Biomarker Screening In Colorectal Cancer

Background:Colorectal cancer (CRC) is a heterogeneous disease associated with diverse phenotypes and genetic features. Differentially expressed gene signatures are associated with the initiation and progression of CRC in different cancer subsets. Poorly differentiated carcinomas are known to have a high-risk of recurrence or metastasis in CRC. However, the genetic characteriatics in this subset tumor are still unclear. The aim of this thesis was to investigate the important gene signature in stage Ⅱ CRC tissues with implications for aggressive malignant phenotypes.Methods:We selected four patients with poorly differentiated adenocarcinoma at stage Ⅱ and performed microarray analysis of carcinoma tissues and normal mucosa isolated from colon distant from the cancer region. A Cancer Array, containing 440 oncogenes and tumor suppressors was untilized to profile mRNA expression and identify differentially expressed genes. Ontology and Ingenuity Pathway Analyses were used to investigate the enriched pathways and their functional relevance of the differentially expressed genes in stage Ⅱ CRC. Genomic alterations of these genes were investigated through The Cancer Genome Atlas database, including 195 CRC specimens. This data was further integrated with published 50 microarray studies from a large number of CRC tumors characterized with prognostic features and The Cancer Genome Atlas database, to identify candidate genes that were significantly associated with the progressive phenotype. Furthermore, immunohistochemical analysis was used to assesse protein expression of cardidate genes and clinicopathological features in an independent cohort of 78 (stage Ⅰ-Ⅳ) CRC specimens.Results:A total of 92 genes that exhibited at least 1.5 fold difference with p<0.05 in mRNA level between the tumor and normal sample was identified. We further identified two clusters, over-expressed gene cluster A and under-expressed gene cluster B in the tumor subgroup by unsupervised hierarchical clustering algorithm. We observed significant enrichment in apoptosis, phosphorylation, cell proliferation, in stage Ⅱ CRC by using GO and DAVID software. Six enriched sub-networks were identified by IPA analysis. They were associated with NFKB, API, STAT3, TP53, HSP90 and CTNNB1 signaling pathways. We generated a list of putative targets of the seven TFs and constructed a transcriptional regulatory gene network including NFKB 1, RelA, TP53, TP63, STAT3, MYC and API by using the newly developed bioinformatics model. Genomic alteration analyses derived from The Cancer Genome Atlas (TCGA) database, identified 24 genes that were altered at 5%～37% across 195 CRC specimens and exhibit consistent expression patterns to our microarray data. Of which 14 genes related to chromosome instability including copy number gain and loss, as well as mRNA expression of those gene directionally correlated to CNVs. Further, a significant high alteration frequency RPN2 and HMGBl was observed in CRC tumors across 9 human cancer types from TCGA. Immunohistochemistry was performed in 78 CRC samples, where RPN2 exhibited significant association with stage Ⅲ/Ⅳ and poorly differentiated tumors, as well as with HMGB1 expression.Conclutions:In this study, using an integrated data analysis based on differentially expressed genes and available genomic and expression datasets, we identified eight gene signatures from stage Ⅱ CRC tumors with implication in late stage aggressive and metastatic potentials. Our study suggests that RPN2 could be potentially used alone or in combination with HMGBl as predictive biomarkers of metastasis in early-stage CRC and may also present as a novel target for RNAi therapeutics.