Identification of driver mutations in gastrointestinal tract cancer

B179 Recent large-scale sequencing studies have identified an average of 81 somatic mutations in a typical human colorectal tumor. The next step must be to distinguish passenger from driver mutations in carcinogenesis. In this study we employ a synthetic DNA transposon, called Sleeping Beauty (SB), to create random somatic mutations in the gastrointestinal (GI) tract of mice, leading to carcinoma development. Tumor tissue is collected and transposon insertions are mapped. Due to the small number of insertions generated relative to genome size, a common insertion site (CIS) in three or more independent tumors likely represents a bona fide oncogenic mutation. This approach utilizes somatic transposition of an SB transposon vector, called T2/Onc, which can induce gain-of-function or loss-of-function mutations upon insertion within or near genes. Thus, T2/Onc is capable of activating oncogenes or disrupting tumor suppressors. In this study, T2/Onc mutations were directed to the GI tract using a Cre controlled SB transposase knock-in allele, in combination with a Cre recombinase transgene driven by the Villin promoter, plus the T2/Onc transposon transgene array. A cohort of 45 mice containing all three transgenes along with 81 control mice containing two of the three transgenes were aged for 18 months and monitored for tumor development. Greater than 55% of experimental animals became moribund before 18 months compared to 16% of control mice. Triple transgenic mice had an average of 14 GI tract polyps per mouse while control mice had no detectable GI tract tumors. In addition to adenomas, multiple foci of gastrointestinal intraepithelial neoplasia (GIN) were identified by H&E staining. Nine experimental animals also had liver tumors in addition to GI tract neoplasia. Linker-mediated PCR was performed on tumor DNA from 63 individual GI tract tumors from 16 mice to determine SB insertion sites. Preliminary analysis of roughly 6,000 mappable insertions identified approximately 350 common insertion sites, each site representing a candidate cancer locus. The most commonly disrupted site was the APC locus, harboring 42 non-redundant insertions. Other known colorectal tumor suppressor genes were identified as a CIS, including Smad4, p53, and PTEN, indicating that this approach can effectively uncover driver mutations in an epithelial cancer. Most of the candidate genes identified have not previously been associated with GI tract tumor formation. We will be testing a subset of identified genes using human cell lines and transgenic mouse models to validate their oncogenic potential. In addition, we will analyze the list of identified genes to find cooperating mutations/pathways present within the same tumor.