Machine Learning-Enabled Insights: Dihydromyricetin’s Novel Role in Inhibiting the TGF-β/ALK5 Signaling Cascade for the Treatment of Pulmonary Fibrosis

Idiopathic pulmonary fibrosis (IPF) denotes a chronic, advancing, and life-threatening lung disorder. Dysregulated cytokines, particularly those in the transforming growth factor-β (TGF-β)-associated signaling pathway, drive the pathological development of IPF. Natural products derived from traditional Chinese medicine hold great potential as promising therapeutic candidates for IPF. This study integrated machine learning (ML) with experimental validation to identify TGF-β/small mother against decapentaplegic (SMAD) pathway inhibitors from natural compounds. An in-house library was screened by means of a dual-luciferase reporter assay, revealing the flavonoid dihydromyricetin (DHM) as the most potent inhibitor. In vitro , DHM suppressed TGF-β1-triggered epithelial–mesenchymal transition (EMT) in A549 cells and fibroblast transdifferentiation in medical research council cell strain 5 (MRC-5) cells. In vivo , DHM attenuated fibrosis and inflammatory responses in a bleomycin (BLM)-triggered pulmonary fibrosis mouse model. Mechanistic studies revealed that DHM targets the type I TGF-β receptor (known as ALK5), reduces its membrane expression, binds directly to the receptor and represses its kinase activity, ultimately downregulating the TGF-β/ALK5 pathway. The present research is the first to report DHM as a TGF-β/SMAD inhibitor identified through ML with therapeutic efficacy against IPF. DHM’s anti-fibrotic effects are mediated through ALK5 blockade, suppressing downstream signaling, EMT, and fibroblast activation. These findings not only highlight DHM’s latent ability to act as a novel remedy for IPF but also underscore the utility of computational approaches in natural product drug discovery.