The Farnham Laboratory

Expression Profiling

Many human cancers are caused by deregulation of specific transcription signaling networks. However, it is not clear whether neoplastic transformation is caused by transcription factors activating a normal set of target genes to higher levels or whether increased amounts of a transcription factor result in activation of a tumor-specific set of target genes.

To address the mechanism by which transcriptional deregulation causes neoplasia, we are pursuing two complementary approaches. First, we are using gene expression profiling and representational difference analysis (a form of subtractive hybridization followed by cloning) to compare normal cells versus cancer cells. Using a mouse model system, we have identified a set of genes which are highly upregulated in liver tumors (Graveel et al,, 2001). Several of the genes we identified are novel and we are actively pursuing the cloning and characterization of the mouse and human cDNAs to these new, tumor-specific mRNAs.

As a complementary approach, we have developed a chromatin immunoprecipitation protocol which allows us to examine binding of oncogenic transcription factors (e.g. E2F, Myc, and b-catenin) to target promoters in normal cells versus cancer cells. Our current studies suggest a complicated specificity in transcription factor/promoter interactions which were not revealed using in vitro assays. Our studies to date have been performed using tissue culture model systems or mouse cancer models. However, it is not clear that tissue culture cells or mice provide the perfect model for human cancers. Therefore, our current goals are to use gene expression profiling and chromatin immunoprecipitation to examine primary tumor samples taken directly from cancer patients. For these studies, we are focusing on cancers of the liver and the colon.

Related Publications