AHA Scientific Sessions 2026 Abstract · Nov 2, 2026
Proteomics for Elucidating Genetic Pathogenesis of Coronary Artery Disease in Diabetes
By Kamineni M, Saadatagah S, Schuermans A et al.
Background: Polygenic risk scores (PRS) for coronary artery disease (CAD) represent summative genetic risk, but performance varies by diabetes status. We sought to determine whether the downstream mediators for CAD vary among those genetically predisposed to CAD with concomitant diabetes. Methods: We analyzed participants with diabetes (HbA1c >= 6.5%) from the UK Biobank (UKB) and the Atherosclerosis Risk in Communities (ARIC) cohort with genetic and Olink (in UKB) and SomaScan (in ARIC) plasma proteomics data available. CAD PRS was computed using an externally-derived validated score. We implemented linear regressions to assess the associations between CAD PRS and each protein, adjusting for age, sex, and the first ten principal components of genetic ancestry. Proteins significant in UKB analyses were evaluated for replication in ARIC, using SomaScan proteomics. Replicated proteins were subsequently assessed for mediation of CAD PRS in association with incident CAD. Results: Among 42,132 participants from the UKB and 13,776 participants from ARIC, 1993 (4.7%) and 2058 (15.0%) had diabetes in the two cohorts, respectively. 16 proteins were significantly associated with CAD PRS in UKB, including CD34 and EPCAM. Of 16 proteins, 10 were available in ARIC. CAD PRS was significantly associated with L1CAM among participants with diabetes in both UKB and ARIC with coefficients of -0.16 (95% CI: [-0.23, -0.10], p-value: 1.83E-06) and -0.05 (95% CI: [-0.08, -0.01], p-value: 0.005) respectively. In mediation analyses, L1CAM did not attenuate the association between CAD PRS and incident CAD. Conclusions: In diabetic patients, CAD genetic risk was associated with circulating proteins involved in neovascularization and inflammation, including CD34 and EPCAM. While L1CAM showed consistent replication across both UKB and ARIC, it did not mediate the relationship between CAD polygenic risk and incident CAD, suggesting that it may represent a biomarker associated with CAD genetic risk rather than a causal mediator.