Through Network Pharmacology Combined with Artificial Intelligence Techniques, Potential Targets of Banxia Xiexin Decoction for the Treatment of Functional Dyspepsia were Identified and Validated.

Abstract

Background: Banxia Xiexin Decoction (BXD) has been shown to exert therapeutic effects on Functional dyspepsia (FD). This study aims to investigate the therapeutic mechanisms of BXD in treating FD.

Methods: Network pharmacology was employed to explore the potential targets of BXD in the treatment of FD. Immunoinfiltration analysis assessed immune activation in FD, with the XGBoost machine learning algorithm used to predict the feature importance of key targets. Deep learning and molecular docking were employed to assess the interactions between active compounds and key targets. Finally, an FD mouse model was established, and Western blotting, immunofluorescence, immunohistochemistry, and Enzyme-linked immunosorbent assay were conducted to validate the findings.

Results: Through network pharmacology analysis and machine learning predictions, three key active compounds were identified. GO enrichment analysis indicated that the mechanism of BXD primarily involves biological processes related to inflammatory responses. Immunoinfiltration analysis suggested that immune activation in FD may be associated with increased mast cell presence. Seven hub genes were identified through PPI analysis, with STAT3 identified as a key feature in XGBoost predictions of FD. In vivo experiments showed that BXD inhibited p- STAT3, alleviated mast cell infiltration and mucosal barrier damage, and enhanced gastrointestinal motility.

Conclusion: BXD may alleviate mast cell infiltration and mucosal barrier damage in FD by inhibiting the expression of p-STAT3, thereby exerting its therapeutic effects.