Dissecting the gene regulatory circuitry underlying chemoresistance in Triple-negative breast cancer

Due to the limited scRNA-seq sample size, we validated and expanded these findings in publicly available bulk RNA-seq data from 397 TNBC patients, identifying a 300-gene signature associated with residual disease (RD) prior to NAC. Building on these findings, we developed a 20-gene predictive model using lasso and elastic-net regression, achieving excellent performance with an area under the curve (AUC) of 0.90 in the training dataset and 0.89 in an independent validation cohort.

We further applied this model to external scRNA-seq datasets to validate its predictive performance and to biologically contextualise chemoresistant phenotypes at the cellular level, although interpretation was limited by the small number of available cells across different populations. To investigate regulatory mechanisms underlying TNBC chemoresistance, the model was applied to publicly available paired RNA-seq and H3K27ac ChIP-seq TNBC patient data and cell lines.

Initial differential binding analysis revealed modest changes between RD and pCR samples, prompting a broader investigation of super-enhancer landscapes and core regulatory circuitry. Super-enhancers were mapped using ROSE, and CRC networks were constructed with CRCmapper, identifying transcription factors, including SP1 and TFAP2C, as key regulators sustaining chemoresistant transcriptional programmes, with subtype-specific differences across basal-like, mesenchymal, and luminal androgen receptor TNBC subtypes.

Functional validation confirmed that knockdown of SP1 and TFAP2C reduced expression of chemoresistance-associated genes, including super-enhancer-regulated targets, and sensitised TNBC cells to chemotherapy. <br><br>Together, these findings provide insights into the gene regulatory mechanisms driving chemoresistance and identify potential biomarkers and therapeutic targets to improve outcomes for TNBC patients.<br>