ABSTRACT
AIM: The aim of this study was to integrate both coding and non-coding available microarray data in the development of colorectal cancer (CRC) with bioinformatics analyses to attain a more inclusive pathobiologic map of their molecular interactions and functions. BACKGROUND: Identification of competing endogenous RNAs (ceRNAs), especially circRNAs, has become a new hotspot in cancer research, although their roles and underlying mechanisms in CRC development remain mostly unknown. METHODS: Microarray data was retrieved from the Gene Expression Omnibus (GEO) database and analyzed. Several bioinformatics tools and databases were applied for further elucidation. Principal component analysis (PCA) was run separately for four datasets. The dysregulated circRNA-miRNA-mRNA, co-expression, and protein-protein interaction (PPI) networks were established. RESULTS: PCA discloses colorectal tumors; normal tissue can be distinguished not only by mRNAs expression profile, but also by both circRNA and miRNA expression profiles. In this study, 14 DE mRNAs, 85 DE miRNAs, and 36 DE circRNAs were identified in CRC tissue and compared with normal tissue. Taking their potential interactions into account, a circRNA-miRNA-mRNA network was constructed. The results disclosed some DE circRNAs with potential oncogenic (circ_0014879) or tumor suppressive (circ_0001666 and circ_0000977) effects. Finally, the PPI network suggests pivotal roles for DOCK2 and PTPRC dysregulation in the progression of CRC, possibly by facilitating tumor escape from immune surveillance. CONCLUSION: The current study proposes a novel regulatory network consisting of DE circRNAs, miRNAs, and mRNAs in CRC development that highlights the roles of DE circRNAs at the upstream of oncotranscriptomic cascade in CRC development, suggesting their potential to be utilized as both prognostic and therapeutic biomarkers.
