Integrating multi-omics analysis to identify potential biomarkers and regulatory networks of ischemic stroke

Ischemic stroke (IS) is a leading cause of death and long-term disability worldwide, with a complex and multifactorial pathophysiology that is still incompletely understood. This study performed untargeted metabolomics and Olink proteomics on serum samples from IS patients and controls, and additionally analyzed single-nucleus RNA sequencing data from rat models of cerebral ischemia to explore cell-type-specific expression patterns. We identified 142 differentially expressed metabolites (FDR-corrected) and 24 proteins that were differentially expressed at nominal significance (P < 0.05). Using machine learning, five metabolites and IL-6 were selected as candidate biomarkers. A combined random forest model incorporating these five metabolites and IL-6 demonstrated good discriminatory performance in the external validation cohort (AUC = 0.85). In prospective analyses of the UK Biobank, higher levels of IL-6 were significantly associated with IS risk, and the hub genes (e.g., PYGL and ENO1) were enriched in pathways such as glycolysis and HIF-1 signaling. Single-nucleus RNA sequencing from rat models of cerebral ischemia revealed cell type-specific expression patterns, with hub genes predominantly expressed in neurons and microglia. These exploratory findings support a hypothetical multi-level regulatory network centered on the "energy crisis-oxidative stress-neuroinflammation" axis. Overall, this study identifies potential circulating biomarker candidates for IS and provides hypothesis-generating insights into its pathophysiological mechanisms, which may aid future research on risk stratification and therapeutic targets.