Analysis of gene expression in human colorectal cancer tissues by cDNA microarray: implications for novel immunotherapy strategies

BIANCHINI M; LEVY EM; ZUCCHINI C; PINSKY V; MACAGNO C; VALVASSORI L; PAOLO CARINCI; MORDOH J

Anahein, California

Congreso; 2005 AACR Annual Meeting; 2005

American Association for Cancer Research

Human colorectal cancer (HCC) is one of the most common malignancies in the Western world. The development and use of molecular-based diagnosis and therapy for human tumors requires detailed molecular characterization of in vivo tumor tissues. In this study we attempted to reveal new tumor markers as well as target genes for novel immunotherapeutic strategies. For this, we used cDNA microarray to profile genes that are differentially expressed in normal and tumor tissues.We examined the global gene expression profiles for 38 human colorectal tissues representing 25 primary colorectal tumors and 13 normal colonic mucosa, obtained from surgical resections. High quality RNA was used for array analysis on a 19.200-element cDNA microarray generated by UHN Microarray Centre. Ten differentially expressed genes were further studied by qRT-PCR.With the purpose to identify a set of genes whose expression is commonly altered in colorectal cancer tissues, we carried out a statistical analysis based on a non parametric t-test (p< 0.001). We identified 1.195 clones, including 445 known expressed genes, many of which appear to reflect primary events important for colon carcinogenesis; for example ATM, COL8A1, CRK, ENC1, SMURF2, ZNF146 transcripts were more abundant in the tumor than in normal tissue. Genes that resulted more highly expressed in normal colon mucosa are also of considerable interest: CA2, DEFB1, DSC3, SDC1, TIMP3, ZNF185. When our analysis was focused on genes associated with immune response, we found 22 transcripts, several of which strongly support the possibility of the acquisition of tumor immune escape mechanisms (down regulation of â2-microglobulin versus HLA-E up regulation). Strikingly, using hierarchical clustering method to group our samples based on similarity in their gene expression pattern, we observed that 151 transcripts were able to separate two major patient clusters. When these clusters were examined with respect to gene function, we observed a significant (p<0.05) representation of genes belonging to immune response family. The genes found in our study could be potent candidates for diagnosis and therapy of colorectal cancer. Remarkably, for the first time we described an in vivo co-regulation of critical factors involved in tumor escape mechanisms in colon cancer. In addition, this escape mechanism appear to be less activated in a subgroup of patients that we have identified using an unsupervised clustering method. Consequently, these new molecular classification elements for HCC, associated with specific changes of several components of the immunological machinery, could provide new insights to explain the different grade of therapeutic responses observed in colorectal cancer patients enrolled in immunotherapy clinical trials. We conclude that future cancer immunotherapy approaches must aim to overcome instances of tumour escape.