Multi-dimensional data-driven computational drug repurposing strategy for screening novel neuroprotective agents in ischemic stroke.

Background: The complexity of biological systems and misconceptions about neuroprotection have hindered the development of neuroprotective drugs for ischemic stroke. This study aims to identify new neuroprotective agents by integrating ischemic stroke transcriptomics with neuronal protection data using a Multidimensional Data-Driven Computational Drug Repositioning strategy (MDCDR). Methods: Three microarray datasets related to ischemic stroke (GSE16561, GSE58294, and GSE22255) were obtained from the GEO dataset and pre - processed to analyze differentially expressed genes (DEGs). The Connectivity Map (CMap) database was used to predict potential drugs. A neuroprotection activity prediction model was constructed by combining six molecular fingerprints with three machine learning algorithms (Random Forest RF, Support Vector Machine SVM, Gradient Boosting Decision Tree GBDT) to screen for potential neuroprotective agents. The efficacy of the screened compounds was evaluated through in vitro experiments on SH-SY5Y cells treated with oxygen-glucose deprivation/reperfusion (OGD/R) and in vivo experiments on middle cerebral artery occlusion/reperfusion (MCAO/R) rat models. Multiple experimental techniques (such as RNA sequencing, DARTS, CETSA, etc.) were used to explore their potential mechanisms of action. Results: The MDCDR strategy screened out 19 potential neuroprotective agents, among which sulbutiamine (SUL) stood out. SUL significantly increased the survival rate, reduced neurological deficit scores, and decreased neuronal loss in MCAO/R rat models, and inhibited cell death in OGD/R - induced cell models. Mechanistic studies revealed that SUL inhibited pyruvate dehydrogenase kinase 2 (PDK2), enhanced mitochondrial function, reduced reactive oxygen species (ROS) levels, thereby suppressing the MAPK signaling pathway and reducing neuronal apoptosis. Silencing PDK2 abolished the protective effect of SUL on OGD/R - treated SH - SY5Y cells. Conclusion: This study successfully developed the MDCDR strategy for screening neuroprotective agents for ischemic stroke. SUL was identified as a promising neuroprotective agent, and PDK2 was a crucial target. This research provides new directions and a theoretical basis for the development of neuroprotective agents against ischemic stroke.