WO2010014043A1 - Methods of preventing, arresting or reversing tumourigenesis and of identifying compounds capable of the same - Google Patents

Abstract

Disclosed are methods of preventing, inhibiting, arresting or reversing tumorigenesis in a cell and of inducing programmed cell death (apoptosis) in a tumor cell. The methods include altering the formation of a complex between RUNX3, or a functional fragment thereof, and at least one of (i) beta-catenin, or a functional fragment thereof, and (ii) a member of the TCF/LEF transcription co-factor family. Disclosed are also methods of diagnosing the risk of tumorigenesis in a cell and of diagnosing the risk of developing a neoplasm in a subject. Such method includes assessing the formation of a complex as defined above. An in-vitro method of identifying a compound capable of altering the formation of the afore defined complex is also disclosed. The method includes contacting the components that form the above complex with each other and adding a compound to the test tube suspected to modulate said complex formation.

Description

METHODS OF PREVENTING, ARRESTING OR REVERSING TUMOURIGENESIS AND OFIDENTIFYINGCOMPOUNDS CAPABLE OFTHE SAME

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application makes reference to and claims the benefit of priority of an application for "Methods Of Preventing, Arresting Or Reversing Tumourigenesis And Of Identifying Compounds Capable Of The Same" filed on July 29, 2008 with the United States Patent and Trademark Office, and there duly assigned serial number 61/084,486. The contents of said application filed on July 29, 2008 is incorporated herein by reference for all purposes, including an incorporation of any element or part of the description, claims or drawings not contained herein and referred to in Rule 20.5(a) of the PCT, pursuant to Rule 4.18 of the PCT.

FIELD OF THE INVENTION

[0002] The present invention relates to the field of prevention, arrest and reversal as well as diagnosis of tumorigenesis. Provided is also a corresponding method of identifying a compound capable of prevention, arrest and reversal of rumorigenesis.

BACKGROUND OF THE INVENTION

[0003] Cancer is a major cause of death worldwide, being the second-leading cause of death in developed countries and even the number one cause of death in e.g. Australia, Japan, Korea, Singapore and the male population of the UK and Spain. The number of people who develop cancer each year is increasing.

[0004] Currently, cancer therapy involves surgery or focuses on the functional or genetic changes associated with the transformation of cells into malignant cells. An ideal anti-cancer drug should selectively kill, or at least inhibit, rapidly proliferating cancerous cells, while leaving non-cancerous cells unaffected. Recent approaches include immunotherapy using antibodies directed to markers of selected types of cancer cells (e.g. US patent application 2005/0244417), the application of agonists to receptors that are expressed on certain types of cancer cells (US patent application 2006/0147456), the application of interferon-containing chitosan-lipid particles (US patent application 2005/0266093), as well as the application of a compound that acts as a cytotoxic agent for a certain type of prostate cancer cells by an unknown mechanism (US patent application 2005/0245559). [0005] It is an object of the present invention to provide an alternative method of preventing, treating and diagnosing tumorigenesis, including carcinogenesis that does not need to involve the above disadvantages.

SUMMARY OF THE INVENTION

[0006] The invention relates to in vitro and in vivo methods of preventing, treating and diagnosing tumorigenesis, as well as a corresponding method of determining whether a compound is a suitable candidate for preventing and treating tumorigenesis. Treating tumorigenesis is understood as including at least one of inhibiting, arresting and reversing tumorigenesis. To emphasize this understanding the terms inhibit, arrest and reverse tumorigenesis are generally used in the following as well as in the appended claims.

[0007] In one aspect the present invention provides a method of preventing, inhibiting, arresting or reversing tumorigenesis in a cell. The method includes altering the formation of a complex between RUNX3, or a functional fragment thereof and one of (i) β-catenin, or a functional fragment thereof, and (ii) a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof and (iii) a combination thereof.

[0008] hi some embodiments altering the complex formation between RUNX3, or a functional fragment thereof and one or more of (i) β-catenin, or a functional fragment thereof, and (ii) a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, includes allowing the phosphorylation status of RUNX3, or a functional fragment thereof, to be altered.

[0009] In a further aspect the present invention provides a method of inducing programmed cell death (apoptosis) in a tumor cell. The method includes altering the formation of a complex between RUNX3, or a functional fragment thereof and one of (i) β-catenin, or a functional fragment thereof, and (ii) a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, in the cell and (iii) a combination thereof.

[0010] hi yet a further aspect the present invention provides a method of diagnosing the risk of tumorigenesis in a cell. The method includes assessing the formation of a complex between

RUNX3, or a functional fragment thereof and one of (i) β-catenin, or a functional fragment thereof, and (ii) a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, in the cell and (iii) a combination thereof.

[0011] In a further aspect the present invention provides a method of diagnosing the risk of developing a neoplasm in a subject. The method includes assessing the formation of a complex between RUNX3, or a functional fragment thereof and one of (i) β-catenin, or a functional fragment thereof, and (ii) a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, and (iii) a combination thereof, in the cell. [0012] Further, the present invention provides an in-vitro method of identifying a compound capable of altering the formation of a complex between RUNX3, or a functional fragment thereof and one of (i) β-catenin, or a functional fragment thereof, and (ii) a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, and (iii) a combination thereof. The method typically includes contacting the components that form said complex with each other. The method typically also includes adding a compound to the test tube suspected to modulate said complex formation. Further, the method typically includes detecting the said complex formation.

[0013] In more detail the invention can be seen to provide the following methods.

[0014] In a first aspect the present invention provides a method of preventing, inhibiting, arresting or reversing tumorigenesis in a cell. In one embodiment the method includes altering the complex formation of β-catenin, or a functional fragment thereof, and RUNX3, or a functional fragment thereof, in the cell. In a further embodiment the method includes altering the complex formation of a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, and RUNX3, or a functional fragment thereof, in the cell, hi another embodiment the method includes altering the complex formation of β-catenin, or a functional fragment thereof, a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, and RUNX3, or a functional fragment thereof, in the cell.

[0015] In a second aspect the invention provides a method of inducing apoptosis in a tumor cell. In one embodiment the method includes altering the complex formation of β-catenin, or a functional fragment thereof, and RUNX3, or a functional fragment thereof, in the cell. In a further embodiment the method includes altering the complex formation of a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, and RUNX3, or a functional fragment thereof, in the cell. In another embodiment the method includes altering the complex formation of β-catenin, or a functional fragment thereof, a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, and RUNX3, or a functional fragment thereof, in the cell.

[0016] In a third aspect the invention provides a method of diagnosing the risk of tumorigenesis in a cell. A method according to this aspect may be a method of identifying a cell having a predisposition to turn tumorigenic. The method includes assessing the complex formation of β-catenin, or a functional fragment thereof, and RUNX3, or a functional fragment thereof, in the cell. In a further embodiment the method includes assessing the complex formation of a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, and RUNX3, or a functional fragment thereof, in the cell. In another embodiment the method includes assessing the complex formation of β-catenin, or a functional fragment thereof, a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, and RUNX3, or a functional fragment thereof, in the cell. [0017] In a fourth aspect the invention provides a method of diagnosing the risk of developing a neoplasm in a subject. In one embodiment the method includes assessing the complex formation of β-catenin, or a functional fragment thereof, and RUNX3, or a functional fragment thereof. In a further embodiment the method includes assessing the complex formation of a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, and RUNX3, or a functional fragment thereof. In another embodiment the method includes assessing the complex formation of β-catenin, or a functional fragment thereof, a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, and RUNX3, or a functional fragment thereof.

[0018] In a fifth aspect the invention provides an in-vitro method of identifying a compound capable of altering the formation of a complex between a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, and RUNX3, or a functional fragment thereof. The method includes contacting the components that form the respective complex with each other. In typical embodiments the method includes adding a compound to the test tube suspected to modulate the complex formation of a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, and RUNX3, or a functional fragment thereof. The method may also include detecting the said complex formation.

[0019] In a sixth aspect the invention provides an in-vitro method of identifying a compound capable of altering the formation of a complex between β-catenin, or a functional fragment thereof, and RUNX3, or a functional fragment thereof. The method includes contacting the components that form the respective complex with each other. In typical embodiments the method includes adding a compound to the test tube suspected to modulate the complex formation between β-catenin, or a functional fragment thereof, and RUNX3, or a functional fragment thereof. The method may also include detecting the said complex formation. [0020] In a seventh aspect the invention provides an in-vitro method of identifying a compound capable of altering the formation of a complex between β-catenin, or a functional fragment thereof, a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, and RUNX3, or a functional fragment thereof. The method includes contacting the components that form the respective complex with each other. In typical embodiments the method includes adding a compound to the test tube suspected to modulate the complex formation between β-catenin, or a functional fragment thereof, a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof. The method may also include detecting the said complex formation. [0021] In an eight aspect the invention provides a method of treating cancer. The method includes the reactivation of RUNX3.

BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The invention will be better understood with reference to the detailed description when considered in conjunction with the non-limiting examples and the accompanying drawings, in which:

[0023] Fig. 1 illustrates the degradation of β-catenin in a non-stimulated healty cell (A) and translocation of β-catenin to the nucleus as well as complex formation with a T cell factor (TCF) or a lymphoid enhancer-binding factor (LEF) upon Wnt signaling (B). RUNX3 binds to the complex of β-catenin and the T cell factor (C), which may be diminished upon RUNX3 phosphorylation via AKT (D).

[0024] Figure 2 depicts the expression of Runx3 in intestinal epithelial cells and up-regulation of β-catenin/Tcf4 activity in the Runx3^~/~ intestine (A: Runx3 immunodetection in wild type and Runx3^~A jejunum; B: Runx3 immunodetection in wt and Runx3^'A proximal colon (left) and wt distal colon (right); C: Hematoxylin and eosin staining of wt and Runx3^'A 40 weeks old mice; D: Detection of proliferating cells in wt and Runx3^'A jejunum and proximal colon (adult; 40 weeks old) and immunostaining with the anti-Ki67 antibody (neonate); E: Number of BrdU (adult) and Ki67 (neonate) positive cells per crypt in wt and Runx3^'A intestine (*P<0.001); F: Relative growth rates of Runx3^+/+ (+/+1 and +/+2) and Runx3^''^~ (-/-1 and -1-2) FID cell lines in the monolayer culture; G: Immunodetection of c-Myc in wt and Runx3^'A intestines; H: Immunodetection of EphrinB and EphB2 in wt and Runx3^'f" jejunal epithelium; I: Paneth cell immunodetection in wt and Runx3^'A small intestine. J: Western analysis of protein expression profiles in wt and Runx3^''^' intestine). [0025] Figure 3 illustrates the formation of a ternary complex of β-catenin, TCF4 and RUNX3 and the attenuation of β-catenin/TCF4 transcriptional activity by RUNX3 (A: Western analysis of RUNX3 expression in 22 human colorectal cancer cell lines; B: Coimmunoprecipitation of exogenously expressed β-catenin, TCF4 and RUNX3 in HCTl 16 cells; C: Two step-coimmunoprecipitation of exogenously expressed β-catenin and TCF4 and RUNX3 in 293 T cells; D: Immunoprecipitation of an endogenous ternary complex of β-catenin/TCF4/RUNX3 in nuclear extracts of HCTl 16 and SW620 cells; E: Binding of Myc-TCF4 and/or HA-β-catenin together with His-tagged RUNX3 to Ni-NTA agarose; F: Immunoprecipitation of wild type β-catenin, Δ45 β-catenin, or S33Y β-catenin, and Flag-RUNX3; G: Reduction of TOPflash (0.1 μg DNA) activity by exogenous RUNX3 (0.1, 0.2, and 0.5 μg DNA) in DLDl cells; H: Coimmunoprecipitation of exogenous RUNX3 (Rl 78Q) and exogenous β-catenin/TCF4 in 293T cells using anti-Flag agarose; I: Reduction of TOPflash (0.1 μg DNA) activity by exogenous wild type (WT) RUNX3 and RIMX3 (Rl 78Q) (0.05 and 0.2 μg DNA) in DLDl cells; J: wild type promoter cyclin Dl activity in presence of exogenous RUNX3 (0.1 and 0.2 μg DNA) in DLDl cells; K: TOP/FOP luciferase activities to indicate sensitivities of Runx3^+/+ (+/+1 and +1+2) and Runx3^'A (-/-1 and -1-2) FID cell lines to Wnt3a stimulation.

[0026] Figure 4 illustrates an attenuation of DNA binding activity of β-catenin/ TCF4 by RUNX3 (A: effect of exogenous RUNX3 on binding of β-catenin/TCF4 to TCF binding sites; B: occupancy of β-catenin/TCF4 at TCF binding sites after RUNX3 knockdown. C: (Upper panel) ChIP assay in DLDl detecting binding of TCF4 to TCF binding sites in the presence RUNX3. (Lower panel) Western analysis of RUNX3 induction in DLDl using ponasterone A; D: Quantification by real-time PCR of ChIP assay for TCF4 (TCF4) and β-catenin (β-catenin) occupancy at TCF binding sites in cyclin Dl (open bars) and c-Myc (gray bars) promoters. E: Western blot analysis of exogenous RUNX3 in DLDl and in HCTl 16 clones expressing antisense RUNX3 DNA (AS-Cl.1 and AS-C1.2); F: Real-time PCR quantification of AXIN2, CD44, and DKKl mRNA in DLDl cells with inducible RUNX3 expression (left panel) and in HCTl 16 cells in which RUNX3 was knocked down; G: TOPflash versus FOPflash activity for DLDl (white) and DLDl expressing RUNX3 (black; panel E) and HCTl 16 (black) and HCTl 16 in which RUNX3 was knocked down (white; panel E); H: TOPflash/FOPflash activity (lower panel) after RUNX3 knockdown (expression in upper pannel) in HCTl 16, SW620, COLO320, SW480 and SW403).

[0027] Figure 5 shows adenomatous polyps in the small intestine of Runx3^+/~ or Apc^Mιn/+ BALB/c mice and progression to adenocarcinoma in Runx3^+/'Apc^Mllj/+ compound mice. A: Hematoxylin and eosin staining of tumors in the small intestine of Runx3^+/~, Apc^Mιn/+ and Runx3^+/'Apc^Mm/+ mice. Boxed regions are enlarged on right. B: Frequency of tumor formation in the small and large intestines of mice with indicated genotypes. C: Number of tumors in the small intestine of individual mice with indicated genotypes. D: Size distribution of polyps in the small intestine of mice with indicated genotypes.

[0028] Figure 6 shows adenomatous polyps in the small intestines of Runx3^+/" BALB/c mice displaying down-regulated Runx3 and up-regulated cyclin Dl and c-Myc. A: Immunodetection of Runx3, β-catenin, cyclin Dl, and c-Myc in adenomatous polyps formed in the small intestine of Runx3^+/'and Apc^Mιn/+ mice. B: Anylysis of very small adenomas formed in jejunum of the compound mice. C, D: Quantification by real-time PCR of Runx3, cyclin Dl, c-Myc, anάp21 mRNA in individual polyps in Runx3^+/~ (C) and Apc^Mm/+ (D) small intestines. E: the ChIP assay detecting the binding of β-catenin/Tcfs to a Tcf consensus site in polyps (Tl and T2) and adjacent normal tissues (Nl and N2) of Runx3^+/~ and Apc^Mιn/+ small intestines. [0029] Figure 7 depicts the down-regulation of RUNX3 expression without nuclear accumulation of β-catenin in human adenomatous polyps. A, B, C: Three patterns of β-catenin and RUNX3 expression in 35 human cases. Type A; nuclear β-catenin with RUNX3 in nuclei (A), type B; membranous β-catenin without RUNX3 expression (B), and type C; membranous β-catenin with RUNX3 expression (C) in T4, T6, and T9 in panel F, respectively. D and E: Up-regulation of cyclinD 1 and c-Myc in adenomas of type A (D) and type B (E). F: Methylation specific PCR (MSP) analysis of the RUNX3 promoter in normal human colon epithelium, DLDl and HCTl 16 cells, and representative of 15 (T1-T15) human adenomatous polyps. G: Methylation status of CpG dinucletide between -70 and -21 relative to the translation initiation site of the RUNX3 exon 1 region. The nucleotide sequence (sense strand) of MSP products from RUNX3 -positive and -negative tumors, DLDl and HCTl 16 cells were shown. The labeled C (box) depicts resistance to bisulfite treatment due to methyaltion. Unmethylated C (not highlighted) was converted to T by the bisulfite treatment. Asterisks on T indicate unmethylated C residues in RUNX3 -negative tumors.

[0030] Figure 8 is an enlargement of immunohistochemistry of adenomas shown in Fig. 7 A, Fig. 7B and Fig. 7C of for better resolution.

[0031] Figure 9 shows RUNX3 inactivation by gene silencing and protein mislocalization with concomitant accumulation of β-catenin in human colorectal cancers. A, B, C: Differential staining patterns of RUNX3 in human colorectal cancers: positive (A), negative (B), and cytoplasmic positive (C). D: Differential staining patterns of RUNX3 in human colorectal cancer cell lines: positive (HCTl 16 and SW480), negative (DLDl and RKO), and cytoplasmic positive (SW403 and CCK81).

[0032] Figure 10 shows the morphology of wt and Runx3^~'^~ epithelium of jejunum and colon stained by hematoxylin and eosin (A). Immunodetection of CD44 (B) and cyclin Dl (C) in wt and Runx3-/- intestines is also shown.

[0033] Figure HA illustrates the mapping of the RUNX3 domain that interacts with TCF4.

Fig. HB illustrates the mapping of the TCF4 domain that interacts with RUNX3. Fig. HC shows an immunoprecipitation using anti-Flag antibody in HCTl 16 cells transfected with Myc-TCFl/ Flag-RUNX3 (left), Lefl/Flag-RUNX3 (center), and Myc-TCF3/Flag-RUNX3

(right).

[0034] Figure 12A shows EMSA analysis of the binding of β-catenin/TCF4 to a TCF binding site (TOP construct) by RUNX3, using nuclear extracts of 293 T cells expressing Myc-TCF4, S33Y β-catenin, Flag-RUNX3, or the vector (mock). Figure 12B shows a corresponding EMSA analysis using Wnt3a-treated Runx3^~'^~ FID cells. *A non-specific band detected in all reactions of Fig. 12A and 12B. Figure 12C shows EMSA analysis of binding of RUNX3/PEBP2β to the RUNX binding site of the IgCa promoter by β-catenin/TCF4, using nuclear extracts from 293T cells expressing Flag-RUNX3, Myc-TCF4, S33Y β-catenin or the vector (mock) and purified PEBP2β protein. * Lower and upper RUNX3-probe and **lower and upper RUNX3/PEBP2β-probe complexes were detected.

[0035] Figure 13 A depicts genotyping of wild type epithelial cells (wt), Runx3^+/' adenomas (T1-T4) and their adjacent normal epithelial cells (N1-N4) in the small intestine of BALB/c mice (upper panel). Quantification of the wild type and knockout alleles of Runx3 in Runx3^+/~ adenomas by real-time PCR (T1-T7) is shown in the lower panel. Fig. 13B depicts genotyping of normal epithelial cells (N) and adenomas in Apc^Min/+ (T1-T4) and Runx3^+/~ (T1-T3) small intestine of BALB/c mice. Fig. 13C depicts methylation-specific PCR (MSP) of the Runx3 promoter region in wild type epithelial cells (wt), Runx3^+/~ normal epithelial cells (N), and Rumc3^+/~ adenomas (Tl -T6) in the small intestine (lower panel). The positions of CpG dinucleotides in the promoter region of Runx3 are depicted in the upper panel. [0036] Figure 14A shows the frequency of tumor formation in small and large intestines.

Figure 14B shows the number of tumors in the small intestine of individual mice. Figure 14C shows the size distribution of polyps in the small intestine of mice. Figure 14C shows stereomicroscopic images of polyps (arrowheads) formed in Apc^{mm +} and Runx3^+/'Apc^{mn +} small intestines of mice.

[0037] Figure 15 depicts the up-regulation of CD44 in adenomas of type A and B (A and B, respectively).

[0038] Figure 16 shows the morphology of small and large intestines of wild type mice reconstituted with R.UHX3^'1' (A-D) or Runx3^{+ ~} (E) bone marrow cells, one year after transplantation.

[0039] Figure 17A depicts the relative proliferation of DLDl and HCTl 16 clones over time. Figure 17B depicts the tumorigenicity of DLDl and HCTl 16 clones (*P<0.01). Figure 17C shows the tumor formation of Runx3^{+ +}and Runx3^{~ '} FID and FIL cell lines in nude mice 60 days after inoculation. Figure 17D depicts the tumor formation of control (C) and Myc-tagged dominant negative TCF4 expressing Runx3^{~ '} FID and FIL cells (indicated as -1-2 in panel C) in nude mice 60 days after inoculation.

[0040] Figure 18 depicts the expression patterns of β-catenin and RUNX3 (type A-C; cf. Fig. 7) and the methylation status of the RUNX3 promoter (M; methylated, U; unmethylated) of 35 human sporadic adenomatous polyps (T1-T35).

[0041] Figure 19 shows the expression pattern of RUNX3 (P, N, and C; cf. Fig. 9) and the methylation status of the RUNX3 promoter (M; methylated, U; unmethylated) of 48 human colorectal cancers.

[0042] Figure 20 illustrates the tumor formation in Runx3+/- mice. [0043] Figure 21 depicts the binding of RUNX3 to Aktl in vitro.

[0044] Figure 22 depicts the formation of an endogenous protein complex in HCTl 16 nuclear extract.

[0045] Figure 23 illustrates the domain mapping of the RUNX3/Akt interaction. [0046] Figure 24 illustrates the domain mapping of the RUNX3/Akt interaction. [0047] Figure 25 is a schematic showing that the kinase domain of Aktl binds to the Runt domain of RUNX3.

[0048] Figure 26 shows that RUTSDG is phosphorylated in vitro. [0049] Figure 27 shows that RUNX3 is phosphorylated by Akt. [0050] Figure 28 shows that RUNX3 is phosphorylated by Akt in DLD-I cells. [0051] Figure 29 shows that the phosphorylation of RUNX3 by Akt reduces the affinity of

RUNX3 for TCF4 (A: Copurification of TCF4 with Runt or the indicated mutant, B: Analysis of the intensity of the bands).

DETAILED DESCRIPTION OF THE INVENTION

[0052] The present invention is based on the surprising finding that RUNX3, a gastric tumor suppressor, forms a complex with β-catenin as well as a complex with a member of the TCF/LEF transcription co-factor family. Further, RUNX3 forms a ternary complex with β-catenin and a member of the TCF/LEF transcription co-factor family, and attenuates the Wnt signaling activity. The inventors have further surprisingly found that the phosphorylation state of RUNX3 is important for the formation of a ternary complex with β-catenin and a member of the TCF/LEF transcription co-factor family. They found that the phosphorylation state of RUNX3 can be altered via a member of the Akt proteins/Zprotein kinase B.

[0053] A significant fraction of human sporadic adenomas and Runx3⁺^- mouse adenomas were found to have inactivated RUNX3 without oncogenic β-catenin accumulation, indicating that RUNX3 inactivation independently induces intestinal adenomas, hi human colon cancers, RUNX3 is frequently inactivated with concomitant β-catenin accumulation, suggesting that adenomas induced by inactivation of RUNX3 progress to malignancy. Taken together, these data demonstrate that RUNX3 functions as a tumor suppressor of colorectal cancer by attenuating Wnt singnaling.

[0054] RUNX3 (runt-related transcription factor 3) is involved in neurogenesis and thymopoiesis and functions as a tumor suppressor gene in gastric cancer. Failure to express RUNX3 because of a combination of hemizygous deletion and DNA hypermethylation of the RUNX3 promoter region has been found in about 60% of primary gastric cancer specimens (Li, Q. L., et al. (2002) Cell 109, 113-124). RUNX3-R122C is a mutation located in the conserved Runt domain that was discovered in a case of gastric cancer and it abolishes the tumor suppressive activity of RUTNRG (ibid.). Subsequent studies have revealed that RUNX3 inactivation is not limited to gastric cancer, and frequent inactivation of RUNX3 due to DNA hypermethylation has been reported in various other cancers, including lung cancer, liver cancer (hepatocellular carcinoma), breast cancer, colon cancer, pancreatic cancer, bladder cancer, bile duct cancer, prostate cancer, and laryngeal cancer. Unlike many tumor suppressors, such as p53, which are inactivated mainly by deletions and mutations, RUNX3 is unique in that it is inactivated primarily by epigenetic silencing, rather than by mutations or deletions. Furthermore, RUNX3 can be reactivated and therefore considered to be a good drug target because mutations in its gene are rare. [0055] RUNX3 activity is closely associated with transforming growth factor β (TGF-β) signaling since the gastric mucosa of Runx3 knockout mice is less sensitive to TGF-β, which induces both cell cycle arrest and apoptosis. To inhibit the growth of a given cell type, TGF-β can employ diverse mechanisms, such as down-regulating c-myc and CDK-2/CDK-4 activity by modulating the functions of pl5INK4B, p21Wafl/Cipl, and p27Kipl. Any genetic or epigenetic alteration of the TGF-β pathway can thus render normal cells vulnerable to tumorigenesis.

[0056] Wnt/β-catenin signaling is an ancient and highly conserved signaling pathway involved in various physiological processes such as development, in particuar embryonic development, tissue regeneration, specification and maintenance of precursor cell and stem cell lineages or stem cell self-renewal. It is also involved in a variety of conditions such as cardiovascular disease, bone malformation, aging, diabetes, neurodegeneration including schizophrenia or Alzheimer disease, acute renal failure and polycystic kidneys, and inflammation. Abnormal Wnt/β-catenin signaling is further known to be associated with cancer. Aberrant Wnt/β-catenin signalling has also been found in ulcerative colitis, where the pathway is activated in early stages of malignant progression (van Dekken, H., et al., Acta Histochemica (2007) 109, 4, 266/272). Aberrant activation of Wnt/β-catenin signaling is for example a major driving force in colon cancer (Vogelstein, B., and Kinzler, K. W. (1998). Identification of c-MYC as a target of the APC pathway. Science 281, 1509-1512; Su, L. K., Kinzler, K. W., Vogelstein, B., Preisinger, A. C, Moser, A. R., Luongo, C, Gould, K. A., and Dove, W. F. (1992). Multiple intestinal neoplasia caused by a mutation in the murine homolog of the APC gene. Science 256, 668-670; van de Wetering, M., Sancho, E., Verweij, C, de Lau, W., Oving, L, Hurlstone, A., van der Horn, K., Batlle, E., Coudreuse, D., Haramis, A. P., et al. (2002). The beta-catenin/TCF-4 complex imposes a crypt progenitor phenotype on colorectal cancer cells. Cell 111, 241-250). More than 90 % of all colorectal cancers include an activating mutation of the Wnt/β-catenin pathway, making this cancer an attractive model for molecular intervention.

[0057] Mutations in Wnt/β-catenin pathway components including APC, Axin, and β-catenin itself are well-established causes of aberrant signaling activation leading to cancer. These genetic defects share in common that they result in the accumulation of β-catenin in the nucleus. In a non-cancerous cell the cytosolic level of β-catenin is kept low in the absence of the Wnt ligand due to phosphorylation by casein kinase 1 and glycogen synthase kinase 3. This phosphorylation occurs while β-catenin is bound to a cytoplasmic destruction complex that includes the tumor suppressor adenomatous polyposis coli (APC), the scaffold protein axin, casein kinase 1 (CKI) and glycogen synthase kinase 3β (GSK3β). Phosphorylated β-catenin is ubiquitinylated and subsequent degraded. Mutations can however lead to an accumulation of cytosolic β-catenin, thereby mimicking constitutive Wnt signaling. Mutations in components of the Wnt/β-catenin pathway, including the β-catenin gene, have been found in various cancer forms such as melanoma, esophageal cancer, thyroid cancer, adenocarcinoma of the small intestine, lung cancer, prostate cancer, liver cancer, gastric cancer, ovarian cancer, uterine cancer, hepatocellular cancer, breast cancer, hair matrix cell tumors (pilomatricomas), desmoid tumors, WiIm' s tumor (kidney), medulloblastoma (the most frequent brain tumors in childhood), synovial sarcoma and endometrial cancer (for an overview see e.g. Giles, RH, et al., Biochim Biophys Acta (2003) 1653, 1-24). Wnt signaling has also been found to play a role in tumor progression and metastasis.

[0058] Non phosphorylated and thus stabilized β-catenin is thought to translocate into the nucleus. In the nucleus β-catenin interacts with downstream effectors that are members of the TCF/LEF transcription co-factor family, e.g. LEFl (lymphoid enhancer-binding factor 1) and TCF (T-cell factor), thereby for example converting LEFl into a transcriptional activator, β-catenin does not interact with DNA itself, but serves as a cofactor of TCF/LEF transcription factors. The TCF and LEF family of transcription factors includes LEFl (LEFl), TCF-I (TCF7), TCF-3 (TCF7L1), and TCF-4 (TCF7L2). The TCF and LEF proteins bind directly to DNA through their high mobility group (HMG) domains and once bound to β-catenin transactivate their target genes. As a result downstream target genes such as Cyclin Dl and Myc are activated, which are genes associated with the regulation of cell proliferation and can thus lead to cell transformation. Further known targets of TCF and LEF include c-jun, multidrug resistance 1 (ABCBl), matrilysin (MMP7), axin 2 or surviving (BIRC5).

[0059] One example of the TCF and LEF proteins is TCF-4, which is expressed commonly in colorectal cancer cells, and has been implicated in the maintenance of undifferentiated intestinal crypt epithelial cells. Suppression of β-catenin-evoked gene transactivation of colorectal cancer cells by dominant-negative TCF-4 is known to switch off genes involved in cell proliferation and to switch on genes involved in cell differentiation. A couple of proteins such as Smads have been reported to interact with the β-catenin and TCF and LEF complexes and modulate their transcriptional activity.

[0060] It is known that biallelic inactivation of APC induces colon adenomas. The present inventors have found that biallelic inactivation of RUNX3 without nuclear/cytoplasmic accumulation of β-catenin also induces colon adenomas. The results suggest that APC and RUNX3 independently function as gatekeepers in colon adenoma development. Wnt signaling is an oncogenic pathway whereas TGF-β is a tumor suppressor pathway. The nuclear effectors of these pathways, β-catenin/TCF4 and RUNX3, respectively, form a ternary complex with diminished DNA binding ability. This ternary complex appears to integrate growth promoting and cytostatic signals for homeostatic balance of growth and differentiation in intestinal epithelial cells.

[0061] Multiple genetic changes have been described in colon carcinogenesis and may result from the triggering of events during the adenoma-carcinoma transition by Wnt signaling activation. In familial adenomatous polyposis coli (FAP), family members have heterozygous mutations in the APC gene, which is a key component of the Wnt pathway that destablizes β-catenin, a nuclear effector of canonical Wnt signaling. Adenomas caused by biallelic inactivation of APC show nuclear and cytoplasmic accumulation of β-catenin. It has been widely surmised that oncogenic Wnt signal initiates colorectal carcinogenesis (Kinzler and Vogelstein, 1996). However, studies of sporadic cases of colorectal cancer and analyses of aberrant crypt foci (ACF) - proposed precursors of adenomas - have suggested that genetic or epigenetic alterations other than inactivation of APC or activation of β-catenin maybe involved (Jass et al., 2002). These alterations could implicate entirely new pathways or alternative mechanisms for the activation of Wnt signaling.

[0062] Wnt and TGF-β superfamily signaling are key pathways that ultimately influence the cell division and cell fate of gut epithelial cells. These pathways are known to be altered in gastrointestinal cancers. In colorectal cancers with stabilized β-catenin, the β-catenin/T cell factor-4 (TCF4) transcription factor complex is constitutively activated. Several components of the TGF-β signaling cascade are bona fide tumor suppressors that inhibit cell growth and cancer development. Inactivation of one of these components, such as the TGF-β receptor type II or Smad4, occurs frequently in gastrointestinal tumors. However, the molecular mechanisms that link the oncogenic Wnt and the tumor suppressive TGF-β pathways in intestinal carcinogenesis have not been fully elucidated.

[0063] RUNX3, a strong gastric tumor suppressor candidate, is inactivated by gene silencing or protein mislocalization in more than 80% of gastric cancers (Li et al., 2002; Ito et al., 2005). More recently, inactivation of RUNX3 was reported in a wide range of other cancer types (Blyth et al., 2005). The RUNXS locus at Ip36, a region that undergoes frequent allelic loss in gastrointestinal cancers, is silenced by hypermethylation of its promoter region in a significant proportion of cancer-derived cell lines and clinical specimens, suggesting that it fulfills a tumor suppressive function in colorectal cancers (Goel et al., 2004; Ku et al., 2004). RUNX3 regulates target gene expression by forming a complex with Smad molecules. The inventors reported earlier that Runx3^" gastric epithelial cells are resistant to the growth-inhibitory and apoptosis-inducing properties of TGF-β, suggesting that RUNX3 is a downstream effector of the TGF-β family signaling pathway. Furthermore, TGF-β regulates nuclear translocation of RUNX3 in gastric epithelial cells (Ito et al., 2005) and activates the transcription of p21 ^ipl and Bim, negative cell cycle regulator and a proapoptotic genes, respectively, in cooperation with RUNX3 and Smads (Chi et al., 2005; Yano et al., 2006; reviewed in Ito, 2008). [0064] In the examples below evidence is provided that RUNX3 directly interacts with

TCF4/β-catenin complex and attenuates Wnt signaling. Since Rnx3 undergoes interactions with both β-catenin and a member of the TCF/LEF transcription co-factor family, complex formation with one of these two binding partners is apparently sufficient to affect Wnt signaling. Phenotypic analysis of Runx3 -deficient mice as well as human specimens suggest that at an early stage of carcinogenesis, in particular colon carcinogenesis, biallelic inactivation of RUNX3, primarily by promoter hypermethylation, induces cancer formation, in particular human colon adenomas, independent of alterations of APC or β-catenin.

[0065] Without being bound by theory the inventors' findings indicate that in a non-cancerous cell there is a balance between two mutually exclusive tumor suppressing effects of RUNX3 activity. The first of these activities is the function of RUTSfX3 as a transcription factor, where it mediates TGF-β-induced growth inhibition and apoptosis. The second activity is the derogation of the stimulation of gene expression via β-catenin in Wnt signalling by the formation of a complex. One application of a method of the invention is restoring a respective balance in a cell where the balance has been interrupted. A further application of a method of the invention is establishing a respective balance in a cell that is carcinogenic or at risk to turn carcinogenic. Yet a further application of a method of the invention is disrupting this balance in favor of one of the two above mutually exclusive tumor suppressing effects of RUNX3 activity. Activation of stimulation of this one of the two above mutually exclusive tumor suppressing effects may be in need in a cell due to e.g. a cellular defect such as a mutation or another dysfunction.

[0066] The inventors' findings further indicate not only that RUNX3 forms a complex, or interacts, with a TCF/LEF transcription factor (which is a nuclear effector of Wnt signaling pathway) but also point to a new link to the Phosphatidylinositol 3-kinase (PI3K)/Akt pathway. This signaling pathway is known to be vital to the growth and survival of cancer cells, and thought to play an important role in tumorigenesis. Activating mutations of the pl lOalpha subunit of PI3K (PIK3CA, with CA standing for "constitutively active") have been identified in a broad spectrum of tumors. The PIK3CA mutation has for example been associated with poor prognosis in colorectal cancer. Such constitutively active mutants of PIK3 activate AKT signaling. 3-Phosphoinositide-dependent kinase 1 (PDKl) is the first node of the PI3K signal output and is required for activation of AKT. It catalyses phosphorylation of phosphatidylino- sitol-4,5-bisphosphate to phosphatidylinositol 3,4,5-trisphosphate (PIP3). PIP3 recruits the serine/threonine protein kinase AKT to the cell membrane, resulting in the phosphorylating of Akt at threonine-308, thereby activating AKT. Accordingly, the activity of Akt is effectively reduced by the action of phosphatases that dephosphorylate PIP3. Akt has been found to be hyperactivated in many tumors, and known to play a major role in cell survival and in resistance to tumor therapy, even though Akt is rarely mutated itself. So far various mechanisms of action have been suspected as Akt's role in tumorigenesis, such as stabilizing Myc and cyclin Dl or by inducing degradation of the cyclin-dependent kinase (Cdk) inhibitor p27^KiP¹, inactivation of pro-apoptotic molecules such as caspase-9 and the BH3-only protein Bad, by triggering the activity of the transcription factor NF-κB or via Foxo transcription factors or GSK3. The present inventors' findings point to a phosphorylation of RUNX3 by AKT, which reduces the affinity of RUNX3 to TCF/LEF transcription factors. Accordingly, the formation of a complex between β-catenin, or a functional fragment thereof, a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, and RUNX3, or a functional fragment thereof, can be modulated by modulating the activation status (via phosphorylation at threonine-308) of Akt.

[0067] The findings by the present inventors can be put into practical use for the prevention, treatment and diagnosis of tumorigenesis, including carcinogenesis. A method according to the invention can in some embodiments be termed a method of preventing, inhibiting, arresting or reversing tumorigenesis in a cell. In some embodiments a respective method can be termed a method of inducing apoptosis in a tumor cell. In yet further embodiments a method according to the invention can be termed a method of diagnosing the risk of tumorigenesis in a cell. Such an embodiment of diagnosis may also be termed a method of diagnosing the risk of developing a neoplasm in a subject. These methods include altering and/or assessing the formation of a complex between RUNX3, or a functional fragment thereof, and one or both of (i) a member of the TCF/LEF transcription co-factor family or a functional fragment thereof, and (ii) β-catenin, or a functional fragment thereof. Typically the assessment of such complex formation serves diagnostic purposes, whereas altering such complex formation serves treatment of a rumor, including a cancer or ulcerative colitis. Using a method of the invention, tumorigenesis may be prevented, inhibited, arrested or reversed - as well as diagnosed or predicted — in any organism, including for instance a mammal, a fish, an amphibian, a bird or a microorganism. A respective microorganism is in some embodiments a cell.

[0068] The present invention also relates to compounds that are able to achieve the modulation of the complex formation as described above. A respective compound may for instance be a nucleic acid molecule, an immunoglobulin, an antagonist or agonist of a cell surface receptor, a compounds that modulates the degree of phosphorylation of one of the components of the above complex, as well as compounds that modulate the intracellular quantity of one or more of the components of the above complex. In this regard, the invention also relates to the use of such compounds for the diagnosis of tumorigenesis.

[0069] Based on assessing the presence of a complex as defined above, the invention also provides a method of identifying a compound that is capable of altering the formation of a respective complex. Such a method may be a method of identifying a candidate compound capable of preventing, inhibiting, arresting or reversing tumorigenesis in a cell and/or of inducing apoptosis in a tumor cell. An alteration such as an enhancement or a reduction of the formation of a complex between Runx3, or a functional fragment thereof, and one or both of (i) a member of the TCF/LEF transcription co-factor family or a functional fragment thereof, and (ii) β-catenin, or a functional fragment thereof indicates that the compound is capable of preventing tumorigenesis in a cell and/or of inducing apoptosis in a tumor cell

[0070] The cell on which a method according to the invention is used may be any cell. The cell may for example be a cell of a tissue. A respective tissue may be any tissue, for example a tissue obtainable or obtained from an organism, such as an animal, e.g. a mammalian species, including a rodent species, an amphibian, e.g. of the subclass Lissamphibia that includes e.g. frogs, toads, salamanders or newts, an invertebrate species, or a plant. Examples of mammals include, but are not limited to, a rat, a mouse, a rabbit, a guinea pig, a squirrel, a hamster, a vole, a hedgehog, a platypus, an American pika, a galago ("bushbaby"), an armadillo, a dog, a lemur, a goat, a pig, a cattle (cow), an opossum, a horse, an elephant, a bat, a woodchuck, an orang-utan, a rhesus monkey, a woolly monkey, a macaque, a chimpanzee, an orang-utan, a tamarin (saguinus oedipus), a marmoset or a human. An illustrative example of a tissue is an organ or a portion thereof, such as adrenal, bone, bladder, brain, skin, cartilage, colon, eye, heart, kidney, liver, lung, muscle, nerve, ovary, spleen, adrenal, liver, lung, pancreas, bladder, prostate, skin, small intestine, spleen, stomach, testicular, thymus, tumor, vascular or uterus tissue, or connective tissue. In some embodiments the cell is obtained or derived from a host organism, which may be any organism. The cell may be directly taken from a respective host organism in form of a sample such as e.g. a biopsy or a blood sample. It may also have been derived from a host organism and subsequently been cultured, grown, transformed or exposed to a selected treatment. In some embodiments the cell may be included in a host organism. It may for instance be present in the blood or in tissue, including in an organ, of the host organism. The host organism from which the cell is derived or obtained, or in which it is included, may be any organism such as a microorganism, an animal, such as a fish, an amphibian, a reptile, a bird, a mammal, including a rodent species, an invertebrate species, e.g. of the subclass Lissamphibia that includes e.g. frogs, toads, salamanders or newts, or a plant.

[0071] The cell may for example be an (e.g. isolated) individual cell or a cell of a cell population. In some embodiments the cell is a somatic cell. Examples of suitable somatic cells, include, but are not limited to a fibroblast, a myeloid cell, a B lymphocyte, a T lymphocyte, a bone cell, a bone marrow cell, a pericyte, a dendritic cell, a keratinocyte, an adipose cell, a mesenchymal cell, an epithelial cell, an epidermal cell, an endothelial cell, a chondrocyte, a cumulus cell, a neural cell, a glial cell, an astrocyte, a cardiac cell, an oesophageal cell, a muscle cell (e.g. a smooth muscle cell or a skeletal muscle cell), a pancreatic beta cell, a melanocyte, a hematopoietic cell, a myocyte, a macrophage, a monocyte, and a mononuclear cell. A somatic cell may be a cell of any tissue, such as the examples above. In some embodiments the cell is a tumor cell, e.g. a cancer cell. A respective tumor cell may also be obtained from an organism, e.g. from a mammal. In other embodiments the tumor cell may be included in a mammal, such as for example a rat, a cow, a pig, and a human. A respective tumor cell may also be cultured and/or be a cell of a cell culture. It may for instance be a cell of a cell line such as a melanoma cell line, e.g. A375, B16 (including B16-F10), BNl, K1735-M2, M14, OCM-I or WM793, colorectal cancer cell line, e.g. SW480, HT29, RKO, LST-Rl, Caco-2, WiDr, GP2d, HCTl 16, LoVo, LS174T, VACO5 HCA7, LS411, C70, LIM1863, SL-174T, SW1417, SW403, SW620, SW837 or VACO4A, a hepatoma cell line, e.g. FHCC-98, H4IIE Hep G2, Hep G2f, Huh-7, PLHC-I, SMMC-7721, SK-Hepl or QGY, a lung cancer cell line, e.g. A549, ABC-I, EBC-I, LC-l/sq,LCD, LCOK, LK-2, Lul35, MS-I, NCI-H69, NCI H157, NCI-N231, NL9980, PCl, PC3, PC7, PC9, PClO, PC14, QG56, RERF-LCMS, RERF-LCAI, RERF-LCKJ, SBC3 or SQ5, oesophageal cancer cell lines A549, EC 109, EC9706 or HKESC-4, a gastric cancer cell line, e,g. BGC823, KATO-III, MGC8O3, MKN-45, SGC7901 or an ovarian cancer cell line, e.g. A2780, C13*, CAOV3, DOV-13, HO8910 (including HO-8910PM), OvCA 3, OvCA 420, OvCA 429, OvCA 432, OvCA 433, OvCar 3, OvCar 5, OvCA 420, OVHM or SKOV-3.

[0072] The cell may in some embodiments be a cell of an organism, which may harbor cancerous tissue, a cell of a tissue, including a cancerous tissue. A cancer cell may for instance be a neuronal, glial, lung, liver, brain, breast, bladder, blood, leukemic, colon, endometrial, stomach, skin, ovarian, fat, bone, cervical, esophageal, pancreatic, prostate, kidney, or thyroid cell. In some aspects a cancer includes, but is not limited to astrocytoma, acute myelogenous leukemia, breast carcinoma, bladder carcinoma, cervical carcinoma, colorectal carcinoma, endometrial carcinoma, esophageal squamous cell carcinoma, glioma, glioblastoma, gastric carcinoma, hepatocellular carcinoma, Hodgkin lymphoma, leukemia, lipoma, melanoma, mantle cell lymphoma, myxofibrosarcoma, multiple myeloma, neuroblastoma, non-Hodgkin lymphoma, lung carcinoma, non-small cell lung carcinoma, ovarian carcinoma, esophageal carcinoma, osteosarcoma, pancreatic carcinoma, prostate carcinoma, squamous cell carcinoma of the head and neck, thyroid carcinoma and urothelial carcinoma. [0073] A cell used in a method of the present invention is typically capable of expressing the protein Runx3, or a functional fragment thereof, in that it includes a nucleic acid sequence encoding Runx3, generally in the form of a functional gene of RUNX3 (whether endogenous or exogenous). In some embodiments the cell expresses Runx3. In some embodiments a respective, for instance endogenous, gene encoding Runx3 is functionally active and expressing Runx3. hi some embodiments an endogenous nucleic acid sequence encoding Runx3 is functionally inactive. In some of these embodiments Runx3 is nevertheless expressed - generally from an exogenous RUNX3 gene. An exogenous gene encoding Runx3 may be introduced by means of recombinant technology, for instance by means of a vector carrying a RUNX3 gene. It may in this regard be advantageous to further use a vector that contains a promoter effective to initiate transcription in the respective host cell (whether of endogenous or exogenous origin).

[0074] The term "vector" relates to a single or double-stranded circular nucleic acid molecule that can be transfected into cells and replicated within or independently of a cell genome. A circular double-stranded nucleic acid molecule can be cut and thereby linearized upon treatment with restriction enzymes. An assortment of nucleic acid vectors, restriction enzymes, and the knowledge of the nucleotide sequences cut by restriction enzymes are readily available to those skilled in the art. A nucleic acid molecule encoding Runx3 can be inserted into a vector by cutting the vector with restriction enzymes and ligating the two pieces together.

[0075] The term "Runx3" refers to any member, including variant thereof, of the Runt-related transcription factor 3 proteins, which are also termed core-binding factor subunit alρha-3, acute myeloid leukemia 2 protein, oncogene AML-2, acute myeloid leukemia 2 protein, Oncogene AML-2, polyomavirus enhancer-binding protein 2 alpha C subunit, polyomavirus enhancer-binding protein 2 alpha C subunit, SL3-3 enhancer factor 1 alpha C subunit, and SL3/AKV core-binding factor alpha C subunit. Examples include, but are not limited to, the mouse protein with the UniProtKB/ TrEMBL accession No. Q921B7, the human protein with the UniProtKB/TrEMBL accession No. Ql 3761, the rat protein with the UniProtKB/TrEMBL accession No. Q91ZK1, the chicken protein with the UniProtKB/TrEMBL accession No. A8QJ84, the protein of the Clearnose skate with the UniProtKB/TrEMBL accession No. Q6SZR4, the protein of the Mongolian gerbil with the UniProtKB/ TrEMBL accession No. Q2MHJ6, the zebrafish protein with the UniProtKB/ TrEMBL accession No. Q9DEA0, the yellowfever mosquito protein with the UniProtKB/TrEMBL accession No. B6S2Q4, the chimpanzee protein encoded by the nucleotide sequence with the EMBL accession No. AY406594 and the protein of the smaller spotted catshark encoded by the nucleotide sequence with the EMBL accession No. DQ990014.

[0076] In some embodiments a cell used in a method of the present invention is capable of expressing the protein β-catenin, also termed CTNNB, or a functional fragment thereof. In some embodiments the cell expresses β-catenin. In some embodiments a gene encoding β-catenin, which may be an endogenous gene, is functionally active and expressing β-catenin. In some embodiments an endogenous nucleic acid sequence encoding β-catenin is functionally inactive, β-catenin may also be expressed from an exogenous β-catenin gene, which may be introduced by means of recombinant technology, e.g. using a vector carrying a β-catenin gene (see also above for Runx3).

[0077] In some embodiments a cell used in a method of the present invention is capable of expressing a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof. In some embodiments the member of the TCF/LEF transcription co-factor family is expressed by the cell. In some embodiments a gene encoding the member of the TCF/LEF transcription co-factor family, which may be an endogenous gene, is functionally active in the cell, thus expressing the member of the TCF/LEF transcription co-factor family. In some embodiments an endogenous nucleic acid sequence encoding the protein is functionally inactive. In some embodiments the member of the TCF/LEF transcription co-factor family is expressed from an exogenous gene encoding the same, which may be introduced by means of recombinant technology, e.g. using a vector carrying a gene of the member of the TCF/LEF transcription co-factor family (cf. also above).

[0078] Members of the Tcf/Lef family are high mobility group (HMG) box transcription factors. The member of the TCF/LEF transcription co-factor family may for instance be Lymphoid enhancer-binding factor 1, abbreviated LEFl or T-cell factor 1, abbreviated TCF-I, also called T-cell-specific transcription factor 1 or Transcription factor 7. It may for instance also be HMG box transcription factor 3 (or simply transcription factor 3), abbreviated TCF-3, or T-cell transcription factor-4 (or simply transcription factor 4), abbreviated TCF-4, which has also been named immunoglobulin transcription factor 2 (ITF-2), SL3-3 enhancer factor 2 (SEF-2) or class A helix-loop-helix transcription factor ME2. [0079] LEFl, Lymphocyte enhancer binding factor 1, may for instance, without being limited thereto, be the mouse protein with the UniProtKB/TrEMBL accession No. Q8BGZ9, the human protein with the UniProtKB/TrEMBL accession No. Q3ZCU4, the zebrafish protein with the UniProtKB/ TrEMBL accession No Q9W7C0, the dog protein with the UniProtKB/TrEMBL accession No. B6VCV6, the rat protein with the UniProtKB/TrEMBL accession No Q9QXN1 or an isoform or variant of such a protein.

[0080] Examples of TCF-I include, but are not limited to, the human protein with the UniProtKB/TrEMBL accession No. P36402, the mouse protein with the UniProtKB/ TrEMBL accession No Q00417, the chicken protein with the UniProtKB/TrEMBL accession No. Q8JHX2, the zebrafish protein with the UniProtKB/ TrEMBL accession No Q9PU63, the protein of the western clawed frog Xenopus tropicalis_with the UniProtKB/TrEMBL accession No. Q7T265 or an isoform or variant of such a protein. Examples of TCF-3 include, but are not limited to, the human protein with the UniProtKB/TrEMBL accession No. Q9HCS4, the protein of the western clawed frog Xenopus tropicalis with the UniProtKB/TrEMBL accession No. Q6YJU5, the chicken protein with the UniProtKB/TrEMBL accession No. Q8 JHX3, the mouse protein with the UniProtKB/TrEMBL accession No Q9Z1 Jl, or an isoform or variant of such a protein. Examples of TCF-4 include, but are not limited to, the human protein with the UniProtKB/TrEMBL accession No. Q9NQB0, the mouse protein with the UniProtKB/TrEMBL accession No. Q56R92, the protein of Xenopus laevis (African clawed frog) with the UniProtKB/TrEMBL accession No. B6C964, or an isoform or variant of such a protein. [0081] In some embodiments a cell used in a method of the present invention is capable of expressing a member of the Akt family, such as Aktl, Akt2 or Akt3. Examples of Aktl include, but are not limited to, the human protein with the UniProtKB/TrEMBL accession No. P31749, the protein of Caenorhabditis elegans with the UniProtKB/TrEMBL accession No Q 17941, the mouse protein with the UniProtKB/TrEMBL accession No A4FUQ9, the rat protein with the UniProtKB/TrEMBL accession No P47196, the protein of the African clawed frog (Xenopus laevis) with the UniProtKB/TrEMBL accession No. Q98TY9, the protein of the fruit fly (Drosophila melanogaster) with the UniProtKB/TrEMBL accession No Q8INB9, or an isoform or variant of such a protein. Examples of Akt2 include, but are not limited to, the human protein with the UniProtKB/TrEMBL accession No. P31751, the zebrafish protein with the UniProtKB/TrEMBL accession No. Q8UUX0, the mouse protein with the UniProtKB/TrEMBL accession No Q60823, the chicken protein with the UniProtKB/TrEMBL accession No Q9PUJ3, the rat protein with the UniProtKB/TrEMBL accession No P47197, or an isoform or variant of such a protein. Examples of Akt3 include, but are not limited to, the human protein with the UniProtKB/TrEMBL accession No. Q9Y243, the mouse protein with the UniProtKB/ TrEMBL accession No Q9WUA6, the rat protein with the UniProtKB/TrEMBL accession No. Q63484, the protein encoded by the Chimpanzee nucleic acid sequence with the UniProtKB/TrEMBL accession No AY399352, the protein encoded by the rhesus monkey (Macaca mulatta) nucleic acid sequence with the UniProtKB/TrEMBL accession No BVl 65988 or an isoform or variant of such a protein.

[0082] The term "nucleic acid" as used herein refers to any nucleic acid molecule in any possible configuration, such as single stranded, double stranded or a combination thereof. Nucleic acids include for instance DNA molecules, RNA molecules, analogues of the DNA or RNA generated using nucleotide analogues or using nucleic acid chemistry, locked nucleic acid molecules (LNA), protein nucleic acids molecules (PNA) and tecto-RNA molecules (e.g. Liu, B., et al., J. Am. Chem. Soc. (2004) 126, 4076-4077). A PNA molecule is a nucleic acid molecule in which the backbone is a pseudopeptide rather than a sugar. Accordingly, PNA generally has a charge neutral backbone, in contrast to for example DNA or RNA. Nevertheless, PNA is capable of hybridising at least complementary and substantially complementary nucleic acid strands, just as e.g. DNA or RNA (to which PNA is considered a structural mimic). An LNA molecule has a modified RNA backbone with a methylene bridge between C4' and OT, which locks the furanose ring in a N-type configuration, providing the respective molecule with a higher duplex stability and nuclease resistance. Unlike a PNA molecule an LNA molecule has a charged backbone. DNA or RNA may be of genomic or synthetic origin and may be single or double stranded. Such nucleic acid can be e.g. mRNA, cRNA, synthetic RNA, genomic DNA, cDNA, synthetic DNA, a copolymer of DNA and RNA, oligonucleotides, etc. A respective nucleic acid may furthermore contain non-natural nucleotide analogues and/or be linked to an affinity tag or a label.

[0083] Many nucleotide analogues are known and can be used in the method of the invention. A nucleotide analogue is a nucleotide containing a modification at for instance the base, sugar, or phosphate moieties. As an illustrative example, a substitution of 2'-OH residues of siRNA with 2'F, 2'0-Me or 2'H residues is known to improve the in vivo stability of the respective RNA. Modifications at the base moiety include natural and synthetic modifications of A, C, G, and TVU, different purine or pyrimidine bases, such as uracil-5-yl, hypoxanthin-9-yl, and 2-aminoadenin-9-yl, as well as non-purine or non-pyrimidine nucleotide bases. Other nucleotide analogues serve as universal bases. Universal bases include 3-nitropyrrole and 5-nitroindole. Universal bases are able to form a base pair with any other base. Base modifications often can be combined with for example a sugar modification, such as for instance 2'-O-methoxyethyl, e.g. to achieve unique properties such as increased duplex stability.

[0084] Preventing, inhibiting, arresting or reversing tumorigenesis as well as inducing apoptosis in a tumor cell by modulating the formation of aforementioned complex can be performed in various ways. Generally this modulation can occur on the level of transcription, on the level of protein turnover, on the functional level by changing the activation state of the respective components of the complex or by a combination of any of these levels of action. A modulation on the level of transcription alters the amount of the respective protein present in the cell and thus available for the complex formation. An increased expression of a respective protein (e.g. Runx3 or β-catenin) may be established by stimulating the expression of a corresponding endogenous protein in the cell. Accordingly, transcription and translation of a respective endogenous gene of the cell encoding the respective DACT protein may be stimulated or a state of inhibition thereof may be reduced or terminated. In terms of the activation status, the ability of RUNX3 to form a complex with β-catenin and/or a member of the TCF/LEF transcription co-factor family may be altered, e.g. increased or reduced, by changing the phosphorylation status of RUNX3, for example at a serine residue, a threonine residue or a tyrosine residue.

[0085] The same result as altering the expression of RUNX3, i.e. an altered, e.g. reduced or increased amount thereof in the cell, is achieved by altering the protein turnover, e.g. by a reduced or increased degradation. As an illustrative example, increasing the amount of RUNX3 in a cell can lead to an increased complex formation between RUNX3 and β-catenin and/or a member of the TCF/LEF family. Thereby transcription of target genes of the TCF/LEF protein may be attenuated. As a result tumorigenesis may be arrested, prevented or reversed. A modulation of the said complex formation on the functional level may include alterations of the components of the complex or a direct interference with the formation of the complex. One embodiment for achieving such and other modulations with consequent effects on the said complex formation includes administering a compound.

[0086] In some embodiments the compound may be a modulator of a member of the family of PI3 kinase enzymes, in particular a class IA PI3 kinase, or a modulator of a lipid phosphatase that hydrolyses phosphatidylinositol 3,4,5-trisphosphate, e.g. to phosphatidylino- sitol-4,5-bisphosphate, such as PTEN (phosphatase and tensin homologue deleted in chromosome 10). A phosphatase such as PTEN counteracts PBK-dependent Akt activation. Two illustrative examples of an inhibitor of PI3 kinase are wortmannin and the compound LY294002. PB kinase is a lipid kinase that phosphorylates phosphatidylinositol 4,5-bis- phosphate to phosphatidylinositol 3,4,5-trisphosphate. Elevated levels of phosphatidylinositol 3,4,5-trisphosphate in a cell are known to activate the serine/threonin kinase Akt (also termed protein kinase B), which translocates to the cytoplasm and to the nucleus (supra). In some embodiments the compound is a general Cyclooxygenase-inhibitor such as Aspirin® or a Cyclooxygenase-2 inhibitor such as NS 398. Uddin S, et al. have recently provided data that suggest that inhibition of Cyclooxygenase-2 results in dephosphorylation and inactivation of Akt (M J Cancer, 2009, JuI 20, epub, "Cyclooxygenase-2 inhibition inhibits PI3K/AKT kinase activity in epithelial ovarian cancer").

[0087] The compound used to modulate the said complex formation can be of any nature. It may for instance be a nucleic acid (see above), a peptide, a peptoid, an inorganic molecule and a small organic molecule. Peptoids can have a much higher cell permeability than peptides (see e.g. Kwon, Y.-U., and Kodadek, T., J. Am. Chem. Soc. (2007) 129, 1508-1509). A peptide may be of synthetic origin or isolated from a natural source by methods well-known in the art. The natural source may be mammalian, such as human, blood, semen, or tissue. A peptide, including a polypeptide may for instance be synthesized using an automated polypeptide synthesizer. Illustrative examples of polypeptides are an antibody, a fragment thereof and a proteinaceous binding molecule with antibody-like functions. Examples of (recombinant) antibody fragments are Fab fragments, Fv fragments, single-chain Fv fragments (scFv), diabodies or domain antibodies (Holt, L. J., et al., (2003) Trends Biotechnol, 21, 11, 484-490). An example of a proteinaceous binding molecule with antibody-like functions is a mutein based on a polypeptide of the lipocalin family (WO 03/029462, Beste et al., (1999) Proc. Natl. Acad. Sd. U.S.A., 96, 1898-1903). Lipocalins, such as the bilin binding protein, the human neutrophil gelatinase- associated lipocalin, human Apolipoprotein D or glycodelin, posses natural ligand-binding sites that can be modified so that they bind to selected small protein regions known as haptens. Examples of other proteinaceous binding molecules are the so-called glubodies (see WO 96/23879), proteins based on the ankyrin scaffold (Mosavi, L.K., et al., (2004) Protein Science 13, 6, 1435-1448) or crystalline scaffold (WO 01/04144) the proteins described in Skerra, (2000) J. MoI. Recognit. 13, 167-187, AdNectins, tetranectins, and avimers. Avimers contain so called A-domains that occur as strings of multiple domains in several cell surface receptors (Silverman, J., et al., (2005) Nature Biotechnology 23, 1556-1561). Adnectins, derived from a domain of human fibronectin, contain three loops that can be engineered for immuno globulin- like binding to targets (Gill, D. S. & Damle, N.K., (2006) Current Opinion in Biotechnology 17, 653-658). Tetranectins, derived from the respective human homotrimeric protein, likewise contain loop regions in a C-type lectin domain that can be engineered for desired binding (ibid.). Peptoids, which can act as protein ligands, are oligo(N-alkyl) glycines that differ from peptides in that the side chain is connected to the amide nitrogen rather than the α carbon atom. Peptoids are typically resistant to proteases and other modifying enzymes and can have a much higher cell permeability than peptides (see e.g. Kwon, Y.-U., and Kodadek,T., (2007) J. Am. Chem. Soc. 129, 1508- 1509). Where desired, a modifying agent may be used that further increases the affinity of the respective moiety for any or a certain form, class etc. of target matter. [0088] The compound may for instance be isolated from a biological or non-biological source or chemically or biotechnologically produced. Examples for such compounds are, without being limited to, small organic molecules or bioactive polymers, such as polypeptides, for instance immunoglobulins or binding proteins with immunoglobulin-like functions, or oligonucleotides. One embodiment of such a compound is a nucleic acid molecule, in particular an RNA or DNA molecule, whereof in particular a non-coding nucleic acid molecule, such as for example an aptamer or a Spiegelmer® (described in WO 01/92655). A non-coding nucleic acid molecule may also be an nc-RNA molecule (see e.g. Costa, FF, Gene (2005), 357, 83-94 for an introduction on natural nc-RNA molecules). Examples of nc-RNA molecules include, but are not limited to, an anti-sense-RNA molecule, an L-RNA Spiegelmer®, a silencer-RNA molecule (such as the double-stranded Neuron Restrictive Silencer Element), a micro RNA (miRNA) molecule, a short hairpin RNA (shRNA) molecule, a small interfering RNA (siRNA) molecule, a repeat-associated small interfering RNA (rasiRNA) molecule or an RNA that interacts with Piwi proteins (piRNA). [0089] The use of small interfering RNAs has become a tool to "knock down" specific genes. It makes use of gene silencing or gene suppression through RNA interference (RNAi), which occurs at the posttranscriptional level and involves mRNA degradation. RNA interference represents a cellular mechanism that protects the genome. SiRNA molecules mediate the degradation of their complementary RNA by association of the siRNA with a multiple enzyme complex to form what is called the RNA-induced silencing Complex (RISC). The siRNA becomes part of RISC and is targeted to the complementary RNA species which is then cleaved. This leads to the loss of expression of the respective gene (for a brief overview see Zamore, PD, & Haley, B (2005) Science 309, 1519-1524). This technique has for example been applied to silencing parasitic DNA sequences, such as the cleavage of HIV RNA, as disclosed in US patent application 2005/0191618.

[0090] A typical embodiment of such a siRNA for the current invention includes an in vitro or in vivo synthesized molecule of 10 to 35 nucleotides, in some embodiments 15 to 25 nucleotides. A respective si-RNA molecule maybe directly synthesized within a cell of interest (including a cell that is part of a microorganism and an animal). It may also be introduced into a respective cell and/ or delivered thereto. An illustrative example of delivering a siRNA molecule into selected cells in vivo is its non-covalent binding to a fusion protein of a heavy-chain antibody fragment (Fab) and the nucleic acid binding protein protamin (Song, E. et al. (2005), Nature Biotech. 23 , 6, 709-717). In an embodiment of the present invention siRNA molecules are used to induce a degradation of mRNA molecules encoding one or more components of the complex the formation of which is to be modulated.

[0091] Another example of a compound used to modulate the said complex formation is a molecule that is able to change the phosphorylation status of cellular components, in particular proteins. Examples of compounds that are known to affect the phosphorylation status of proteins are broad-spectrum kinase inhibitors, serine/threonine kinase inhibitors, tyrosine kinase inhibitors, tyrosine phosphorylation stimulators or tyrosine phosphatase inhibitors. As an illustrative example a protein kinase inhibitor (see also below) or protein kinase activator in form of a synthetic small organic compound maybe used for this purpose. Recent overviews on protein kinase inhibitors have for instance been given by Dancey & Sausville (Nature Reviews Druz Discovery 2, 4, 296-313 (2007)), Thaimattam et al. (Current Pharmaceutical Desisn 13, 2751-2765 (2007)) and Liao (J. Med. Chem 50, 3, 409-424 (2007)). In some embodiments a respective compound is capable of altering the degree of phosphorylation of RUNX3 or a functional fragment thereof. In some embodiments a respective compound is capable of altering the degree of phosphorylation of Akt/protein kinase B. Illustrative examples of an inhibitors of Akt are the low molecular weight organic compounds A-443654, KP372-1, VQD-002 or phosphatidylinositol) analogs. As an illustrative example, phosphorylation of Akt may lead to activation of Akt, thereby causing phosphorylation of RUNX3. As a result, the formation of a complex between RUNX3 , or a functional fragment thereof and at least one of (i) β-catenin, or a functional fragment thereof, and (ii) a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, maybe attenuated.

[0092] An illustrative selection of a compound that is able to change the phosphorylation status of cellular components is a modulator of the degree of tyrosine phosphorylation, of serine phosphorylation or of threonine phosphorylation of cellular proteins. This selection is based on the inventive finding that a change of the phosphorylation status of tyrosine residues in the cell has an effect on the efficiency of the complex formation of RUNX3 with β-catenin and/or a member of the TCF/LEF transcription factor family. The use of a compound that changes the phosphorylation status of threonine, serine or tyrosine residues in the cell is therefore also an embodiment of a method of altering the complex formation between RUNX3 , β-catenin and/or a member of the TCF/LEF transcription factor family.

[0093] Of the above mentioned compound groups a suitable compound identified and used in the present invention may be selected from tyrosine kinase inhibitors, a large number of which are commercially available such as tyrphostins, quinazolines, quinoxalines, quinolines, 2-phenylaminopyrimidines, flavonoids, benzoquinoids, aminosalicylates or stilbenes (which are described in e.g. WO 9618738, WO 03035621 and references cited therein, for an example of their experimental identification see e.g. US 6,740,665). Examples of tyrphostins are AG213, AG490, AG 879, AG 1295, AG 1478, AG 1517, AGL 2043, tyrphostin 46 and methyl 2,5-dihydroxycinnamate. Quinazolines are for instance PD153035, PD 156273, gefitinib or lapatinib; quinoxalines are for example PD153035 or ZD1839. An example for a quinoline is 5-methyl-5H-indolo[2,3-β]quinoline, an example for a 2-phenylaminopyrimidine is imatinib, examples for flavonoids are genistein or quercetin, an example for a benzoquinoid is herbimycin A, an example for an aminosalicylate is lavendustin A, and an example for a stilbene is piceatannol. Other suitable compounds may include a receptor tyrosine kinase inhibitor such as the tyrphostin erbstatin, an EGFR specific receptor tyrosine kinase inhibitor such as WHI-P97 or the tyrphostin AG 592, a tyrosine phosphorylation stimulator such as aurin tricarboxylic acid or a tyrosine phosphatase inhibitor such as sodium pervanadate or isoxazole carboxylic acids.

[0094] A further example of such a compound modulating the tyrosine phosphorylation of a RUNX3 protein is an agonist or antagonist for a cell surface molecule that is able to induce the regulation of a protein kinase or protein phosphatase. Examples of such cell surface molecules are receptor tyrosine kinases, membrane receptors with associated tyrosine kinase activity, and G protein coupled receptors, the signal transduction of which are interconnected with pathways regulating protein kinases and phosphatases. Examples for a receptor tyrosine kinase are a receptor for a platelet derived growth factor, a receptor for erythropoietin, a receptor for tumor necrosis factor, a receptor for leukaemia inhibitory factor, a receptor for an interferon, a receptor for insulin, a receptor for an insulin-like growth factor, a receptor for an interleukin, a receptor for a fibroblast growth factor, a receptor for a granulocyte-macrophage colony stimulating factor, a receptor for a transforming growth factor, or a receptor for an epidermal growth-factor (EGF). Such receptors are known to possess the ability to phosphorylate tyrosine residues of various proteins and to be themselves able to regulate further factors inside the cell that possess a similar effect (see e.g. Pazin MJ, Williams LT, Trends in Biochemical Sciences 17 (10), 1992, 374-378, for the EGF receptor see e.g. Janmaat ML, Giaccone G, Oncologist 8 (6), 2003, 576-586). The terms "agonist" and "antagonist" in this context therefore refer to the ability of the cell surface molecule to produce such effects and the modulation of this ability.

[0095] One embodiment of such an agonist or antagonist is a proteinaceous molecule that binds to a molecule on the cell surface, which is able to induce the regulation of a tyrosine kinase or tyrosine phosphatase. Examples of such proteinaceous binding molecules are immuno- globulins, (recombinant) immunoglobulin fragments such as Fab fragments, Fv fragments, single-chain Fv fragments (scFv), diabodies, triabodies (Iliades, P., et al., FEBS Lett (1997) 409, 437_441)₅ decabodies (Stone, E., et al., Journal of Immunological Methods (2007) 318, 88-94) and other domain antibodies (Holt, L. J., et al., Trends Biotechnol. (2003), 21, 11, 484-490). Single-chain Fv fragments are for instance fusions of variable regions from one heavy chain and one light chain of an immunoglobulin molecule. An example of a proteinaceous binding molecule with antibody-like functions is a mutein based on a polypeptide of the lipocalin family (WO 2003/029462; WO 2005/019254; WO 2005/019255; WO 2005/019256; Beste et al, Proc. Natl. Acad. Sd. USA (1999) 96, 1898-1903). Lipocalins, such as the bilin binding protein, the human neutrophil gelatinase-associated lipocalin, human Apolipoprotein D, human tear lipocalin, or glycodelin, posses natural ligand-binding sites that can be modified so that they bind to selected small protein regions known as haptens. Other non-limiting examples of further proteinaceous binding molecules so-called glubodies (see WO 96/23879), proteins based on the ankyrin scaffold (Mosavi, L.K., et al., Protein Science (2004) 13, 6, 1435-1448) or the crystalline scaffold (WO 2001/04144), the proteins described by Skerra (J. MoI. Recognit. (2000) 13, 167-187), AdNectins, tetranectins, avimers and peptoids. Avimers contain so called A-domains that occur as strings of multiple domains in several cell surface receptors (Silverman, J, et al., Nature Biotechnology (2005) 23, 1556-1561). Adnectins, derived from a domain of human fibronectin, contain three loops that can be engineered for immuno- globulin-like binding to targets (Gill, D. S. & Damle, N.K., Current Opinion in Biotechnology (2006) 17, 653-658). Tetranectins, derived from the respective human homotrimeric protein, likewise contain loop regions in a C-type lectin domain that can be engineered for desired binding (ibid.). Peptoids, which can act as protein ligands, are oligo(N-alkyl) glycines that differ from peptides in that the side chain is connected to the amide nitrogen rather than the α carbon atom. Peptoids are typically resistant to proteases and other modifying enzymes and can have a much higher cell permeability than peptides (see e.g. Kwon, Y. -U., and Kodadek, T., J. Am. Chem. Soc. (2007) 129, 1508-1509). Where desired, a modifying agent may be used that further increases the affinity of the respective moiety for any or a certain form, class etc. of target matter. [0096] As already explained, in some embodiments a method according to the invention is a method of identifying a candidate compound that is capable of preventing tumorigenesis in a cell and/or of inducing apoptosis in a tumour cell. Such a method may include introducing the compound into a cell that is capable of expressing Runx3 or a functional fragment thereof and one or more of (i) β-catenin, or a functional fragment thereof, and (ii) a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof. Further the method includes determining the above complex formation, i.e. between Runx3 and (i) β-catenin, and/or (ii) a member of the TCF/LEF transcription co-factor family (including a functional fragments of the respective proteins). An alteration in the complex formation is an indication that the compound is capable of preventing tumorigenesis in a cell and/or of inducing apoptosis in a tumour cell. [0097] For some embodiments of the invention, compounds may be used in form of a library. Examples of such libraries are collections of various small organic molecules, chemically synthesized as model compounds, or nucleic acid molecules containing a large number of sequence variants. A method of identifying a compound according to the invention may be carried out as a screening method, including a high-throughput method. In a respective method a library of compounds may for example be screened to identify candidate compounds capable of altering the said complex formation. In embodiments where a plurality of candidate compounds are analysed according to a method of the present invention in order to identify a compound capable of preventing, inhibiting, arresting or reversing tumourigenesis, such an embodiment may typically called a screening process. These candidate compounds may be analysed independent from each other, e.g. concurrently, consecutively or in any way out of phase. In in vivo or ex vivo embodiments the candidate compounds may for example be added to a cell culture medium or be administered to an organism, for example a mouse or a fruit fly. In some in vitro embodiments any number of steps of analysing a plurality of candidate compounds may for example be carried out automatically- also repeatedly, using for instance commercially available robots. For such purposes any number of automation devices may be employed, for instance an automated read-out system, a pipetting robot, a rinsing robot, or a fully automated screening system. As an illustrative example, the process may be an in-vitro screening process, for example carried out in multiple- well microplates (e.g. conventional 48-, 96-, 384- or 1536 well plates) using one or more automated work stations. Hence, in some embodiments the invention provides a process of high-throughput screening. The method may also be carried out using a kit of parts, for instance designed for performing the present method

[0098] Yet other related methods are in-vivo methods that include providing a host organism. Any desired host organism may be provided as long as it is capable of accommodating and growing a tumour cell, e.g. a cancer cell. Examples of a host organism include, but are not limited to, a mammal, a fish, an amphibian and a bird. For examples of a suitable mammal see above. Any desired cancer cell maybe used for this purpose (see above for examples).The method further includes introducing a cancer cell into the host organism. Furthermore the method includes the use of a compound as described above, i.e. a compound that is suspected to be capable of altering the formation of a complex between Runx3, or a functional fragment thereof, and one or both of (i) a member of the TCF/LEF transcription co-factor family or a functional fragment thereof, and (ii) β-catenin, or a functional fragment. In some embodiments the cancer cell includes the compound. Accordingly the compound may be introduced into the cancer cell before introducing the same into the host organism. In some embodiments the compound is administered to the host organism, before, after or concurrently with introducing the cancer cell therein. Typically the compound is introduced into the cancer cell at a certain stage of the method. The method further includes monitoring the growth of tumours in the host organism. [0099] In some embodiments methods of prognosis and diagnosis according to the present invention include detecting the presence of one of the above complexes.

[0100] Some methods and uses according to the invention include or aim at inducing apoptosis in a tumor cell. Apoptosis is a programmed cell death and typically a mechanism in a multicellular organism to remove undesired cells. Where a cell's capability to undergo or initiate apoptosis is impaired or abolished, a damaged cell is able to proliferate in an unchecked manner, thereby developing into a cancer cell. An apoptotic cell shows a characteristic morphology, by which it can be identified under a microscope. By inducing apoptosis in a tumor cell, a corresponding method may also be used as a therapy for the treatment or prevention of cancer.

[0101] If desired, the progress of apoptosis in a tumor cell may be monitored, for example by propodium iodide staining or flow cytometry analysis, mitochondrial dysfunction or caspase 3 activation. Typically the method of the invention triggers an apoptotic cell death response involving mitochondria disruption and caspase activation. Non-cancerous cells however show only a marginal cell death response, if any at all. Besides determining apoptosis in a respective cell in some embodiments a method according to the present invention may include determining cell viability in a respective cell. Respective methods are well established in the art.

[0102] Some methods according to the present invention are methods of controlling tumorigenesis. These methods include in particular methods of preventing, inhibiting, arresting or reversing tumourigenesis. Tumourigenesis may for example be carcinogenesis, including the formation of malignant forms of carcinomas. Accordingly, the method may for example be included in a treatment or prevention of a proliferative disease or disorder, such as cancer.

[0103] The present invention encompasses inter alia the assessment of one of the above named complexes in a cell for diagnostic, prognostic, and therapeutic purposes. Based on the inventors' findings the invention also provides methods of identifying a compound that is capable of preventing, inhibiting, arresting or reversing tumorigenesis, including carcinogenesis, in a cell and/or of inducing apoptosis in a tumor cell. Some of these methods are in vivo or ex vivo methods. Some of the methods are in-vitro methods of identifying a respective compound. The compound may be capable of influencing the formation of one of the above complexes. Some methods according to the invention include exposing the components of this complex to each other, whether in-vitro or in-vivo. One such method is an in-vitro method, which includes contacting the components that form, or are suspected to form, a complex with each other. The compound may be capable of altering the complex formation between the components thereof, e.g. between β-catenin, or a functional fragment thereof, and RUNX3, or a functional fragment thereof. In some embodiments a respective method includes contacting the compound and the components of the respective complex such as β-catenin, or a functional fragment thereof, and RUNX3, or a functional fragment thereof, hi such embodiments the formation of a complex is detected. [0104] In some embodiments the method further includes detecting the formation of the complex. Any suitable method of detecting a complex formation may be used. A detection method may for instance include electrophoresis, HPLC, flow cytometry, fluorescence correlation spectroscopy or a modified form of these techniques. Methods such as immunoprecipitation or copurification using a chromatography technique may be carried out under native conditions, i.e. conditions where at least substantially no denaturation of the proteins of interest occurs. Other techniques involve a measurement of the biomolecular binding itself. Such measurements may for instance rely on spectroscopic, photochemical, photometric, fluorometric, radiological, enzymatic or thermodynamic means. An enhancement or a reduction of the formation of a complex as named above indicates that the compound may be capable of preventing, inhibiting, arresting or reversing tumorigenesis in a cell and/or of inducing apoptosis in a tumor cell.

[0105] Assessing the formation or presence of said complex may include a measurement of the binding of one or more of its components. Such measurements may for instance rely on spectroscopic, photochemical, photometric, fluorometric, radiological, enzymatic or thermodynamic means. An example for a spectroscopic detection method is fluorescence correlation spectroscopy. A photochemical method is for instance photochemical cross-linking. The use of photoactive, fluorescent, radioactive or enzymatic labels respectively represent illustrative examples for photometric, fluorometric, radiological and enzymatic detection methods. An example for a thermodynamic detection method is isothermal titration calorimetry. Some of these methods may include additional separation techniques such as electrophoresis or HPLC. In detail, examples for the use of a label include a compound as a probe or an immunoglobulin with an attached enzyme, the reaction catalysed by which leads to a detectable signal. An example of a method using a radioactive label and a separation by electrophoresis is an electrophoretic mobility shift assay.

[0106] In some embodiments forming a complex as defined above includes the translocation of RUNX3 into the nucleus, in particular from the cytoplasm of the cell to the nucleus. Likewise, releasing a respective complex may include the transfer of RUNX3 from the nucleus to another compartment or organelle of the cell, in particular the cytoplasm. Without wishing to be bound by theory it is believed that one or more so far unknown factors may be responsible for arranging RUNX3 in the cell at a location that differs from the nucleus, in particular the cytoplasm. A compound according to the invention may accordingly affect the cellular location of RUNX3 and thereby influence the formation the formation of a complex of RUNX3, or a functional fragment thereof, with β-catenin, or a functional fragment thereof, and/or a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof. A method of the invention, including a diagnostic and/or a therapeutic method may include determining the cellular location of RUNX3. A location of RUNX3 outside the nucleus may for instance be an indication that the respective cell bears a risk or has a predisposition of turning tumorigenic, including cancerogenic. Accordingly, an individual in whom such a cellular location has been identified may have an increased risk of developing a neoplasm, such as a tumor, including cancer or ulcerative colitis.

[0107] As noted above, the present invention also relates to a method of diagnosing the risk of developing a neoplasm, such as a tumor, including cancer, in a subject. A respective tumor may for example be a breast tumor, a lung tumor, a colorectal tumor, a tumor of the urinary bladder or a tumor of the fallopian tube (also termed oviduct). Likewise, a respective cancer may for instance be breast cancer, lung cancer, colorectal cancer, cancer of the urinary bladder or cancer of the fallopian tube (also termed oviduct), including one of the corresponding carcinomas. An illustrative example of a carcinoma of lung cancer is a non-small cell lung carcinoma. The method includes determining the presence, possibly including determining the amount thereof, of one of the above named complexes. The measurement may in some embodiments be carried out in a sample, such as a tissue sample or a cell sample, from the subject. In some embodiments the method may also include comparing the results of measuring the presence and/or the amount of a omplex as described above. For a respective control measurement a sample may be used, in which the above described complex formation is known to be on a customary ("normal") level. In typical embodiments an altered complex formation as compared to the control measurement indicates that the subject suffers from or is at risk of developing a neoplasm. [0108] Further methods of the invention are methods, both in-vivo and in-vitro methods, of identifying a compound capable of altering the said complex formation, i.e. a complex between RUNX3, or a functional fragment thereof, and one or both of (i) a member of the TCF/LEF transcription co-factor family or a functional fragment thereof, and (ii) β-catenin, or a functional fragment thereof. The compound may be capable of preventing, inhibiting, arresting or reversing tumorigenesis, including carcinogenesis. The compound may be capable of altering the forming of the afore described complex. Typically these methods include exposing the components of this complex to each other in presence of the compound of interest, whether in-vitro or in-vivo. In some embodiments the method further includes detecting the formation of the complex.

[0109] The invention is further illustrated by the following figures and non limiting examples. As one of ordinary skill in the art will readily appreciate from the disclosure of the present invention, other compositions of matter, means, uses, methods, or steps, presently existing or later to be developed that perform substantially the same function or achieve substantially the same result as the corresponding exemplary embodiments described herein may likewise be utilized according to the present invention.

[0110] As noted above, an almost ideal model for studying the Wnt/β-catenin signaling pathway, pathogenic alterations thereof and for intervention therein is colorectal cancer. Those skilled in the art will thus appreciate that results obtained with this model can easily be transferred to any other disorder related to aberrant Wnt/β-catenin signalling. Since some embodiments of the methods of the invention include detecting or modulating a complex of RUNX3 with β-catenin, the present examples are based on this model.

EXEMPLARYEMBODIMENTS OFTHE INVENTION

[0111] After the priority date of the present application the exemplary data and conclusions disclosed herein have been disclosed in Cancer Cell (2008) 14, 3, 226-37 (including an accompanying erratum in: Cancer Cell (2009) 15, 3, 240) the contents of which is/are incorporated herein in its/their entirety for all purposes.

[0112] Fig. 1. Simplified schematic of important features of the Wnt/β-catenin pathway: (A) In the absence of Wnt signal β-catenin levels are regulated by the formation of a multiprotein complex, defining a cytosolic destruction complex. The complex includes the tumour suppressor adenomatous polyposis coli (APC), a scaffold protein axin, casein kinase 1 (CKI)) and glycogen synthase kinase 3β (GSK3β). It causes β-catenin phosphorylation, thereby marking it for ubiquitinylation and consequently degradation via the proteasome. (B) Upon binding of Wnt to the Frizzled (Fz) receptor and a low-density-lipoprotein (LDL) receptor related protein such as the LDL receptor related protein 6 (LRP6), DvI is recruited to this receptor, leading to its activation, whereupon a cascade of events is triggered. As a result hypophosphorylated β-catenin is stabilized, accumulates and translocates to the nucleus. There it forms a complex with a T cell factor (TCF) or a lymphoid enhancer-binding factor (LEF), thereby activating transcription of numerous genes, including c-MYC, cyclin Dl, gastrin or matrilysin. (C) RUNX3 binds to the complex of β-catenin and the T cell factor, thereby attenuating transcriptional activity. (D) It is contemplated that activation of AKT leads to RUNX3 phosphorylation, whereby the complex between RUNX3, β-catenin and the T cell factor according to the method of the invention may be abrogated, or at least weakened.

[0113] Figure 2. Expression of Runx3 in intestinal epithelial cells and up-regulation of β-catenin/Tcf4 activity in the RunxS^'1' intestine. (A) Immunodetection of Runx3 in wild type (wt) and RunxS^'1' jejunum. Note that Runx3 expression is greatly reduced in Paneth cells; β-catenin is detected in the nuclei of these cells

(arrowheads).

(B) Immunodetection of Runx3 in wt and RunxS^'1' proximal colon (left) and wt distal colon

(right). (C) Hematoxylin and eosin staining of wt and RunxS^''^' jejunum and proximal colon of mice at

40 weeks of age.

(D) Detection of proliferating cells in wt and RunxS^''^' jejunum and proximal colon by BrdU incorporation (adult; 40 weeks old) and immunostaining with the anti-Ki67 antibody

(neonate). (E) Number of BrdU (adult) and Ki67 (neonate) positive cells per crypt in wt and RunxS^''^' intestine (*P<0.001).

(F) Relative growth rates of Runx3^+/+ (+/+1 and +1+2) and RunxS^''^' (-/-1 and -1-2) FID cell lines in the monolayer culture.

(G) Immunodetection of c-Myc in wt and RunxS^''^' intestines. (H) Immunodetection of EphrinB and EphB2 in wt and RunxS^''^' jejunal epithelium.

(I) Detection of Paneth cells with an anti-lysozyme antibody in wt and RunxS^{' '} small intestine.

Arrowheads show mispositioning of Paneth cells in RunxS^''^' small intestine.

(J) Western analysis of protein expression profiles in wt and RunxS^''^' intestine.

Specimens were counterstained with hematoxylin (A, B, G, and H). Scale bars are equal to 50 μm (B, G, H) and 100 μm (A, C, D, and I).

[0114] Figure 3. Detection of a β-catenin/TCF4/RUNX3 ternary complex and the attenuation of β-catenin/TCF4 transcriptional activity by RUNX3.

(A) Western analysis of RUNX3 expression in 22 human colorectal cancer cell lines. (B) Exogenously expressed β-catenin, TCF4 and RUNX3 coimmunoprecipitates. HCTl 16 cells were transfected with Myc-TCF4 and the vector (lane 1), Myc-TCF4 and Flag-RUNX3 (lane 2), Flag-RUNX3 and the vector (lane 3), Flag-RUNX3 and Myc-TCF4 (lane 4), Myc-TCF4, Flag-RUNX3 and the vector (lane 5), Myc-TCF4, Flag-RUNX3 and HA-β-catenin (lane 6). Proteins were immunoprecipitated with anti-Flag agarose (lanes 1 and 2), anti-Myc (lanes, 3 and 4), and anti-HA (lanes 5 and 6), and the immunoprecipitates subjected to Western blot analysis using anti-Myc, anti-β-catenin, anti-HA, and anti-Flag antibodies. *Murine IgG was detected (lanes 1 and 2). **Anti-β-catenin (for endogenous β-catenin; lanes 1-4) and anti-HA (for exogenous HA-β-catenin; lanes 5 and 6) antibodies were used.

(C) Ternary complex of exogenously expressed β-catenin and TCF4 and RUNX3 detected by two step-coimmunoprecipitation. 293T cells were transfected with Myc-TCF4, S33Y β-catenin, and control vector (lane 1), Myc-TCF4, S33Y β-catenin, and Flag-RUNX3 (lane 2). Proteins were immunoprecipitated with anti-Flag agarose (1st IP). Bound proteins were eluted with the Flag peptide and immunoprecipitated with normal mouse IgG (lane 3) or anti-Myc antibody (lane 4) (2nd IP). The first and second immunoprecipitates were subjected to Western analysis using anti-Myc, anti-β-catenin, and anti-Flag antibodies. *A non-specific band (lanes 3 and 4). (D) An endogenous ternary complex of β-catenin/TCF4/RUNX3 in HCTl 16 and SW620 but not in SW480 cells. Nuclear extracts were immunoprecipitated with anti-dephospho-β-catenin (activated β-catenin) antibody, anti-TCF4 antibody, anti-RUNX3 antibody, and normal murine IgG. The immnunoprecipitates were subjected to Western blot analysis using anti-dephospho-β-catenin, anti-TCF4, and anti-RUNX3 antibodies. (E) Interaction of in vitro translated His-tagged RUNX3 with in vitro translated Myc-TCF4 and/or HA-β-catenin, as revealed by pull-down assay with Ni-NTA agarose. Western analysis was performed using anti-HA, anti-His, and anti-Myc antibodies.

(F) Oncogenic β-catenins have a higher affinity for RUNX3 than wild type β-catenin. HCTl 16 cells were transfected with wild type β-catenin, Δ45 β-catenin, or S33Y β-catenin, together with Flag-RUNX3 or control vector. Proteins immunoprecipitated with anti-Flag agarose were subjected to Western analysis using anti-β-catenin antibody.

(G) Reduction of TOPflash (0.1 μg DNA) activity by exogenous RUNX3 (0.1, 0.2, and 0.5 μg DNA) in DLDl cells.

(H) Interaction of RUNX3 (R178Q) with β-catenin/TCF4. 293T cells were transfected with Myc-TCF4, S33Y β-catenin together with control vector (lane 1), Flag-RUNX3 (lane 2), or

Flag-RUNX3 (R178Q) (lane 3). Proteins were immunoprecipitated with-anti-Flag agarose.

Immunoprecipitates were subjected to Western analysis using anti-Myc, anti-β-catenin, and anti-Flag antibodies.

(I) Reduction of TOPflash (0.1 μg DNA) activity by exogenous wild type (WT) RUNX3 and RUNX3 (R178Q) (0.05 and 0.2 μg DNA) in DLDl cells.

(J) Exogenous RUNX3 (0.1 and 0.2 μg DNA) reduces the activity of the wild type cyclin Dl promoter (CyclinDl-WT; 0.1 μg) but not that of a variant promoter with a mutated TCF binding site (CyclinDl-mTCF; 0.1 μg).

(K) Differential sensitivities of Runx3^+/+ (+/+1 and +/+2) and RunxS^''^' (-/-1 and -1-2) FID cell lines to Wnt3a are evaluated by their relative TOP/FOP luciferase activities (arbitrary units).

Cells were stimulated by the medium containing 20% or 50% of conditioned medium of Wnt3a-expressing L cells.

[0115] Figure 4. RUNX3 attenuates the DNA binding activity of β-catenin/TCF4.

(A) Exogenous RUNX3 attenuates the binding of β-catenin/TCF4 to TCF binding sites of the cyclin Dl and c-Myc promoters. DLDl clones expressing exogenous RUNX3 (+) or control vector (-) (see panel E) were subjected to ChIP analysis using antibodies against TCF4 (lanes 3 and 4), dephospho-β-catenin (lanes 6 and 7), or normal murine IgG (lanes 5 and 8). DNA precipitates were amplified by PCR (35 or 37 cycles).

(B) Enhanced occupancy of β-catenin/TCF4 at TCF binding sites of the cyclin Dl and c-Myc promoters following RUNX3 knockdown. HCTl 16 clones expressing antisense DNA against RUNX3 (AS-Cl.1; lanes 1 and 5, and AS-C1.2; lanes 2 and 6) or control vector (control; lanes 3, 4, and 7) (see panel E) were subjected to ChIP. analysis using anti-TCF4 (lanes 5-7) or normal murine IgG (lane 4) antibodies. DNA precipitates were amplified by PCR (35 or 37 cycles).

(C) (Upper panel) RUNX3 displaces TCF4 from TCF binding sites of the cyclin Dl and c-Myc promoters in a dose-dependent manner. DLDl stable transfectant in lower panel was subjected to ChIP assay with the indicated antibodies and PCR cycles. (Lower panel) Western analysis of RUNX3 induction in DLDl stably transfected with inducible RUNXS following stimulation with 20μM ponasterone A (Pon A).

(D) Quantification by real-time PCR of ChIP assay for TCF4 (TCF4) and β-catenin (β-catenin) occupancy at TCF consensus sites of the cyclin Dl (open bars) and the c-Myc (gray bars) promoters. The relative occupancy in RUNX3 -expressing DLDl as compared to the control clone (panel E). TCF4 binding in RUNX3-knockdowned HCTl 16 clones (AS-Cl.1 and AS-C1.2) were compared with those in the control clone (panel E). PCR amplification of ChIP using primer sets (designed in the cyclin Dl and c-Myc promoters) which do not flank any TCF consensus sites were less than 0.0001 in all trials. (E) Protein expression profile in DLDl clones transfected with a vector (control) or Flag-RUNX3 (RUNX3) and in HCTl 16 clones expressing antisense RUNX3 DNA (AS-Cl.1 and AS-C1.2) or control vector (control), as detected by Western blot analysis.

(F) Left panel: Real-time PCR quantification of AXIN2, CD44, and DKKl mRNA in DLDl cells with inducible RUNX3 expression. The amount of mRNA 24h after induction by Pon A was normalized against non-induced samples. Right panel: The amount of mRNA of the same set of genes in HCTl 16 cells in which RUNX3 is knocked down (average of AS-Cl.1 and AS-C1.2; see panel E) was normalized to the control clone.

(G) TOPfiash versus FOPflash activity (arbitrary units) for DLDl (white) and DLDl expressing RUNX3 (black; panel E) and HCTl 16 (black) and HCTl 16 in which RUNX3 is knocked down (white; panel E).

(H) Differential effects of RUNX3 knock-down in HCTl 16, SW620, COLO320, SW480 and SW403. (Upper panel) Western analysis showing RUNX3 expression in cells treated with 3 independent RUNX3 (shl-3) and the control (control) shRNAs. (Lower panel) TOPflash/FOPflash activity in cells after RUNX3 knockdown was normalized to the controls.

[0116] Figure 5. Adenomatous polyps in the small intestine of Runx3^+/~ or Apc^Mιn/+ BALB/c mice and progression to adenocarcinoma in Runx3^+/'Apc^Mιn/+ compound mice.

(A) Hematoxylin and eosin staining of tumors in the small intestine of Runx3^+/~, Apc^Mιn/+ and Runx3^+/'Apc^Mm/+ mice at 65 weeks of age. Runx3^+/'Apc^Mιn/+ mice developed adenocarcinoma. Boxed regions are enlarged on right. A scale bar is equal to 1 mm.

(B) Frequency of tumor formation in the small and large intestines of mice with indicated genotypes at 65 weeks of age. Tumors larger than 0.2 mm in diameter were counted.

(C) Number of tumors in the small intestine of individual mice with indicated genotypes at 65 weeks of age (*¹P<0.01 and *²P<0.05).

(D) Size distribution of polyps in the small intestine of mice with indicated genotypes at 65 weeks of age. Four size classes were designated: greater than 5.0 mm in diameter, between 2.5 and 5.0 mm, between 1.0 and 2.5 mm, and between 0.2 and 1.0 mm.

[0117] Figure 6. Adenomatous polyps in the small intestines of Runx3^+/' BALB/c mice displaying down-regulated Runx3 and up-regulated cyclin Dl and c-Myc.

(A) Immunodetection of Runx3, β-catenin, cyclin Dl, and c-Myc in adenomatous polyps formed in the small intestine of Runx3^+/~ and Apc^Mιn/+ mice at 65 weeks of age. Dashed lines indicate the border between normal and adenomatous cells (marked by A).

(B) Very small adenomas formed in jejunum of the compound mice were analyzed. When Runx3 is down-regulated, β-catenin was not activated (upper panels). In contrast, when β-catenin was activated, Runx3 was not down-regulated (lower panels).

Specimens were counterstained with hematoxylin (A, B). Scale bars are equal to 100 μm (A, B). (C and D) Quantification by real-time PCR of Runx3, cyclin Dl, c-Myc, and p21 mRNA in individual polyps (Tl -3) in Runx3^+/~ (C) and Apc^Mυτ/+ (D) small intestines, normalized to the values of adjacent normal epithelium.

(E) Enhanced binding of β-catenin/Tcfs to the Tcf consensus site of the cyclin Dl promoter as revealed by the ChIP assay in polyps (Tl and T2) and adjacent normal tissues (Nl and N2) of Runx3^+/~ and Apc^Min/+ small intestines. Tl and T2 represent 2 pools of 3-4 polyps each from 1-2 mice with adenomas to provide sufficient material for the material for the ChIP assay. DNA fragments precipitated by anti-β-catenin antibody or control IgG were amplified by PCR (33 cycles). The Gαpdh promoter region was amplified as a negative control.

[0118] (F and G) Real-time PCR quantification of ChIP to assay for bound β-catenin/Tcfs at Tcf consensus site of the cyclin Dl (F) and the c-Myc (G) promoters, normalized to the inputs. [0119] Figure 7. Down-regulation of RUNX3 expression without nuclear accumulation of β-catenin in human adenomatous polyps.

(A-C) Three patterns of β-catenin and RUNX3 expression in 35 human cases. Type A; nuclear β-catenin with RUNX3 in nuclei (A), type B; membranous β-catenin without RUNX3 expression (B), and type C; membranous β-catenin with RUNX3 expression (C) in T4, T6, and T9 in panel F, respectively. Enlargement of a part of panels A-C is shown in Figure 8. (D and E) Up-regulation of cyclinDl and c-Myc in adenomas of type A (D) and type B (E). 4 slides each of adenoma and the adjacent normal epithelium shown in A and D (type A) and in B and E (type B) are all serial sections. Specimens were counterstained with hematoxylin. A scale bar is equal to 100 μm.

(F) Methylation specific PCR (MSP) analysis of the RUNX3 promoter in normal human colon epithelium, DLDl and HCTl 16 cells, and representative of 15 (Tl-Tl 5) of a total 35 human adenomatous polyps (see Fig. 18).

(G) Methylation status of CpG dinucletide between -70 and -21 relative to the translation initiation site of the RUNX3 exon 1 region. The nucleotide sequence (sense strand) of MSP products from RUNX3 -positive and -negative tumors, DLDl and HCTl 16 cells were shown. The labeled C (highlighted with a box) depicts resistance to bisulfite treatment due to methyaltion. Unmethylated C (not highlighted) was converted to T by the bisulfite treatment. Asterisks on T indicate unmethylated C residues in RUNX3 -negative tumors. Unmethylated C (not highlighted) was converted to T by the bisulfite treatment. Asterisks on T indicate unmethylated C residues in RUNX3 -negative tumors.

[0120] Figure 8. Enlargement of immunohistochemistry of adenomas shown in A, B and C of Figure 7 for better resolution.

[0121] Figure 9. RUNX3 inactivation by gene silencing and protein mislocalization with concomitant accumulation of β-catenin in human colorectal cancers.

(A-C) Differential staining patterns of RUNX3 in human colorectal cancers: positive (A), negative (B), and cytoplasmic positive (C), and summarized as P, N, and C, respectively in the entries of Fig. 19. Specimens were counterstained with hematoxylin. A scale bar is equal to 100 μm. (D) Differential staining patterns of RUNX3 in human colorectal cancer cell lines: positive (HCTl 16 and SW480), negative (DLDl and RKO), and cytoplasmic positive (SW403 and CCK81).

[0122] Figure 10. (A) Morphology of wt and Runx3^'A epithelium of jejunum and colon stained by hematoxylin and eosin. Three representative Runx3^{~ '} jejunums and colons (7-1-3) from individual adult mice at 30-40 weeks of age are shown. Inflammation was observed in a severe case of colon hyperplasia (-/-3 colon).

(A and B) Immunodetection of CD44 (B) and cyclin Dl (C) in wt and Runx3^'A intestines. Scale bars are equal to 500 μm (A) and 50 μm (B, C). [0123] Figure 11. (A) Mapping of the RUNX3 domain that interacts with TCF4. HCTl 16 cells were transfected with Flag-tagged RUNX3 derivatives and Myc-tagged full-length TCF4. Proteins were immunoprecipitated with anti-Flag agarose and subjected to Western blot analysis using anti-Myc and anti-Flag antibodies.

(B) Mapping of the TCF4 domain that interacts with RUNX3. HCTl 16 cells were transfected with Myc-TCF4 derivatives and Flag-full-length RUNX3. Proteins were immunoprecipitated with anti-Myc antibody and subjected to Western blot analysis using anti-Flag and anti-Myc antibodies.

(C) Interaction between RUNX3 and TCFs (TCFl, Lefl, and TCF3). HCTl 16 cells were transfected with Myc-TCF1/Flag-RUNX3 or Myc-TCFl /the vector (left), Lefl/Flag-RUNX3 or Lefl /the vector (center), and Myc-TCF3/Flag-RUNX3 or Myc-TCF3/the vector (right). Proteins were immunoprecipitated with anti-Flag antibody and subjected to Western blot analysis using anti-Myc, anti-Lefl, or anti-Flag antibodies. * Murine IgG is detected

[0124] Figure 12. (A) Attenuation of the binding of β-catenin/TCF4 to the TCF binding site of the TOP construct by RUNX3, as revealed by EMSA using nuclear extracts prepared from 293T cells expressing Myc-TCF4, S33Y β-catenin, Flag-RUNX3, or the vector (mock). All reactions contained the same amount of proteins, as normalized to mock extract. One dose of RUNX3 extract (Xl) is the same as the amount of TCF4 extract (1 μg protein). The activity of RUNX3 in the extract was confirmed by EMSA using a probe with a RUNX3 site (see panel B). Unlabeled probes were added at an 8-fold excess relative to labeled probes for competition. *A non-specific band detected in all reactions.

(B) A larger amount of β-catenin/Tcf4-DNA complex (arrow) was observed in the nuclear extract from Wnt3a-treated Runx3^~'^~ FID cells than Runx3^+/+ FID cells as revealed by EMSA (right). Both nuclear extracts expressed comparable levels of dephosphorylated β-catenin and Tcf4 as revealed by Western blot analysis (left). Unlabeled probes were added at a 15-fold excess relative to labeled probes for competition. *A non-specific band.

(C) Attenuation of binding of RUNX3/PEBP2β to the RUNX binding site of the IgCa promoter (Hanai et al., 1999) by β-catenin/TCF4, as revealed by EMSA using nuclear extracts from 293T cells expressing Flag-RUNX3, Myc-TCF4, S33Y β-catenin or the vector (mock) and purified PEBP2β protein. All reactions contained the same amount of proteins normalized to the mock extract. One dose of extract expressing Myc-TCF4 or S33Y β-catenin (Xl) is the same as the amount of RUNX3 extract. Unlabeled probes were added at a 10-fold excess relative to labeled probes for competition. *Lower and upper RUNX3-probe and **lower and upper RUNX3/PEBP2β-probe complexes were detected. [0125] Figure 13. (A) (Upper panel) Wild type epithelial cells (wt), Runx3^+/~ adenomas

(T1-T4) and their adjacent normal epithelial cells (N1-N4) in the small intestine of BALB/c mice were genotyped. (Lower panel) Quantification by real-time PCR of the wild type and knockout alleles of Runx3 in Runx3^+/~ adenomas (T1-T7). The ratio of heterozygosity in Runx3^+/~ tumors to that in adjacent normal epithelial cells was calculated as relative amount of wild type allele per knockout allele in T1-T7. The average of the ratio in seven Runx3^+/~ tumors was 1.03 ± 0.22.

(B) Normal epithelial cells (N) and adenomas in Apc^MW+ (Tl -T4) and Runx3^+A (Tl -T3) small intestine of BALB/c mice were genotyped by PCR and Hindlϊl digestion for distinguishing wt and mt allele of Ape as previously reported (Luongo et al., 1994). (C) (Upper panel) Positions of CpG dinucleotides in the promoter region of Runx3. (Lowerpael) Methylation of the Runx3 promoter region. Wild type epithelial cells (wt), Runx3^+/~ normal epithelial cells (N), and Runx3^+/~ adenomas (Tl -T6) in the small intestine were examined by methylation-specific PCR (MSP) using primers specific to the M1-M3 regions. Methylated (M) and unmethylated (U) DNA were detected. El, a mouse gastric cancer cell line with Runx3 promoter hypermethylation (Guo et al., 2002) was used as a positive control.

[0126] Figure 14. (A) Frequency of tumor formation in small and large intestines. Analysis of adenomatous polyps induced in mice of the BALB/c:C57/B6 background at 25 weeks of age. Tumors larger than 0.2 mm in diameter were counted. (B) Number of tumors in the small intestine of individual mice. (C) Size distribution of polyps in the small intestine of mice.

(D) Stereomicroscopic images of polyps (arrowheads) formed in Apc^mn/+ and Runx3^+/'Apc^mn/+ small intestines of mice. A scale bar is equal to 5 mm.

[0127] Figure 15. Up-regulation of CD44 in adenomas of type A and B (A and B, respectively; see Figure 7). A scale bars is equal to 100 μm. [0128] Figure 16. Morphology of small and large intestines of wild type mice reconstituted with Runx3^~A (A-D) or Runx3^+/~ (E) bone marrow cells, one year after transplantation. Jejunums and colons of five individual mice were stained by hematoxylin and eosin. Chimerisms of Runx3^''^' or Runx3^+/' bone marrow cells were; A, 99.8%; B, 99.8%; C, 99.9%; D, 29.3%; E, 99.0%. Scale bars are equal to 500 μm.

[0129] Figure 17. (A) Relative proliferation (arbitrary units) of DLDl and HCTl 16 clones (see Figure 4E). Cells were counted at indicated times.

(B) Tumorigenicity of DLDl and HCTl 16 clones (see Fig. 4E). Cells (5 x 10⁶) were injected into nude mice subcutaneously. The tumors were obtained and weighed when they grew beyond one gram (*P<0.01).

(C) Tumor formation of Runx3^+/+and Runx3^~f~ FID and FIL cell lines (see Fig. 3K) in nude mice 60 days after inoculation. Cells (3 x 10⁶) of each line were subcutaneously injected into the mice. Number of mice bearing tumors is indicated above the number of mice used. (D) Tumor formation of control (C) and Myc-tagged dominant negative TCF4 (DN: van de Wetering et al., 2002) expressing Runxi^'1' FID and FIL cells (indicated as 7-2 in panel C) in nude mice 60 days after inoculation. Cells (3 x 10⁶) were subcutaneously injected into the mice. Number of mice bearing tumors is indicated above the number of mice used.

[0130] Figure 18. Expression patterns of β-catenin and RUNX3 (type A-C; see the text and Figure 7) and the methylation status of the RUNX3 promoter (M; methylated, U; unmethylated as revealed by MSP) of 35 human sporadic adenomatous polyps (T1-T35) are summarized.

[0131] Figure 19. Expression pattern of RUNX3 (P, N, and C; see the text and Figure 9) and the methylation status of the RUNX3 promoter (M; methylated, U; unmethylated as revealed by

MSP) of 48 human colorectal cancers are summarized, β-catenin was accumulated in the nuclei/cytoplasm of all cases except for No. 26, 33, and 42 marked in blue. DNA was not available in 11, 12, and 27 cases (n.a.).

[0132] Figure 20 illustrates the tumor formation in Runx3+/- mice.

[0133] Figure 21 depicts the binding of RUNX3 to Aktl in vitro. 293 cells were transfected with Flag-tagged RUNX3 and myc-tagged AKTl. Immunoprecipitation (IP) was performed with anti-Myc polyclonal antibody and immunoblot was performed by either Flag monoclonal antibody or Myc monoclonal antibody. Myc epitope-tagged wild type Aktl (wt), constitutively active (CA) and dominant negative (DN) Aktl were used for immunoprecipitation and for direct immunoblotting. Immunoprecipitation was performed by using anti-RUNX3 polyclonal antibody from Active Motif (Carlsbad, CA₅ USA). [0134] Figure 22 depicts the formation of a homologous protein complex in HCTl 16 nuclear extract. Endogenously expressed RUNX3 and endogenously expressed AKT interaction was analyzed using immunoprecipitation using a nuclear extract from colorectal cancer-derived HCTl 16 cells. Immunoprecipitation was performed by using either anti-RUNX3 polyclonal antibody (Active Motif) or anti-AKTl polyclonal rabbit antibody from Cell Signaling. Immunoblotting was carried out by using either anti-RUNX3 monoclonal antibody, R3-5G4 generated in house or AKTl polyclonal antibody.

[0135] Figure 23 illustrates the domain mapping of the RUNX3/Akt interaction, using immunoprecipitation with an anti-myc immunoglobulin (polyclonal). 293 cells were transfected with the indicated deletion constructs of Flag-tagged RUNX3 and Myc-tagged Aktl. Immunoblotting (lower panel) was carried out with an anti-Flag immunoglobulin.

[0136] Figure 24 illustrates the domain mapping of the RUNX3/Akt interaction with a

FLAG immunoglobulin (monoclonal). 293 cells were transfected with the indicated series of deletion constructs of Myc-tagged Aktl and with Flag-tagged RUNX3. Immunoprecipitation was performed with a Flag monoclonal immunoglobulin. Immunoblotting was carried out using an anti-Myc immunoglobulin.

[0137] Figure 25 is a schematic illustrating the regions involved. The kinase domain of Aktl and the Runt domain of RUNX3 interact. [0138] Figure 26 shows that RUNX3 is phosphorylated by Aktl in vitro. Within the highly conserved region of RUNX3, there is a typical consensus amino acid sequence motif known to be the target of phosphorylation by AKT. This figure shows that RUNX3 is a substrate of phosphorylation activity of AKT. A: His-RUNX3, B: His-RUNX3 (T151A), C: His-RUNX3 (T14A). The specifity of the Akt substrate phospho-(Ser/Thr) antibody is indicated at the bottom of the figure. His-tagged full length RUNX3 was prepared in house, commercially available AKTl kinase was from Cell Signaling (Danvers, MA, USA) and GSK3β, a known substrate of Aktl (Cell Signaling).

[0139] Figure 27 shows that exogenously expressed RUNX3 is phosphorylated by endogenously expressing Akt. Endogenous proteins in 293 cells were immunoprecipitated using an immunoglobulin specific to proteins phosphorylated at an AKT-substrate specific phosphorylation site. Immunoblotting was performed by means of a Flag monoclonal immunoglobulin. Lane 2 shows endogenously phosphorylated RUNX3. Lanes 3 and 4 depict a reduced level of phosphorylation of RUNX3 due to knock down of Aktl by shRNAl (lane 3) or shRNA 2 (lane 4). Lane 5 is a positive control in which an exogenously expressed activated form of Aktl strongly phosporylates RUNX3.

[0140] Figure 28 shows that endogenous RUNX3 is phosphorylated by Akt in DLD-I cells. This phosphorylation is inhibited by an inhibitor of AKT kinase. DLD-I (PI3KCA -/wt) overexpressing Flag-RUNX3 was pre-incubated in 0.5 % serum prior to treatments. PI3KCA indicates the catalytic subunit of phosphatidylinositol 3-kinase (PI3K). Lysates were immunoprecipitated with either rabbit IgG or AKT-substrate polyclonal antibody.

[0141] Figure 29 shows that the phosphorylation of RUNX3 by Akt reduces the affinity of RUNX3 for TCF4 (A: Copurification of TCF4 with Runt or the indicated mutant, B : Analysis of the intensity of the bands).

[0142] As shown above, the complex formation of RUNX3 with the transcription factor TCF4 attenuates TCF4 activity, thereby reducing Wnt signaling activity. The threonine residue of the target of phosphorylation site in RUNX3 was substituted by alanine (A), aspartic acid (D) or glutamic acid (E) (indicated above the lanes in Fig. 29 A, see the legend of Fig 29B). Substitution with alanine completely abolished the phosphorylation site. On the other hand the substitution with either aspartic acid or glutamic acid introduces a negative charge, thereby providing RUNX3 with a resemblance to a phosphorylated protein. The presence of the acidic amino acids D or E is thought to mimick the negative charge of a phosphorylation. The data shown here indicate that a substitution introducing alanine results in even stronger interactions with AKT when compared to wild type RUNX3. On the other hand, those substituted with D or E show significantly reduced activity to interact with AKT. These results suggest that phosphorylation of RUNX3 by AKT leads to a dissociation of the complex between RUNX3 and TCF4, thereby resulting in a reduction of the ability of RUNX3 to attenuate Wnt singnaling. These data add a new aspect to the well established role of Wnt signaling as an oncogenic pathway. The oncogene AKT is able to inactivate the function of the tumor suppressor RUNX3, thus increasing the oncogenic activity of the Wnt pathway.

RESULTS

Expression of Runx3 in intestinal epithelial cells and up-regulation of β-catenin/Tcf4 activity in the Runx3^'A intestine

[0143] The Runx3 protein was immunohistochemically detected in the epithelial cells of the small and large intestines (Figure 2A, B). Runx3 is expressed in all epithelial cell types in the small intestine except for Paneth cells, where maturation is induced by Wnt signaling and accompanied by nuclear accumulation of β-catenin (Fig. 2A; van Es et al., 2005; Andreu et al., 2005). The present inventors reported that Runx3^"A mice of the C57BL/6J background die soon after birth due to starvation (Li et al., 2002). However, some Runx3^'A mice of the BALB/c background (less than 3% of all neonates) survive for about a year. These survivors were analyzed at the adult stage.

[0144] Adult and neonate itønxJT^7" epithelia of the small and large intestines show a high proliferation activity (Fig. 2D, E), and epithelia of adult Runx3^~A mice showed hyperplasias (Fig. 2C and Figure 1 OA). Cell lines obtained from RunxS^'1' small and large intestinal epithelium (FID and FIL cell lines, respectively) grew faster on monolayer cultures than those from Runx3^+/+ mice (Fig. 2F for FID; data not shown for FIL). No tumors were formed in Runx3^~/~ intestine before they died at around 10 months of age. However, small adenomas developed in Runx3^+/" intestine at about 65 weeks of age. Analyses of Runx3^+/~ intestine will be described below. [0145] The proliferation of intestinal epithelial cells is known to be accelerated by the

Wnt-β-catenin/Tcf4 pathway. Therefore, the inventors examined whether Wnt signaling is activated in Runx3^{~ '} intestinal epithelial cells. Target genes known to be positively regulated by β-catenin/Tcf4, such as CD44, cyclin Dl, c-Myc, Conductin, and EphB2 (He et al, 1998; Tetsu and McCormick, 1999; Batlle et al., 2002; van de Wetering et al., 2002) were up-regulated in the ileum, jejunum and colon (Fig. 2G, H, J and Fig. 1OB, C), while a gene negatively regulated by β-catenin/Tcf4, EphrinBl (Batlle et al., 2002), was down-regulated in epithelial cells of the Runx3^'u intestine (Fig. 2H, J).

[0146] It has been demonstrated that the EphB/EphrinB system controls the positioning of epithelial cells within the small intestinal mucosa (Batlle et al., 2002). Enhancement of β-catenin/Tcf4 activity in the 4£">defϊcient small intestine where EphB/EphrinB system is dysregulated also causes displacement of epithelial cells (Sansom et al., 2004; Andreu et al., 2005). This phenomenon can be clearly recognized by the random localization of Paneth cells, which are normally tightly clustered at the bottom of the gland (Batlle et al., 2002). The inventors observed that Paneth cells were distributed throughout the villi in the Runx3^'/" small intestine (Figure 21). It is noteworthy that the levels of β-catenin and Tcf4 were not noticeably altered in Runx3^~/~ compared to wild type epithelial cells of the small and large intestines (Figure 2J), indicating that the increase in Wnt signaling activity in Runx3^'A intestinal epithelial cells is not due to an increase in the levels of these proteins. It is worth noting that Runx3^'/' FID cells showed higher sensitivity to stimulation by Wnt3a than that of Runx3^+/+ cells (Fig. 3K). This higher TOP/FOP activity in Runx3^'A cells after Wnt3a treatment was supported by an electrophoresis mobility shift assay (EMSA) which showed higher amounts of the β-catenin/Tcf4 complex bound to the TCF site after Wnt3a treatment (Figure 12B). These results suggest that higher Wnt signaling activity observed in Runx3^~'^~ epithelial cells of intestine is not simply due to increased cell population which has higher Wnt activity.

[0147] Taken together, Runx3 appears to regulate Wnt signaling activity negatively.

RUNX3 forms a ternary complex with β-catenin and TCF4

[0148] Since β-catenin and Tcf4 levels were unaffected in Runx3^~f~ intestinal epithelial cells

(Fig. 2J)₃ the possibility that RUNX3 directly inhibits the function of β-catenin/TCF4 was examined. For the mechanistic analysis of their observations, the inventors examined 22 well-characterized colon cancer cell lines for the expression of RUNX3 (Figure 3A). Only 8 cell lines, HCTl 16, SW480, COLO320, SW403, SW837, CCK81, SW620 and RCMl, express RUNX3 at various levels. It was found that exogenously-expressed TCF4, β-catenin, and RUNX3 could be co-immunoprecipitated in HCTl 16 cells (Fig. 3B) and 293T cells (data not shown) and that they formed a ternary complex as revealed by the two step-coimmunoprecipitation (Figure 3C). The endogenous ternary complex was also detected in DNA-free nuclear extracts of HCTl 16 and SW620 cells, which express dephosphorylated (activated) β-catenin and TCF4 as well as RUNX3 in the nucleus (Figure 3D). RUNX3 that is co-immunoprecipitated with TCF4 migrates slower than those immunoprecipitated with anti-RUNX3 or anti-β-catenin. The nature of this RUNX3 subspecies is unknown at the moment. Of note, RUNX3 and TCF4 did not interact in SW480 cells (see below for significance of this results). Direct interactions between RUNX3 and TCF4 and between RUNX3 and β-catenin were also observed in a cell-free system (Fig. 3E). Since β-catenin and TCF4 directly interact, the obtained results suggest that each component in the ternary complex interacts directly with each of the other components. Furthermore, mapping experiments revealed that the Runt domain of RUNX3 and the HMG box of TCF4, which are DNA binding domains, are required for interaction between RUNX3 and TCF4 (Figure HA, B). This result is consistent with an earlier report that the interaction between Runx2 and LEFl occurs through their respective DNA binding domains (Kahler and Westendorf, 2003). RUNX3 was also found to interact with TCFl, TCF3, and Lefl (Figure 11C). Interestingly, comparatively larger amounts of the oncogenically active β-catenin mutants Δ45 and S33Y (Morin et al., 1997), interacted with RUNX3 than wild type β-catenin (Figure 3F), suggesting either that RUNX3 specifically interacts with an activated form of β-catenin or that the mutant forms are more stable.

RUNX3 attenuates the trans activational potential of β-catenin/TCF4 in Wnt signaling [0149] To elucidate the consequence of the interaction of RUNX3 with β-catenin/TCF4, the transactivation activity of β-catenin/TCF4 was examined using a TOP/FOPflash reporter system. In DLDl, increasing amounts of exogenous RUNX3 progressively repressed the relatively high TOP activity (Fig. 3G). In the same cells, the basal level of the cyclin Dl promoter significantly depended on the presence of a TCF binding site (Fig. 3 J; Lin et al., 2000).

Increasing amounts of exogenous RUNX3 progressively repressed wild type promoter activity to a level comparable to that of a promoter lacking the TCF binding site (Fig. 3J). It was confirmed that the reporter construct lacks RUNX binding sites. These results clearly demonstrate that RUNX3 represses cyclin Dl promoter activity by inhibiting the transactivational potential of β-catenin/TCFs.

[0150] RUNX3 up-regulates _p21^WAF1/Cipl and inhibits cell growth (Chi et al., 2005). It seems, therefore, RUNX3 has two functions: one as a transcription factor with DNA binding activity and the other as an attenuator of β-catenin/TCF4 without involving DNA binding. RUNX mutant, RUNX3(R178Q) lacks DNA binding ability and hence transactivation activity (Inoue et al., 2007). This mutant was found to interact with β-catenin and TCF4 (Figure 3H) and, similar to wild type RUNX3, attenuate β-catenin/TCFs transactivation (Figure 31), suggesting strongly that RUNX3 has a role other than DNA binding transcription factor and that RUNX3 directly inhibits the activity of β-catenin/TCF4.

RUNX3 attenuates the DNA binding activity of β-catenin/TCF4

[0151] A ChIP assay using DLDl cells revealed that β-catenin/TCF4 efficiently binds to consensus TCF binding sites in the cyclin Dl and c-Myc promoters (Fig. 4A, lanes 4 and 7), as previously reported (Nateri et al., 2005). When RUNX3 was stably expressed in DLDl cells, the ability of β-catenin/TCF4 to bind either promoter was greatly reduced (Fig. 4A, lanes 3 and 6, and Fig. 4D) and this was accompanied by the reduction of c-Myc and cyclin Dl proteins (Fig. 4E) and the TOP/FOP luciferase activity (Fig. 4G). In contrast, when RUNX3 expression was inhibited by antisense RUNX3 in HCTl 16 cells, the binding of β-catenin/TCF4 to these promoters was enhanced (Fig. 4B, D) with concomitant up-regulation of c-Myc and cyclin Dl (Fig. 4E) and TOP/FOP luciferase activity (Fig. 4G). Since the levels of TCF4 and dephosphorylated β-catenin were unaffected in both cases (Fig. 4E), the inhibition of β-catenin/TCF4 binding to these promoters depended mainly on the level of RUNX3 expression (Fig. 4C). Consistent with these observations, exogenous expression of RUNX3 down-regulated target genes of canonical Wnt signaling, AXIN2, CD44, and DKKl in DLD 1 , while knock-down of RUNX3 in HCTl 16 cells achieved the opposite effect (Fig. 4F). Using 3 independent shRNAs to knockdown RUNX3 in HCTl 16, SW620, COLO320, SW480, and SW403, the consistent inhibitory effect of RUNX3 on the TOP/FOP luciferase activity in HCTl 16, SW620 and COLO320 cells was observed (Fig. 4H). However, knockdown of RUNX3 in SW480 and SW403 did not change the TOP/FOP ratio appreciably. Interestingly, RUNX3 in SW480 did not interact with TCF4 (Fig. 3D). In the case of SW403, RUNX3 is mislocalized to the cytoplasm (Figure 9D). These results further strengthen the conclusion that RUNX3 attenuates Wnt activity by interacting with β-catenin/TCF4 (see Discussion).

[0152] Additionally, an electrophoretic mobility shift assay (EMSA) was performed using a DNA probe containing a consensus TCF site. The results showed that RUNX3 inhibited the binding of TCF4 and β-catenin/TCF4 to this site in a dose-dependent manner (Figure 12A). Thus, it can be concluded that RUNX3 inhibits the transactivation of cyclin Dl, c-Myc, and possibly other Wnt target genes by inhibiting the DNA binding ability of β-catenin/TCF4. Conversely, it was also found that RUNX3 loses its affinity for a consensus RUNX site when bound to β-catenin/TCF4 (Figure 12C). These results suggest that RUNX3 and TCF4 mutually inhibit their respective DNA binding activities through interactions involving their DNA binding domains.

Adenomatous polyps and adenocarcinomas are induced in the Runx3^+/~ and Runx3^^'Apc^Mιn/+ intestines, respectively

[0153] Since the Wnt signaling pathway is a well-known oncogenic pathway and RUNX3 is a nuclear effector of the well-established TGF-β tumor suppressor pathway, an intriguing possibility is that oncogenicity in intestinal epithelial cells reflects the relative activities of these two antagonistic pathways. The gastrointestinal tract (GIT) epithelium of Runx3^+/~ mice was indistinguishable from that of Runx3^+/+ mice for a year. However, at around 65 weeks of age, small adenomas developed in the small intestine at a frequency comparable to that of Apc^Mm/+ mice of the BALB/c background (54% of Runx3^+/~ mice and 64% of Apc^Mn/+ mice at 65 weeks; Figure 5A, B). All intestinal tumors tested from Runx3^+/~ (n=54) and Apc^Mm/+ (n=22) mice were adenomatous polyps. Adenomatous polyps were found not only in Runx3^+/~ small intestine with BALB/c background, but also in those with other genetic background, such as C3H/HeJ at 65 weeks of age (55% of Runx3^+/" mice, 6 out of 11 mice, developed 2.2 ± 0.9 tumors).

[0154] Loss of heterozygosity (LOH) of the Ape allele is observed in the adenomatous polyps in Apc^Mw/+ mice, indicating that biallelic inactivation of Ape allele, with ensuing nuclear accumulation of oncogenic β-catenin, occurs frequently (Polakis, 2000). Similarly, the inventors found that LOH of the Ape allele in adenomatous polyps of their cases (Figure 13B). In the case of adenomatous polyps in Runx3^+/~ mice, down-regulation of Runx3 expression was observed (Figure 6 A, C). However, LOH of the Runx3 allele was not detected in Runx3^+/~ polyps (Figure 13A). Instead, CpG island methylation of the Runx3 promoter (Guo et al., 2002) in Runx3^+/' polyps was detected, suggesting that methylation of the Runx3 promoter is one of the causes of Runx3 repression in adenoma cells (Figure 13C).

[0155] Interestingly, adenomatous polyps found in both Runx3^+/~ and Apc^Mιn/+ small intestines displayed up-regulation of cyclin Dl and c-Myc (Figure 6 A, C, D). In both cases, down-regulation of p21 was also observed (Figure 6C, D), which is consistent with earlier report that p21 can be repressed by β-catenin/TCF4 (van de Wetering et al., 2002). In both types of tumors, DNA binding of β-catenin/Tcfs on the Tcf consensus site in cyclin Dl and c-Myc promoters was enhanced despite the fact that obvious nuclear accumulation of β-catenin was observed only in Apc^Mιn/+ mice (Figure 6E-G). These results show that the nuclear accumulation of oncogenic β-catenin in Apc^Mm/+ mice and the down-regulation of Runx3 in Runx3^+I" mice induce a common phenotype (up-regulation of c-Myc and cyclin Dl) and cause a comparable level of tumorigenicity in the experimental system used here. To confirm that the tumorigenicity of cells deficient in Runx3 activity is dependent on the oncogenic activity of β-catenin/Tcf4, Runx3^~'^~ FID and FIL cells were used. Only Runx3^~'^~ FID and FIL cells, but not Runx3^+/+ cells, formed tumors in inoculated mice (Figure 17C). Runx3^'/' FID and FIL cells stably expressing a dominant negative form of TCF4 (van de Wetering et al., 2002) showed that the tumorigenicity of Runx3^~'^~ FID and FIL cells was indeed attenuated by inhibition of β-catenin/Tcf4 (Figure 17D).

[0156] Since RUNX3 and β-catenin/TCF4 antagonize each other, inactivation of Runx3 in Apc^Mm/+ mice should enhance carcinogenic activity in the intestine. In the Runx3^+/' and Apc^Mιn/+ mice, adenomatous polyps (0.2 mm or larger in diameter) were induced in the small intestine, but not in the large intestine. In the Runx3^+/'Apc^Min/+ mice, on the other hand, invasive adenocarcinomas (generally larger than 5 mm in diameter) were induced in small intestine (Figure 5 A, D) and adenomatous polyps formed in the large intestine (33.3% of 21 mice; 1 or 2 tumors per mouse; Figure 5B) at 65 weeks of age. The frequency of tumor formation, the number of tumors per mouse, and tumor size were all increased in the small intestine of Runx3^+/-Apc^Min/+ mice (Figure 5B-D).

[0157] Examination of very small adenomas collected from Runx3^+/'Apc^^l^⁺ mice showed either nuclear accumulation of β-catenin or loss of expression of RunxS, but not both changes simultaneously (Figure 6B). The result is again consistent with the interpretation that adenomas were induced by biallelic inactivation of either the Ape or the Runx3 allele. This observation explains why the number of tumors in the Runx3^+/'Apc^Mιn/+ small intestine was nearly equal to the sum of Apc^Mm/+ and Runx3^+/~ tumors (Figure 5C). When larger adenomas or adenocarcinomas with reduced Runx3 expression were examined, β-catenin was also accumulated in most of the cases, suggesting that alteration of both genes results in stronger activation of Wnt pathway.

[0158] The extent of tumorigenicity conferred by the Apc^+/~ genotype varies in different strains, with a higher incidence in C57/B6 mice than others (Shoemaker et al., 1997). The effect of C57/B6 genetic background on the tumorigenicity of Runx3^{+ '} and Apc^{Mm +} was examined using the Fl mouse line in the BALB/c:C57/B6 background. Although tumorigenicity of Apc^Mml+ was much enhanced as expected, that of Runx3^+/~ was not observed within 25 weeks after birth. Nevertheless, the Runx3^+/~ status significantly promoted Apc^Mm/+ tumorigenicity. The numbers of tumors in the intestine of Apc^Mιn/+ and Rumc3^+/~Apc^Mιn/+ mice were comparable (Figure 14B) at about 25 weeks of age. However, tumor size and the frequency of tumor formation in the small and large intestines, respectively, were enhanced in Runx3^+/'Apc^{Mιn +} mice of the BALB/c:C57/B6 background (Figure 14A, C, D).

[0159] In summary, as in the case of the Apc^Mm/+ mouse, the Runx3^+/~ mouse is an excellent model for studying oncogenesis, especially the model of intestinal oncogenesis particularly in its early stages.

RUNX3 is frequently down-regulated in human adenomatous polyps without accumulation of β-catenin

[0160] Since biallelic inactivation of Runx3 and Ape independently formed adenomatous polyps in mice, sporadic human colon adenomatous polyps were examined to see whether similar observations can be made. 10 of 35 (29%) cases tested showed down-regultation of RUNX3 without nuclear/cytoplasmic accumulation of β-catenin (type B in Figure 7B and Figure 8B), while 15 of 35 (43%) cases showed accumulation of β-catenin without inactivation of RUNX3 (type A in Figure 7 A and Figure 8A). Remaining cases showed membranous β-catenin without inactivation of RUNX3 (type C in Figure 7C and Figure 8C). In both A and B types of adenomas, up-regulation of cyclin D 1 and c-Myc was observed consistently (Figure 7D, E). Cyclin Dl and c-Myc were up-regulated in 100% (13/13) and 92% (12/13), respectively, of type A polyps and 80% (8/10) and 100% (10/10), respectively, of type B polyps. A target gene of canonical Wnt signaling, CD44, was also up-regulated in 100% (10/10) of both A and B types (Figure 15 A, B). In type C, on the other hand, cyclin Dl and c-Myc were up-regulated in 67% (6/9) and 67% (6/9), respectively (data not shown). Therefore, Wnt target genes are up-regulated in two thirds of the type C cases, suggesting that another mechanism of Wnt signaling activation exists. For the remaining one third cases, alteration of signaling pathway other than Wnt might have to be explored.

[0161] In agreement with the mouse cases, methylation of the R UNX3 promoter region was detected in all cases that showed down-regulation of RUNX3 (Figures 6F, G and 17). With the methylation specific PCR (MSP) technique, however, the unmethylated allele was frequently observed in addition to the methylated allele (Figure 7F). This is because DNA wasisolated from small tumors which contain some healthy (normal) surrounding tissues in which RUNX3 promoter is not methylated. In some adenomas in which expression of RUNX3 was detected, methylation was also detected (Fig. 18). This is most likely due to a partial methylation of the promoter region. It is likely that RUNX3 is biallelically inactivated in human colon adenomatous polyps where RUNX3 expression is lost without β-catenin accumulation (type B), since mono- -allelic inactivation of Runx3 per se does not eliminate the expression of Runx3 in mice.

RUNX3 is frequently inactivated in human colorectal cancers with concomitant accumulation of β-catenin

[0162] As colon adenomas do not always progress to carcinomas (Kinzler and Vogelstein, 1996), the inventors examined whether RUNX3 is inactivated in colon carcinoma specimens.

Consistent with earlier reports by others (Kinzler and Vogelstein, 1996; Bodmer, 2006), nuclear/cytoplasmic accumulation of β-catenin was observed in 45 out of 48 (94%) cancer specimens tested (Figure 9A-C, and Fig. 19). On the other hand, RUNX3 was found to be inactivated by gene silencing in 14 out of 48 (29%) cases (Figure 9B and 'N' in Fig. 19), and by the cytoplasmic expression of RUNX3, an another mechanism of RUNX3 inactivation (Ito et al., 2005), in 7 out of 48 (15%) cases of colon cancer (Figure 9C and entry 'C in Fig. 19).

RUNX3, therefore, is inactive at least in 44 % of colon cancers. Most of these specimens showed nuclear/cytoplasmic accumulation of β-catenin (Fig. 19). This is in contrast to human colon adenomas in which there were none that showed nuclear/cytoplasmic accumulation of β-catenin and RUNX3 inactivation simultaneously so far tested (Figure 7). These results altogether suggest that adenomas induced by inactivation of RUNX3 will progress to carcinomas and, during this progression period, nuclear/cytoplasmic accumulation of β-catenin appears to take place. Even in the specimens where RUNX3 is expressed in the nuclei, RUNX3 may not necessarily be functional as a cytostatic protein as mentioned above. RUNX3 was not detected in 14 out of 22 human colorectal cancer-derived cell lines (Fig. 3A) and excluded from the nucleus in SW403 and CCK81 cells (Figure 9D). Therefore, the RUNX3 inactivation by gene silencing and protein mislocalization is prevalent in human colorectal cancer samples (44%; 21/48) and cell lines (73%; 16/22, 77%; 17/22 including the SW480 case without interaction between RUNX3 and TCF4). Cell growth inhibitory and tumor suppressive effects of RUNX3 in human colorectal cancer-derived cell lines, DLDl (RUNX3 -negative) and HCTl 16 (RUNX3 -positive), both of which are β-catenin-activated cell lines, were confirmed by exogenous expression of RUNX3 and knock-down of RUNX3, respectively (Figure 17A, B).

[0163] Contrary to the present observations, there are reports that there is no Runx3 expression in the epithelial cells of intestines and that the hyperplasia observed in the Runx3^"/' large intestine was due to passive effects of inflammation caused by loss of Runx3 function in leukocytic populations (Brenner et al., 2004). To confirm that loss of Runx3 function in intestinal epithelial cells is responsible for their tumorigenic potential, the tumorigenicity of Runx3^'/' small and large intestinal epithelial cells was examined by two independent methods. In the first method, wild type mice were irradiated and then transplanted with the bone marrow cells of Runx3^'A mice to prepare mice whose leukocytes are Runx3^'A but epithelial cells are Runx3^+/+. Histological examination of small and large intestines of these mice after one year of the transplantation revealed neither hyperplasia nor dysplasia (Figure 16 A-E). In the second method, the tumorigenicity of Runx3^+/+ and Runx3^~/~ FID and FIL cells was compared in nude mice. As mentioned earlier and shown in Figure 17C, only Runx3^'/' cells formed tumors. The results of these two sets of experiments establish that the tumorigenic activity of intestinal epithelial cells of Runx3^{' '} mice is epithelial cell-autonomous. The present results, however, do not necessarily exclude the possibility that leukocytes or other mesenchymal cells may exercise additional effects on the growth of epithelial cells when the intestinal epithelium is encompassed by these cells as reported for Smad4 (Kim et al., 2006; Pan et al., 2007).

DISCUSSION

[0164] Through in vivo tumorigenesis studies on mice with heterozygously inactivated

Runx3 and Ape, it was found that biallelic inactivation of each of these genes can independently induce small intestinal adenomas. In this study, more than half of sporadic human colon adenomas were found to be induced by a mechanism(s) not involving the oncogenic activity of β-catenin. About a half of the sporadic colon adenomas that did not show accumulated β-catenin expressed RUNX3 at greatly reduced level due to hypermethylation of the RUNX3 promoter. It can be concluded that biallelic inactivations of APC and RUNX3 are independently able to induce human colon adenomas.

[0165] The evidence presented here suggests that colon adenomas initiated by inactivation of RUNX3 are likely to progress to malignancy. While RUNX3 was observed to be inactivated in 77 % of colon cancer-derived cell lines, more than half of the colon cancer specimens showed nuclear localization of RUNX3, apparently suggesting that RUNX3 is active in these cases. However, RUNX3 may be inactive in some of these cases (see below). Further studies are required to reveal how extensively RUNX3 is inactivated in colon cancer. [0166] In the adenomas developed in Runx3^+/~ intestine, it was observed that the Wnt target genes are up-regulated without concomitant accumulation of β-catenin. The results presented here suggest that Runx3 functions as a "brake" of Wnt signaling pathway and inactivation of Runx3 per se would not constitutively "switch on" Wnt pathway. In fact the inventors observed that the amounts of TCF4 and β-catenin were comparable in Runx3^+/+ and Runx3^~'^~ intestine with no nuclear accumulation of β-catenin. It can be assumed that Wnt signaling is abnormally activated in Ruwc3^'A epithelial cells only when the Wnt ligand stimulates cells in vivo. Thus, without ligand stimulation, the Wnt pathway would be at the physiological state. Overall, the level of enhancement of Wnt signaling would be moderately high. In the case of APC inactivation or β-catenin activation, Wnt signaling will be "switched on" constitutively and the target gene expression (output of Wnt pathway) would be abnormally high but somewhat curbed if Runx3 is present. When nuclear accumulation of β-catenin and inactivation of RUNX3 occur in the same cells simultaneously, the output would be very strong and oncogenic. The results of the present analysis of colon cancer specimens are consistent with this prediction.

[0167] Of the 5 RUNX3 positive cell lines that were tested, knockdown of RUNX3 in HCTl 16, SW620 and colo320 resulted in the increase of TOP/FOP ratios. In the case of

HCTl 16, it was confirmed that the cell growth and tumorigenicity were accelerated when

RUNX3 was reduced (Figure 17A, B). Therefore, RUNX3 in these cell lines has cytostatic ability. However, knockdown of RUNX3 in SW480 and SW403 did not show significant increase of the TOP/FOP ratio due to a lack of interaction between RUNX3 and TCF4 and protein mislocalization, respectively (Fig. 4H), emphasizing the importance of the interaction between these two transcription factors for attenuation of Wnt signaling by RUNX3.

Preliminary results suggest that the interaction between RUNX3 and TCF4 is regulated by another signaling pathway that appears to be defective in SW480. It is possible, therefore that while some colon cancer specimens show RUNX3 nuclear localization, RUNX3 may lack the ability to interact with TCF4. This possibility must be further investigated.

[0168] When adenomatous polyps were induced in the Runx3^+f~ intestine, the remaining intact Runx3 allele was silenced in majority of the polyps tested by Runx3 promoter methylation. This result suggests that inactivation of Rwvc3 begins at an early stage of carcinogenesis.

Recently, a distinct trait referred to as CpG island methylator phenotype (CIMP) has been suggested as one of the mechanisms to lead to sporadic colorectal cancer (Toyota et al., 1999).

As many as 35-40% of colorectal cancer were reported to be classified as CIMP (Goel et al.,

2007) and RUNX3 was identified as one of the five in a marker panel that most strongly satisfies the CIMP (Weisenberger et al., 2006).

[0169] The transgenic mice over-expressing DNA methyltransferase Dnmt3b show the enhancement of tumorigenesis in Apc^Min/+ mice (Linhart et al., 2007). In tumors developed in these mice, tumor suppressor genes, Sfrp2, Sfrp4 and Sfrp5 are methylated and silenced, whereas genes often methylated in cancer cells, Mlhl, Mgmt, Cdkn2b, Ape, RbI, VhIh and Brcal are not. Considering that Sfrp, a secreted frizzle-related protein, interferes with Wnt signaling and that Linhart et al also noted reduced Runx3 expression in these tumors, their report strongly supports the conclusion drawn here in that Runx3 is a downstream attenuator of Wnt signaling cascade. It is intriguing to note that attenuators of Wnt signaling at upstream and downstream ends are apparently targeted for methylation and silenced during oncogenic development. It would be important to study whether Runx3 is a target of Dnmt3b during the early stage of carcinogenesis.

[0170] Somewhat surprisingly, Runx3^~'^~ mice never developed epithelial tumors, unlike Runx3^+/' mice. There is a precedent, however: certain threshold levels of PU.1, C/EBPα or GATA-I expression are required for carcinogenesis (reviewed by Rosenbauer et al., 2005). This observation would be in line with those observed in the leukemia cases.

[0171] Since β-catenin and TCF4 are downstream effectors of the canonical Wnt signaling pathway and RUNX3 is a nuclear effector of the TGF-β superfamily, these results suggest that proliferation signals transmitted by the Wnt pathway and anti-proliferation signals transmitted by the TGF-β family pathway culminate in the ternary complex formed by β-catenm/TCF4 and RUNX3. This ternary complex would function to harmoniously co-regulate gut development and maintain the normal proliferative state of the gut epithelium, m this study, it was shown that the DNA binding negative mutant of RUNX3 can attenuate β-catenin/TCF4 activity. Therefore, a function of RUNX3 was uncovered that is independent of the activity as a transcription factor. This activity, as an attenuator of Wnt signaling pathway, seems to have a profound role in gut development, homeostasis and cancer.

[0172] Gut development, and intestinal stem cell maintenance and differentiation are regulated by interactions between key signaling pathways. The observations made in this study provide a significant insight into the interaction between the Wnt and TGF-β superfamily pathways in intestinal tumorigenesis.

EXPERIMENTAL PROCEDURES

Cell lines, mouse lines and RUNX3 shRNAs

[0173] Colorectal cancer cells were maintained in DMEM medium supplemented with 10% fetal bovine serum. Cells were transfected with pcDNA3, pcDNA-Flag-RUNX3, or pEF-BOS-neo-RUNX3-AS as described previously (Ito et al., 2005). Stable transfectants were selected in the presence of 0.5 mg/ml G418 (GIBCO). An ecdysone-inducible Flag-RUNX3 clone of DLDl was established as described previously (Yamamura et al., 2006). shRNAs targeting RUNX3 (shl : gcccagagaagatgagtctat, SEQ ID NO: 1 ; sh2: aagcagctatgaatccattgt, SEQ ID NO: 2; sh3: tcagtagtgggtaccaatctt, SEQ ID NO: 3) and the control shRNA were obtained from SuperArray Bioscience (Maryland). pGeneClip-hMGFP (Promega) was used as the vector for the transfection. Sorted cells with the GFP expression were subjected to the Western blot analysis, and were transfected with reporter plasmids; TOP/FOPflash (Upstate). [0174] Immnunocytochemistry to detect RUNX3 in colorectal cancer cell lines was performed using anti-RUNX3 (MBL; R3-6E9) antibody as described previously (Ito et al., 2005).

[0175] Mouse small and large intestinal epithelial cell lines, FID and FIL, respectively, were established from isolated intestinal epithelium of 16.5 dpc Rwvc3^+/+p53^'A and RunxS^'^pSS^'^ fetuses in C57BL/6J background and maintained as described previously for similarly obtained mouse gastric epithelial cell lines (Li et al., 2002).

[0176] To generate i?«nx3-deficient and Apc^Mm/+ mice of the BALB/c background, male

Runx3^+/~ mice of the C57BL/6J background (Li et al., 2002) and C57BL/ 6 J -Min/+ mice

(Jackson Laboratory) were back-crossed with wild type BALB/c female mice for at least six generations. AU animal studies were done in accordance with the guidelines of the Institutional

Animal Care and Use Committee in Singapore. [0177] To generate Runx3^+A and Apc^min/+ mice of the BALB/c:C57/B6 background (Fl), female Runx3^{+ "}Apc^{mιn +} mice of the BALB/c background were crossed with wild type male mice of the C57/B6 background. Rwu3^+/~, Apc^min/+, and Runx3^+/'Apc^min/+ mice of the

BALB/c:C57/B6 background used in this study were offspring of Fl Apc^mιn/+ male and Fl Runx3^+/~ female mice.

Bone marrow reconstitution assay

[0178] Bone marrow cells were collected from Runx3^'A and Runx3^+/~ neonate C57BL/6J mice and transplanted into wild type EGFP -transgenic C57BL/6J mice after irradiation. Chimerism was measured by the donor-derived marker in peripheral bloods of the recipient at 6 weeks and one year after transplantation. Studies were done in accordance with the guidelines of

Institutional Animal Care and Use Committee in Singapore.

Histological analysis

[0179] Tissues were fixed with 10% formalin (for human tissues) or 4% paraformaldehyde

(for mouse tissues), embedded in paraffin, and cut into 5 μm sections. Anti-RUNX3 (MBL; R3-1E10 for mouse; Yano et al., 2006 and MBL; R3-6E9 for human; Ito et al, 2005), anti-Ki67 (DAKO; M7249), anti-c-Myc (Santa Cruz; sc-764 for mouse and Upstate; 06-340 for human), anti-EphrinBl (Santa Cruz; sc-910), anti-EphB2 (R&D; AF467), anti-lysozyme (DAKO; A0099), anti-cyclin Dl (Zymed; 13-4500 for mouse and Novocastra; NCL-CYCLIN Dl-GM for human), and anti-β-catenin (Santa Cruz; sc-7199) antibodies were used on rehydrated sections pretreated with Target Retrieval Solution (DAKO). Anti-CD44 (ENDOGEN; MA-4405 for mouse and Santa Cruz; sc-7297 for human) and anti-cyclin Dl (Zymed; 13 -4500) antibodies were used for the immunodetection on rehydrated sections pretreated with a Target Retrieval Solution (DAKO). An Envision™+ system (HRP/DAB) (DAKO) was used for visualization. An Envision™+ system (HRP/DAB) (DAKO) was used for visualization. BrdU incorporation and cellular proliferation were detected with the BrdU Labeling and Detection Kit II (Roche).

Human polyps and cancers

[0180] All the polyps tested were adenomas resected during colonoscopic examination in the context of regular screening. They were left and right sided in approximately similar proportions. The carcinomas were surgically resected. None of the patients with adenomas or carcinomas had a clinical presentation (including family history) reminiscent of inherited carcinoma, neither had histological features to suggest an inherited patter. Indeed, from both clinical and pathological viewpoints, the adenomas and carcinomas were sporadic in nature. Furthermore, the patients had no other predisposing factor and, in particular, no inflammatory bowel disease.

Western blot analysis and immunoprecipitation

[0181] 30 μg of proteins extracted from colorectal cancer cell lines, and mouse intestinal tissues with or without reduction were analyzed by Western blot using anti-RUNX3 (MBL; R3-5G4; Ito et al., 2005), anti-dephosphorylated β-catenin (Alexis; ALX-804-260 for human and Upstate; 05-665 for mouse), anti-TCF4 (Upstate; 05-511), anti-CD44 (ENDOGEN; MA-4405), anti-c-Myc (Santa Cruz; sc-40 for human and Santa Cruz; sc-764 for mouse), anti-cyclin Dl (BD biosciences; 556470), anti-Conductin (Santa Cruz; sc-8570), anti-EphrinB 1 (R&D; AF473), anti-EphB2 (R&D; AF467), and anti-β-actin (Sigma, A5441) antibodies.

[0182] Whole cell extracts of HCTl 16 or 293T cells expressing 6Myc-tagged TCF4, Flag-tagged RUNX3, Flag-tagged RUNX3(R178Q) (Inoue et al, 2007), HA-tagged β-catenin, β-catenin (WT), β-catenin (Δ45; Morin et al., 1997), or β-catenin (S33Y; Morin et al., 1997) were immunoprecipitated with anti-Myc (Santa Cruz; sc-789), anti-HA (Santa Cruz; sc-805), or anti-β-catenin (BD Biosciences; 610153) antibodies with Protein G Sepharose 4 Fast Flow (Amersham) or anti-FLAG M2 agarose (Sigma; A2220), followed by Western blot analysis using anti-Myc (Santa Cruz; sc-40), anti-HA (Santa Cruz; sc-7392), anti-β-catenin (BD Biosciences; 610153), or anti-FLAG (Sigma; F7429) antibodies. For two step-coimmnunoprecipitation, 3X FLAG peptide (Sigma; F4799) was used to elute proteins from anti-FLAG M2 agarose.

[0183] For detection of endogenous protein interactions, nuclear extracts were prepared from colorectal cancer cell lines using NE-PER Nuclear and Cytoplasmic Extraction Reagents (PIERCE) and treated with DNase I (Promega). Immunoprecipitation was performed using anti-RUNX3 (MBL; R3-5G4), anti-TCF4 (Upstate; 05-511), or anti-dephosphorylated β-catenin (Alexis; ALX-804-260) antibodies or mouse normal IgG with Protein G Sepharose 4 Fast Flow, followed by Western blot analysis using the same antibodies as for immunoprecipitation. [0184] HA-tagged β-catenin, 6Myc-tagged TCF4, and 6His-tagged RUNX3 were translated in vitro using the TNT® T7 Quick Coupled Transcription/Translation System (Promega; Ll 170). Proteins pulled down by Ni-NTA agarose (QIAGEN; 30210) were revealed by Western blot analysis using anti-HA (Santa Cruz; sc-7392), anti-Myc (Santa Cruz; sc-40), and anti-His (Clontech; 631212) antibodies.

Reporter assay

[0185] The cyclin Dl promoter construct with a mutated TCF binding site

(CyclinDl-mTCF) was made by mutagenizing the TCF consensus sequence located near the nucleotide -80 (from CTTTGATC to CTTTGGCC) in DlΔ-944pXP2 (CyclinDl-WT; Herber et al., 1994) using the QuikChange XL Site-Directed Mutagenesis Kit (Stratagene). DLDl cells were transfected with reporter plasmids; TOP/FOPflash (Upstate), CyclinDl-WT, or CyclinDl-mTCF, along with pRL-TK (Promega) and effector plasmids; pcDNA3, pcDNA-Flag-RUNX3, or pcDNA-Flag-RUNX3 (Rl 78Q) using FuGENE 6 (Roche). Luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega) and normalized to the luciferase activity expressed by pRL-TK. In order to examine the β-catenin/Tcfs dependent reporter activities in response to Wnt3a stimulation, FID cells were transfected with TOP/FOPflash and treated with conditioned media collected from Wnt3a-expressing or parental L cell cultures (Shibamoto et al., 1998) for 24h.

ChIP

[0186] ChIP was performed using the Chromatin hnmunoprecipitation (ChIP) Assay Kit (Upstate) and anti-TCF4 (Santa Cruz; sc-13027), anti-dephosphorylated β-catenin (Alexis;

ALX-804-260) antibodies, or mouse normal IgG. The following primers were used for PCR amplification of DNA fragments containing the TCF consensus site: 5'-aggcgcggcggctca gggatg-3¹, SEQ ID NO: 4, and 5'-actctgctgctcgctgctact-3', SEQ ID NO: 5, for the human cyclin

Dl promoter (Nateri et al., 2005); 5'-ttgctgggttattttaatcat-3', SEQ ID NO: 6, and 5'-actgtttgacaaaccgcatcc-3', SEQ ID NO: 7, for the human c-Myc promoter (Nateri et al., 2005);

5'-cctcccccttttctctgccc-3', SEQ ID NO: 8, and 5'-cctctggaggctgcaggactttgc-3', SEQ ID NO: 9, for the murine cyclin Dl promoter; 5'-aatgcacagcgtagtattcagg-3', SEQ ID NO: 10, and

5'-aaaccgttaaccccttcctcc-3', SEQ ID NO: 11, for the murine c-Myc promoter. The following primers were used for negative controls; 5'-cgtcttcaccaccatggaga-3', SEQ ID NO: 12, and 5'-cggccatcacgcgacagttt-3', SEQ ID NO: 13, for human GAPDH gene (Nateri et al., 2005);

5'-ggggttgctgtgtcactaccg-3', SEQ ID NO: 14, and 5'-cagagacctgaatgctgcttcc-3', SEQ ID NO: 15, for murine Gapdh gene.

[0187] Quantitative PCR was performed using SYBR® Green PCR Kit (QIAGEN) and 7500 Fast Real-Time PCR System (Applied Biosystems). The following primers were used for the quantitative PCR amplification of DNA fragments of human gene promoters containing the TCF consensus site and not containing any TCF sites, respectively: 5'-ccctcccgctcccattc-3', SEQ ID NO: 16 / 5'-tacaggggagttttgttgaagttg-3', SEQ ID NO: 17, and 5'-gcagtcgctgagattctt tgg-3 ', SEQ ID NO: 18 / 5 '-agaatgggcgcatttcca-3 ', SEQ ID NO: 19, for the cyclin Dl promoter, and 5'-cccgtctagcacctttgatttc-3\ SEQ ID NO: 20 / 5'-tgttgcaaaccggcgc-3', SEQ ID NO: 21, and 5'-cggcagcccgagactgt-3', SEQ ID NO: 22 / 5'-tcagaagagacaaatcccctttg-3\ SEQ ID NO: 23, for the c-Myc promoter.

Quantitative RT-PCR

[0188] Quantitative RT-PCR was performed using RNeasy Kit (QIAGEN), Omniscript RT Kit (QIAGEN), 7500 Fast Real-Time PCR System (Applied Biosystems), and TaqMan® Gene

Expression Assays (Applied Biosystems; Hs00610344_ml for human AXIN2;

Hs00153304_ml for human CD44; Hs00183740_ml for human DKKl; Hs99999903_ml for human β-actin for normalization; Mm00490666_ml for murine Runx3; Mm00432359_ml for murine cyclin Dl; Mm00487803_ml for murine c-Myc; Mm00432448_ml for murine p21; Mm99999915_gl for murine Gapdh for normalization).

Methylation specific PCR (MSP) and bisulfite sequencing

[0189] Genomic DNA extracted by proteinase K digestion from rehydrated sections of human tissues or DLDl and HCTl 16 cells was treated with sodium bisulfite using the CpGenome DNA Modification Kit (Chemicon). PCR for experiments depicted in Fig. 7 was performed using primer sets; 5'-ataatagcggtcgttagggcgtcg-3', SEQ ID NO: 24, and 5'-gcttctact ttcccgcttctcgcg-3', SEQ ID NO: 25, for methylated DNA and 5'-ataatagtggttgttagggtgttg-3', SEQ ID NO: 26, and 5'-acttctactttcccacttctcaca-3', SEQ ID NO: 27, for unmethylated DNA as previously described (Homma et al., 2006). The PCR products were subjected to sequencing using each primer. Further primer sets used in experiments depicted in Figures 12, 17 and 18 for detection of methylated DNA were 5-ataaagagaaattaggcgc-3, SEQ ID NO: 28, and 5-ataaccc tcgaaaaacgcg-3, SEQ ID NO: 29 (M3), 5-gatgtttgtttaggtcgtagcggtc-3, SEQ ID NO: 30, and 5-ccaaactcgaaattcgccgta-3, SEQ ID NO: 31 (M2), and 5-tgcgattggttgcgtttcgc-3, SEQ ID NO: 32, and 5-cgaaaatacgcataccgcg-3, SEQ ID NO: 33 (Ml). Primer sets used in experiments depicted in Figures 12, 17 and 18 for detection of unmethylated DNA were 5-ataaagagaaattaggtgt -3, SEQ ID NO: 34, and 5 -ataaccctcaaaaaacaca-3, SEQ ID NO: 35 (M3), 5-tgtttgtttaggttgtagtggt tgt-3, SEQ ID NO: 36, and 5-cccccaaactcaaaattcaccata-3, SEQ ID NO: 37 (M2), and 5-tgtgattgg ttgtgttttgt-3, SEQ ID NO: 38, and 5-caaaaatacacataccaca-3, SEQ ID NO: 39 (Ml).

Immunoprecipitation and pull-down assay

[0190] For mapping of the interaction domains between RUNX3 and TCF4, proteins were immunoprecipitated from whole cell extracts of HCTl 16 or 293 T cells co-expressing either Flag-tagged RUNX3 derivatives with 6Myc-tagged full-length TCF4 or 6Myc-tagged TCF4 derivatives with Flag-tagged full-length RUNX3 using anti-Flag M2 agarose (Sigma; A2220) or anti-Myc (Santa Cruz; sc-789) with Protein G Sepharose 4 Fast Flow (Amersham), respectively, followed by Western blot analysis using anti-Myc (Santa Cruz; sc-40) or anti-Flag (Sigma; F7429) antibodies. [0191] For detection of interaction between RUNX3 and TCFl, Lefl, or TCF3, proteins were immunoprecipitated from whole cell extracts of HCTl 16 or 293 T cells expressing Flag-tagged RUNX3 with Myc-tagged TCFl, Lefl, or Myc-tagged TCF3 using anti-Flag antibody (Sigma; F7429), followed by Western blot analysis using anti-Myc (Santa Cruz; sc-789), anti-Lefl (Upstate; 05-602) or anti-Flag antibodies.

EMSA

[0192] EMSA was performed using the LightShift Chemiluminescent EMSA kit and a Chemiluminescent Nucleic Acid Detection Module (PIERCE). Each binding reaction (15 μl) contained 50 ng/μl poly (dl dC), 75 frnol labeled probe, and 3 μg nuclear extracts in the buffer supplied in the kit. Nuclear extracts were prepared from Runx3^+/+ and Runxi^'1' FID cells treated with 50% of Wnt3a conditioned medium (see Fig. 3K) and from 293T cells expressing Flag-tagged RUNX3, β-catenin (S33Y; Morin et al., 1997), or 6Myc-tagged TCF4 exogenously, or cells harboring pcDNA3, using NE-PER Nuclear and Cytoplasmic Extraction Reagents (PIERCE). Nuclear extracts and 0.1 μg 6His-tagged recombinant PEBP2β protein were incubated on ice for 15 min before addition of labeled probe. The binding mixture was incubated on ice for 20 min and resolved in a 5% polyacrylamide gel in 0.5x TBE buffer. For super-shift of bands, anti-Myc (Santa Cruz; sc-40), anti-dephosphorylated β-catenin (Alexis; ALX-804-260), anti-TCF4 (Santa Cruz; sc-13027), anti-RUNX3 (R3-5G4), anti-PEBP2β (MBL; D127-3), or mouse normal IgG were added after the binding reaction. The following 5' biotinylated oligonucleotides were used as labeled probes; 5'-gggggtaagatcaaagggggta-3', SEQ ID NO: 40 (TOP), 5'-gggggtaaggccaaagggggta -3', SEQ ID NO: 42 (FOP), IgCa-WT: 5'-acagccagaccacaggccagac -3', SEQ ID NO: 41, and IgCa-MT: 5'-acagccagaccctcggccagac-3', SEQ ID NO: 42. Anti-dephosphorylated β-catenin (Alexis; ALX-804-260), anti-TCF4 (Upstate; 05-511), and anti-RUNX3 (MBL; R3-1E10) antibodies were used in Western blot analysis for FID cells. Allelic loss analysis

[0193] To prepare genomic DNA for allelic loss analysis, tissues were microdissected using PALM MembraneSlides (P.A.L.M.) under a stereomicroscope and digested in 50 μl lysis buffer [5OmM Tris-HCl (pH 7.5), 10OmM NaCl, and 20 niM EDTA] containing 200 μg/ml proteinase K at 50⁰C for 3 hours, followed by heat inactivation at 95°C for 10 min. For Ape alleles, DNA amplified from 1 μl genomic DNA solution was digested by HindUI to produce 123 bp (Ape wt allele) and 144 bp (Ape mt allele) fragments, as reported previously (Luongo et al, 1994). For typing Runx3 alleles, the primer sets 5'-gactgtgcatgcacctttcaccaa-3', SEQ ID NO: 43, (forward) and 5'- atgaaacgccgagttaacgccatca -3', SEQ ID NO: 44 (reverse for Runx3 mt allele; 519 bp) or 5'- tagggctcagtagcacttacgtcg -3\ SEQ ID NO: 45 (reverse for Runx3 wt allele; 169 bp) were used in PCR reactions containing genomic DNA. Quantitative PCR was performed using SYBR® Green PCR Kit (QIAGEN) and 7500 Fast Real-Time PCR System (Applied Biosystems).

Statistics

[0194] Statistical evaluation was done by unpaired Student' s t-test or Mann- Whitney U-test. Data are given as mean ± s.e.m. P < 0.05 was considered statistically significant.

[0195] The listing or discussion of a previously published document in this specification should not necessarily be taken as an acknowledgement that the document is part of the state of the art or is common general knowledge.

[0196] The inventions illustratively described herein may suitably be practiced in the absence of any element or elements, limitation or limitations, not specifically disclosed herein.

Thus, for example, the terms "comprising", "including," containing", etc. shall be read expansively and without limitation. Additionally, the terms and expressions employed herein have been used as terms of description and not of limitation, and there is no intention in the use of such terms and expressions of excluding any equivalents of the features shown and described or portions thereof, but it is recognized that various modifications are possible within the scope of the invention claimed. Thus, it should be understood that although the present invention has been specifically disclosed by exemplary embodiments and optional features, modification and variation of the inventions embodied therein herein disclosed may be resorted to by those skilled in the art, and that such modifications and variations are considered to be within the scope of this invention.

[0197] The invention has been described broadly and generically herein. Each of the narrower species and subgeneric groupings falling within the generic disclosure also form part of the invention. This includes the generic description of the invention with a proviso or negative limitation removing any subject matter from the genus, regardless of whether or not the excised material is specifically recited herein.

[0198] Other embodiments are within the following claims. In addition, where features or aspects of the invention are described in terms of Markush groups, those skilled in the art will recognize that the invention is also thereby described in terms of any individual member or subgroup of members of the Markush group.

LIST OF REFERENCES CITED

[0199] Those references that have above only been indicated by means of first author and year are as follows:

Andreu, P., Colnot, S., Godard, C, Gad, S., Chafey, P., Mwa-Kawakita, M., Laurent-Puig, P., Kahn, A., Robine, S., Perret, C, et al. (2005). Crypt-restricted proliferation and commitment to the Paneth cell lineage following Ape loss in the mouse intestine. Development 132, 1443-1451.

Batlle, E., Henderson, J.T., Beghtel, H., van den Born, M.M.W., Sancho, E., HuIs, G, Meeldijk, J., Robertson, J., van de Wetering, M., Pawson, T., et al. (2002). β-catenin and TCF mediate cell positioning in the intestinal epithelium by controlling the expression of EphB/EphrinB. Cell 111, 251-263.

Blyth, K., Cameron, E.R., and Neil, J.C. (2005). The RUNX genes; gain or loss of function in cancer. Nature Rev. Cancer 5, 376-387.

Bodmer, W.F. (2006). Cancer genetics: colorectal cancer as a model. J. Hum. Genet. 51, 391-396.

Brenner, O., Levanon, D., Negreanu, V, Golubkov, O., Fainaru, O., Woolf, E., and Groner, Y. (2004). Loss of Runx3 function in leukocytes is associated with spontaneously developed colitis and gastric mucosal hyperplasia. Proc. Natl. Acad. Sci. USA 101, 16016-16021.

Chi, X.Z., Yang, J.O., Lee, K. Y, Ito, K., Sakakura, C, Li, Q.L., Kim, H.R., Cha, E.J., Lee, Y.H., Kaneda, A., et al. (2005). RUNX3 suppresses gastric epithelial cell growth by inducing p2l^WAF/Cipl expression in cooperation with transforming growth factor β-activated SMAD. MoI. Cell. Biol. 25, 8097-8107.

Goel, A., Arnold, C.N., Tassone, P., Chang D.K., Niedzwiecki, D., Dowell, J.M., Wasserman, L., Compton, C, Mayer, RJ., Bertagnolli, M.M., et al. (2004). Epigenetic inactivation of RUNX3 in microsatellite unstable sporadic colon cancers. Int. J. Cancer 112, 754-759.

Goel, A., Nagasaka, T., Arnold, C.N., Inoue, T., Hamilton, C, Niedzwiecki, D., Compton, C, Mayer, R.J., Goldberg, R., Bertagnolli, M.M., et al. (2007). The CpG island methylator phenotype and chromosomal instability are inversely correlated in sporadic colorectal cancer. Gastroenterology 132, 127-138.

Guo, W.H., Weng, L.Q., Ito, K., Chen, L.F., Nakanishi, H., Tatematsu, M., and Ito, Y. (2002). Inhibition of growth of mouse gastric cancer cells by Runx3, a novel tumor suppressor. Oncogene 21, 8351-8355.

Hanai, J.-i., Chen L.R, Kanno, T., Ohtani-Fujita, N., Kim, W.Y., Guo, W.-H., Imamura, T., Ishidou, Y., Fukuchi, M., Shi, M.-J., et al. (1999). Interaction and functional cooperation of PEBP2/CBF with Smads. J Biol. Chem. 274, 31577-31582.

He, T.-C, Sparks, A.B., Rago, C, Hermeking, H., Zawel, L., de Costa, L.T., Morin, P.J., Vogelstein, B., and Kinzler, K. W. (1998). Identification of c-MYC as a target of the APC pathway. Science 281, 1509-1512.

Herber, B., Truss, M., Beato, M., and Muller, R. (2004). Inducible regulatory elements in the human cyclin Dl promoter. Oncogene 9, 1295-1304.

Homma, N., Tamura, G, Honda, T, Matsumoto, Y., Nishizuka, S., Kawata, S., and Motoyama, T. (2006). Spreading of methylation within RUNX3 CpG island in gastric cancer. Cancer Sci. 97, 51-56.

Inoue, Ki., Ito, K., Osato, M., Lee, B., Bae, S.C., and Ito, Y. (2007). The transcription factor Runx3 represses the neurotrophin receptor TrkB during lineage commitment of dorsal root ganglion neurons. J. Biol. Chem. 282, 24175-24184.

Ito, K., Q. Liu, M. Salto-Tellez, T. Yano, K. Tada, H. Ida, C. Huang, N. Shah, M. Inoue, A. Rajnakova, et al. (2005). RUNX3, a novel tumor suppressor, is frequently inactivated in gastric cancer by protein mislocalization. Cancer Res. 65, 7743-7750.

Ito, Y. (2008). RUNX genes in development and cancer: regulation of viral gene expression and the discovery of RUNX family genes. Adv. Cancer Res. 99, 33-76. Jass, J.R., Whitehall V.L.J., Young, J., and Leggett B.A. (2002). Emerging concepts in colorectal neoplasia. Gastroenterology 123, 862-876.

Kahler, R. A. and Westendorf, JJ. (2003). Lymphoid enhancer factor- 1 and β-catenin inhibit Runx2-dependent transcriptional activation of the osteocalcin promoter. J. Biol.Chem. 278, 11937-11944.

Kim, B. G, Li, C, Qiao, W., Mamura, M., Kasperczak, B., Anver, M., Wolfraim, L., Hong, S., Mushinski, E., Potter, M., et al. (2006). Smad4 signalling in T cells is required for suppression of gastrointestinal cancer. Nature 441, 1015-1019.

Kinzler, K. W. and Vogelstein, B. (1996). Lessons from hereditary colorectal cancer. Cell 87, 159-170.

Ku, J.L., Kang, S.B., Shin, Y.K., Kang, H.C., Hong, S.H., Kim IJ., Shin, J.H., Han, I.O., and Park, J.G. (2004). Promoter hypermethylation downregulates RUNX3 gene expression in colorectal cancer cell lines. Oncogene 23, 6736-6742.

Li, Q.L., Ito, K., Sakakura, C, Fukamachi, H., Inoue, K.i., Chi, X.Z., Lee, K. Y, Nomura, S., Lee, CW. Han, S.B. et al. (2002). Causal relationship between the loss of RUNX3 expression and gastric cancer. Cell 109, 113-124.

Lin, S.Y., Xia, W, Wang, J.C., Kwong, K. Y, Spohn, B., Wen, Y, Pestell, R.G, and Hung, M.C. (2000). β-Catenin, a novel prognostic marker for breast cancer; Its roles in cyclin Dl expression and cancer progression. Proc. Natl. Acad. Sci. USA 97, 4262-4266.

Linhart, H.G., Lin, H., Yamada, Y, Moran, E., Steine, EJ., Gokhale, S., Lo, G, Cantu, E., Ehrich, M., He, T., et al. (2007). Dnmt3b promoters tumorigenesis in vivo by gene-specific de novo methylation and transcriptional silencing. Genes Dev. 21, 3110-3122.

Luongo, C, Moser, A.R., Gledhill, S., and Dove, WF. (1994). Loss of Apc⁺ in intestinal adenomas from Min mice. Cancer Res. 54, 5947-5952.

Morin, P.J., Sparks, A.B., Korinek, V., Barker, N., Clevers, H., Vogelstein, B., and Kinzler, K. W. (1997). Activation of β-catenin-Tcf signaling in colon cancer by mutations in β-catenin or APC. Science 275, 1787-1790.

Nateri, A.S., Spencer-Dene, B., and Behrens, A. (2005). Interaction of phosphorylated c-Jun with TCF4 regulates intestinal cancer development. Nature 437, 281-285.

Polakis, P. (2000). Wnt signaling and cancer. Genes & Dev. 7^, 1837-1851.

Pan, D., Schomber, T., Kalberer, C.P., Terracciano, L.M., Hafen, K., Krenger, W, Hao-Shen, H., Deng, C, and Skoda, R.C. (2007). Normal erythropoiesis but severe polyposis and bleeding anemia in $mad4-dificient mice. Blood 110, 3049-3055.

Rosenbauer, R, Koschmieder, S., Steidl, U., and Tenen, D.G (2005). Effect of transcription- factor concentrations on leukemic stem cells. Blood 106, 1519-1524.

Sansom, O.J., Reed, K.R., Hayes, A.J., Ireland, H., Brinkmann, H., Newton, I.P., Batlle, E., Simon-Assmann, P., Clevers, H., Nathke, I.S., et al. (2004). Loss of Ape in vivo immediately perturbs Wnt signaling, differentiation, and migration. Genes Dev. 18, 1385-1390.

Shibamoto, S., Higano, K., Takada, R., Ito, R, Takeichi, M., and Takada, S. (1998). Cytoskeletal reorganization by soluble Wnt-3a protein signaling. Genes Cells 3, 659-670.

Shoemaker, A.R., Gould, K.A., Luongo, C, Moser, A.R. and Dove, W.R (1997). Studies of neoplasia in the Min mouse. Biochem. Biophis. Acta 1332, F25-F48.

Tetsu, O. and McCormick, R (1999). β-Catenin regulates expression of cyclin Dl in colon carcinoma cells. Nature 398, 422-426.

Toyota, M., Ahuja, N., Ohe-Toyota, M., Herman, J.G., Baylin, S.B., and Issa, J.P.J. (1999). CpG island methylator phenotype in colorectal cancer. Proc. Natl. Acad. Sci. USA 96, 8681-8686. van de Wetering, M., Sancho, E., Verweij, C, de Lau, W., Oving, L, Hurlstone, A., van der Horn, K., Batlle, E., Coudreuse, D., Haramis, A.P., et al. (2002). The β-catenin/TCF-4 complex imposes a crypt progenitor phenotype on colorectal cancer cells. Cell 111, 241-250.

van Es, J.E., Jay, P., Gregorieff, A., van Gijn, M.E., Jonkheer, S., Hatzis, P., Thiele, A., van den Born, M., Begthel, H., Brabletz, T, et al. (2005). Wnt signalling induces maturation of Paneth cells in intestinal crypts. Nature Cell Biol. 7, 381-386.

Weisenberger, D.J., Siegmund, K.D., Campan, M., Young, J., Long, T.I., Faasse, M.A., Hong Kang, G, Widschwendter, M., Weener, D., Buchanan, D., et al. (2006). CpG island methylator phenotype underlies sporadic microsatellite instability and is tightly associated with BRAF mutation in colorectal cancer. Nature Genet. 38, 787-793.

Yamamura, Y, Lee W.L., Inoue, K.i., Ida, H., and Ito, Y. (2006). RUNX3 cooporates with Fox03a to induce apoptosis in gastric cancer cells. J. Biol. Chem. 281, 5267-5276.

Yano, T, Ito, K., Fukamachi, H., Chi, X.Z., Wee, H.-J., Inoue, K.i., Ida, H., Bouillet, P., Strasser, A., Bae, S.C., et al. (2006). The RUNX3 tumor suppressor upregulates Bim in gastric epithelial cells undergoing TGF-β-induced apoptosis. MoI. Cell. Biol. 26, 4474-4488.

Claims

WHAT IS CLAIMED IS:

1. A method of preventing, inhibiting, arresting or reversing tumori genesis in a cell, the method comprising altering the formation of a complex between RUNX3, or a functional fragment thereof and at least one of (i) β-catenin, or a functional fragment thereof, and (ii) a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof.

2. The method of claim 1, wherein the cell is a rectal cell, a colon cell, an adrenal cell, a bone cell, an epithelial cell, a nerve cell, a brain cell, a cell of cartilage, a cell of the eye, a cell of the heart, a kidney cell, a liver cell, a lung cell, a muscle cell, an ovary cell, a cell of the pancreas, a prostate cell, a skin cell, a cell of the small intestine, a spleen cell, a stomach cell, a testicular cell, a thymus cell, a vascular cell, a cell of the uterus or a cell of connective tissue.

3. The method of claims 1 or 2, wherein said tumorigenesis is carcinogenesis.

4. A method of inducing apoptosis in a tumor cell, the method comprising altering the formation of a complex between RUNX3, or a functional fragment thereof and at least one of (i) β-catenin, or a functional fragment thereof, and (ii) a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, in the cell.

5. The method of claim 4, further comprising determining apoptosis in the tumor cell.

6. The method of claims 4 or 5, wherein the tumor is one of cancer and ulcerative colitis.

7. The method of claim of any one of claims 1-6, wherein the cell is obtained from or is comprised in a host organism.

8. The method of claim 7, wherein the cell is cultured.

9. A method of diagnosing the risk of tumorigenesis in a cell, the method comprising assessing the formation of a complex between RUNX3, or a functional fragment thereof and at least one of (i) β-catenin, or a functional fragment thereof, and (iϊ) a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, in the cell.

10. The method of claim 9, wherein an alteration of said complex formation is an indication of an increased risk that the cell will become tumorigenic.

11. The method of any one of claims 1 - 14, wherein said complex formation is modulated by a compound that modulates the phosphorylation status of cellular components.

12. The method of claim 11 , wherein the compound alters the degree of phosphorylation of RUNX3 or a functional fragment thereof.

13. A method of diagnosing the risk of developing a neoplasm in a subject, the method comprising assessing the formation of a complex between RUNX3, or a functional fragment thereof and at least one of (i) β-catenin, or a functional fragment thereof, and (ii) a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, in the cell.

14. The method of claim 13 , wherein the neoplasm is a tumor.

15. The method of any one of claims 1 - 14, wherein the member of the TCF/LEF transcription co-factor family is one of LEFl, TCF-I, TCF-3, and TCF-4.

16. An in-vitro method of identifying a compound capable of altering the formation of a complex between RUNX3, or a functional fragment thereof and at least one of (i) β-catenin, or a functional fragment thereof, and (ii) a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof, comprising:

(a) contacting the components that form said complex with each other,

(b) adding a compound to the test tube suspected to modulate said complex formation, and (c) detecting the said complex formation.

17. The method of claim 16, wherein the member of the TCF/LEF transcription co-factor family is one of LEFl, TCF-I, TCF-3, and TCF-4.

18. The method of claim 16 or 17, wherein the detection is performed by a suitable spectroscopic, photochemical, photometric, fluorometric, radiological, enzymatic or thermodynamic method.

19. The method of any one of claims 16 - 18 for the in-vitro screening for potential compounds that are useful for preventing, inhibiting, arresting or reversing tumorigenesis due to their alteration of the complex formation between RUNX3, or a functional fragment thereof and at least one of (i) β-catenin, or a functional fragment thereof, and (ϋ) a member of the TCF/LEF transcription co-factor family, or a functional fragment thereof.

20. The method of any one of claims 16 - 19, further comprising: comparing the obtained results with those of a control measurement.

21. The method of claim 20, wherein the control measurement comprises the use of a compound that does not alter the formation of said complex.

22. The method of any one of claims 16 - 21, wherein the cell is a cancer cell.