Analysis of cardiac magnetic resonance imaging in 36,000 individuals yields genetic insights into dilated cardiomyopathy

Abstract

Dilated cardiomyopathy (DCM) is an important cause of heart failure and the leading indication for heart transplantation. Many rare genetic variants have been associated with DCM, but common variant studies of the disease have yielded few associated loci. As structural changes in the heart are a defining feature of DCM, we report a genome-wide association study of cardiac magnetic resonance imaging (MRI)-derived left ventricular measurements in 36,041 UK Biobank participants, with replication in 2184 participants from the Multi-Ethnic Study of Atherosclerosis. We identify 45 previously unreported loci associated with cardiac structure and function, many near well-established genes for Mendelian cardiomyopathies. A polygenic score of MRI-derived left ventricular end systolic volume strongly associates with incident DCM in the general population. Even among carriers of TTN truncating mutations, this polygenic score influences the size and function of the human heart. These results further implicate common genetic polymorphisms in the pathogenesis of DCM.

Fig. 1. Manhattan plots of genome-wide association discovery analyses of cardiac MRI phenotypes.

For each cardiac MRI phenotype, the −log10(P value) is graphed on the y axis at each chromosomal position on the x-axis. P is the BOLT-LMM P value. The nearest gene to each genome-wide significant lead SNP is labeled at each locus, except when a cardiomyopathy-related gene is present within 500 kb of the lead SNP. SNPs are colored blue near loci that have previously been observed in common genetic analyses using cardiac traits (TTE or cardiac MRI). SNPs are colored red near loci that were previously unreported.

Fig. 2. PheWAS highlights the connection between a polygenic score for LVESVi and dilated cardiomyopathy.

The polygenic score derived from LVESVi was applied to PheCodes (panel a) and curated disease phenotypes (panel b) in the UK Biobank. Each of the curated phenotypes is defined in Supplementary Table 1. For both panels (a) and (b), the x-axis represents the identifying code for the disease phenotype. The y-axis represents the −log10 of the P value of the association between the polygenic score and the phenotype in a logistic model adjusted for age at enrollment, the genotyping array, sex, and the first five principal components of ancestry. Triangles oriented upward represent betas that are concordant with the LVESVi PRS (e.g., a higher LVESVi PRS corresponds with a higher risk of DCM), and the reverse is true for downward-oriented triangles. The three most strongly associated phenotypes in each panel are labeled in the figure. The triangles in panel (a) are colored by three-digit PheCode. The triangles in panel (b) are colored red if positively correlated with the LVESVi polygenic score, and blue if negatively correlated. The PheWAS plots for all seven cardiac MRI phenotypes are available in Supplementary Fig. 7.

Fig. 3. The LVESVi polygenic score influences the risk for incident dilated cardiomyopathy.

The cumulative DCM incidence (defined as 1 minus the Kaplan–Meier survival estimate) is plotted for individuals in the bottom tenth percentile (blue), middle 80% (gray), and top tenth percentile (red) for the LVESVi polygenic score. The 95% confidence intervals (derived from the standard error of the cumulative hazard) are represented with lighter colors. The x-axis represents the number of years since enrollment in the UK Biobank. The y-axis represents the cumulative incidence of DCM.