=78542. B) Methylation status of vDMR

Figure 3:

FluB and SARS-CoV-2 infection are associated with differential responses of epithelial and immune methylomes. A) Fold change compared to control sample of presence of regions annotated to specific cell lineages in replicated vDMR sets. B) Fold change compared to control sample of presence of genomic regions annotated to specific immune cell lineages in hypermethylated vDMR for FluB and SARS-CoV-2. C) -log₁₀ p-value fisher’s exact test for a positive association between infection and presence of genomic regions annotated to specific immune lineages in hypermethylated vDMR. Dashed line represents p-value of 0.05.

Figure 4:

Functional analysis of replicated vDMR in adolescent infection. A) UMAP projection of cells obtained from pooled adolescent nasal mucosa samples infected with FluB, SARS-CoV-2 and uninfected age-matched controls. B-C) Expression level of genes “modules” derived from genes nearest vDMR for Flu B (B) and SARS-CoV-2 (C). D) Percent of total cells in each cell type for uninfected and infected pooled samples. E) Heatmaps of cell-cell interactions between all cell types for FluB infected, SARS-CoV-2 infected, and uninfected controls.