Establishing plausibility of cardiovascular adverse effects of immunotherapies using Mendelian randomisation

Abstract

Immune checkpoint inhibitors (ICIs) and Janus kinase inhibitors (JAKis) have raised concerns over serious unexpected cardiovascular adverse events. The widespread pleiotropy in genome-wide association studies offers an opportunity to identify cardiovascular risks from in-development drugs to help inform appropriate trial design and pharmacovigilance strategies. This study uses the Mendelian randomization (MR) approach to study the causal effects of 9 cardiovascular risk factors on ischemic stroke risk both independently and by mediation, followed by an interrogation of the implicated expression quantitative trait loci (eQTLs) to determine if the enriched pathways can explain the adverse stroke events observed with ICI or JAKi treatment. Genetic predisposition to higher systolic blood pressure (SBP), diastolic blood pressure (DBP), body mass index (BMI), waist-to-hip ratio (WHR), low-density lipoprotein cholesterol (LDL), triglycerides (TG), type 2 diabetes (T2DM), and smoking index were associated with higher ischemic stroke risk. The associations of genetically predicted BMI, WHR, and TG on the outcome were attenuated after adjusting for genetically predicted T2DM [BMI: 53.15% mediated, 95% CI 17.21%-89.10%; WHR: 42.92% (4.17%-81.67%); TG: 72.05% (10.63%-133.46%)]. JAKis, programmed cell death protein 1 and programmed death ligand 1 inhibitors were implicated in the pathways enriched by the genes related to the instruments for each of SBP, DBP, WHR, T2DM, and LDL. Overall, MR mediation analyses support the role of T2DM in mediating the effects of BMI, WHR, and TG on ischemic stroke risk and follow-up pathway enrichment analysis highlights the utility of this approach in the early identification of potential harm from drugs.

Keywords: GWAS; Mendelian randomisation; blood pressure; diabetes; eQTL; pleiotropy; reactome; stroke.

Figure 1

Associations of genetically predicted exposure [systolic blood pressure (A), diastolic blood pressure (B), serum low-density lipoprotein cholesterol (C), triglycerides (D), body mass index (E), waist-to-hip ratio (F), type 2 diabetes (G), smoking (H)] on ischemic stroke before and after adjusting for genetically predicted mediators. The y-axis details the genetically predicted mediator for which adjustments were made. SBP, systolic blood pressure; DBP, diastolic blood pressure; BMI, body mass index; WHR, waist-to-hip ratio; LDL, low-density lipoprotein cholesterol; TG, triglycerides; CKD, chronic kidney disease; T2DM, type 2 diabetes; OR, odds ratio; SD; standard deviation.

Figure 2

The effects of genetically predicted exposures on ischaemic stroke. The Sankey plot illustrates the direct and mediated effects of the risk factors on the risk of ischaemic stroke. SBP, systolic blood pressure; DBP, diastolic blood pressure; BMI, body mass index; WHR, waist-to-hip ratio; LDL, low-density lipoprotein cholesterol; TG, triglycerides; T2DM, type 2 diabetes.

Figure 3

Pathway enrichment analysis. The Sankey plot illustrates enriched pathways (middle column) in the list of genes that had expression levels associated with the genetic instruments for the considered exposure traits (left column). The right column details drugs involved in the enriched pathways. SBP, systolic blood pressure; DBP, diastolic blood pressure; WHR, waist-to-hip ratio; LDL, low-density lipoprotein cholesterol; TG, triglycerides; T2DM, type 2 diabetes. SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; TNF, tumor necrosis factor; SSRI, selective serotonin reuptake inhibitor; HCV, hepatitis C virus; GI drugs, gastrointestinal drugs; JAK Inhibitors, Janus kinase inhibitors; PD-1/PDL-1, programmed cell death protein 1 and programmed death ligand 1.