Journal of Translational Medicine (2020-07-01)

Comprehensive analysis and establishment of a prediction model of alternative splicing events reveal the prognostic predictor and immune microenvironment signatures in triple negative breast cancer

  • Shanshan Yu,
  • Chuan Hu,
  • Lixiao Liu,
  • Luya Cai,
  • Xuedan Du,
  • Qiongjie Yu,
  • Fan Lin,
  • Jinduo Zhao,
  • Ye Zhao,
  • Cheng Zhang,
  • Xuan Liu,
  • Wenfeng Li

DOI
https://doi.org/10.1186/s12967-020-02454-1
Journal volume & issue
Vol. 18, no. 1
pp. 1 – 17

Abstract

Read online

Abstract Background Triple-negative breast cancer (TNBC) is widely concerning because of high malignancy and poor prognosis. There is increasing evidence that alternative splicing (AS) plays an important role in the development of cancer and the formation of the tumour microenvironment. However, comprehensive analysis of AS signalling in TNBC is still lacking and urgently needed. Methods Transcriptome and clinical data of 169 TNBC tissues and 15 normal tissues were obtained and integrated from the cancer genome atlas (TCGA), and an overview of AS events was downloaded from the SpliceSeq database. Then, differential comparative analysis was performed to obtain cancer-associated AS events (CAAS). Metascape was used to perform parent gene enrichment analysis based on CAAS. Unsupervised cluster analysis was performed to analyse the characteristics of immune infiltration in the microenvironment. A splicing network was established based on the correlation between CAAS events and splicing factors (SFs). We then constructed prediction models and assessed the accuracy of these models by receiver operating characteristic (ROC) curve and Kaplan–Meier survival analyses. Furthermore, a nomogram was adopted to predict the individualized survival rate of TNBC patients. Results We identified 1194 cancer-associated AS events (CAAS) and evaluated the enrichment of 981 parent genes. The top 20 parent genes with significant differences were mostly related to cell adhesion, cell component connection and other pathways. Furthermore, immune-related pathways were also enriched. Unsupervised clustering analysis revealed the heterogeneity of the immune microenvironment in TNBC. The splicing network also suggested an obvious correlation between SFs expression and CAAS events in TNBC patients. Univariate and multivariate Cox regression analyses showed that the survival-related AS events were detected, including some significant participants in the carcinogenic process. A nomogram incorporating risk, AJCC and radiotherapy showed good calibration and moderate discrimination. Conclusion Our study revealed AS events related to tumorigenesis and the immune microenvironment, elaborated the potential correlation between SFs and CAAS, established a prognostic model based on survival-related AS events, and created a nomogram to better predict the individual survival rate of TNBC patients, which improved our understanding of the relationship between AS events and TNBC.

Keywords