A chromatin-informed transcriptional regulatory framework to stratify patients and guide therapy selection in triple-negative breast cancer

Triple-negative breast cancer is an aggressive and heterogeneous breast cancer subtype with few effective targeted therapies and frequent resistance to chemotherapy. Here, we integrate transcriptional regulatory network inference with chromatin accessibility across a large-scale multi-system collection of primary tumors, patient-derived xenografts and model cell lines to quantify transcription factor activity and identify regulators that underpin triple-negative breast cancer identity. This approach prioritizes 94 high-confidence triple-negative breast cancer transcription factors whose activity capture inter-tumor heterogeneity and independently stratify patient outcome across clinical endpoints. Linking transcription factor activity to pharmacogenomic drug sensitivity profiles identifies reproducible drug-transcription factor associations across independent datasets, including NFE2L3 and CBFB activity as predictors of sensitivity to mTOR inhibition, which we validate in everolimus-treated triple-negative breast cancer patient-derived xenograft models. Collectively, we provide a transcriptional and chromatin-informed framework to capture triple-negative breast cancer regulatory state and expand transcription factor guided precision medicine to this breast cancer subtype. Triple-negative breast cancer (TNBC) is an aggressive, heterogeneous subtype with limited treatment options. Here, transcriptional network inference integrated with chromatin accessibility across primary tumours and models defines a transcriptional and chromatin-informed framework underlying TNBC identity.