Genetic associations of circulating plasma proteins with cardiometabolic diseases

Figures

• 

  Fig. 1. Overview of this MR study design. After identifying that the top SNP with the strongest association signal in the cis-pQTL study was used as the single genetic instrument, SMR analysis comprehensively investigated the causal association of 2,011 plasma proteins with 6 major CVDs and 19 metabolic phenotypes. Plasma proteome data were obtained from two large-scale pQTL studies, including the UK Biobank Pharma Proteomics Project (UKB-PPP; N = 34,557) and the deCODE Health study (N = 35,559). Furthermore, we further explored potential therapeutic targets and evaluated the druggability of identified proteins using colocalization analysis. Additionally, a phenome-wide association study (PheWAS) was used to profile the possible side effects and indications of candidate proteins after their development into drugs. AF, atrial fibrillation; CAD, coronary artery disease; HF, heart failure; VTE, venous thromboembolism; PAD, peripheral artery disease; BMI, body mass index; WHR, waist hip ratio; SBP, systolic blood pressure; DBP, diastolic blood pressure; PP, pulse pressure; FI, fasting insulin; FG, fasting glucose; 2hGlu, two-hour glucose; HbA1c, glycated hemoglobin levels; LDL,…

• 

  Fig. 2. SMR and colocalization analysis on the associations between plasma proteins and the risk of six kinds of cardiovascular diseases. A Manhattan plot for SMR analysis. Proteins above the dotted line were those with false discovery rate corrected P < 0.05. B Forest plot of the SMR analysis, displaying the odds ratios (OR) for the associations between plasma proteins and the outcome variables. C Colocalization analysis of associations between plasma proteins and five CVDs in both discovery and replication studies. Full names of proteins: DUSP13B, dual-specificity phosphatase13B; FBP1, fructose-1,6-bisphosphatase 1; GUSB, beta-glucuronidase; LRIG1, leucine rich repeats and immunoglobulin like domains 1; SPON1, spondin 1; TNFSF12, tumor necrosis factor superfamily member 12; ANGPTL4, angiopoietin-like 4; APOH, apolipoprotein H; ASGR1, asialoglycoprotein receptor 1; COL6A3, collagen type VI alpha 3 chain; COMT, catechol-O-methyltransferase; GAS6, growth arrest specific 6; GSTT2B, glutathione S-transferase theta 2B; HGFAC, hepatocyte growth factor activator; INHBC, inhibin beta C chain; ITIH3, inter-alpha-trypsin inhibitor heavy chain H3; LAYN, layilin; MIF, macrophage migration…

• 

  Fig. 3. Multiple analyses on the associations between plasma proteins and the six kinds of metabolic phenotypes. A Heatmap summarized beta value in SMR analysis (βSMR) for associations between CVD-related proteins and 17 metabolic phenotypes. Each protein was split vertically: the left half shows the βSMR from the discovery study, and the right half shows the βSMR from the replication study. B Colocalization results for protein-metabolic-trait pairs across 17 phenotypes grouped into seven categories, shown for both discovery and replication datasets. Posterior support for a shared causal variant was reported as PPH4 (e.g., high support PPH4 ≥ 0.8; medium support 0.5 ≤ PPH4 < 0.8). C Random-effects IVW MR results for six proteins (PCSK9, LRIG1, INHBC, APOH, GUSB, and DUSP13B) versus cardiovascular outcomes. Only associations meeting P < 0.05 were displayed. ORs and their 95% CIs were shown. D Random-effects IVW MR results for the same six proteins versus six metabolic traits (anthropometric phenotypes, glycemic phenotypes, liver-related enzyme phenotypes, kidney-related phenotypes, lipidemic phenotypes and blood pressure). Displayed associations satisfy P < 0.05. βSMR were presented

• 

  Fig. 4. Phenotypes significantly associated with six candidate proteins. A, B KEGG pathway enrichment for 189 significant proteins in the discovery set (A) and 49 CVD-related proteins (B). C Integrated result of SMR analysis of six colocalized candidate proteins, CVDs and six metabolic phenotypes. The direction of the associations was indicated by βSMR. D Significance degree of SMR associations for PCSK9, APOH, DUSP13B, GUSB, INHBC and LRIG1 proteins with CAD, PAD, AF, and metabolic traits. E Phenome-wide SMR analysis of prioritized plasma proteins using FinnGen data. Only associations related to metabolism and CVDs that meet the thresholds of Pmulti < 0.05 and PHEIDI > 0.01 were displayed