Integrating network toxicology, machine learning, and molecular dynamics simulations to reveal tanshinone iia's dual mechanisms in TNBC and doxorubicin-induced cardiotoxicity.

Doxorubicin (Dox)-induced cardiotoxicity remains a critical barrier to optimizing breast cancer (BC) treatment, highlighting the urgent need to dissect its toxicological mechanisms and develop toxicity-mitigating combination strategies; here, we address this gap by integrating network toxicology, molecular dynamics simulations, bioinformatics, and machine learning to unravel how tanshinone IIA (Tan IIA) alleviates Dox cardiotoxicity while identifying its key targets for combating triple-negative breast cancer (TNBC). Our analyses reveal that Tan IIA regulates 13 core targets of Dox cardiotoxicity-with enrichment in pathways including canonical cancer and small cell lung cancer pathways-and that six of these targets exhibit high binding affinity for Tan IIA or Dox; notably, machine learning prioritized the histone methyltransferase EZH2 as the central target for Tan IIA's anti-TNBC activity, and we further show EZH2 is highly expressed in breast invasive carcinoma (BRCA) tissues and correlates positively with infiltration of immune cells (e.g., B cells, CD4⁺ T cells) and expression of immune-related molecules (including immunosuppressors and MHC-associated antigen-presenting molecules). Collectively, these findings demonstrate that Tan IIA may mitigate Dox cardiotoxicity via modulation of targets such as APAF1, AR, and TERT (and their associated signaling cascades) while targeting EZH2 to exert anti-TNBC effects, providing a mechanistic framework for repurposing Tan IIA to improve the safety and efficacy of Dox-based BC therapy.