Association of genetically proxied cancer-targeted drugs with cardiovascular diseases through Mendelian randomization analysis

Fig. 1. Study overview. Drug target genes were identified from both the DrugBank and ChEMBL databases, and cis-eQTLs were extracted from the eQTLGen consortium to serve as genetic instruments. Positive control analyses were conducted to validate the genetic instruments. Summary GWASs for cardiovascular diseases were obtained from the FinnGen dataset. For each exposure, we performed MR analyses utilizing the SMR method and multiple sensitivity analyses. eQTL = expression quantitative trait loci; SMR = summary-data-based Mendelian randomization; GAWS = genome-wide association study

Fig. 2. Causal associations between cancer-targeted therapies and cardiovascular diseases. A Mendelian randomization study estimated ORs for the effect of a 1 SD decrease in the expression of (A) VEGFA, (B) FLT1, (C) FGFR1, and (D) MAP2K2 on 19 cardiovascular diseases. The results were obtained from SMR analysis at a Bonferroni correction p value < 0.00033 (8 exposures and 19 outcomes) and a HEIDI test p value > 0.01 to avoid linkage disequilibrium. The red points indicate SMR P values < 0.00033 and HEIDI test p-value > 0.01, the yellow points indicate SMR 0.00033 < P values < 0.05 and HEIDI test p values > 0.01, and the blue points indicate other values. OR = odds ratio; CI = confidence interval; HEIDI = heterogeneity in the dependent instrument

Fig. 3. Colocalization analysis between significant drug target genes and CVD outcomes. The figure depicts colocalization between the drug target gene eQTL signal and the outcome GWAS. The purple diamond points indicate the lead SNP with the minimal P value in the eQTL. The eQTL P values were extracted from the eQTLGen consortium (31,684 samples), and the stroke GWAS P values were extracted from the FinnGen database (39,818 cases, 27,1817 controls). eQTL: expression quantitative trait loci; GWAS: genome-wide association study