Integrated Multiomics Approach to Identify Genetic Underpinnings of Heart Failure and Its Echocardiographic Precursors

Abstract

Background: Heart failure (HF) may arise from alterations in metabolic, structural, and signaling pathways, but its genetic architecture is incompletely understood. To elucidate potential genetic contributors to cardiac remodeling and HF, we integrated genome-wide single-nucleotide polymorphisms, gene expression, and DNA methylation using a transomics analytical approach. Methods: We used robust rank aggregation (where the position of a certain gene in a rank order list [based on statistical significance level] is tested against a randomly shuffled rank order list) to derive an integrative transomic score for each annotated gene associated with a HF trait. Results: We evaluated ≤8372 FHS (Framingham Heart Study) participants (54% women; mean age, 55±17 years). Of these, 62 (0.7%) and 35 (0.4%) had prevalent HF with reduced ejection fraction and HF with preserved left ventricular ejection fraction, respectively. During a mean follow-up of 8.5 years (minimum–maximum, 0.005–18.6 years), 223 (2.7%) and 234 (2.8%) individuals developed incident HF with reduced ejection fraction and HF with reduced ejection fraction, respectively. Top genes included MMP20 and MTSS1 (promotes actin assembly at intercellular junctions) for left ventricular systolic function; ITGA9 (receptor for VCAM1 [vascular cell protein 1]) and C5 for left ventricular remodeling; NUP210 (expressed during myogenic differentiation) and ANK1 (cytoskeletal protein) for diastolic function; TSPAN16 and RAB11FIP3 (involved in regulation of actin cytoskeleton) for prevalent HF with reduced ejection fraction; ANKRD13D and TRIM69 for incident HF with reduced ejection fraction; HPCAL1 and PTTG1IP for prevalent HF with reduced ejection fraction; and ZNF146 (close to the COX7A1 enzyme) and ZFP3 (close to SLC52A1 —the riboflavin transporter) for incident HF with reduced ejection fraction. We tested the HF-related top single-nucleotide polymorphisms in the UK biobank, where rs77059055 in TPM1 (minor allele frequency, 0.023; odds ratio, 0.83; P =0.002) remained statistically significant upon Bonferroni correction. Conclusions: Our integrative transomics approach offers insights into potential molecular and genetic contributors to HF and its precursors. Although several of our candidate genes have been implicated in HF in animal models, independent replication is warranted.