The Construction of Bone Metastasis-Specific Prognostic Model and Co-expressed Network of Alternative Splicing in Breast Cancer

Background: Breast cancer (BRCA) ranks among the top most common female malignancies and was regarded as incurable when combined with bone and distant metastasis. Alternative splicing events (ASEs) together with splicing factors (SFs) were considered responsible for the development and progression of tumors. Methods: Datasets including RNA sequencing and ASEs of BRCA samples were achieved from TCGA and TCGASpliceSeq databases. Then, a survival model was built including 15 overall-survival-associated splicing events (OS-SEs) by Cox regression and Lasso regression. The co-expressed SFs of each bone-and-distant-metastasis-related OS-SE were discovered by Pearson correlation analysis. Additionaly, Gene Set Variation Analysis (GSVA) was performed to identify the downstream mechanisms of the key OS-SEs. Finally, the results were validated in different online platforms. Results: A reliable survival model was established (the area under ROC = 0.856) and CIRBP was found co-expressed with FAM110B ( R = 0.320 , P < 0.001 ) associated with the fatty acid metabolism pathway. Conclusions: Aberrant SF, CIRBP, regulated a specific ASE, exon skip (ES) of FAM110B, during which the fatty acid metabolism pathway played an essential part in tumorigenesis and prognosis of BRCA.