Integrative analysis of the canonical TGF?-Smad signaling pathway in breast cancer (Preprint)
Dan Zhou,Weiwei Tang,Yun Zhang,Hanxiang An
DOI: https://doi.org/10.2196/preprints.8978
2017-01-01
Abstract:BACKGROUND Breast cancer deaths are mainly attributed to metastasis, while epithelial to mesenchymal transition (EMT) plays a pivotal role in promoting cancer cell metastasis. EMT is activated by various stimuli, resulting in different therapeutic responses. The TGFβ signaling pathway is a canonical driver of EMT, and many molecular drugs are developed to target this pathway. However, the alternations of this pathway in BC remain elusive. OBJECTIVE The aim of this study was to investigate the relationship between the canonical TGFβ-Smad signaling pathway and breast cancer pathogenesis. METHODS Herein, we investigated the mutational, transcriptional, epigenetic and post-transcriptional alternations of 14 core members in this pathway through bioinformatics approach by analyses of 56,496 BC patients from various public databases, including COSMIC, Oncomine, DiseaseMeth and Starbase. A protein-protein interaction network was derived from Cytoscape software. Associations of these selected members with relapse free survival (RFS), overall survival (OS), distant metastases free survival (DMFS) and post progression survival (PPS) were performed using the Kaplan-Meier plotter online tool. RESULTS Our integrated analyses revealed that Smad4 was the most frequently mutated member of the TGFβ-Smad pathway in BC with a mutation frequency of 0.85% (30/3537). The expression levels of TGFβ1, TGFβ3, Smad1, Smad2 and Smad7 were significantly higher, whereas TGFβ2, ALK1, ALK2, ALK5, TGFβR2, Smad3, Smad4, Smad5 and Smad6 were downregulated in BC compared with normal subjects (P<0.001). Evaluation of epigenetic alteration identified that the promoters of ALK1 (P=0.000e+00), ALK2 (P=2.22e-16) and ALK5 (P=1.625e-12) were hypermethylated in 1,274 BC samples. In contrast, Smad2 (P=6.56e-04) was significantly hypomethylated in BC. Although all 14 members had potential diagnostic significances with different prediction power (AUC ranging from 0.788 to 1.00) in various subtypes of BC, only TGFβ3 could predict RFS (HR=0.69, 95% CI [0.62-0.77]), OS (HR=0.69, 95% CI [0.56-0.86]), DMFS (HR=0.8, 95% CI [0.66-0.97]) and PPS (HR=0.76, 95% CI [0.6-0.97]) for patients with BC in an independent manner. Additionally, 11 out of 14 members were associated with RFS in luminal A BC. CONCLUSIONS Our results indicated that TGFβ-Smad pathway core members could serve as novel diagnostic and prognostic biomarkers in BC.