Mendelian Randomization and Transcriptome Analysis Identify Ischemic Stroke Biomarkers With Putative Relevance to Cerebrospinal Fluid
Background Circulating proteins have been associated with the pathogenesis of ischemic stroke (IS), yet its biomarkers remain underutilized. Using plasma protein GWAS data with putative relevance to CSF, this study integrated mendelian randomization (MR) and transcriptomics to identify potential IS biomarkers. Methods A two‐sample MR analysis was undertaken to determine the genetic association between circulating protein levels and IS. The identification of differentially expressed genes (DEGs) in the GSE268634 and GSE262257 datasets was carried out via the transcriptomic analysis. Candidate biomarkers overlapping MR‐derived genes (MRGs) and DEGs underwent functional enrichment, protein‐protein interaction (PPI), and machine learning (LASSO/SVM‐RFE) screening. Biomarker mechanisms were assessed via gene set enrichment analysis (GSEA), immune infiltration, and hypothesis‐generating drug prediction. The validation included RT‐qPCR and immunohistochemistry in MCAO/R rats. Results The MR analysis identified 157 circulating protein‐related MRGs with suggestive genetic associations with IS. Transcriptomics identified 4144 DEGs, and 46 overlapping with MRGs. Functional enrichment highlighted their roles in cell adhesion and immune responses. Machine learning identified six candidate biomarkers, among which CDH7, MGAT4C, and ITPKC exhibited both high diagnostic accuracy (AUC > 0.7) and consistently differential expression, and were therefore prioritized as putative biomarkers. GSEA revealed that CDH7 and MGAT4C were positively correlated, whereas ITPKC was negatively correlated with the calcium signaling pathway. Immune infiltration analysis showed that CDH7 and MGAT4C were negative, whereas ITPKC was positively correlated with immune cells. Computationally predicted drugs including genistein and pioglitazone may alleviate IS damage, though this requires experimental confirmation. RT‐qPCR and immunohistochemistry indicated markedly high CDH7 and MGAT4C expression, whereas low ITPKC expression was in MCAO/R rats. Conclusion CDH7, MGAT4C, and ITPKC are genetically associated and transcriptionally altered candidates derived from circulating protein‐related analyses for IS, warranting further investigation.