Application of open-access databases to determine functional connectivity between resveratrol-binding protein QR2 and colorectal carcinoma

Colorectal cancer (CRC) is a major cause of cancerassociated deaths worldwide. Recently, oral administration of resveratrol (trans-3,5,4'-trihydroxystilbene) has been reported to significantly reduce tumor proliferation in colorectal cancer patients, however, with little specific information on functional connections. The pathogenesis and development of colorectal cancer is a multistep process that can be categorized using three phenotypic pathways, respectively, chromosome instability (CIN), microsatellite instability (MSI), and CpG island methylator (CIMP). Targets of resveratrol, including a high-affinity binding protein, quinone reductase 2 (QR2), have been identified with little information on disease association. We hypothesize that the relationship between resveratrol and different CRC etiologies might be gleaned using publicly available databases. Aweb-based microarray gene expression data-mining platform, Oncomine, was selected and used to determine whether QR2 may serve as a mechanistic and functional biotarget within the various CRC etiologies. We found that QR2 messenger RNA (mRNA) is overexpressed in CRC characterized by CIN, particularly in cells showing a positive KRAS (Kirsten rat sarcoma viral oncogene homolog) mutation, as well as by the MSI but not the CIMP phenotype. Mining of Oncomine revealed an excellent correlation between QR2 mRNA expression and certain CRC etiologies. Two resveratrol-associated genes, adenomatous polyposis coli (APC) and TP53, found in CRC were further mined, using cBio portal and Colorectal Cancer Atlas which predicted a mechanistic link to exist between resveratrol -> QR2/TP53 -> CIN. Multiple web-based data mining can provide valuable insights which may lead to hypotheses serving to guide clinical trials and design of therapies for enhanced disease prognosis and patient survival. This approach resembles a BioGPS, a capability for mining web-based databases that can elucidate the potential links between compounds to provide correlations of these interactions with specific diseases.