Distinct pathway-based effects of blood pressure and body mass index on cardiovascular traits: comparison of novel Mendelian Randomization approaches

Abstract

BackgroundMendelian randomization (MR) leverages genetic variants as instrumental variables to determine causal relationships in epidemiology. However, challenges persist due to heterogeneity arising from horizontal pleiotropy. On the other hand, exploration of the biological underpinnings of such heterogeneity across variants can enhance our understanding of disease mechanisms and inform therapeutic strategies. Here, we introduce a new approach to instrument partitioning based on enrichment of Mendelian disease categories and compare it to a method based on genetic colocalisation in contrasting tissues. MethodsWe employed one-sample and two-sample MR methodologies using blood pressure (BP) exposure SNPs grouped by proximity to Mendelian disease genes affecting the renal system or vasculature, or body mass index (BMI) variants related to mental health and metabolic Mendelian disorders. We then compared the causal effects of Mendelian-partitioned SNPs on cardiometabolic outcomes with subsets inferred from gene expression colocalisation in kidney, artery (for BP), adipose, and brain tissues (for BMI). Additionally, we assessed whether effects from these groupings could emerge by chance using random SNP subset sampling. ResultsOur findings suggest that the causal relationship between systolic BP and coronary heart disease is predominantly driven by SNPs associated with vessel- related Mendelian diseases over renal. However, kidney-oriented SNPs showed more pronounced effect size in the colocalization-based analysis, hinting at a multifaceted interplay between pathways in the disease aetiology. We consistently identified a dominant role of Mendelian vessel and coloc artery exposures in driving the negative effect of diastolic BP on left ventricular stroke volume and positive effect of systolic BP on type 2 diabetes. We also found higher causal estimates for metabolic versus mental health SNPs when dissecting BMI pathway contribution to atrial fibrillation risk using Mendelian disease. In contrast, brain variants yielded higher causal estimates than adipose in the colocalization method. ConclusionsThis study presents a novel approach to dissecting heterogeneity in MR by integrating clinical phenotypes associated with Mendelian disease. Our findings emphasize the importance of understanding tissue-/pathway- specific contributions in interpreting causal relationships in MR. Importantly, we advocate caution in interpreting pathway-partitioned effect size differences without robust statistical validation.