Autophagic molecular network in IS pathogenesis: A multi-omics Mendelian randomization study

This study aimed to identify autophagy-related genes with a causal role in ischemic stroke (IS) risk using a multi-omics Mendelian randomization (MR) approach. We integrated summary-level data for blood-derived DNA methylation (mQTLs), expression (eQTLs), and protein (pQTLs) quantitative trait loci for 594 autophagy-related genes with large-scale IS genome-wide association studies for discovery (GCST006908) and validation (FinnGen). Summary-data-based MR (SMR) with colocalization analysis was performed, supplemented by brain eQTL analysis, protein-protein interaction networking, and drug prediction. The analysis identified 113 mQTLs, 38 eQTLs, and 9 pQTLs associated with IS risk, with colocalization supporting shared causal variants for a subset. Validation in FinnGen confirmed 13 mQTLs and 2 eQTLs, notably involving genes like CDKN1A, RRAGD, and SLC35D3. Multi-omics integration revealed regulatory cascades - for example, methylation at cg17245862 influencing DYNLT1 expression and protein levels. Brain-specific eQTL analysis supported 10 genes, including DYNLT1. Protein-protein interaction network analysis highlighted hub genes such as CDKN1A, PARK7, and LEP. Drug prediction suggested N-Acetyl-L-cysteine and acetaminophen as potential modulators targeting these genes. This study provides evidence for a potential causal role of autophagy-related genes and molecular traits in IS development, notably DYNLT1. Further research focusing on robustly identified candidates is warranted.