Loss-of-function variants in founder population highlight atrial myopathy as susceptibility to atrial fibrillation

Abstract Background Atrial fibrillation (AF) is the most common cardiac arrhythmia, and it's associated with genes responsible for cardiac conduction, cardiogenesis, and cardiac structure. The Finnish population has evolved in relative isolation and undergone several bottlenecks, resulting in enrichment of deleterious variants. This facilitates identification of possible causal genes through protein-truncating variants that likely would not have been identified in other populations. Purpose To identify novel genetic associations with large effect on AF risk. Methods We accessed publicly available summary statistics on AF (ICD-10 code I48) from the FinnGen project R6, containing 28,670 AF cases and 135,821 controls. Then, we annotated all variants in the summary statistics and extracted variants that were predicted with a high effect impact (n=7,113). To account for multiple testing, a false discovery rate cutoff <10% was applied. The UK Biobank was assessed for comparison of allele frequencies. Afterwards, we investigated protein and single-cell RNA expression of the genes of interest in human atrial tissue. Human atrial tissue was obtained by us from 7 individuals and used for mass spectrometry-based proteomics, while single-nucleus RNA sequencing (snRNAseq) data from human hearts was acquired by Tucker [1]. Results In a genome-wide association study (GWAS) focusing on loss-of function (LOF) variation, we report two LOF variants in the structural genes SYNPO2L and CTNNA3 with much higher allele frequencies compared to non-Finnish Europeans (85-fold and 80-fold enrichment, respectively). The variants increase the risk of AF considerably, which is emphasized as the two variants show the highest effect sizes of all GWAS variants ever associated with AF (SYNPO2L; odds ratio [OR] = 2.79, P-value = 1.32x10–8 and CTNNA3; OR = 2.43, P-value = 9.40x10–7), exceeding that of most clinical risk factors. We accessed phenome-wide association study (PheWAS) results on both variants. There was no association with other phenotypes for the SYNPO2L variant whereas the CTNNA3 variant showed suggestive association with valvular heart disease (P-value = 2.95x10–5). SYNPO2L and CTNNA3 exhibited high protein and RNA expression levels in atrial tissue and were predominantly expressed in cardiomyocytes (Fig. 1). Conclusion We identified novel associations between LOF variation in the structural genes SYNPO2L and CTNNA3 and AF. Our study showed how genetic examination of a European subpopulation facilitates discovery of genetic variants and pathophysiological understanding. The results underline the importance of thoroughly investigating subpopulations as unique variants with large effect sizes can be identified. Funding Acknowledgement Type of funding sources: Public hospital(s). Main funding source(s): This work was supported by the Research Foundation at Rigshospitalet and the Hallas-Møller Emerging Investigator Novo Nordisk.