Network pharmacology integrated with HTS2 deciphers the anti-inflammatory mechanism of pentacyclic triterpene from Sanguisorba officinalis

Herbal medicines exert therapeutic effects via multiple constituents, yet conventional studies often overlook low-abundance, highly active compounds. To address this, we propose an integrated framework combining network pharmacology and high-throughput sequencing-based high-throughput screening (HTS²), coupling in silico target and pathway prediction with high-throughput transcriptomic validation to systematically link natural compounds to their molecular targets. Applying this framework to Sanguisorba officinalis, twenty-three isolated compounds were first screened using network pharmacology to predict anti-inflammatory activities and candidate targets. They were then profiled by HTS² to identify compound-specific transcriptional signatures and pathway modulation. Integrating the network pharmacology and HTS² results, a representative pentacyclic triterpene, 3β,6β-dihydroxy-urs-12,19(29)-dien-28-oic acid (compound 17), was identified, with 5-lipoxygenase-activating protein (FLAP) as its primary molecular target. This compound exhibited the strongest anti-inflammatory activity: it directly bound FLAP, reduced leukotriene B₄ production, suppressed MAPK signaling, and downregulated COX-2 and iNOS expression. Target engagement and downstream effects were validated by molecular docking, molecular dynamics simulations, CETSA, Western blotting, ELISA, and in vivo zebrafish assays. Collectively, this study introduces a generalizable network pharmacology-HTS² strategy for mechanistic deconvolution of natural product mixtures, enabling mechanism-driven discovery of bioactive compounds from traditional medicines.