[Construction of a ubiquitination-related gene prognostic model for breast cancer based on taxane treatment response].

OBJECTIVES: Taxanes are first-line cchemotherapeutic agents for breast cancer; however, the development of resistance often leads to treatment failure. Ubiquitination plays a critical regulatory role in tumor progression and drug resistance. This study aims to construct a prognostic model based on ubiquitination-related genes (URGs) according to taxane treatment response, to predict patient prognosis and drug sensitivity, identify potential therapeutic targets, and provide a basis for individualized treatment of breast cancer.

METHODS: Breast cancer samples treated with taxanes were obtained from The Cancer Genome Atlas (TCGA) and classified into taxane-sensitive and taxane-resistant groups as the training cohort. The GSE25055 dataset from the Gene Expression Omnibus (GEO) was used as the validation cohort. A comprehensive list of URGs was retrieved from the Integrated Annotations for Ubiquitin and Ubiquitin-like Conjugation Database (iUUCD), and differentially expressed URGs between sensitive and resistant groups were identified. Univariate Cox regression analysis was performed to screen prognostic genes, followed by least absolute shrinkage and selection operator (LASSO) Cox regression with 10-fold cross-validation to identify core genes and construct a risk score model. Patients were stratified into high- and low-risk groups based on the risk score. Kaplan-Meier survival analysis and receiver operating characteristic (ROC) curves were used to evaluate predictive performance. A nomogram integrating risk score and clinical factors was developed and validated. Drug sensitivity differences between risk groups were analyzed using the Genomics of Drug Sensitivity in Cancer (GDSC) and the pRRophetic R package, with estimation of half-maximal inhibitory concentration (IC50). Tissue microarray analysis and cellular experiments were conducted to validate the key gene E4F1.

RESULTS: A total of 59 differentially expressed URGs were identified between taxane-sensitive and resistant groups. Through univariate Cox and LASSO-Cox analyses, 4 core genes (E4F1, CBLL1, RNF13, and TRIM59) were selected to construct the prognostic model. In the training cohort (TCGA), the areas under the ROC curve (AUCs) for predicting 1-, 3-, and 5-year overall survival (OS) were 0.896 (95% CI 0.754 to 1.000), 0.866 (95% CI 0.723 to 1.000), and 0.893 (95% CI 0.768 to 1.000), respectively. In the validation cohort (GSE25055), the corresponding AUCs were 0.620 (95% CI 0.467 to 0.685), 0.624 (95% CI 0.508 to 0.692), and 0.618 (95% CI 0.497 to 0.719), indicating that the model could effectively distinguish high- and low-risk patients. The nomogram integrating risk score and TNM stage showed improved predictive performance, with AUCs of 0.992 (95% CI 0.976 to 1.000), 0.890 (95% CI 0.754 to 1.000), and 0.897 (95% CI 0.786 to 1.000) for 1-, 3-, and 5-year OS, respectively, and good calibration. Drug sensitivity analysis identified 73 agents with significant IC50 differences between the two groups. The high-risk group exhibited increased sensitivity to Bryostatin-1, PHA-665752, and Salubrinal (all P<0.05). High-risk patients were significantly associated with both higher T stage and more advanced overall clinical stage (both P<0.05). Experimental validation demonstrated that E4F1 was significantly overexpressed in breast cancer cells and tissues, and correlated with shorter OS (all P<0.05). Knockdown of E4F1 significantly inhibited invasion and migration of breast cancer cells, and this inhibitory effect was further enhanced when combined with paclitaxel (all P<0.05).

CONCLUSIONS: The prognostic model based on E4F1, CBLL1, RNF13, and TRIM59 is an effective tool for predicting prognosis and drug sensitivity in breast cancer patients and may facilitate individualized treatment strategies. Among these, E4F1 may function as a potential oncogene and serve as a therapeutic target for enhancing taxane sensitivit.