PR002647 (Project)

Description:Acute ischemic stroke (AIS) urgently requires reliable diagnostic biomarkers to mitigate pre-hospital delays and enhance thrombolysis utilization. We performed integrated untargeted proteomic, metabolomic, and lipidomic analyses on serum samples from a discovery cohort (AIS patients, n=52; healthy controls [HC], n=50). Proteomics identified 295 differentially abundant proteins, implicating blood-brain barrier disruption (downregulated structural anchors: FLNA, ACTB, ACTN1; Rho GTPases: RAC1, CDC42), thromboinflammatory activation (upregulated platelet receptors: ITA2B, ITB3, GP1BA; inflammasome components: CRP, SAA) and complement dysregulation (C1S, C4BPA). Metabolomics and lipidomics revealed 134 altered species, demonstrating: an energy crisis (depleted citrate cycle intermediates, aberrant glycolysis), oxidative stress (elevated glutamate/L-pyroglutamate; depleted antioxidants), and pathological lipid remodeling (accumulated diacylglycerols/sphingomyelins; reduced lysophosphatidylcholines/sterols).multiomics integration revealed that AIS pathogenesis involves synergistic glycolysis-triggered energy crisis, glutathione-dependent oxidative collapse, and catastrophic lipidome disarray. Machine learning (RF/LASSO) derived a diagnostic panel comprising FA9, APOC2, SAMP, C1S, OSTP, PCSK6, CCN2, PDGFC, HPSE, ANGL3, glutamate, glycerate, and succinate. This panel achieved exceptional diagnostic accuracy in an independent validation cohort (AIS=46, HC=48; AUC=0.998, 95% CI: 0.955-1.000), significantly outperforming established markers (PCSK9, succinate alone) and showing strong correlation with NIHSS scores. Our study establishes a mechanistically grounded, high-performance multiomics biomarker signature for AIS diagnosis.