Genetic evidence prioritizes circulating proteins for heart failure beyond shared BMI-related genetic liability

Heart failure (HF) and body mass index (BMI) share substantial genetic architecture, which may lead genetically informed target discovery to preferentially identify adiposity-related pathways. We sought to identify circulating proteins associated with HF beyond this shared genetic component.

We applied GWAS-by-subtraction to overall HF, nonischemic HF, and nonischemic HF with reduced or preserved ejection fraction to derive BMI-related and BMI-subtracted HF components. We then performed proteome-wide cis -pQTL Mendelian randomization and colocalization using four independent proteomic cohorts, followed by tissue-specific eQTL colocalization, cardiac transcriptomic annotation, and druggability assessment.

Compared with the original HF phenotypes, the BMI-subtracted components showed attenuated genetic correlations with BMI (0.045-0.147) while retaining 28 independent loci for overall HF and nine for nonischemic HF. Across 19,930 protein-HF tests, 11 associations involving nine proteins were prioritized by the Mendelian randomization and colocalization analyses. For example, a 1-SD increase in genetically predicted CELSR2 abundance was associated with lower overall HF risk (odds ratio, 0.96 [95% CI, 0.94-0.98]; P =8.6×10 ^-7 ), whereas a 1-SD increase in genetically predicted CSF3 abundance was associated with higher nonischemic HF risk (odds ratio, 1.32 [95% CI, 1.18-1.48]; P =2.0×10 ^-6 ). CELSR2 and TMEM106B colocalized with cis -eQTLs in failing left ventricular myocardium, and DAG1 showed cardiomyocyte enrichment with concordant downregulation in failing hearts.

We identified nine circulating proteins associated with HF beyond the genetic component shared with BMI. These findings extend the range of genetically supported pathways implicated in HF and nominate candidate proteins for further mechanistic and therapeutic investigation.

What Is New?: GWAS-by-subtraction separated HF genetic susceptibility into BMI-related and BMI-subtracted components, allowing proteome-wide target discovery to focus on HF associations beyond the genetic component shared with BMI.Proteome-wide Mendelian randomization and colocalization prioritized nine circulating proteins.Integration with cardiac molecular data provided additional support for selected proteins, including failing-left-ventricular eQTL colocalization for CELSR2 and TMEM106B and cardiomyocyte expression with concordant myocardial downregulation for DAG1.What Are the Clinical Implications?: Accounting for shared BMI-related genetic liability may extend human genetics-guided target discovery to HF pathways that could be overlooked when conventional HF GWAS are used as the disease outcome.The findings illustrate that therapeutic interpretation must consider the genetically supported direction of effect. The LPA association was consistent with Lp(a) lowering, whereas the CSF3 and RSPO3 results did not support straightforward repurposing of existing pharmacological strategies.The prioritized proteins provide candidates for further functional and therapeutic investigation, but replication and experimental validation are needed before their clinical relevance can be established.