Integrative proteogenomic analyses identify plasma proteins that impact the risk of ischemic stroke

BACKGROUND: Understanding the proteins implicated in the pathogenesis of ischemic stroke is important for elucidating disease mechanisms and informing prevention strategies. In this study, we aim to identify plasma proteins with a potentially causal effect on risk of ischemic stroke by integrating the largest available genetic data sources for plasma proteins, ischemic stroke and its risk factors. METHODS: We use genome-wide association (GWA) statistics to identify cis protein quantitative trait loci for over 3,500 proteins across two proteomic platforms. Subsequently, we perform two-sample Mendelian randomization (MR) to assess the potentially causal relationships between plasma proteins and a) risk of ischemic stroke and its subtypes, and b) well-established cardiovascular risk factors. Downstream analyses include Bayesian colocalization, phenome-wide associations and interrogation of biological databases. RESULTS: We identify 21 proteins with evidence of potentially causal associations with risk of ischemic stroke or its subtypes at 5% false discovery rate, with 16 supported further by colocalization. Four proteins (CEP85, KNG1, MMUT, SPATA20) represent findings not previously implicated in ischemic stroke through MR, or through cognate genes in GWA studies. Integration of evidence from phenome-wide MR, animal models and tissue-specific gene expression highlights agonists of MMUT, CEP85 and GRK5, and inhibitors of F11 and KNG1 as the most promising for further consideration as targets for prevention of ischemic stroke. CONCLUSIONS: Our study provides the most comprehensive data integration to date supporting the identification and causal relevance of plasma proteins for ischemic stroke and implicating a number of potential therapeutic targets.