Project Status
Microarray analysis performed at our research site has identified a number of genes genes that are differentially regulated between colorectal tumors and normal colonic mucosa. Among the genes upregulated in CRC (202 genes, ³ 3 fold) are known targets of the Wnt pathway but also a large number of other genes that have not yet been linked to Wnt signaling. Our bioinformatic analysis shows that TCF/b-catenin binding sites indicative of Wnt target genes are significantly overrepresented in genes upregulated in CRC as compared to genes which are not changed or downregulated (2.1 vs 0.4 binding sites per 1kb of upstream promoter region, Fig. 1).

By microarray analysis we could identify 208 genes to be downregulated (³ 3 fold) after expression of a dominant-negative TCF-4 mutant in a colorectal cancer cell line (DLD1). By matching genes upregulated in CRC with those downregulated by dominant-negative TCF we could narrow down the list of potential Wnt targets to 151 genes. Thus the spectrum of genes regulated by Wnt signaling in CRC appears to be much broader than suggested by the current list of known Wnt target genes. Since activation of the Wnt pathway is an essential and early step of colorectal tumorigenesis these novel target genes are excellent candidates for diagnosis and drug targeting.

In the analyses we noticed that several components of the Wnt pathway are upregulated in the tumors. For instance, we could show that conductin, a negative regulator of the pathway, is upregulated in CRC and liver tumors  and is a target gene of TCF/b-catenin (5). We have also analysed the nuclear factors pontin and reptin, which were shown to interact with b-catenin and to differentially regulate its activity: while pontin activates transcription by TCF/b-catenin, reptin represses transcription. We have cloned the promoters of both genes and found that pontin, but not reptin promoter activity is directly increased by TCF/b-catenin. This indicates that a positive feedback loop exists in CRC which superactivates the Wnt pathway through the selective upregulation of pontin. Of note, pontin is a cofactor of the c-myc protein and might thus also act further downstream in the Wnt signaling cascade (Kurniati et al., Manuscript in preparation).

We also found that several genes with a function in mitosis (e.g. cdc2, mad2) are upregulated in CRC. Specifically we have analysed the mad2 gene and could show that its promoter is activated by TCF/b-catenin. Since mad2 is a component of the mitotic spindle check point machinery which controls the fidelity of chromosomal seggregation our data indicate that Wnt signaling directly impinges on mitotic processes. This finding is of importance given the frequent observation of chromosomal instability in CRC.

Rationale
We assume that many of the genes transcriptionally altered in CRC as compared to normal mucosa are not causally involved in cancer formation but are rather indirectly regulated as a general consequence of tumor growth. To distinguish the “important” genes from the “bystanders” we concentrate on genes directly regulated by aberrant activation of the Wnt pathway.

Identification of candidate Wnt target genes
As exemplified above, our strategy is to compare changes in gene expression in CRC vs. normal mucosa to those in cellular systems where we can inhibit or activate Wnt signaling in a controlled manner. Genes that match in such an analysis will be further studied in promoter and functional experiments. We apply the following criteria for narrowing down genes from our initial comparison (i.e. the 151 genes upregulated in CRC and downregulated by dom.-neg. TCF):

(i) Likelihood to be involved in a signaling pathway or major regulatory step in cancer from known function or sequence.

We focus on genes with a putative role in signal transduction events, such as growth factors, kinases, phosphatases, transcription factors or genes with a putative function in other tumor-related phenotypes, e.g. invasion, angiogenesis, and cell cycle regulation. Additional interesting candidates are components of the Wnt pathway, which might have negative or positive feedback roles as seen with conductin or pontin. There might also be genes that are upregulated in CRC but -  due to cell line specificities - not repressed by dom.-neg. TCF in the DLD1 system. To avoid neglecting such genes we will also perform microarray analysis of alternative cellular models  for Wnt pathway regulation. For instance we have established previously that Wnt conditioned medium can induce stabilization of b-catenin and expression of conductin in breast carcinoma cells (5).

(ii) Regulation of promoter activity by TCF/b-catenin complexes.

For analysis of direct transcriptional regulation of a given gene by TCF/b-catenin we have cloned promoter sequences by genomic PCR and generate luciferase reporter constructs. The putative transcription start sites will be taken from the literature, if available, or determined by computer-assisted alignment of EST clones with genomic sequences. The reporters are tested for activation by TCF/b-catenin or for repression by dominant-negative TCF. Promoters which show clear regulation are mutated at the putative TCF sites to validate that the regulation depends on direct interaction with TCF. In addition chromatin IP assays will be performed to demonstrate TCF/b-catenin binding to promoters in vivo. By these analyses we expect to  to obtain criteria for further selection of genes for functional assays.

We will also analyze genes that are downregulated in CRC and upregulated by dom.-neg. TCF. These genes are thought to be indirectly controlled by Wnt signaling but might nevertheless have important functions. A paradigm for this type of regulation is the p21CIP gene which is repressed by c-myc in response to Wnt signaling.

For the functional analysis we will either upregulate the encoded proteins by exogenous expression of the respective cDNAs or suppress the expression of endogenous proteins by RNA interference technology. As read-out systems we will use assays that allow to determine effects on cell proliferation and apoptosis on a single cell basis (e.g.  BrdU incorporation and staining with annexinV, repectively). 

We will also determine whether individual genes can affect invasiveness. Here we will use a collagen invasion assay where transfected cells can be traced after co-transfection of a GFP marker plasmid. For selected genes, stable transfectants will be generated that will also allow in vivo analysis of tumor growth and metastasis formation in collaboration with the CancerNet project “Animal tumor models for the evaluation of candidate gene function in tumor metastasis” (J. Sleeman). 

In specific cases it might also be possible to interfere with protein function with blocking antibodies or to test the activity of secreted proteins by use of recombinant proteins or conditioned medium of transfected cells.
 
A specific focus will be on the role and regulation of genes involved in mitosis or regulation of S-Phase which we have frequently found in our microarray analyses. Here we have established a variety of assays such as mitotic checkpoint analysis, FISH for analyzing chromosomal content, and assays for microtubule dynamics. Evidently, in the end we will have to focus on a rather limited number of attractive candidates for detailed analysis.