Phenylpropanoid Amides from Buteae monospermae Semen as Hepcidin-Modulating Candidates for Anemia of Inflammation: Network Pharmacology, Deep Learning Prediction, and In Vitro Validation

Anemia of inflammation (AI) is largely driven by interleukin-6 (IL-6)-induced hepcidin overproduction, which disrupts systemic iron homeostasis. Buteae monospermae Semen (BMS), the seeds of Butea monosperma (Lam.) Kuntze and a traditional Tibetan medicine, has long been used for blood-related disorders, yet its active constituents against AI remain unclear. Here, an integrated strategy combining ADMET screening, network pharmacology, deep learning-based affinity prediction, molecular docking, molecular dynamics simulation, and in vitro validation was applied to 58 BMS-derived compounds. Forty-three candidates with acceptable ADMET profiles were retained, and their predicted targets were enriched in inflammation- and iron metabolism-related pathways, particularly IL-6/JAK2/STAT3 signaling. Two phenylpropanoid amides, compounds 16 and 21, were prioritized and shown to interact favorably with hepcidin-related regulatory targets. In IL-6-stimulated HepG2 cells, both compounds dose-dependently reduced hepcidin levels (EC50 = 2.486 and 14.49 μM), with therapeutic indices of 4.40 and 33.7, respectively. These findings identify BMS-derived phenylpropanoid amides as hepcidin-modulating candidates for AI and demonstrate an effective computational–experimental strategy for discovering bioactive compounds from traditional medicines.