A ten long noncoding RNA-based prognostic risk model construction and mechanism study in the basal-like immune-suppressed subtype of triple-negative breast cancer.

Abstract

According to the Fudan University Shanghai Cancer Center (FUSCC) system, triple-negative breast cancer (TNBC) is divided into four stable subtypes: (I) luminal androgen receptor, (II) immunomodulatory, (III) basal-like immune-suppressed (BLIS), and (IV) mesenchymal-like. However, the treatment outcomes of the corresponding targeted therapies are unsatisfactory, especially for the BLIS subtype. Therefore, we aimed to identify the key long noncoding RNAs (lncRNAs) to construct a prognostic model for BLIS subtype and discover potential targets to explore potential therapeutic strategies in this study. The FUSCC cohort was used to establish a prognostic risk model via least absolute shrinkage and selection operator (LASSO) and Cox regression analysis. The Cancer Genome Atlas (TCGA) cohort was then used to evaluate and verify the model. To understand the functional aspects of the model, functional, immune landscape, mutation, and drug sensitivity analyses were performed between high- and low-risk groups. Ten prognostic-related lncRNAs identified, including C5ORF66-AS2, DIO3OS, FZD10-DT, LINC00393, LNC-ERI1-32, LNC-FOXO1-2, LNC-SPARCL1-1, HCG23, LNC-MMD-4 and LNC-TMEM106C-6, were selected for risk score system construction. The results showed that the model constructed could divide the patients with BLIS subtype into two groups of high and low risk, and patients with higher risk scores had shorter recurrence-free survival. In addition, drug sensitivity analysis identified 3 compounds, including BMS-754807, cytochalasin b, and linifanib, that could have a potential therapeutic effect on patients with the BLIS subtype. The risk prognosis model showed good prognostic value for the BLIS subtype patients, and the ten lncRNAs may be potential therapeutic targets.