TGFB-related signature of EMT identifies basal subtype of breast cancer

Background : Cancer is a major health problem world‐wide and most deaths are due to tumour metastasis to other organs. Epithelial‐to‐mesenchymal transition (EMT) is a physiological process by which epithelial cells undergo a phenotypic transformation to become more motile, and obtain characteristics of fibroblastoid cells. This normal developmental process is subverted in cancer and facilitates cancer progression and metastasis. TGFb is one of the inducers of EMT. Methods : Four published microarray datasets of TGFb‐stimulated cancer cell lines were used. The quality of the datasets was validated using packages in R. To obtain a TGFb‐related signature of EMT, a meta‐analysis of the data was undertaken. We used the Limma package in R/Bioconductor to obtain test statistics in each of the datasets. The signature of TGFb‐driven EMT was then obtained using two. Methods : vote counting (Venn method) and product of rank (PR). GO enrichment analysis was performed. Finally the signature was assessed in breast cancer patient data as well as in breast cancer cell line data from the TCGA database and GOBO online tool. Results : Using vote counting a total of 25 genes with q‐value   1 were identified, while using PR, we obtained 194 genes. 22 genes overlapped between the two methods, six of which were also reported by other studies. GO analysis showed that the PR signature was involved in developmental processes, wound healing, cell adhesion and migration. The up‐regulated genes of the PR signature could identify basal‐B breast cancer cell lines and could distinguish between basal and luminal subgroups. Also the expression level of these genes was higher in triple negative (TN) compared to HER2+ and hormone receptor (HR)+ subgroups. The down‐regulated gene‐set was an indicator of basal‐B breast cancer cell lines but could not distinguish between basal‐A and luminal subgroups. The TGFb‐related signature of EMT could identify subset of patients with more aggressive forms of breast cancer (basal and TN). These genes can be further analyzed for diagnostic and therapeutic purposes.