Nano and molecular mechanism effects of CsV vascular endothelial functions using machine learning

Chikusetsusaponin V's (CsV) effects on Endothelial Nitric Oxide Synthase (eNOS) and vascular endothelial functions were the focus of this investigation. Various CsV doses were introduced into in vitro cultures of Bovine Aortic Endothelial Cells (BAECs), Human Umbilical Vein Endothelial Cells (HUVECs), and animal models. To evaluate the impact of CsV on endothelial activities, vascular stiffness, eNOS mRNA and protein expression, and Nitric Oxide (NO) production were examined using quantitative PCR (qPCR), Western Blotting (WB), and B-ultrasound imaging. Protein mass spectrometry, molecular docking, bioinformatics, and network pharmacology were further applied to predict upstream transcription factors and molecular interactions regulating eNOS activity. Complementarily, advanced machine learning (ML) models including neural networks, Random Forests (RF), Support Vector Machines (SVM), and RF–Fuzzy Logic were employed to predict endothelial responses under varying CsV conditions. The neural network achieved the highest predictive accuracy (86.36%), while the RF–Fuzzy Logic model demonstrated superior precision (90.91%) and recall (83.33%). Feature importance analysis identified impedance modulus and water contact angle (WCA) as critical determinants of CsV-induced endothelial regulation. These findings provide nano- and molecular-level insights into the mechanisms by which CsV modulates endothelial function, integrating experimental assays with ML-driven predictions to inform potential therapeutic strategies for cardiovascular diseases.