[Integrating single-cell analysis and epharmalib reverse virtual screening to predict novel vascular endothelial cell targets of dapagliflozin in treating diabetic cardiomyopathy].

<b>Objective:</b> To investigate endothelial cell heterogeneity in diabetic cardiomyopathy (DCM) and identify potential therapeutic targets of dapagliflozin in cardiac vascular endothelial cells. <b>Methods:</b> ePharmaLib reverse virtual screening was performed on 15 148 protein crystals to identified the binding interactions between human-derived proteins and dapagliflozin. Subsequently, single-cell RNA sequencing data (PRJNA1069235) from wild-type mice (control group) and db/db mice (DCM group) were integrated, then dimensionality reduction and clustering analysis were performed to identify endothelial cell subpopulations in the heart tissue of DCM mice, followed by functional annotation. Cell-cell communication analysis was explored to investigate fibroblast-endothelial cell interactions. The Agilent Mouse ceRNA Microarray chip was used to perform transcriptomic analysis of heart tissue from mice fed a high-fat diet and treated with dapagliflozin. Intersection analysis of reverse virtual screening results, single-cell RNA sequencing results and chip analysis data was performed to identify common differentially expressed genes. In vitro, human umbilical vein endothelial cells were divided into blank control group, tumor necrosis factor-α (TNF-α) group (TNF-α 10 mg/L), dapagliflozin low concentration group (TNF-α 10 mg/L+dagliazin 2 μmol/L), dapagliflozin medium concentration group (TNF-α 10 mg/L+dagliazin 5 μmol/L) and dapagliflozin high concentration group (TNF-α 10+dagliazin 10 μmol/L). Western blot and real-time reverse transcriptase polymerase chain reaction were used to detect the expression of inflammatory factors and differential genes. <b>Results:</b> ePharmaLib reverse virtual screening identified 168 human-derived proteins with potential binding affinity to dapagliflozin, and single-cell analysis identifiedf 6 types of endothelial cell subpopulations. Compared with the control group, the abundance of capillary endothelial cells was significantly lower in DCM group, while the abundance of microvascular and venous endothelial cells was significantly higher (<i>P</i> all<0.05). Cell-cell communication analysis showed significant expression of Pgf-Vegfr1 ligand-receptor pair. In addition, 15 differentially expressed genes were identified by intersection analysis of 168 dapagliflozin-binding proteins. Including Bcl2, Baz2b, Nos3, Ephb4, Cdk8, Pparg, Pde2a, Fgfr2, Fto, Stk24, Dlg1, Gsk3b, Pdpk1, Fas and Tnks2. Notably, Baz2b, Pparg, Fto and Gsk3b were differentially expressed in all cell subpopulations. Six differential genes, including Pde7a, Dlg1, Gsk3b, Nampt, Met and Adk, were obtained by the intersection of the chip analysis data with the virtual screening results of dapagliflozin. In vitro, compared to the human umbilical vein endothelial cells of TNF-α group, the expression levels of interleukin-6, interleukin-1β and p-P65 proteins and messenger RNA of Bcl2, Nos3, Cdk8, Pde2a, Dlg1, Pdpk1, Tnks2, Baz2b, Pparg, Fas, Pde7a and Nampt were significantly lower than dapagliflozin high concentration group (<i>P</i> all<0.05). <b>Conclusions:</b> Dapagliflozin may inhibit endothelial cell inflammatory responses and improve endothelial dysfunction in DCM by regulating key genes such as Dlg1, Bcl2, Nos3, Pde7a and Nampt.