Proteomic Pathways Linking Type 2 Diabetes Genetic Clusters to Cardiovascular Diseases: A Network Mendelian Randomisation Study.

BACKGROUND: Genetic clusters related to Type 2 diabetes (T2D) have differential impact on cardiovascular diseases (CVDs), although the underlying mechanisms, such as proteomic perturbation, remain unexplored. We conducted a network Mendelian randomisation (MR) study to identify proteomic mediators linking T2D genetic clusters and CVDs. METHODS: We assessed the associations between genetic liability to T2D and eight genetic clusters (N = 2 535 601) with 12 CVDs (N ≤ 1 946 349). For significant pairs, we performed a two-step MR (mediation) to identify proteins mediating these associations. Inverse-variance weighted method was the main analysis, with sensitivity analyses including genetic colocalisation. Robustness was evaluated in European-specific analyses as well as using Olink-measured proteomic data. False discovery rate (FDR) was used to correct for multiple comparisons. RESULTS: Obesity cluster was linked to most CVDs (except haemorrhage strokes), with similar patterns for body fat, metabolic syndrome and lipodystrophy clusters. Majority of the protein signals (2739 out of 4907) were related to obesity cluster although glycemia-related clusters were only associated with a few proteomic signals (≤ 13). Amongst 16 proteins supported by genetic colocalisation (PPH4 > 0.8), the proportion mediated ranged from 1.2% to 36.8%. APOC3, ADH1B, F11 and KLKB1 appeared to be shared across the different genetic clusters, implicating their shared role in CVDs risk. Sensitivity and replication analyses gave similar estimates. CONCLUSION: This proteogenomic MR study delineates protein-mediated pathways linking cluster-stratified T2D genetic predisposition to cardiovascular risk, which can inform preventive strategies by stratifying individuals based on distinct biological pathways represented by these genetic clusters.