Wnt signaling inside the nucleus

Formation of the β‐catenin–TCF‐4 transcriptional complex in colorectal cancer

The TCF and LEF proteins bind directly to DNA through their high mobility group domains, but binding alone is not sufficient for gene transactivation. TCF and LEF transactivate their target genes only when coupled with β‐catenin. The target genes of TCF and LEF identified so far include c‐jun (JUN),^{( 17 )} c‐myc (MYC),^{( 18 )} cyclin D1 (CCND1),^{( 19 )} multidrug resistance 1 (ABCB1),^{( 20} , ^{21 )} matrilysin (MMP7),^{( 22} , ^{23 )} axin2 (AXIN2),^{( 24} , ^{25 )} survivin (BIRC5),^{( 26 )} and many others (http://www.stanford.edu/~rnusse/pathways/targets.html).

T‐cell factor 4 is a member of the TCF and LEF family of transcription factors, which comprises LEF1 (LEF1), TCF‐1 (TCF7), TCF‐3 (TCF7L1), and TCF‐4 (TCF7L2). Only TCF‐4 is expressed commonly in colorectal cancer cells,^{( 27 )} and has been implicated in the maintenance of undifferentiated intestinal crypt epithelial cells.^{( 28 )} Suppression of β‐catenin‐evoked gene transactivation of colorectal cancer cells by dominant‐negative TCF‐4 switches off genes involved in cell proliferation and switches on genes involved in cell differentiation.^{( 29 )} Induction of dominant‐negative TCF‐4 has been reported to restore the epithelial cell polarity of a colorectal cancer cell line and convert the cell line into a single layer of columnar epithelium,^{( 30 )} indicating that colorectal cancer cells still require accumulation of the β‐catenin protein, thereby transactivating the target genes of TCF‐4, for maintenance of cell proliferation, depolarization, and dedifferentiation.

Protein components of the β‐catenin and TCF and LEF transcriptional complexes

Groucho/transducin‐like enhancer protein,^{( 31 )} C‐terminal binding protein,^{( 32} , ^{33 )} CBP/p300,^{( 34} , ^{35 )} Smads,^{( 36 )} NEMO‐like kinase,^{( 37 )} Chibby,^{( 38 )} and other proteins^{( 39} , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³ , ⁵⁴ , ^{55 )} have been reported to interact with the β‐catenin and TCF and LEF complexes and modulate their transcriptional activity (Table 1). Fig. 2 shows the proposed binding sites of β‐catenin and TCF and LEF on some of these proteins. In a series of recent proteomic studies we also identified FUS/TLS,^{( 56 )} PARP‐1,^{( 57 )} Ku70,^{( 58 )} Ku80,^{( 58 )} Topo IIα,^{( 14 )} and SF1^{( 59 )} as putative components of the β‐catenin–TCF‐4 nuclear complex (Fig. 2; Table 1). Among these proteins, Topo IIα and PARP‐1 are enhancers, and SF1, FUS/TLS, and Ku70 are repressors of the β‐catenin and TCF and LEF transcriptional complexes. The identification of these proteins allowed us to clarify some novel functional aspects of the complexes in DNA damage recognition and pre‐mRNA splicing.

DNA damage and Wnt signaling

We identified PARP‐1 as a coactivator of the β‐catenin and TCF‐4 complex.^{( 57 )} PARP‐1 is an enzyme that catalyzes the polyADP‐ribosylation of a variety of acceptor proteins in response to DNA damage and plays an important role in the maintenance of genome integrity.^{( 60} , ^{61 )} PARP‐1 is cleaved by caspase‐3 during the course of apoptosis.^{( 60} , ^{61 )} Cleavage by caspase‐3 is likely to inhibit the interaction between PARP‐1 and TCF‐4.^{( 57 )} Removal of PARP‐1 from the β‐catenin and TCF‐4 complex may inhibit gene transcriptional activity. Ku70 is another protein that interacts physically with a domain of TCF‐4 containing the high mobility group box and functions as an inhibitor of TCF and LEF transcriptional activity.^{( 58 )} In response to DNA damage, PARP‐1 polyADP ribosylates its own automodification domain and inhibits its interaction with TCF‐4. PARP‐1 competes with Ku70 for binding to TCF‐4. Dissociation of PARP‐1 from TCF‐4 allows Ku70 to interact with TCF‐4. Binding of Ku70 to TCF‐4 seems to inhibit the interaction between TCF‐4 and β‐catenin and attenuate the transcriptional activity of TCF‐4 (Fig. 3). Increased expression of PARP‐1 and decreased expression of Ku70 in colorectal adenocarcinoma cooperatively enhances TCF‐4/β‐catenin‐mediated gene transactivation and may contribute to colorectal carcinogenesis.

Pre‐mRNA splicing and Wnt signaling

The FUS/TLS protooncogene product^{( 56 )} and SF1^{( 59 )} are corepressors of the β‐catenin and TCF‐4 complex. FUS/TLS was identified originally as the N‐terminal part of a fusion‐gene product produced by chromosomal translocation t(12;16)(q13;p11) in malignant myxoid liposarcoma.^{( 62} , ^{63 )} FUS/TLS participates in splicing processes as well as in transcription. FUS/TLS modulates 5′‐splice site selection in pre‐mRNA splicing.^{( 64 )} SF1 recognizes the intron branch point sequence UACUAAC in the pre‐mRNA during splicesome assembly and regulates the pre‐mRNA splicing reaction.^{( 65 )} In addition to FUS/TLS and SF1, nuclear β‐catenin interacts with several classes of RNA‐binding proteins including DEAD‐box RNA helicases, small nuclear ribonucleoprotein‐related proteins, and heterogeneous nuclear ribonucleoproteins and regulates pre‐mRNA splicing.^{( 56 )} Heterogeneous nuclear ribonucleoproteins bind nascent RNA polymerase II transcripts and make up the earliest detectable complex assembled on pre‐mRNA during the splicing reaction.^{( 66 )}β‐Catenin cDNA transfection induced the expression of a splicing variant of ER‐β that lacks exons 5 and 6 (ER‐βΔ5‐6). ER‐βΔ5‐6 inhibits the ligand‐dependent transcriptional activity of wild‐type ER‐β in a dominant‐negative manner.^{( 56 )} ER‐βΔ5‐6 was detected in most of the 16 human colon cancer samples examined by Lee et al.,^{( 67 )} indicating that ER‐βΔ5‐6 may play a specific role in colorectal carcinogenesis.

SF1 interacts with the β‐catenin and TCF‐4 complex and regulates pre‐mRNA splicing as well as gene transactivation.^{( 59 )} SF1 is responsible for the induction of known cancer‐related splice variants, including Wnt‐induced secreted protein‐1v (WISP1v)^{( 68 )} and fibroblast growth factor receptor‐3 (FGF3R)‐ATII,^{( 69 )} as well as ER‐βΔ5‐6. Human cancer expresses a large variety of alternatively spliced mRNA transcripts. Alternatively spliced RNA transcripts with high tumor specificity may be considered candidate cancer biomarkers and therapy targets.

Dynamic regulation of the protein composition of the β‐catenin–TCF‐4 nuclear complex during intestinal epithelial cell differentiation

The protein composition of the TCF‐4‐containing nuclear complex is not fixed, but is regulated dynamically by endogenous programs associated with intestinal epithelial cell differentiation. The expression level of the PARP‐1, Topo IIα, FUS/TLS, and SF1 proteins was regulated in each according to the differentiation status of intestinal epithelial cells (Fig. 4).

PARP‐1 was expressed intensely in intestinal adenoma cells of FAP patients and multiple intestinal polyposis mice.^{( 57 )} The expression of PARP‐1 was closely associated with the accumulation of β‐catenin and with the undifferentiated status of intestinal epithelial cells. The protein level of PARP‐1 was highest in undifferentiated cells at the bottom of normal intestinal crypts, and gradually decreased along the axis of cell differentiation. Removal of PARP‐1 from the β‐catenin and TCF‐4 complex may inhibit gene transcriptional activity during the course of differentiation and death of intestinal epithelial cells.

Expression of Topo IIα is limited to the nuclei of epithelial cells in the supra‐basal to middle region of normal large‐intestinal crypts,^{( 14 )} where the stem and transit amplifying cells are believed to reside and DNA replication and cell proliferation are most active. The expression level of Topo IIα in colorectal adenoma cells of FAP patients is almost equal to that of normal proliferating cells. The expression of Topo IIα is markedly increased in adenocarcinoma cells of patients with FAP and sporadic cancer in comparison with neighboring normal intestinal epithelial cells. Importantly, the Topo IIα protein is colocalized with β‐catenin in the nuclei of colorectal cancer cells, but not in the nuclei of normal intestinal epithelial cells.

The expression of FUS/TLS was also closely associated with the accumulation of β‐catenin and with the undifferentiated status of intestinal epithelial cells.^{( 56 )} The expression level of FUS/TLS decreased gradually from the bottom to the surface of the intestinal crypts (Fig. 4a). FUS/TLS was expressed strongly in the nuclei of colorectal adenocarcinoma cells in both FAP and sporadic cases.

In contrast to PARP‐1, Topo IIα, and FUS/TLS, the expression of SF1 is correlated with the differentiation status of intestinal epithelial cells and is inversely correlated with tumorigenesis (Fig. 4b). Sodium butyrate, a differentiation inducer of colorectal cancer cells, increases the expression of SF1.^{( 59 )} SF1 has been identified as one of the proteins whose expression is upregulated by the induction of dominant‐negative TCF‐4. The expression of SF1 is regulated by the β‐catenin–TCF‐4 complex and, conversely, SF1 regulates its transcriptional activity. The induction of SF1 protein in intestinal villous epithelial cells may reflect a differentiation‐associated negative‐feedback mechanism that prevents continuity of the β‐catenin and TCF‐4 transcriptional activity. Furthermore, SF1 is known to be a component of the complex and to negatively regulate β‐catenin‐evoked gene transactivation and cell proliferation.^{( 59 )} Consistently, Sf1 ^+/– mice show increased susceptibility to colon tumorigenesis induced by azoxymethane in comparison with wild‐type (Sf1 ^+/+) mice.^{( 70 )} The tumors developing in Sf1^+/– mice show higher expression levels of known target genes of TCF/LEF than the tumors developing in Sf1^+/+ mice.^{( 70 )}

Development of drugs targeting the β‐catenin–TCF‐4 complex

Several attempts have been made to identify small molecules that interfere with the β‐catenin and TCF/LEF nuclear complexes. Lepourcelet et al. established a high‐throughput screening method to search for small‐molecule antagonists targeting the β‐catenin and TCF‐4 interaction.^{( 71 )} Two fungal derivatives were found to suppress the transcriptional activity of the β‐catenin and TCF‐4 complex. Although these molecules also inhibited the interaction between β‐catenin and the APC gene product, the feasibility of targeting protein–protein interactions in cancer drug discovery was confirmed.

Emami et al. screened a secondary structure‐templated small‐molecule library of 5000 compounds and identified one compound, named ICG‐001, which was found to bind specifically to CBP and inhibit its interaction with β‐catenin.^{( 72 )} Waltzer and Bienz first reported that Drosophila CBP binds to dTCF and represses dTCF,^{( 34 )} whereas mammalian CBP, or its closely related homolog p300, has been reported to bind to β‐catenin and function as a coactivator. The region necessary for the interaction with p300 was reported to include the N‐terminal transactivation domain and the first two armadillo repeats of β‐catenin,^{( 35 )} although different results were reported by Hecht et al.^{( 73 )} and Takemaru and Moon.^{( 74 )} CBP and p300 function as transcriptional coactivators by modifying regional chromatin structure and by recruiting the basal transcription machinery to gene promoters. CBP and p300 have been shown to play different roles in transactivation of the BIRC5 gene by β‐catenin.^{( 75 )} ICG‐001 effectively reduces the growth of colorectal cancer cells and induces apoptosis. CBP is known to function as a coactivator of a large variety of transcription factors, and the specificity and clinical utility of ICG‐001 remains to be clarified. However, this study revealed for the first time the feasibility of targeting molecules associated with the β‐catenin–TCF‐4 complex.

An inhibitor of the polyADP‐ribosylation activity of PARP‐1, 3‐aminobenzamide, did not affect TCF/LEF transcriptional activity,^{( 57 )} indicating that the enzymatic activity of PARP‐1 is not required for augmentation of gene transactivation.

Topo II is a known target of chemotherapeutic agents used widely in current oncological practice.^{( 76} , ^{77 )} Topo II is an enzyme that catalyzes topological changes in the DNA double helix, and its catalytic activity is essential for the process of DNA replication and mitosis.^{( 78 )} We have demonstrated that Topo IIα is a functional component of the β‐catenin and TCF‐4 complex, and that its transcriptional activity is inhibited by two structurally and mechanistically distinct Topo II inhibitors, merbarone and etoposide.^{( 14 )} Because Topo IIα is an enhancer of the β‐catenin and TCF‐4 complex, and interaction of Topo IIα with β‐catenin is specific to colorectal cancer, Topo IIα seems to be an ideal candidate for drug targeting of the Wnt signaling pathway. However, a relatively high dose of merbarone or etoposide was necessary to shut down the TCF/LEF transcriptional activity completely. A new drug targeting the interaction of Topo IIα with β‐catenin as well as the catalytic activity of Topo IIα might suppress aberrant Wnt signaling more effectively than the currently available Topo II inhibitors. High‐throughput screening of agents that antagonize the interaction, and crystal structure analysis of the mutual binding interfaces of the β‐catenin and Topo IIα proteins, might facilitate drug discovery.

Conclusion

Combination chemotherapeutic regimens including oxaliplatin and irinotecan have significantly improved the survival of patients with colorectal cancer. Molecular therapy targeting vascular endothelial growth factor and epidermal growth factor receptor is now beginning to be incorporated into clinical trials. Despite these advances, colorectal cancer is still a major threat to human health, and identification of new drug targets is necessary in order to increase therapy options. The feasibility of targeting regulatory molecules associated with the β‐catenin and TCF transcription complex has now been confirmed. The discovery of ICG‐001 may represent a milestone that could lead to the development of new therapeutics with higher efficacy and specificity for colorectal cancer.