Exploration of Drug-Target Interaction Against Gastric Cancer-Derived Exosomes Using a Predicted GPX4 Protein Structure Based on AlphaFold2

The gastric cancer-derived extracellular vesicle miR-431-5p can induce ferroptosis in gastric cancer cells by downregulating the levels of SLC7A11 and GPX4, thereby inhibiting the occurrence and development of gastric cancer. Currently, drug design based on this mechanism is not yet fully developed. Therefore, focusing on the critical ferroptosis inhibition target GPX4 within this pathway, we utilized emerging AI tools for complete protein prediction and drug design. We used AlphaFold2 to predict the complete protein structure of GPX4, visualized its model using PyMOL, and analyzed its binding sites. Based on the existing drug-binding model, we then performed molecular docking with the predicted GPX4 protein using AutoDock Vina. Compared with existing models, we found that the drug-binding sites differed, which warrants further experimental analysis to explore the reasons behind these differences. This study suggests that using AI tools to assist in drug design is feasible and offers reference value for future drug development.