In Vitro and In Vivo Validation of Endothelium-Derived Potential Therapeutics for Myocardial Ischemia/Reperfusion Injury Identified by an AI-Enhanced Single-Cell and Virtual-Cell Paradigm

Myocardial ischemia/reperfusion (MI/R) injury affects heart attack outcomes. Endothelial cells dysfunction immediately after MI/R, but the key molecules and how to block them remain unclear. We combined single-cell atlas analysis, AI simulation, and experimental single-cell RNA sequencing data from mouse MI/R; we did quality control, cell annotation, hdWGCNA, and differential gene screening to identify endothelial genes. We constructed a protein network with STRING, predicted structure with AlphaFold3, and used AutoDock for molecular docking to find potential drugs. Virtual knockout simulations were used to check gene deletion effects. The compound andrographolide (AG) was tested in in vitro and in vivo MI/R models by measuring cell viability, inflammation, pathway activity, infarct size, and cardiac function. Single-cell analysis showed that S100 calcium binding protein A8 (S100A8) is an important element in vascular inflammation. It promotes inflammation by interacting indirectly with Cluster of differentiation 14 (CD14). Molecular docking showed that AG binds stably to S100A8. In vitro, AG reduced endothelial injury and blocked the IL-17 pathway. In vivo, AG reduced infarct size, improved cardiac function, and lowered S100A8 and IL-17 pathway proteins. Using single-cell analysis, AI, and experiments, we showed that S100A8 is related to MI/R injury. Andrographolide protects microvasculature via the S100A8 pathway, offering a promising treatment approach and new insights into heart injury mechanisms.