Identification of potential therapeutic targets for coronary artery disease based on protein array and Mendelian Randomization analysis

Abstract

Abstract Background Currently, coronary artery disease (CAD) is a main cause of mortality worldwide. Exposing the unknown pathogenic mechanism and exploring potential therapeutic targets of CAD is a critical issue in current CAD research. We aimed to identify potential therapeutic target for CAD using 2-sample Mendelian randomization (MR) approach among differential expression proteins between CAD and control group. Methods Serum samples from 16 first-onset CAD patients diagnosed by coronary angiography and 10 non-CAD controls were collected and detected using protein array. We identified differentially expressed proteins based on “DEseq2” R package and found the core module using MCODE plugin in Cytoscape. Proteins in core modules were used as candidate protein for causal inference. Genetic instrument variables (IV) of candidate proteins were extracted from Age, Gene/Environment Susceptibility-Reykjavik Study (AGES-RS). Two-sample MR was performed to detect the causal association between the candidate proteins and CAD. For the identified protein causally associated with CAD, sensitivity analyses were used to confirm the finding, including using other MR approach, using more stringent instrument variables, and external cohorts for verification. Network MR was performed to explore whether metabolic risk factors for CAD mediated the risk of identified proteins, including indices of lipid metabolism, glucose metabolism, obesity and blood pressure. Mediation analysis was used to measure the effect of mediation. Results A total of 81 differentially expressed protein were found in CAD group compare to control group, among which 3 core functional modules were found. In five candidate proteins with available instrumental variables, only GP73 was causally associated with CAD risk. Per 1 SD increment in genetically determined GP73 level was associated with 12% increased CAD risk (odds ratio: 1.12; 95% confidence interval: 1.06–1.18, FDR<0.001). Sensitivity analysis yielded consistent results. Network MR analysis shows that across metabolic risk factors profiles for CAD, GP73 level was causally associated with low-density lipoprotein cholesterol (LDL-c), glycated hemoglobin (HbA1c) and waist-to-hip ratio (WHR). Besides, LDL-c, HbA1c and WHR serves as mediators in the causal pathway from GP73 to CAD and transmitted 40.4%, 1.4%, and 5.5% of the total effects, respectively. Conclusion Based on protein array and bioinformatics analysis, this study screened out 3 modules that play core functions in the differential expressed proteins. Among the 19 proteins in the core module, there was a causal relationship between circulating GP73 levels and incident CAD, and the causal effect is mainly mediated by lipid metabolism. The results suggest that GP73 can not only be used as a novel biomarker of CAD, but also may provide a new potential therapeutic target for the management of CAD. Funding Acknowledgement Type of funding sources: None. Figure 1