Genomic Structural Equation Modeling Reveals Shared Genetic Architecture and Pleiotropic Hub Genes of Sepsis-Induced Cardiomyopathy.

Background

Sepsis-induced cardiomyopathy (SICM) is a life-threatening complication driven by inflammatory cascades. Current genetic studies are restricted to single-trait analyses that cannot capture the shared genetic architecture spanning from immune dysregulation to structural myocardial damage.

Methods

We applied genomic structural equation modeling to integrate genome-wide association study (GWAS) summary statistics for six phenotypes-sepsis, cardiac troponin I, left ventricular ejection fraction (LVEF), left ventricular diastolic strain rate, right ventricular peak ejection rate, and heart failure-constructing a latent factor for the shared genetic basis of SICM-related phenotypes. Downstream analyses included multivariate GWAS, fine-mapping (SuSiE/FINEMAP), sparse canonical correlation analysis-based transcriptome-wide association study (sCCA-TWAS) with FOCUS prioritization, MAGMA gene-set enrichment, cell-type enrichment (CELLECT), spatial transcriptomic mapping (gsMap), and stratified LD score regression (S-LDSC).

Results

The model showed adequate fit (CFI = 0.936), with left ventricular diastolic strain rate and LVEF anchoring the factor most strongly (λ = 0.811 and 0.636, respectively). Multivariate GWAS identified 4220 lead variants, of which 4197 did not reach genome-wide significance in any constituent single-trait GWAS. Cross-referencing sCCA-TWAS with FOCUS fine-mapping prioritized 39 genes spanning inflammatory transduction, gap junction remodeling, proteostatic defense, and energy sensing. AMPK signaling was recurrently captured across fine-mapping and transcriptome-wide analyses. CELLECT identified cardiac muscle cells as the sole significant cell type.

Conclusions

This study provides the first integrative multi-trait genetic framework for the shared genetic basis of SICM-related phenotypes, identifying AMPK as a recurrently captured pleiotropic hub at the inflammation-metabolism intersection and providing a foundation for future biomarker and mechanistic investigations.