Population-specific and cross-ancestry genome-wide association study identifies shared genetic architecture and 6 new risk loci including CAMK2D associated for Brugada syndrome

Abstract Introduction The Brugada syndrome (BrS) is an inherited arrhythmia characterized by coved-type ST-segment elevation in the right precordial leads and an increased risk for sudden cardiac death. Genome-wide association studies (GWAS) of BrS, mainly performed in European ancestry, have illustrated its complex genetic architecture as an polygenic disease besides a monogenic trait due to mutations in the major responsible cardiac sodium channel gene SCN5A. Clinical manifestations of BrS including the disease prevalence and the incidence of sudden death are known to be variable between Asian and European populations, however, its underlying mechanisms are largely unknown. Objectives We aimed to identify novel BrS-associated loci and to assess ancestry-dependent genetic heterogeneity that may underlie the variable clinical manifestations of BrS patients across different populations. Methods We performed a genome-wide meta-analysis of BrS in the Japanese ancestry (940 cases and 1,634 controls), followed by a cross-ancestry meta-analysis of Japanese and European GWAS (total of 3,760 cases and 11,635 controls) to identify novel risk loci and compared the allelic effects between two ancestries. We characterized the novel loci based on fine-mapping of potential causal variants, and expression and splicing quantitative trait associations using the Genotype-Tissue Expression (GTEx) database. Results The Japanese-specific genome-wide meta-analysis identified a novel association signal near the ZSCAN20 (lead variant, rs2336244; P=1.0E-08) besides previously reported 4 association signals at the SCN5A-SCN10A and HEY2 loci. Cross-ancestry genome-wide meta-analysis identified 17 association signals including 6 novel loci. The effect directions were consistent for 94.1% of the variants (16 of 17; P for sign test = 0.00027) and their allelic effects were highly correlated across ancestries (Pearson’s R=0.91 [P=2.9E-07]). The genetic risk score (GRS) derived from European-ancestry GWAS of BrS was significantly associated with the risk of BrS in the Japanese populations (odds ratio, 2.12 [95% confidence interval, 1.94–2.31]; P=1.2E-61), suggesting the shared genetic architecture across ancestries. The functional characterization of the 6 novel loci showed that one lead SNP on CAMK2D promotes an alternative splicing resulting in an isoform switch of calmodulin kinase II that favors up-regulation of the pro-inflammatory isoform δ9 and down-regulation of the pro-survival isoform δ3. Conclusions Our results identified 6 novel susceptible loci associated with BrS and revealed shared genetic architecture across ancestries.GWAS Manhattan plotShared genetic architecture