Proteomic signatures characterized by inflammatory signaling and metabolic processes have divergent associations with 30-year incidence of coronary artery disease in type 1 diabetes: the Pittsburgh Epidemiology of Diabetes Complications cohort.

Abstract

Background: While inflammation is hypothesized to be a possible driver of the excess coronary artery disease (CAD) observed in type 1 diabetes, prior biomarker studies have focused on candidate proteins, overlooking the complex systemic processes underlying atherosclerosis. We thus applied a high-throughput proteomics approach to identify proteins prospectively associated with CAD in type 1 diabetes. We further hypothesized inflammatory proteins would partially mediate the effect of glycemic exposure on CAD risk.

Methods: We measured 184 circulating proteins (Olink CVD II and III panels) in earliest available serum samples from the Pittsburgh Epidemiology of Diabetes Complications cohort of childhood-onset (< 17 years) type 1 diabetes (n = 417, baseline mean age 30, mean diabetes duration 22 years). Protein modules were identified using weighted correlation network analysis. Prospective associations between modules (per eigenvalue SD) and 30-year incident hard CAD (hCAD; CAD death, nonfatal myocardial infarction, revascularization, or blockage ≥ 50%) were estimated in Cox models. Causal mediation analysis estimated mediation of the HbA1c-hCAD association by the modules.

Results: Incident hCAD occurred in 151 participants (36.2%). We identified three protein modules: one comprising 56 proteins, enriched for TNF/Notch signaling, T-cell activation, and necroptosis; another comprising 33 proteins, enriched for lipid catabolism, epithelial cell apoptosis, vitamin D response; and a third comprising 33 proteins, enriched for hemostasis, wound healing, and vascular growth factors. After adjusting for traditional covariates and HbA1c, greater abundance of proteins in the first module was associated with increased hCAD incidence (HR = 1.35, 95% CI: 1.06-1.72, p = 0.014) while the second module was protective (HR = 0.82, 95% CI: 0.70-0.96, p = 0.017). The third module was not associated with hCAD (HR = 1.03, 95%CI: 0.86-1.23, p = 0.751). The first module mediated 14% of the total HbA1c effect on hCAD.

Conclusions: In this prospective analysis of a type 1 diabetes cohort, a protein signature characterized by inflammatory signaling partially mediated the effect of hyperglycemia on hCAD, while a signature characterized by metabolic processes including lipid catabolism and hormone response was associated with hCAD protection independent of glycemia. These findings move beyond single-marker associations, identifying distinct protein signatures that warrant further study as intervention targets for CAD prevention in type 1 diabetes.