Unraveling the dual role of fibroblast growth factor 5 (FGF5) in cardiorenal diseases through multi-omics causal inference.

Abstract

Background: As populations age worldwide, multimorbidity has become an increasingly prominent public health challenge. Among age-associated conditions, cardiovascular diseases (CVDs) are among the most prevalent and life-threatening, and they frequently coexist with other chronic disorders, particularly chronic kidney disease (CKD). Pleiotropic molecules may provide an important entry point for elucidating comorbidity mechanisms and identifying therapeutic opportunities. However, the clinical translational value of different pleiotropic candidates exhibits substantial heterogeneity, and their directional effects remain poorly understood.

Methods: In this study, we integrated four large-scale protein quantitative trait loci (pQTL) cohorts from European and East Asian populations, and applied a dual-window Bayesian colocalization strategy to systematically screen for shared genetic signals across 2719 plasma proteins and 34 age-related diseases. We further conducted causal inference and integrative functional characterization of the core candidate protein via Mendelian randomization (MR), observational cross-sectional analysis, genetic fine-mapping, cis-regulatory element analysis and single-cell transcriptomics.

Results: Following systematic cross-ancestry screening of 2719 plasma proteins, FGF5 was identified as the core candidate protein, with its circulating pQTL signals showing strong colocalization with CVDs, CKD, and cardiorenal-related phenotypes. Notably, this genetic regulatory effect of FGF5 exhibited marked renal tissue specificity, with its pQTL signals colocalizing strongly with expression quantitative trait loci (eQTL) signals exclusively in the renal cortex, glomeruli, and renal tubules (PP.H4 > 0.97). MR analysis suggested that genetically predicted elevation in circulating FGF5 levels was significantly associated with an increased risk of CVD, with 42.9-66.7% of this association potentially mediated by blood pressure. In contrast, MR analyses suggested a potentially protective effect of elevated FGF5 levels on CKD, with 8.4-11.8% of the association potentially mediated by uric acid. Clinical observational analyses provided complementary support for this bidirectional association pattern, with plasma FGF5 levels showing a positive linear association with CVD risk and a nonlinear U-shaped association with CKD prevalence. Genetic fine-mapping prioritized rs2903657 and rs17004846 as the most likely candidate causal variants within the same strongly linked cis-regulatory haplotype. Phenome-wide association scan (PheWAS) further supported the pleiotropic involvement of these two variants across blood pressure, renal function, and metabolic traits. Single-cell transcriptomic analysis further showed that FGF5 was preferentially enriched in proximal tubular epithelial cells in the injury-repair state.

Conclusions: This study systematically elucidates the potential antagonistic pleiotropic effects of FGF5 across cardiorenal diseases. These findings provide clues for understanding the shared genetic mechanisms of cardiorenal multimorbidity and offer genetic evidence to guide future functional translation and precision intervention strategies.