Network Pharmacology Identifies IKBKB as the Primary Target of Sinomenine in Osteoarthritis, with Accuracy...

Background/Objective: Osteoarthritis lacks disease-modifying therapy, and NF-κB signaling is among the best-supported intervention points in cartilage. Sinomenine suppresses this pathway in chondrocytes, but only its downstream consequences are known, and network pharmacology rarely measures its own accuracy. Methods: Targets from six prediction resources were intersected with osteoarthritis genes from five databases and with cartilage and synovial differentially expressed genes. Candidates underwent topological analysis and unrestricted enrichment; prioritized targets were examined by docking and triplicate 100 ns molecular dynamics. Benchmarking used a curated bioactivity reference set, with a variant correcting input-validation overlap. Predictions were tested by thermal shift, limited proteolysis, NMR, thermophoresis, and kinase and silencing assays in interleukin-1β-challenged chondrocytes and in mice after medial meniscus destabilization. Results: Integration yielded 96 candidates, 34 transcriptionally corroborated, giving a network of 78 nodes and 241 edges with 12 consensus hubs. NF-κB signaling ranked first among 74 enriched pathways (adjusted p = 2.7 × 10^-20), and convergent criteria prioritized IKBKB, RELA, TNF, PTGS2 and MMP9. Benchmarking gave an area under the curve that fell from 0.81 to 0.74 and sensitivity from 0.73 to 0.57 after correcting for circularity. Sinomenine raised the melting temperature of IKBKB by 4.3 °C, bound it at 8.7 μM and inhibited it ATP-competitively, whereas RELA and CHUK showed no engagement. Nuclear p65 fell from 71.3% to 34.8%, IKBKB silencing abolished most of the MMP13 response, and cartilage damage scores fell from 5.0 to 2.5 at 40 mg/kg. Conclusions: Sinomenine engages IKBKB directly and acts principally at the IKK-dependent activation step. The generating pipeline proved only moderately accurate once circularity was corrected, which argues for reporting such metrics routinely.