Genome-wide association study of BNT162b2 vaccine-related myocarditis identifies potential predisposing functional areas in Hong Kong adolescents

Abstract

Vaccine-related myocarditis associated with the BNT162b2 vaccine is a rare complication, with a higher risk observed in male adolescents. However, the contribution of genetic factors to this condition remains uncertain. In this study, we conducted a comprehensive genetic association analysis in a cohort of 43 Hong Kong Chinese adolescents who were diagnosed with myocarditis shortly after receiving the BNT162b2 mRNA COVID-19 vaccine. A comparison of whole-genome sequencing data was performed between the confirmed myocarditis cases and a control group of 481 healthy individuals. To narrow down potential genomic regions of interest, we employed a novel clustering approach called ClusterAnalyzer, which prioritised 2,182 genomic regions overlapping with 1,499 genes for further investigation. Our pathway analysis revealed significant enrichment of these genes in functions related to cardiac conduction, ion channel activity, plasma membrane adhesion, and axonogenesis. These findings suggest a potential genetic predisposition in these specific functional areas that may contribute to the observed side effect of the vaccine. Nevertheless, further validation through larger-scale studies is imperative to confirm these findings. Given the increasing prominence of mRNA vaccines as a promising strategy for disease prevention and treatment, understanding the genetic factors associated with vaccine-related myocarditis assumes paramount importance. Our study provides valuable insights that significantly advance our understanding in this regard and serve as a valuable foundation for future research endeavours in this field.

Keywords: BNT162b2 vaccine; Clustering; Genetic risk predisposition; Vaccine side effect; Vaccine-induced myocarditis.

Fig. 1

Schematics of ClusterAnalyzer. Based on summary statistics of SNPs tested for BNT162b2 vaccine-related myocarditis association, (i) SNPs with p < 0.05 were selected and (ii) clustered by genomic distance. P-value consistencies were evaluated for each signal within each cluster and quantified as Mean Absolute Error (MAEs). Clusters containing at least one signal with an MAE < 3.0 were prioritized

Fig. 2

Manhattan and quantile-quantile (QQ) plots. Manhattan plot of (A) initial SNP associations and (B) prioritized SNP associations using ClusterAnalyzer. The horizontal axis in each plot represents the SNP location from chromosome 1 to chromosome 22, while the vertical axis represents significance level in negative logarithm. The red line corresponds to p = 1.0 × 10^− 5. Green line corresponds to p = 5.0 × 10^− 8. QQ plots are also provided for (C) initial SNP associations and (D) prioritized SNP associations

Fig. 3

Linkage-association plots and regional association plots generated by LocusZoom for chr4:81,892,037–81,965,272 (A, B) and chr11:40,473,809 − 40,585,162 (C, D). SNPs with different direction of Odds Ratio (OR) compared to the top SNP are not shown. Dotted lines represent predicted log association P values based on Linkage Disequilibrium (LD, R²)

Fig. 4

Visualization of gene set enrichment analysis results generated by Cytoscape, using g:Profiler for prioritized genes. Shades of red indicate false discovery rates, calculated based on g:SCS (Set Counts and Sizes, PMID: 17478515). Pathway clustering was carried out with AutoAnnotate, an app in Cytoscape, using the default settings