Cellular Model of Colon Cancer Progression Reveals Signatures of mRNAs, miRNA, lncRNAs, and Epigenetic Modifications Associated with Metastasis

摘要

Here, we developed and comprehensively characterized a cellular model of colon cancer progression consisting of four defined derivatives of a colon cancer cell line that resulted from consecutive epithelial-mesenchymal and mesenchymal-epithelial transitions (EMT/MET) and phenotypically recapitulate the metastatic cascade. Initial EMT was induced by prolonged exposure to IL6, a cytokine also generated by the tumor-stroma. Genome-wide characterization of transcriptional (mRNA, miRNA, and lncRNA) and epigenetic (DNA methylation, H3K4me3, H3K79me3, and H3K27me3 histone modifications) profiles of the cell derivatives, combined with correlative analyses of expression, methylation, and clinical data from the TCGA-COAD database gave insights into the molecular basis of their phenotypic changes. The signatures characterizing invasive, mesenchymal-like cell states as well as the metastases-derived epithelial-like state showed significant association with metastasis, positive nodal status, and poor survival of colon cancer patients. Global hypomethylation of gene-regulatory regions was observed during tumor progression, with the lowest degree of methylation present in cells isolated from metastases. Upregulation of an axon-guidance-related gene signature was the most significant feature of metastatic tumor cells and was also found in primary tumors from colon cancer patients with distant metastases. Furthermore, the microRNAs miR-99a, miR-100, and miR-125b showed elevated expression in mesenchymal-like cells, associated with poor survival, and promoted migration and invasion. Finally, elevated expression of H19 lncRNA due to promoter demethylation was observed in cells isolated from metastases and was associated with poor survival of colon cancer patients. In the future, our results may be further exploited for the discovery and evaluation of novel metastasis-associated mechanisms and biomarkers. (C) 2017 AACR.