199Genetic architecture of left ventricular phenotypes derived from 17,000 CMR studies in the UK Biobank population imaging cohort

Abstract

Background: Although the contribution of modifiable risk factors on LV remodelling is well-established, the genetic basis of LV phenotypes is unclear at present. Purpose: To discover genetic variants associated with LV imaging phenotypes by conducting genome-wide association studies (GWASs). Methods: The study consisted of 16,923 European UK Biobank participants (mean ± SD age: 62.5 ±7.5 years; 45.8% men) without prevalent myocardial infarction or heart failure. All LV parameters were measured from the short-axis bSSFP images acquired with a 1.5T Siemens scanner. The first 5000 studies were manually evaluated while the remainder were automatically analysed with a deep convolutional neural network. Genotyping was done using bespoke Affymetrix arrays, augmented by imputation (∼ 7 million variants at minor allelic frequency ≥ 5%). We performed GWASs of six LV traits - LV end-diastolic volume (LVEDV), LV end-systolic volume (LVESV), LV stroke volume (LVSV), LV ejection fraction (LVEF), LV mass (LVM) and LV mass to end-diastolic volume ratio (LVMVR) - using a linear mixed model adjusted for age, sex, height, weight, systolic blood pressure, phenotype-derivation method, genotype array type and imaging centre. The threshold for genome-wide significance was set at p < 1 x 10-8 to additionally account for multiple phenotypes. We identified candidate genes at genome-wide significant loci using multiple orthogonal lines of evidence including presence of coding variant, gene expression data, chromatin interaction analyses and knockout models. Results: We discovered nine unique genome-wide significant loci, seven of which were novel (Figure 1A). The TTN, MTSS1 and BAG3 loci were associated with more than one LV trait (Figure 1B). Extensive in-silico bioinformatic analyses prioritised seven candidate causal genes (TTN, BAG3, MTSS1, GRK5, ALPK3, NMB, CDKN1A) associated with LV phenotypes based on two or more independent sources of evidence (Figure 1C). The TTN gene encodes a large abundant protein of striated muscle instrumental for cardiogenesis. BAG3 is a cyto-pro-tective protein that regulates HSP70 family of molecular chaperones and appears to be critical for maintenance of myotube survival. The MTSS1 gene is involved in actin filament organisation while GRK5, a G protein-coupled receptor kinase, has been shown to modulate β-adrenergic receptor signalling pathway. ALPK3 gene promotes cardiomyocyte differentiation and the polymorphism of NMB gene is known to be associated with obesity. Lastly, CDKN1A functions as a regulator of cell cycle progression and has been shown to play a role in cardio-myocyte hypertrophy. Conclusions: In these largest GWASs to-date of CMR LV phenotypes, we report nine genetic loci, and indicate several candidate genes and key biological pathways which not only improve our understanding of the genetic architecture of prognostically important LV phenotypes but also shed light on potential novel therapeutic targets for LV remodelling.