Transcriptome-wide analysis and modelling of prognostic alternative splicing signatures in invasive breast cancer: a prospective clinical study.

ABSTRACT

Aberrant alternative splicing (AS) has been highly involved in the tumorigenesis and progression of most cancers. The potential role of AS in invasive breast cancer (IBC) remains largely unknown. In this study, RNA sequencing of IBC samples from The Cancer Genome Atlas was acquired. AS events were screened by conducting univariate and multivariate Cox analysis and least absolute shrinkage and selection operator regression. In total, 2146 survival-related AS events were identified from 1551 parental genes, of which 93 were related to prognosis, and a prognostic marker model containing 14 AS events was constructed. We also constructed the regulatory network of splicing factors (SFs) and AS events, and identified DDX39B as the node SF gene, and verified the accuracy of the network through experiments. Next, we performed quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) in triple negative breast cancer patients with different responses to neoadjuvant chemotherapy, and found that the exon-specific expression of EPHX2, C6orf141, and HERC4 was associated with the different status of patients that received neoadjuvant chemotherapy. In conclusion, this study found that DDX39B, EPHX2 (exo7), and HERC4 (exo23) can be used as potential targets for the treatment of breast cancer, which provides a new idea for the treatment of breast cancer.