Abstract
BACKGROUND: Epicardial adipose tissue (EAT) is a metabolically active visceral fat depot contributing to coronary atherosclerosis, yet the molecular mechanisms underlying EAT-related coronary artery disease (CAD) in type 2 diabetes mellitus (T2DM) remain unclear. Previously, we identified novel circulating miRNAs targeting fatty acid metabolism in T2DM-CAD. This study aimed to investigate whether EAT may explain the association between dysregulated hsa-miR-4505, hsa-miR-4743-5p, hsa-miR-4750-3p and CAD in T2DM patients and whether it can detect diabetic atherosclerosis alone or in a multi-modal combination. METHODS: Seventy-six patients with T2DM and/or CAD along with eighteen healthy controls were enrolled in the study. All participants underwent transthoracic echocardiography to assess EAT thickness on the free wall of the right ventricle at end-systole and bioelectrical impedance analysis for body composition determination. Spearman's rank correlation and multivariate linear regression accounting for relevant clinical confounders were used to explore the associations between EAT and miRNAs. To further investigate whether EAT acts as an intermediary between miRNA and CAD in T2DM, causal mediation analysis was employed. The receiver operating characteristics curves were generated to evaluate the diagnostic performance of the combined models built using multivariate logistic regression. RESULTS: The median EAT thickness was significantly higher in T2DM-CAD patients compared to T2DM subjects and controls (p < 0.0001). The bivariate analysis showed a positive correlation between triglyceride concentration and EAT thickness, and a negative one with hsa-miR-4750-3p expression. After multivariable adjustment, hsa-miR-4750-3p (β = - 0.445, p = 0.003) emerged as a standalone predictor of EAT thickness. Logistic regression analysis identified enlarged EAT, up-regulated hsa-miR-4505, hsa-miR-4743-5p and down-regulated hsa-miR-4750-3p to be independently associated with higher CAD risk in T2DM. Adding miRNAs to EAT improved CAD detection in T2DM (AUC = 0.988), outperforming both EAT (AUC = 0.869), clinical factors (AUC = 0.829), and their combination (AUC = 0.901). The mediation analysis revealed that EAT accounted for 48.79% of the total effect of hsa-miR-4750-3p on CAD in T2DM. CONCLUSIONS: These findings suggest that the proposed miRNA-EAT regulatory axis may be involved in the pathogenesis of diabetic atherosclerosis, with EAT appearing to partially mediate the relationship between hsa-miR-4750-3p and CAD. The integrated approach linking EAT and miRNAs holds potential for CAD risk stratification in T2DM.