Proteogenomic analysis in population biobanks identifies new therapeutic targets across heart failure subtypes

A Henry,C Finan,J Gratton,N Paliwal,S Chopade,S Jassal,A K Hedman,E B Fauman,A Malarstig,A D Hingorani,R T Lumbers

DOI: https://doi.org/10.1093/eurheartj/ehae666.3731

IF: 39.3

2024-10-29

European Heart Journal

Abstract:Background Heart failure (HF) represents a multifactorial clinical syndrome with differential response to existing therapies across its subtypes. Proteogenomic analysis of genetic variants associated with changes in circulating protein level and risk of heart failure subtypes provides an opportunity for a systematic identification of proteins with subtype-specific causal effects as novel therapeutic targets. Purpose To identify proteins as potential therapeutic targets for specific heart failure subtypes by leveraging proteomics and genomics data in population biobanks. Methods We evaluated the causal effects of 2,773 proteins on 5 HF-related outcomes, including overall HF, non-ischaemic HF (ni-HF), ni-HF with reduced / preserved ejection fraction (ni-HFrEF / ni-HFpEF), and dilated cardiomyopathy (DCM). For each protein-phenotype pair, we performed a two-sample Mendelian randomisation (MR) analysis using protein-specific genetic association estimates derived from protein quantitative trait locus (pQTL) studies in 114,145 individuals from deCODE, UK Biobank, Fenland study, and SCALLOP consortium and genome-wide association studies (GWAS) of HF outcomes in 145,795 individuals by HERMES Consortium. We further prioritised candidate protein targets based on replication of results across datasets, model parameters, assays, and colocalization of causal genetic variants. Finally, we profiled the causal effects of identified proteins on 32 related cardiac traits and risk factors to explore potential mechanisms of actions and on-target side effects. Results We identified 493 proteins associated with at least one HF outcome at a false discovery rate (FDR) < 1% in the primary MR analysis. We further prioritised 44 proteins with concordant effect direction across datasets, model parameters, and assays or colocalization in at least one dataset. All 44 identified proteins showed cross association with at least one of 35 tested traits other than HF or cardiomyopathies at FDR <1%, including coronary artery disease (18 proteins), body mass index (11 proteins), type 2 diabetes (6 proteins), systolic / diastolic blood pressure (13 proteins), and chronic kidney disease (9 proteins). Conclusion We characterised the causal associations of 2,773 circulating proteins across HF-related outcomes and prioritised 44 proteins by integrating proteomics and genomics data in population biobanks. These findings extend our understanding of the molecular mechanism underlying different forms of HF and may inform the development of new therapeutic strategies.Study design

cardiac & cardiovascular systems