Stephania cepharantha Hayata (Baiyaozi) attenuates cirrhosis-associated intestinal barrier injury by restoring mitochondrial function and suppressing PI3K/AKT signaling.

ETHNOPHARMACOLOGICAL RELEVANCE: Baiyaozi, the dried tuberous root of Stephania cepharantha Hayata, has long been used in traditional Chinese medicine for acute hepatitis, bacillary dysentery, acute gastroenteritis, abdominal pain, and bleeding disorders, indications that collectively correspond to hepatic inflammation, intestinal mucosal injury, and the bleeding tendency characteristic of cirrhosis-associated intestinal barrier injury. This ethnopharmacological correspondence provides a rationale for evaluating Baiyaozi (Stephania cepharantha Hayata) as a candidate therapy for the cirrhotic intestinal barrier, a setting in which current microbiota-targeted strategies act only indirectly on the intestinal epithelium and no traditional medicine has been shown to act directly on its cell-intrinsic driver, mitochondrial dysfunction. AIM OF THE STUDY: To identify mitochondria-related therapeutic targets for cirrhosis-associated intestinal barrier injury from human intestinal transcriptomic data, to identify Baiyaozi as a candidate medicine through reverse network pharmacology, and to validate its efficacy and underlying mechanism in vivo, focusing on the restoration of mitochondrial function and suppression of PI3K/AKT signaling. MATERIALS AND METHODS: We performed differential expression analysis on GSE285291 (human duodenal and ascending colon tissues; 28 cirrhotic patients, 8 healthy controls) and intersected DEGs with 2,719 mitochondrial dysfunction-related genes (MDRGs) to identify MDRDEGs. A three-step machine learning pipeline (Random Forest, SVM-RFE, LASSO) constructed a diagnostic model. Hub genes were identified by PPI analysis. Candidate TCMs were predicted via Coremine and TCMSP, and binding affinities assessed by molecular docking. The chemical constituents of the Baiyaozi decoction were characterized by UHPLC-Q-TOF-MS/MS. A CCl4-induced murine cirrhosis model validated the intestinal protective effects of Baiyaozi. RESULTS: 177 MDRDEGs were identified. The five-gene diagnostic model (JAK2, NDUFA2, PIK3CA, PRKAA1, SHMT2) achieved AUC = 0.996. JAK2 and PIK3CA were hub genes. Among TCM candidates, Baiyaozi alone met pharmacokinetic screening thresholds, and its active alkaloids exhibited strong binding affinities to both JAK2 and PIK3CA (-6.7 to -10.5 kcal/mol). UHPLC-Q-TOF-MS/MS tentatively identified 14 alkaloids in the decoction, including cepharanthine, tetrandrine, and crebanine. Crebanine, also predicted in silico, showed the strongest JAK2 affinity (-9.9 kcal/mol), while cepharanthine and tetrandrine exhibited the highest PIK3CA affinities (-10.5 and -10.0 kcal/mol). In vivo, Baiyaozi attenuated colonic mucosal injury, restored tight junction proteins, rescued mitochondrial markers (COXIV, ATP5A, TFAM, NDUFS1), normalized ATP and mtDNA copy number, and suppressed aberrant p-PI3K/p-AKT activation. CONCLUSIONS: This study establishes the first mitochondria-centric diagnostic model for cirrhosis-associated intestinal mucosal injury and demonstrates that Baiyaozi exerts intestinal barrier-protective effects through concomitant restoration of mitochondrial energy homeostasis and suppression of PI3K/AKT signaling, a mechanism distinct from the microbiota-mediated pathways reported for other TCMs.