Shikonin as a Parkin-dependent hypoxia-selective agent against triple-negative breast cancer identified by machine learning

Triple-negative breast cancer (TNBC) poses significant therapeutic challenges due to a lack of targeted therapies and the frequent development of resistance to conventional solutions. This study delved into programmed cell death (PCD) pathways, particularly apoptosis and necroptosis under hypoxic microenvironments, to meet the demand for novel therapeutic strategies. A machine learning model was trained on a combined dataset of necroptosis-related genes and hypoxia-related risk score (HRRS) genes from 1,092 breast cancer patients. SHAP (SHapley Additive exPlanations) was used to evaluate feature importance within the model. Leveraging shikonin’s established role as a necroptosis inducer, its efficacy in hypoxic TNBC microenvironments was evaluated in MDA-MB-231 cells through cell viability assays, protein expression profiling related to necroptosis, apoptosis analysis, and transcriptomic profiling. Mechanistically, co-immunoprecipitation and ubiquitination assays were performed to investigate Parkin-HIF-1α interactions, while siRNA-mediated Parkin knockdown coupled with CCK-8 assays assessed shikonin cytotoxicity dependence on Parkin. Machine learning identified pellino E3 ubiquitin protein ligase 1 (PELI1) and TNF receptor superfamily member 10b (TNFRSF10B) as top predictive contributors. RNA-sequencing analysis revealed that hypoxic TNBC cells relied on adaptive pathways (e.g., glycolysis, erythropoiesis, angiogenesis), which were disrupted by shikonin via cell death activation. Mechanistically, Parkin mediated HIF-1α ubiquitination and degradation, a process amplified under hypoxia. Our findings demonstrate that shikonin overcomes hypoxia-induced therapeutic resistance through dual induction of necroptosis and apoptosis in TNBC. Mechanistically, Parkin mediates shikonin’s anti-cancer effects by promoting ubiquitination and degradation of HIF-1α, effectively disrupting hypoxic adaptation. These results nominate shikonin as a promising hypoxia-selective agent, and suggest that combinatorial strategies targeting PELI1 and TNFRSF10B pathways may enhance therapeutic efficacy against resistant TNBC.