摘要Background:Osteoporosis is a chronic bone disease characterized by bone loss and decreased bone strength.However,current anti-resorptive drugs carry a risk of vari-ous complications.The deep learning-based efficacy prediction system (DLEPS) is a forecasting tool that can effectively compete in drug screening and prediction based on gene expression changes.This study aimed to explore the protective effect and potential mechanisms of cinobufotalin (CB),a traditional Chinese medicine (TCM),on bone loss.Methods:DLEPS was employed for screening anti-osteoporotic agents according to gene profile changes in primary osteoporosis.Micro-CT,histological and morphologi-cal analysis were applied for the bone protective detection of CB,and the osteogenic differentiation/function in human bone marrow mesenchymal stem cells (hBMMSCs) were also investigated.The underlying mechanism was verified using qRT-PCR,Western blot (WB),immunofluorescence (IF),etc.Results:A safe concentration (0.25 mg/kg in vivo,0.05 μ M in vitro) of CB could ef-fectively preserve bone mass in estrogen deficiency-induced bone loss and promote osteogenic differentiation/function of hBMMSCs.Both BMPs/SMAD and Wnt/β-catenin signaling pathways participated in CB-induced osteogenic differentiation,further regulating the expression of osteogenesis-associated factors,and ultimately promoting osteogenesis.Conclusion:Our study demonstrated that CB could significantly reverse estro-gen deficiency-induced bone loss,further promoting osteogenic differentiation/function of hBMMSCs,with BMPs/SMAD and Wnt/β-catenin signaling pathways involved.