The molecular phenotyping of human tumors has yielded potentially important therapeutic prognostic information. In colon cancer, for example, chromosome 18 deletion identifies a subgroup of patients at high risk for relapse following surgical resection, and who may benefit from adjuvant chemotherapy. By contrast, the microsatellite instability (MSI) phenotype identifies patients who may be at lower risk of progression and may have higher cure rate compared to similar stage cancers. In the past few years, CpG island hypermethylation and silencing of genes has emerged as one of the most common molecular changes in neoplasia. Recent work has put in evidence a novel pathway in colorectal tumorigenesis termed CpG Island Methylator Phenotype (CIMP), which effects about 50 percent of the cases. CIMP appears to lead to tumor formation/progression through the simultaneous methylation of multiple genes. In colon cancer, CIMP is associated with a distinct genetic profile, with a high (30 percent) rate of MSI, a high (90 percent) rate of K-RAS mutations in MSI-negative cases, and a low (25 percent) rate of p53 mutations. These data suggest that CIMP may add an important layer of information in the molecular phenotyping of malignances for prognostic purposes. In this grant, we propose to explore this issue in colon cancer using large series of uniformly treated patients entered on ECOG protocols. By determining the methylation and mutation/deletion profiles of 780 cases, we determine specifically (1) the prognostic/predictive significance of CIMP and gene-specific methylation for relapse rates in limited stage disease (following primary treatment), response rate to chemotherapy in advanced stage disease an overall survival in both groups and (2) the interactions between genetic and epigenetic events in neoplasia and the prognostic/predictive significance of combined epigenetic and genetic profiling. Ultimately, this work aims at establishing that methylation profiling may carry the same clinical implications as that already established for genetic profiling in neoplasia.
