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

Mendelian randomization (MR) leverages trait associated genetic variants as instrumental variables (IVs) to determine causal relationships in epidemiology. However, genetic IVs for complex traits are typically highly heterogeneous and, at a molecular level, exert effects on different biological processes. Exploration of the biological underpinnings of such heterogeneity 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 (pathway-partitioned) and compare it to an existing method based on genetic colocalization in contrasting tissues (tissue-partitioned).

We employed individual- and summary-level MR methodologies using SNPs grouped by pathway informed by proximity to Mendelian disease genes affecting the renal system or vasculature (for blood pressure (BP)), or mental health and metabolic disorders (for body mass index (BMI)). We compared the causal effects of pathway-partitioned SNPs on cardiometabolic outcomes with those derived using tissue-partitioned SNPs informed by colocalization with gene expression in kidney, artery (BP), or adipose and brain tissues (BMI). Additionally, we assessed the likelihood that estimates observed for partitioned exposures could emerge by chance using random SNP sampling.

Our pathway-partitioned findings suggest the causal relationship between systolic BP and heart disease is predominantly driven by vessel over renal pathways. The stronger effect attributed to kidney over artery tissue in our tissue-partitioned MR hints at a multifaceted interplay between pathways in the disease aetiology. We consistently identified a dominant role for vessel (pathway) and artery (tissue) driving the negative directional effect of diastolic BP on left ventricular stroke volume and positive directional effect of systolic BP on type 2 diabetes. We also found when dissecting the BMI pathway contribution to atrial fibrillation that metabolic-pathway and brain-tissue IVs predominantly drove the causal effects relative to mental health and adipose in pathway- and tissue-partitioned MR analyses, respectively.

This 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 pathway-/tissue-specific contributions to complex exposures when interpreting causal relationships in MR. Importantly, we advocate caution and robust validation when interpreting pathway-partitioned effect size differences.