AU2003265009A1 - REGULATING Beta-CATENIN LEVELS - Google Patents

Description

WO 2004/024901 PCT/NL2003/000640 The invention relates to the fields of biology and medical science. More in particular the invention relates to means and methods for differentiating and de-differentiating cells. The invention further relates to means and methods for maintaining an undifferentiated phenotype in a population of cells or a subpopulation thereof. 5 The Wnt signal induces the nuclear translocation of transcriptionally active 3 catenin (Catnb) through interference with a multi-protein complex, composed of GSK3B, AXIN1, AXIN2, and APC, capable of earmarking free cytoplasmic P-catenin for degradation through the ubiquitin/proteasome pathway 12 . One of the key 10 components of this complex is the APC tumor suppressor, which serves as the scaffold to which P-catenin, AXIN1 and AXIN2 bind 3 . Mutations in the APC gene are responsible for familial adenomatous polyposis (FAP), an autosomal dominant predisposition to the development of hundreds to thousands of adenomatous polyps in the colon and rectum 4 . Moreover, APC mutations initiate the majority of sporadic 15 colorectal cancers 3,5 . Several mutant alleles of the mouse Apc gene have been generated (Fig. la). Notably, they display different developmental defects and different degrees of tumor predisposition and incidence 6 . The ApcMin allele encodes for a stable truncated protein of 850 amino acids, deleting all the P-catenin binding and downregulating motifs. 20 Heterozygous Apc+

/

Min mice develop multiple intestinal tumours whereas Apc Mi 'ni i n embryos die at very early developmental stages 7 ,s . Two different Apc mouse models, i.e. ApcSS8T and ApcG16 3 8N, were generated in our laboratory by introducing a selection cassette at codon 1638 of the endogenous Apc gene in the sense and anti-sense transcriptional orientation, respectively 9 . Heterozygous Apc+ 1 63 8 N mice develop 25 approx. 5-6 intestinal polyps in addition to a broad spectrum of extra-intestinal manifestations including desmoids, epidermal cysts, pilomatricomas and breast tumors 10, . Homozygosity for the Apc1*N allele is embryonic lethal'o , 1l . In contrast, ApC16 3 S8T/ 163 8T animals are viable and tumor-free s. Similar genotype-phenotype correlations have been established in man between APC 30 mutations and the degree of severity of the corresponding FAP phenotype 12 . However, the molecular and cellular mechanisms underlying these phenotypic differences are WO 2004/024901 PCT/NL2003/000640 2 not understood. In fact, due to the high complexity and turnover of the intestinal epithelium as biological system, little is known on the cancer-causing cellular mechanisms triggered by loss of APC function. Alterations in proliferation rates, apoptosis, cell migration and differentiation have been postulated. 5 In the present invention we demonstrate that the p-catenin level in a population of cells determines the differentiation capacity of that population of cells. A high p catenin level in a population of cells has the effect that the cell population displays at least one aspect of an undifferentiated phenotype. One the other hand the reverse 10 situation, a reduction of the level of p-catenin in a population of cells has the effect that that population demonstrates at least one aspect of a differentiated phenotype. By way of example this mechanism is illustrated herein by the capacity of pluripotent stem cells carrying different Apc and p-catenin alleles to differentiate into specialized cell lineages. Genetic and molecular evidence that the ability and sensitivity of ES 15 cells to differentiate into the three germ layers is inhibited by increased dosages of f catenin made available for Wnt signaling by specific Apc defects; these range from a severe differentiation blockade in Wnt-deficient Apc alleles, to more specific neuroectodermal, dorsal mesodermal and endodermal defects in more hypomorphic alleles. Accordingly, a targeted oncogenic mutation in the gene coding for p-catenin 20 (Catnb) also affects ES cell differentiation potential. Expression profiling of the wild type and Apc-mutant teratomas supports the differentiation defects at the molecular level and pinpoints a large number of downstream structural and regulating genes. Chimeric experiments showed that this effect is cell autonomous. One conclusion that is drawn from the invention as presented herein is that constitutive activation of the 25 APC/p-catenin signaling pathway results in differentiation defects in tissue homeostasis and that this underlies tumorigenesis in the colon and other self renewing tissues. Now that we have identified this mechanism, some findings in the art are better 30 explained. For instance, the observation that loss of Tcf4 function in mouse leads to the depletion of intestinal stem cells can now better be interpreted as being indicative of a role of the Wnt/p-catenin signal transduction pathway in epithelial stem cell WO 2004/024901 PCT/NL2003/000640 3 maintainance 8 . The present invention also fits well with recent, computer modeling studies wherein tumor initiation in the colon is correlated with crypt stem cell overproduction 14 . 5 Thus in one aspect the invention provides a method for modulating differentiation in a population of cells comprising modulating the presence of P-catenin in said population of cells. It has been found that the level of P-catenin in said population is a major determinant of differentiation state of cells said population. Particularly when stem cells are present in said population of cells. Thus in one embodiment the 10 presence of P-catenin is modulated in a population comprising stem cells. This embodiment is particularly useful for modulating differentiation of stem cells in said population. By modulating the presence of p-catenin upward in said population, the relative number of less differentiated cells in said population increases. On the other hand by modulating the presence of p-catenin downwards in said population, the 15 relative number of more differentiated cells increases. This aspect of the invention can be used in a variety of ways. For instance, for altering the number of stem cells in said population. Modulating the level of P-catenin upward leads to an increased number of stem cells in said population, whereas a downward modulation leads to a decreased number of stem cells. A population of cells wherein the number of less differentiated 20 cells is increased is said to display at least one aspect of a dedifferentiated phenotype. A population of cells wherein the number of differentiated cells is increased is said to display at least one aspect of a differentiated phenotype. Thus one aspect of a population of cells exhibiting a dedifferentiation phenotype is an increased number of undifferentiated cells, i.e. an increased number of stem cells in the population. In the 25 present invention it was also found that when the presence of 1-catenin is modulated downward, that the level of downward modulation is at least in part decisive in determining the differentiation pathway that is entered by stem cells in the population. Alternatively, by manipulating the presence of P-catenin upward in stem cells it is possible to prevent for instance ectodermal differentiation whereas meso 30 and dermal differentiation is at least partly impaired. Thus the type of differentiation is a reflection of the precise modulation of P-catenin.. The type of differentiation is at WO 2004/024901 PCT/NL2003/000640 4 least in part further influenced by the type of stem cells present in the population of cells. In the art the term "stem cell" is used for a variety of different cell types. In the 5 present invention the term "stem cell" is used to define a cell with the property to form daughter cells having the same properties as the mother cell (i.e. having the capacity to self-renew). A further property of a stem cell as used herein is the property to differentiate into a more specialized cell having a reduced capacity to self-renew. Stem cells are thought to divide 'asymmetrically' i.e. the daughter cell 10 differentiate while the 'mother' retain its pluripotency and self-renewing capacity. There are many different stem cells. Non-limiting examples are stem cells capable of forming an entire individual (so-called) embryonal stem cells and more restricted stem cells such as hemopoietic stem cells with the capacity to form an entire hemopoietic system. The normal differentiation process can at least in part be 15 affected in tumorigenesis and/or other developmental defects. Some tumors have regained the property of unlimited self-renewal which is as used herein a property of an undifferentiated phenotype. By expanding the size of the stem cell compartment, a differentiation defect increases the size of the target cell population where additional somatic mutations will occur eventually leading to malignancy 20 Apart from the differentiation potential, also the potency of self-renewal may vary between different stem cells. In general more primitive stem cells comprise a potent capacity to self-renew, whereas more differentiated stem cells have only limited self renewal capacity. Differentiation of stem cells typically occurs via the generation of a number of identifiable intermediate cells having an intermediate cell division and self 25 renewal capacity. Such intermediate stem cells are considered less primitive. As used herein the term stem cells both refers to primitive omnipotent stem cells capable of unlimited self-renewal and cell division and less potent stem cells with a reduced capability of self-renewal and cell division potential. Typically, stem cells are capable of at least 5 and more preferably, at least 10 cell divisions wherein at least one of the 30 daughter cells has the same property as the mother cell, in a suitable in vitro or in vivo environment.

WO 2004/024901 PCT/NL2003/000640 5 The presence of p-catenin in a population of cells can be modulated in a variety of ways. p-catenin is produced by a cell,thus one way of modulating the presence of 3 catenin is to modulate the expression of the gene encoding P-catenin. -catenin is also turned over in a cell as a protein. Thus, another way of modulating the presence of P 5 catenin is to modulate the turn-over of the 3-catenin protein in the cell. One way of modulating the turn-over is by modulating phosphorylation of 0-catenin. Phosphorylation earmarks p-catenin for turn-over by a proteasome complex via a multiprotein complex comprising GSK3B, axin/conductin and APC. As P-catenin is part of the Wnt-signalling pathway, the presence of 0-catenin is preferably modulated 10 by altering the activity of the Wnt signalling pathway of the cell and/or the cellular environment surrounding said cell. Of course a combination of methods can also be used. In one embodiment the activity of a Wnt-signalling pathway is modulated by modulating the activity of a multiprotein complex comprising GSK3B, axin/conductin and APC. Preferably, said activity of a multiprotein complex is at least partially 15 compromised. The population of cells can be derived from several sources. Preferably, the presence of P-catenin is modulated in the population in vitro. The population or a part thereof can subsequently be transplanted. In vivo, modulation of the presence of P-catenin is 20 also possible. The population of cells preferably comprises embryonal stem cells, stem cells of the GI tract, the hemopoietic system or the skin, or a combination thereof. In a particularly preferred embodiment the population of cells comprises embryonal cells or cells of the GI tract, as these cells, and particularly the stem cells therein, are very responsive to modulation of differentiation via the modulation of the presence 0 25 catenin in said population. Differentiation is a turning point in state of the stem cell. This turning point is reached when the presence of P-catenin reaches a threshold in the cell. When the presence of p-catenin in a differentiated cell, or less primitive stem cell reaches above 30 a threshold the cell gains at least a proliferation aspect of a dedifferentiation phenotype. When the presence of 1-catenin in a stem cell reaches below a certain threshold the cell gains at least the reduced proliferation potential of a less primitive WO 2004/024901 PCT/NL2003/000640 6 stem cell or differentiated phenotype. Thus, in a preferred embodiment of the invention said P-catenin presence is affected to a level that is below or above a threshold associated with a stem cell/differentiation phenotype in said population of cells, preferably, a proliferation threshold. 5 Protein activity of a multiprotein complex comprising GSK33, axin/conductin and APC can be modulated in a variety of ways. In a preferred embodiment, protein activity is provided through a nucleic acid sequence encoding said at least part of said protein activity. In a preferred embodiment said nucleic acid sequence is provided in a 10 gene delivery vehicle. A gene delivery vehicle preferably comprises a virus based delivery vehicle, preferably an adenovirus, an adeno-associated virus or retroviral vector based gene delivery vehicle. Currently, many different hybrid gene delivery systems are available. These are of course also within the scope of the invention. Thus the invention further provides a gene delivery vehicle comprising a nucleic acid 15 sequence encoding at least one protein from a multiprotein complex comprising GSK33, axin/conductin and APC for use in enhancing differentiation in a population of stem cells. By diminishing the activity of a Wnt-signalling pathway it is possible to diminish the 20 presence of P-catenin in the population of cells. Thus in one aspect the invention provides a Wnt signalling pathway antagonist for use in enhancing differentiation in a population of stem cells. The antagonist is preferably proteinaceous. Candidate antagonists are Dickkopf protein activity and secreted frizzled related proteins or a functional part, derivative and/or analogue thereof. By enhancing the protein activity 25 of a multiprotein complex comprising GSK313, axin/conductin and APC, it is possible to enhance differentiation in a population of cells comprising stem cells. Thus in another embodiment the invention provides a protein capable of enhancing the protein activity of a multiprotein complex comprising GSK33, axin/conductin and APC for use in enhancing differentiation in a population of stem cells. In a preferred 30 embodiment said protein is active APC or a functional part, derivative and/or analogue thereof. Preferably, said APC comprises a mutant APC comprising a modified APC activity.

WO 2004/024901 PCT/NL2003/000640 7 The invention further provides a gene delivery vehicle comprising a nucleic acid sequence encoding a wnt signalling pathway antagonist for use in enhancing differentiation in a population of stem cells. In a preferred embodiment, said antagonist is provided through a nucleic acid sequence encoding said at least part of 5 said antagonist. In a preferred embodiment said nucleic acid sequence is provided in a gene delivery vehicle. A gene delivery vehicle preferably comprises a virus based delivery vehicle, preferably an adenovirus, an adeno-associated virus or retroviral vector based gene delivery vehicle. Currently, many different hybrid gene delivery systems are available. These are of course also within the scope of the invention. 10 In yet another aspect the invention provides a pharmaceutical formulation suitable for enteral administration comprising P-catenin inhibiting activity. Enteral administration can of course also be achieved through oral formulations such as but not limited to, tablet or capsules comprising an appropriate enteric coating. By 15 varying the enteric coating and/or an another aspect of the formulation, delivery of the active component can be targeted to certain parts of the GI tract. In a preferred embodiment said p-catenin inhibiting activity is provided through a gene delivery vehicle comprising a nucleic acid sequence encoding at least one protein from a multiprotein complex comprising GSK313, axin/conductin and APC, or a nucleic acid 20 sequence encoding a Wnt signalling pathway antagonist. In another preferred embodiment said pharmaceutical formulation comprises at least one protein of a multiprotein complex comprising GSK38, axin/conductin and APC or a Wnt signalling pathway antagonist. In a preferred embodiment, a pharmaceutical formulation of the invention is used for the treatment of cancer of the GI-tract, wherein said composition 25 decreases the presence of P-catenin in a population of cells comprising tumor cells. In another preferred embodiment a pharmaceutical formulation is used for at least in part preventing the occurrence of tumors of the GI-tract, wherein said formulations is provided to individuals at risk of the development of said tumors. 30 WO 2004/024901 PCT/NL2003/000640 8 Brief description of the drawings Figure 1. Expression analysis of truncated Apc proteins in mutant ES cell lines. 5 a. Schematic representation of the proteins expressed by the Apc-mutant alleles employed in this study. b. Western blot analysis of total protein lysates from Apc mutant ES cell lines with a N-terminal APC monoclonal antibody (Ab-1, see Methods) recognizing amino acids 1 29 of APC. Lane 1, wild type Apc+l+; lane 2, heterozygous Apc+/16sN; lane 3, 10 heterozygous Apc+1638T; lane 4, heterozygous Apc+l1572T; lane 5, Apc+/mn; lane 6, wild type Apc

+

'

+. c. Immunoprecipitation analysis of ApclSsN and Apc 16 8 T ES cells. Cellular extracts were immunoprecipitated with the Apc N-terminal polyclonal antibody AFPN and detected with an APC monoclonal antibody (Ab-1, Oncogene Research Products) 15 recognizing amino acids 1-29 of APC (see Materials and Methods). Lane 1, homozygous Apc638T 16 38T ES cells; lane 2, heterozygous Apc+/l6s 3 sT; lane 3, wild type; lane 4, Apc 16 3 8N1638N Lanes 5-9: determination of the expression level of the Apc1638 truncated protein in homozygous Apc 163 8sN ES cells. ApclGssNI1638N ES cells were mixed with decreasing 20 percentages of wild type ES cells, and subsequently immunoprecipitated (see Materials and Methods). Lane 5, 10%; lane 6, 5%; lane 7, 2%; lane 8, 1%; lane 9, 0,5% wild type ES cells. In view of the observation that truncated Apc1638 and full-length Apc are expressed at a 1:1 ratio in Apc+1638T ES cells (Fig. lb, lane 3) and immunoprecipitate with equal efficiency from total cellular lysates (Fig. le, lane 2), 25 the Apc1638 protein in ApclG3SN/1638N ES cells was estimated to be expressed at about 2% of the endogenous Apc levels in wild type cells (lane 7). d. P-catenin/Tcf reporter assays performed in wild type, Apce 6 3 8N /1 6 3 8 ssN and ApcMitNin ES cell lines. Undifferentiated ES cells were transfected with either the pTOPFLASH (black bars) or the pFOPFLASH (gray bars) luciferase reporter construct, together 30 with the CMV-galactosidase construct as internal control 9,15. Normalized pTOPFLASH and pFOPFLASH levels are indicated for each cell line in triplicate transfections as shown.

WO 2004/024901 PCT/NL2003/000640 9 Figure 2. In vivo differentiation analysis of teratomas derived from wild type Ape mutant ES cells. Panel a. Low magnification view of histological sections stained with haematoxylin/eosin (H/E) showing the heterogeneous differentiation profile of 5 teratomas derived from wild type ES cells. Panel b. Low magnification view of histological sections stained with haematoxylin/eosin (H/E) showing the more homogeneous and undifferentiated histological features of teratomas derived from homozygous Apc 1 3 s 8 N ES cells. Panel c. Detail of homozygous Apcs 163a N teratomas stained with an antibody against 10 lysozyme (Novocastra) identifying Paneth cells (p) within a non-ciliated secretory epithelium, and with alcian blue identifying secretory goblet cells (g), both indicative of intestinal differentiation. Panel d. Histological section of a homozygous Apc 16 8 T teratoma stained with Alcian blue and heamatoxylin shows cartilage/bone (blue) differentiation. 15 Panels e-f. H/E stained sections of compound heterozygous Apc 16 8N15 7 2T teratomas show a high degree of vascularization (e) and a ciliated secretory epithelium (f) absent in Apcl638N/1638N teratomas. Panels g-p. Immuno-histochemical staining of paraffin sections from wild type (g-k) and homozygous Apc 163

BN/

16 38N (l-p) teratomas with neural markers NCAM (g,1), 20 neurofilament (h,m), synaptic vesicles (i,n), GFAP (j,o) and smooth muscle (h,p). The slides were lightly counterstained with haematoxylin (for details see Material and Methods). Panels q-t. Expression and subcellular localization of P-catenin in teratomas derived from Apc mutant ES cells. Paraffin sections were stained for p-catenin by immuno 25 histochemistry and counterstained with haematoxylin (for details see Materials and Methods). (q) Wild type teratoma show mainly cytoplasmic and membraneous 3 catenin staining. (r) Apc 16 3 8 sNls 63 sT teratomas with clear increase and heterogeneity of intracellular E-catenin. (s) Apc 6 sN/1572T teratomas with moderate nuclear P-catenin staining. (t) Apc i 8 MsN/1 6 8SN teratomas displaying strong nuclear P-catenin staining in 30 cells undergoing mesenchymal condensation and epithelial differentiation. Scale bars = 0,1 mm.

WO 2004/024901 PCT/NL2003/000640 10 Figure 3. Analysis of chimeric teratomas reveals the cell autonomous nature of the differentiation defect of Apc-mutant ES cells. Panel a. Stereo-microscopic view of a wild type Apc teratoma solely composed of ES cells targeted with a Rosa26-pgeo construct, showing uniform LacZ expression (blue). 5 Panel b. Stereo-microscopic view of a chimeric teratoma composed of wild type Rosa26-pgeo (blue) and non-tagged (white) wild type ES cells showing uniform amalgamation of the two cell types, resulting in a light blue appearance with small patches of unicellular contribution. Panel c. Stereo-microscopic view of a chimeric teratoma composed of tagged wild type 10 (blue) and Apc 1 6 38sN/1638N (white) ES cells showing a non-uniform distribution of the two cell types. Panel d. Stereo-microscopic view of a chimeric teratoma composed of tagged wild type (blue) and Apc1638N

I

1572T (white) ES cells showing the same sorting mechanism as in (C). 15 Panel e. Histological section of chimeric Rosa26-pgeo-Apc + l+ (blue) / Apc 16 38N/ 1 6 8 sN (red) ES cells stained for p-galactosidase activity (turcoise blue) and counterstained with basic fuchsin (red nuclei), showing contribution of both cell types in smooth muscle (Sm) and of exclusively Rosa26-pgeo cells in cartilage, (c). Panels f-h. Histological sections of chimeric Rosa26-Pjgeo-Apc*'+/non-tagged Apc+I+ES 20 teratomas (f) and of Rosa26-pgeo-Apc

+

'/

+ / non-tagged Apc 1 638N/ 16 38N teratomas (g and h) stained for I-galactosidase activity (turcoise blue cytoplasm), immuno-stained against P-catenin (brown) and counterstained with heamatoxylin (blue nuclei). Note that P catenin is homogeneously expressed in the chimeric epithelium composed of the same cell-type (f). As shown in (g) the Apc-mutant epithelium expresses high dosages of P 25 catenin and has differentiated to non-ciliated epithelium (white arrowhead), whilst the wild type part containing relatively low dosages of P-catenin, has differentiated to ciliated epithelium (black arrowhead). This nicely demonstrates the correlation between the Apc defect, increased P-catenin dosage and differentiation. (h) nuclear translocation of P-catenin in mutant Apc cells is not prevented by surrounding wild 30 type Apc cells. White bar = 1mm; black bar = 0,1 mm.

WO 2004/024901 PCT/NL2003/000640 11 Figure 4. In vitro differentiation analysis of ApcMin and Apc 16 38N mutant ES cells. Wild type, homozygous Apc6 s 8

N

/1G 3a N and Apc M -

M

n ES cells were cultured for two days with LIF. Differentiation was allowed by withdrawing LIF and colonies were 5 monitored at 4, 6 and 8 days after the start of the cultures. a. Examples of wild type, Apc l 6 a 8N/163 a 8N, and ApcuMintin ES cell colonies after 8 days of culture. The colonies were subdivided in three categories: completely undifferentiated with no mesenchymal cells protruding from the edge of the ES colony (white arrows); partly differentiated with mesenchymal cells surrounding a core of undifferentiated 10 cells (stippled arrows), and completely differentiated composed exclusively of mesenchymal cells (black arrows). Number (% of total) of completely undifferentiated (white bars), partly differentiated (stippled bars) and completely differentiated (black bars) wild type and Apc mutant ES colonies after 4, 6 and 8 days of culture in absence of LIF. Each bar was 15 determined from counting over 200 colonies from independently cultured plates and randomly selected stereo-microscopic fields. Results were confirmed by two independent experiments. b. P-catenin immunohistochemical staining of in vitro differentiated colonies showing absent, moderate, and strong P-catenin nuclear accumulation in wild type, 20 ApcI6sNm163 8 N, and Apc m in m in ES cells, respectively. Figure 5. Biochemical and immunohistochemical analysis of P-catenin mutant teratomas. a. L-catenin/Tcf reporter assays performed in wild type (Apc+/+Ip-cat+l+), ApC1638N/ 16 3 8 N , 25 Catnblox(exs) (equivalent to wild type), and Catbnoexa (mutant) ES cell lines. Undifferentiated ES cells were transfected with either the pTOPFLASH (black bars) or the pFOPFLASH (gray bars) luciferase reporter construct, together with the Renilla Luciferase construct as internal control 9,15. Normalized pTOPFLASH and pFOPFLASH levels are indicated for each cell line from triplicate transfections. 30 b. Differentiation profiles of P-catenin mutant teratomas. Panels a and b: hematoxylin-eosin staining of paraffin embedded sections of teratomas derived from generated Catnblox(exa) and Catbnoexa ES cells, respectively.

WO 2004/024901 PCT/NL2003/000640 12 Panel c: expression and subcellular distribution of P-catenin in teratomas derived from Catbnuexs ES cells. Paraffin sections were stained for P-catenin by immuno histochemistry and counterstained with haematoxylin (see Methods). Overexpression and nuclear localization of P-catenin is observed, among others, for example in 5 mesenchymal cell types (white arrowhead). Abbreviations: c, ciliated epithelium; n, neural lineages; b, bone; nc, non-ciliated epithelium; sin, smooth muscle. Histological observations were confirmed by immunohistochemistry with lineage-specific antibodies (not shown). 10 Figure 6. Expression profiling analysis of the Apc-mutant teratomas. a. Two-dimensional agglomerative cluster analysis of Apc-mutant teratomas compared with wild type (Apc'1+), where a total of 1484 genes exhibited a fold change > 2 with a p-value less than 0.01. Each row represents a specific genotype/teratoma and each column a single gene. As shown in the colour bar, red indicates up 15 regulation, green down regulation, and black no change. b. Expression analysis of a selected group (n=300) of differentially expressed genes in the 4 different mutant Apc genotypes. For this analysis, genes differentially expressed with a fold change +-5 and a p-value < 0.01 between the Apcl 163 8 N/1638N and Apc* l+ arrays are displayed. Also, genes differentially expressed between the 20 Apcles 16 3 8 T /s and Apc + 1 * arrays with a fold change +-2 and a p-value < 0.01 were excluded from this analysis. The rationale for the latter is that the Apc 168a T1638T and Apc+/ + genotypes do not differ both in terms of signaling activity (as measured by TOP FLASH) and differentiation (as measured by teratoma assay). Hence, genes differentially expressed between these two ES cell lines are not likely to contribute to 25 the differentiation defect due to Wnt signaling activation and could be excluded to reduce background noise.

WO 2004/024901 PCT/NL2003/000640 13 Examples Methods 5 Animals The experiments on mice as here reported were approved by the local animal experimental committee of the Leiden University, and by the Commission Biotechnology in Animals of the Dutch Ministry of Agriculture (permission no. VVA/BD 01.168). 10 Constructs. The gene targeting Rosa26-p-geo construct used to tag wild type ES cells by homologous recombination fuses the ubiquitously expressed protein encoded by the endogenous Rosa26 locus with a hybrid P-galactosidase-neomycine resistance protein (P-geo). The construct is a derivative of the gene trap construct employed by Friedrich 15 and Soriano 8 is and was constructed and kindly provided by J-H Dannenberg and H. te Riele (in preparation). Embryonic stem (ES) cell lines Heterozygous Apc 168 sN, and ApcMin mice on a mixed C57BL/6J x CD1 background were 20 mated and blastocysts were harvested 3.5 days after fertilization. The flushed pre implantation blastocysts were plated on MEFs-coated 96-wells dishes and cultured to isolate undifferentiated ES lines according to Rudnicki et al.

20 . Six independently isolated Apo min n, 4 Apc 16 8N/1638N, 4 littermate Apc+IMin and two wild type Apc

+

/' ES lines were obtained and employed in the differentiation experiments. Only early 25 passage (p4-p5) embryo-derived ES lines were employed for the experiments described here. All remaining cell lines (Apcl 6 3 8 N/1 6 3 8 N , Apce63 8 N

/

16 3 8 T, Ap c 1 572T/1638N, Apc

I

GBST

16 SST) were derived by gene targeting in the E14 ES cell line (129/Ola), as previously described 9 . The latter were previously established and are therefore characterized by higher passage numbers (~p20). To exclude that the observed defects result from in 30 vitro acquired mutations, at least two independently generated clones for each genotype were employed for the teratoma assays. The Catnblox(exs) and Catnbexe ES lines were obtained by Lox-gene targeting and by Cre-mediated deletion of exon 3 as previously described 2 2

.

WO 2004/024901 PCT/NL2003/000640 14 Generation of teratomas Single-cell suspensions were prepared from semi-confluent undifferentiated ES cultures grown on MEFs-coated 9 cm culture-dishes, using Ix Trypsin/EDTA (Life 5 Technologies). Cells were washed three times with a large volume of PBS prior to injection. A total of 4x10 6 cells in 200 01 PBS were injected subcutaneously in the flank of a syngenic mouse. C57BL/6J x CD1 recipient mice were employed for the blastocyst-derived ES lines (Apc168N and ApcMin; see above). p-catenin and Apc mutant ES cells derived by gene targeting in the E14 ES cell line 9 ,22 were injected in 10 129/Ola recipient mice. Chimeric teratomas were induced by mixing of 2x10 6 cells of each cell line shortly before injection. After 3 weeks, the mice were sacrificed and teratomas were removed, washed with PBS, fixed overnight in 4% PFA and processed for paraffin based histology using standard techniques. 15 Western blot and imnnunoprecipitation analysis. For the detection of the various mutant Apc proteins by western analysis, total protein lysates and immunoprecipitates were derived from freshly cultured ES cell lines and resolved on agarose gels as described previously 9 . Quantification of residual amounts of the Apc1638 protein was performed by mixing homozygous Apc16 8 8N ES 20 cells with given percentages of wild type ES cells prior to immunoprecipitation with the N-terminal AFPN polyclonal antibody 9 , immunoblotting and detection with an APC monoclonal antibody (Ab-1, Oncogene Research Products) recognizing amino acids 1-29 of APC. 25 In vitro differentiation of ES cells. ES cells (10 4 cells/well) were plated on 12-well tissue culture plates and cultured in ES culture medium containing LIF (Life technologies) for two days. Spontaneous differentiation was allowed by keeping the ES clones in culture for 2 days in the presence and for 6 days in the absence of LIF without passaging. Induced 30 differentiation was achieved by plating 2 days old embryoid bodies on gelatin-coated 12-well culture plates and LIF withdrawal. Differentiation to mesodermal, neurectodermal, and throphectodermal derived lineages was achieved by addition of WO 2004/024901 PCT/NL2003/000640 15 DMSO (0.05%, 0.1%, 0.5%, 1% or 2%), retinoic acid (0.1 and 0.5 uM, Sigma) and FGF4 (25 ng /ml) respectively, as previously decribed 19,20. Immunohistochemical analysis of in vitro differentiated ES colonies. 5 Differentiated embryoid bodies attached to 12-well dishes were fixed with 1% PFA for 10 min, washed for 5 min. in PBS, dehydrated in methanol and stored at -20'C until use. For immunohistochemistry, plates were re-hydrated in PBS and blocked in 5% non-fat dried milk (NFDM) in PBS for two hours at room temperarture. The primary antibody was diluted in 5% NFDM 2 hours prior to use. Incubation with the primary 10 antibody was performed at 4o C overnight under gentle shaking. The next day, the plates were washed 5 times for 30 min. at room temperature with PBS containing 0.1% Tween-20 (PBT), after which the plates were incubated overnight with the secondary antibody diluted in PBS containing 5% NFDM. The following day the plates were washed thoroughly with PBT, equilibrated in alkaline phosphatase buffer for 15 15 min and stained with BCIP and NBT according to the manufacturer's recommendations (Roche inc.). Antibodies The primary antibodies .used for immunohistochemistry were: mouse anti-B-catenin 20 (Transduction Laboratories); anti-AFP (Novocastra); anti-NCAM (5B8, from Developmental Studies Hybrydoma Bank (DSHB)); anti-neurofilament (2H113, DSHB); anti-synaptic vesicles (SV-2, DSHB); anti-Glial Fibrillary Acidic Protein (GFAP, DAKO corp.), anti-smooth muscle actin (clone 1A4, Neomarkers Ab-1), anti-striated muscle (A4.1025, DSHB). The AFPN polyclonal was previously described 9 . 25 For conventional immuno-histochemistry and Western analysis, a goat anti-mouse IgG/IgM - peroxidase conjugate was used. A goat anti-mouse IgG/IgM alkaline phospatase conjugate was employed as the secondary antibody for the immunohistochemical analysis of in vitro differentiated ES cells. All secondary antibodies were purchased from Jackson ImmunoResearch Laboratories. 30 TOP/FOP flash reporter assays. Twenty hours before transfection, 105 cells/well were plated on 12-well tissue culture plates coated by a feeder cell layer. Cells in each well were transfected with 500 ng WO 2004/024901 PCT/NL2003/000640 16 pTopflash or pFopflash vector (kindly provided by Dr. Hans Clevers), and 500 ng LacZ vector (or alternatively 25 ng Renilla Luciferase) using Lipofectamine 2000 (LifeTechnologies) as recommended by the manufacturer. Luciferase activities were measured after 24 hours in a luminometer (Lumat LB 9507) and normalized for the 5 transfection efficiency by either measuring the -galactosidase activity as described previously 9 , or with the Dual Luciferase Reporter Assay system (Promega Corp.) according to the manufacturer's instruction. Expression profiling by oligonucleotide microarray and data analysis 10 Total RNA (200 O]g) was extracted from snap-frozen teratomas by standard procedures. The poly A + fraction (-2 Og) was isolated by Dynabeads® biomagnetic separation (Dynal Biotech) and directly employed to probe the GeneChip® Murine Genome U74A Set microarray (Affymetrix, Inc.) according to the protocol provided by the manufacturer (see also http://www.affymetrix.com). 15 The microarray data were analyzed using the Rosetta Resolver v3.0 Gene Expression Data Analysis System. The input data for the system was the Affymetrix GeneChip generated CEL files. The CEL files contain the 7 5 th percentile pixel intensity of a given feature. The CEL file also contains the standard deviation (SD) of the pixels, and the number of pixels of a given feature. The Affymetrix array contains perfect 20 match (PM) and corresponding mismatch probes (MM), which are used to control for background and nonspecific hybridization. The arrays generally contain 16 probe pairs per probe set. The data were processed by an error model in Rosetta Resolver v3.0 Gene Expression Data Analysis System. The error-weighted PM/MM differences were considered 25 outliers, thus not used for computation, if it were greater than 3 SDs away from the mean of the PM/MM intensity difference for a given probe set. If a probe pair was excluded on one array, then it was excluded from the differential calculation between 2 arrays. Differential expression was determined by correcting for intra-array gain adjustment, normalizing between arrays, and correcting for non 30 linearity. The intra-array gain adjustment was calculated by dividing the array into 64 sectors and providing gain adjustments for each sector. For differential expression between 2 arrays, the sectors of the 2 arrays were normalized sector-by-sector. Then a set of invariant probe sets was detected, by calculating the error-weighted log ratio WO 2004/024901 PCT/NL2003/000640 17 between the 2 arrays. The probe sets, whose probe pairs exhibited a logo ratio closest to zero, were utilized for the non-linearity adjustment. A p-value was calculated for each probe set, by utilizing the error-normalized differential expression between arrays. The error model assumed the log ratio statistic followed a normal 5 distribution. A probability density function of the error-normalized differential was utilized, with the null hypothesis: the sequence was not differentially expressed. The agglomerative clustering algorithm was utilized for clustering genes and arrays into a hierarchical structure. The error-weighted correlation without mean subtraction, based on log ratio was used as the similarity measure. When merging 10 the two closest objects, a heuristic criterion of average-linkage was used to redefine the between-cluster similarity measure. Only sequences that met the following criteria were utilized for clustering: a fold change > 2 and a p-value less than 0.01. The color displays given in Fig. 6a, represent the logo ratio between the 2 arrays, as red when the sequence is upregulated relative to the control; green when down 15 regulated; and black when the expression is close to zero.

WO 2004/024901 PCT/NL2003/000640 18 Results Characterization of Apc-mutant ES cell lines To further elucidate the molecular basis of the phenotypic consequences of different mouse Apc mutations, and to evaluate their effect on stem cell differentiation, we 5 generated different ES cell lines carrying the four Apc mutant alleles depicted in Fig. la in different combinations, and analyzed their expression level and P-catenin-driven transcriptional activity. ES lines were generated either by two rounds of gene targeting 9 , or by direct isolation from pre-implantation blastocysts (see Methods). Apart from the above mentioned mutations, we also included Apc 152 7T, a new allele 10 obtained by deleting 197 nucleotides from the 3'-end of the original Apc 16 asT targeting construct and thereby removing the only residual SAMP repeat (the Axin binding domain) 9 (Fig. la). Western blot analysis shows that all the alleles depicted in Fig. la but one, Apc16G8N, stably express the predicted truncated protein at a 1:1 ratio with the full-length wild type protein (Fig. lb). In contrast, no truncated protein could be 15 detected in heterozygous Apc+/ 1 638 N mice or ES cells (Fig. lb). To determine whether the Apc 1

"

6 sN mutation represents a true null allele, we attempted the identification of minimal amounts of the truncated protein in homozygous Apc 16 8 8N cells by immunoprecipitation (IP) analysis (Fig. Ic). A weak 182 kDa band, i.e. of the same size as the Apc6 3 8T truncated protein, was observed. The 20 quantification of the truncated Apc protein when compared to its full-length equivalent, was performed by competition IP assays and revealed that the Apc638N allele expresses the mutant protein at approximately 2% of the endogenous level in ES cells (Fig. 1c). Therefore, the Apcl 1 6 3 sN mutation represents a leaky hypomorphic variant of the Apc 16 8 T allele. 25 The above differences in expression levels of truncated Apc proteins are reflected by the corresponding levels of transcriptionally active nuclear P-catenin measured by the TCF/P-catenin responsive reporter assay, TOPFLASH 1 5 . Previous analysis revealed an increasing gradient of P-catenin regulatory activity among the different genotypes with Apc s 8 N1 63 8sN ES cells showing the highest reporter activity, followed by 30 Apce 16 sN/1572T r Apce 63 sNl 163 ssT, and Apcas 1638T 1638T, the latter being comparable to wild type ES cells 9 . As the initial analysis did not include Apc

M

in, we isolated both Apciian and Apc 163 8Nl16 8N ES cell lines directly from blastocysts of mixed genetic background C57BL6/CD1 (see Methods). Although direct comparison between the former WO 2004/024901 PCT/NL2003/000640 19 (performed with 129-01a ES lines) and the latter reporter assays is not feasible due to the different genetic backgrounds, ApcMi n/ M in ES cells showed considerably higher (twofold) TOPFLASH reporter levels than Apc i 6 SN1638

G

N ES cells (Fig. ld). Hence, our collection ofApc-mutant ES cell lines show a gradient of 3-catenin 5 regulatory activity useful to study its dosage-dependent consequences on different cellular functions. In vivo differentiation analysis of Apc-mutant ES cell lines by teratoma formation. To evaluate the differentiation potential of the Apc-mutant cell lines in an in 10 vivo assay, undifferentiated ES cells were injected subcutaneously in syngenic mice to induce formation of teratomas. These are benign tumors derived from pluripotent stem cells and composed of well differentiated tissues of ecto-, meso- and endodermal origin i 6 . Differentiation profiles of the resulting teratomas were first investigated by histological and immuno-histochemical analysis and in comparison with those 15 obtained with wild type (Apc + /+ ) ES cells (Fig. 2, panel a). Ecto-, meso- and endodermal differentiation defects in hypomorphic Apc genotypes. Teratomas derived from independent Apc1 63 a8N1638NES cell lines both obtained by two 20 rounds of gene targeting and blastocyst-derived, showed severe differentiation defects (Fig. 2, panels b-c). Several differentiation types, namely neural, bone, cartilage, and ciliated epithelia, were absent. As neural differentiation is sometimes difficult to recognize by morphological criteria, four different neural markers were employed to identify neuro-ectodermal cellular types, namely neurons, astrocytes, and Schwann 25 cells (N-CAM and GFAP), neurofilaments (2H3), and synaptic vescicles (SV-2). These antibodies failed to stain homozygous Apc i 6 8 N sections, in contrast with wild type teratomas where approximately 50-75% of the cells were positively stained (Fig. 2, panels g-j and l-o). Differentiation to striated muscle (adult myosin, A4.1025) was also severely affected and detectable only in a minority of the sections (not shown). In 30 contrast, smooth muscle cellular types (smooth muscle actin, 1A4) were abundant in Apc i Gs6N teratomas (Fig. 2, panels k and p). Other cell types and tissues positively identified in the Apc1 63 8N teratomas included simple non-ciliated epithelia, keratinized epithelia, and non-ciliated secretory epithelia including Paneth cells, the latter WO 2004/024901 PCT/NL2003/000640 20 indicative of intestinal differentiation (Fig. 2, panel c). Hence, homozygous Apc 16 s8N ES cells display severe differentiation defects that affect the neurectodermal, dorsal mesodermal, and endodermal lineages. In contrast, teratomas derived from homnozygous Apc638ST1G6 8T or from compound Apc 163 s8N1638T ES cells were 5 indistinguishable from wild type teratomas at both the immuno-histochemical and histological levels (Fig. 2, panel d; data not shown), indicating that the differentiation defects observed in the Apc 16 38N/1638N teratoma is directly correlated to the quantitative expression of the 182 kDa truncated protein. The decreased dosage of the otherwise Wnt/P-catenin proficient Apc1638 truncated 10 protein 9 in homozygous Apc 6 8 s N cells suggests that a quantitative defect in the ability of Apc to control P-catenin levels underlies the above differentiation abnormalities. We analyzed P-catenin expression in the teratomas by immunohistochemistry (Fig. 2, panels q-t). A very intense and abundant nuclear staining was observed in the Apc 16 38N/1638N teratomas. The nuclear staining was not observed in every cell but was 15 clearly present in structures undergoing mesenchymal condensation and epithelial differentiation, a process known to be mediated by Wnt signaling 17 (Fig. 2, panel t). Notably, compound heterozygous Apcl 6 38N/1638T teratomas where the Apc1638 protein is expressed at approximately 50% of the level in homozygous Apc 163 8T cells, did not show clear nuclear P-catenin accumulation but displayed differences in P-catenin 20 cytoplasmic staining patterns between flanking epithelial sheets, suggesting a more subtle defect of its downregulation by Apc (Fig. 2, panel r). These abnormal staining patterns were never observed in Apc1638T

/

1638T or wild type teratomas. The low dosage (2%) of the Apc1638 protein in homozygous Apc 1 6 38 N teratomas seems therefore insufficient to prevent nuclear accumulation of P-catenin during differentiation. On 25 the other hand, a 50% increase in the dosage of this truncated protein, as expressed by the Apc 16 8N/1638T teratomas, is sufficient to prevent P-catenin nuclear accumulation. The above results point to a direct relationship between nuclear accumulation of 3 catenin and differentiation. To test this hypothesis, we analyzed the differentiation potential of Apc 1 572T, a truncated Apc protein almost identical to Apc 16 s 8 T except for the 30 deletion of the last residual SAMP repeat (Axin binding) (Fig. la). The Apc 15 72T protein is stably expressed (Fig. lb) but is affected in its abililty to downregulate P catenin due to failure to bind Axin. The Apc 1 572T mutation was targeted in ES cells WO 2004/024901 PCT/NL2003/000640 21 both in the heterozygous (Apc+/16572T) as well as compound Ap 16 38NI/1572T forms, in order to allow comparison with Apc 1 G38N/168T and Apc 16 8N/16 s 38N cells. Teratomas derived from Apc+/1572T ES cells displayed normal differentiation profiles. However, compound heterozygous Apc 1 63 a N/1572T teratomas were severely affected in their differentiation 5 potential when compared with the Apc 1 638N/1688T and wild type counterparts. The differentiation defects of Apc 1 GS8N/1572T ES cells closely resembled those observed in Apc 16 8N/1638N teratomas, with the only exception of their capacity to differentiate into ciliated epithelia (Fig. 2, panels e and f). Furthermore, nuclear translocation of p catenin was also evident although frequency and intensity were somewhat lower than 10 observed in the homozygous Apc16 1 8N teratomas (Fig. 2, panel s). Thus, removal of the only Axin binding motif present in the Apc1638T protein results in nuclear translocation of p-catenin and in a dramatic decrease in differentiation potential. Expression profiling of Apc-mutant teratomas by oligonucleotide microarrays. 15 The differentiation defect observed in Apc-mutant teratomas by immunohistochemical analysis is likely to result from profound changes in gene expression patterns. We analyzed the teratomas by oligonucleotide microarrays encompassing probes for approx. 12,000 mouse genes and expressed sequence tags (ESTs) (GeneChip® Murine Genome U74A Set, Affymetrix, Inc). Poly-A + RNA was isolated from teratomas 20 derived from wild type (Apc+/'), Apcl138N/1638N, Apc 1 8 38N/1638T, Apcl1638N/1572T, and Apc 16 38T/ 1 6 38T ES cell lines, and hybridized to the oligonucleotide microarrays. Gene expression profiles from the Apc-mutant teratomas were employed to select for tissue specific genes up- and down-regulated when compared with wild type tumors. As shown in Table 1, a strong correlation was observed between the morphological and 25 immunohistochemical characterization of the mutant teratomas and their tissue specific expression profiles. In particular, a large collection of neuroectodermal specific probes were shown to be consistently downregulated among the Apc-mutant teratomas (Table la), in agreement with previous hystological and immunohistochemical observations (Fig. 2). Mesoderm-derived lineages like bone and 30 cartilage were also absent in Apc-mutant teratomas. Accordingly, bone- and cartilage specific genes were significantly downregulated (Table lb). Other mesodermal lineages, like smooth and striated muscle were respectively present and absent in the Apc-mutant teratomas, as also shown by the differential up- and down-regulation of WO 2004/024901 PCT/NL2003/000640 22 the specific gene markers (Table 1c). Specific endodermal lineages were also differentially represented in mutant teratomas: while non-ciliated epithelia of intestinal type were present (Fig. 2), ciliated epithelia as those of the respiratory tract were under-represented or absent. Microarray analysis confirmed the upregulation of 5 several intestinal markers (e.g. defensin related cryptdins; Table Id), whereas lung specific genes were downregulated (Table le). Notably, upregulation up to 100 fold of embryonic a- and P-like globin genes was observed in the Apc-mutant tissues (Table if). 10 The Apc differentiation defect is cell autonomous To investigate the cellular nature of the differentiation defect due to loss of Apc function, we generated chimeric teratomas composed of either Apc1638N/1638N or ApC1638N/1572T cells mixed with Apc 1 * ES cells, the latter tagged with a constitutively expressed P-galactosidase gene (Rosa26-p-geo) 18 . First, we generated chimeric 15 teratomas with wild type ES cells mixed with their Rosa26-P-geo targeted counterpart. As shown in Fig. 3, panels a and b, this results in a rather homogeneous mixture of blue and unstained cells with a light blue macroscopic appearance when compared with teratomas made exclusively of wild type Rosa26-p-geo ES cells. When the latter were employed together with the Apc-mutant ES cells to generate 20 teratomas, all differentiated cell lineages were found in the resulting chimeric tumors, thus indicating that the homozygous Apc 16 3 8 N and compound Apcl63 8 N /157 2 T ES cells did not affect the differentiation potential of wild type ES cells (Fig. 3). Tissues never observed in Apcl 6 8 8N/1638N or Apcel s 63N/15 l 72T teratomas were entirely composed of wild type cells, whereas other tissues contained both mutant and wild type cells, thus 25 showing that the Apc differentiation defect is cell autonomous. Accordingly, P-catenin staining was strongly elevated and often nuclear in homozygous ApclGSSN cells, but cytoplasmic and membrane-bound in wild type cells even in chimeric epithelia where mutant and wild type cells are in direct contact with each other (Fig. 3, panels e-h). Thus, the Apc mutation is cell autonomous with respect to differentiation and p 30 catenin downregulation capacity. Notably, a cell-sorting phenomenon was observed when the chimeric teratomas were analyzed macroscopically (Fig. 3, panels c and d). This non-uniform distribution of wild type and mutant cells was observed with both the homozygous Apc 16 38N and WO 2004/024901 PCT/NL2003/000640 23 compound Apc16s8N1572 r ES lines and is likely to be due to homotypic cell-cell recognition. Further studies will elucidate the molecular and cellular mechanisms underlying this phenomenon. 5 ApcMin/M i ES cells fail to form teratomas. The above results are indicative of an Apc dosage-dependent defect in ES cell differentiation possibly due to partial loss of -catenin regulation. The Apcmin allele encodes for a short truncated protein deprived of all the P-catenin binding and downregulating domains and therefore unable to control Wnt/p-catenin signaling. We 10 attempted teratoma formation using independent blastocyst-derived ApcM

O

iniVin ES cell lines, together with littermates-derived Apc l

'

+ and Apc+

/M

in ES lines as controls (see Methods). Notably, all attempts to generate teratomas with Apcuin/vin ES cell lines failed in our in vivo assay, whereas their Apc

+

'

+ and Apc+Min counterparts gave rise to teratomas with normal differentiation patterns. 15 In vitro differentiation analysis of ApcMin/Mi and Apc668N/1638N ES cells The inability of Apclin' m in ES cell lines to form teratomas indicates a severe differentiation defect possibly due to the complete loss of P-catenin downregulating activity in this allele. To study in more detail the cellular nature of the most severe 20 Apc differentiation defects, we carried out in vitro differentiation analysis ofApcu i ftin and Apc63 a NI16 38 N ES cell lines by simply withdrawing leukemia inhibiting factor (LIF) from the culture medium and by analyzing cell morphology at different time intervals when compared with wild type cells (Fig. 4). In general, Apc

M

i

'

iMin and Apc 16 3SN1638N ES cells differentiate at a much lower rate than their wild type counterparts (8 days vs. 2 25 days in wild type ES cells; see Fig. 4a). Although differentiation was eventually observed in both mutant genotypes, they appeared to maintain their undifferentiated state in culture for a significantly longer period than the wild type and heterozygous controls. In order to identify the cell types formed upon differentiation, ES colonies were analyzed by immuno-histochemical staining with different lineage-specific 30 antibodies. Apci63s a

N

1 s 688N colonies formed, as in teratomas, massive amounts of smooth muscle but failed to form neuroectodermal derivatives. ApcMin/

M

in colonies formed large amounts of visceral- and parietal-endoderm (identified by morphology and by antibodies against alpha-fetoprotein and vimentin respectively) but failed to WO 2004/024901 PCT/NL2003/000640 24 differentiate into smooth muscle cells or nerve cells. Notably, after prolonged culture ofApcMin cells, only extensive epithelial sheets of parietal endoderm were present (data not shown). Next, we tried to modulate in vitro differentiation of the Apcin and Apc 16 8N ES cells 5 towards neuroectodermal, mesodermal and trophectodermal lineages by complementing the culture medium with retinoic acid, DMSO and Fgf4 respectively19, 2 0 . Compared to the wild type ES line, both ApcMinlMin and Apc638N/16 8 8N ES cells did not respond upon these stimuli as no change in their morphology or expression of tissue-specific markers could be observed (data not shown), thus 10 indicating that the differentiation defects cannot be rescued by external factors. In vitro differentiated ApeMin and Apc 16 3 sN ES clones were also stained for p-catenin and compared to Apc

+

*

+ ES cells (Fig. 4b). An intense nuclear signal was present in most ApcMiniin cells. Apc 16 3 8N/ 168 8N cells also displayed nuclear P-catenin though the signal intensity was considerably weaker and less abundant than in the ApcMin clones. 15 Wild type ES cells did not show any P-catenin nuclear accumulation. Excessive P catenin accumulation in the nucleus has been shown to result in programmed cell death 21 which may explain the inability of Apc M I ni

M

in to form teratomas. Differentiation impairment in -catenin mutant ES cells. 20 The last evidence that the differentiation defect observed in Apc-mutant ES cells is due to the loss of the P-catenin regulating function, was obtained by repeating the teratoma differentiation assay with ES cells carrying a Cre-Lox mediated deletion of P-catenin exon 322. The latter encompasses all phosphorylation serine/threonine target residues of P-catenin which allow its earmarking for ubiquitilation and proteolytic 25 degradation. Two ES lines were employed: in Catnblox(ex) exon 3 of the p-catenin gene is flanked by LoxP sites without compromising functionality of the corresponding llele, whereas Catnb exs was obtained by Cre-mediated deletion of exon 322. The two ES lines were first analyzed by TOPFLASH reporter assay and compared with wild type and Apc16 3 8N/1638N ES cell lines (Fig. 5a). Catnbexs ES cells showed approximately 3 fold 30 higher reporter levels than in Apc168N/ 1 GSsN. Direct comparison between Catnbexs and ApcMinmin ES cells was not feasible as these lines were only available in different WO 2004/024901 PCT/NL2003/000640 25 genetic backgrounds. TOPFLASH reporter levels in CatnboxexS) and wild type cells were indistinguishable. Next, Catnboxexls) and Catnbex 3 ES cells were assayed for their differentiation potential by teratoma formation. Teratomas were successfully derived from both cell lines. 5 Catnblox(x) differentiation profiles were indistinguishable from those obtained with wild type ES cells wehereas Catnbex 3 cells formed teratomas with limited differentiation patterns, similar to those observed in Apc 1 38N/1638N. In particular, ectoderm-derived cell lineages were absent whereas an abundance of smooth muscle and non-ciliated epithelia was observed (Fig. 5b). Nuclear P-catenin localization (Fig. 10 5b) and other immunohistochemical observations obtained with the same set of antibodies employed for the analysis of Apc-mutant teratomas (not shown), confirmed these aberrant differentiation patterns in the Catnbex 3 teratomas. Hence, the differentiation defect observed in Apc-mutant ES cells is due to improper -catenin regulation. Moreover, different dosages of residual P-catenin directly 15 correlate with differences in differentiation potential and with the intensity and frequency of P-catenin nuclear accumulation. Taken together, the data point to a dosage-response effect between Apc mutations affecting the P-catenin downregulating function, the increase in P-catenin mediated transcriptional activity and the differentiation defects in ES cells. 20 Expression profiling and data mining. Loss of P-catenin regulation is likely to result in multiple changes in the transcription of several Wnt downstream genes. To investigate the changes in gene expression that accompany the differentiation of wild type and Apc-mutant ES cells, and to pinpoint 25 potential key regulatory genes, we analyzed the Apc-mutant teratomas by Affymetrix microarray technology. Expression data from the Apc-mutant teratomas were compared to the wild type tumors by an unsupervised, hierarchical clustering algorithm using the Rosetta Resolver (Rosetta Inpharmatics Inc., Kirldand, Washington) (see Methods). As shown in the dendogram in Fig. 6a, this analysis 30 correctly clustered the Apc-mutant teratomas characterized by aberrant differentiation (Apcl 6 38N/1638N and Apc1638N/157 2 T ) from those that were indistinguishable from the wild type tumors at the immuno-histochemical level (Apc 16 38N/1638T and Apc1638T/1638T). It should also be noted that the majority of genes up- or down-regulated WO 2004/024901 PCT/NL2003/000640 26 in Ap 1 a l8Ni s a 1638N, Apc G a 38N/1572T, and, to a lesser extent, in Apcl 1G 8N/168T, were in general not changed in Apcl168T/l63

S

T, as clearly visible in Fig. 6a. The latter confirms and underlines that the Apc-mutant genotypes resulting in the constitutive activation of Wnt signaling are characterized by distinct gene expression profiles when compared 5 with teratomas derived from Wnt-proficient ES lines (as judged by TOP-FLASH reporter assays). Among the differentially expressed entries, both structural tissue specific (Table 1) and regulatory (Table 2) genes were identified. The latter belong to well-known signal transduction pathways such as Wnt (Table 2a), transforming growth factor beta (Table 2b), fibroblast growth factor (Table 2c), and retinoic acid 10 (Table 2d). Although the elucidation of the signal transduction pathways underlying the stem cell differentiation caused by Apc mutations is beyond the scope of the present study, the pattern of differential expression among these genes appears to be in agreement with the differentiation defects observed in the Apc-mutant teratomas. For example, the up-regulation of the bone morphogenic proteins 2 and 4 (Bmp2 and 15 Bmp4, Table 2b) may explain some of the observed differentiation defects. Bmp's are morphogenetic signaling proteins belonging to the Tgf-3 superfamily originally isolated for their capacity to induce ectopic bone formation 23 ,24 . Bmp's signal through heteromeric complexes of type I and type II transmembrane Ser/Thr kinase receptors thus triggering the expression of downstream target genes. Among the latter, the 20 homeobox genes Msxl and Msx2 have been previously reported to be induced by Bminp2/4 25 and are accordingly upregulated among the Apc-mutant teratomas (Table 2b). Expression of Msxl (and Msx2) is known to interfere with the differentiation process by blocking cell cycle exit through Cyclin D 1 upregulation 26 . Though the latter was not significantly upregulated in our data set, the homologue Cyclin D3 was (Table 25 2a). The same is true for Tbx2, a known modulator (both activator and repressor) of bone development whose activity largely depend on the cellular context 2 7 . Tbx2 expression is also induced by Binp2 2 8 and accordingly in our data set (Table 2b). Notably, Dickkopf genes (Dkkl and Dkk2) were also found to be upregulated among the Apc-mutant teratomas (Table 2a). Dkkl and Dkk2 encode secreted proteins that 30 act as potent inhibitors of Wnt signaling and are involved in head induction in Xenopus embryogenesis and are highly expressed during murine development in mesodermal tissues that mediate epithelial-mesenchyme transitions 2 1 9 . The upregulation of the Dkk's inhibitors is in apparent contradiction with the similar WO 2004/024901 PCT/NL2003/000640 27 behavior of several Wnt-related genes including inducers/ligands (Wnt-ligands), receptors (frizzled), transcription factors (Lefl), and downstream targets (Wisp's, cyclins, etc) (Table 2a). However, the complexity and tissue heterogeneity of the teratomas represent clear confounding factors as we cannot ascertain which tissues 5 contribute to which differential gene expression patterns. To visualize the Apc/p-catenin dosage-dependent control of gene expression, we selected a subset of 300 highly differentially expressed genes in Apcl 68 8N1638N (fold changes >5 and <-5) by subtracting those differentially expressed in Apc 16 aTl 163 s 8T when compared with wild type (fold changes >2 and <-2), and followed their behavior 10 in the 4 genotypes. Fig. 6b shows the results of this trend analysis across the different Apc mutations. A clear gradient of transcriptional response is observed starting from the Wnt-proficient alleles (Apc1638T/1638T) and gradually increasing (or decreasing) in more severely affected genotypes (Apc168 a N 1638T, Apc 16 38 a a N1572T, Apc168aNs163aN). The latter clearly indicates that different dosages of P-catenin made available for Wnt signaling 15 by specific Apc defects result in different target gene expression responses and consequently in different degrees of ES differentiation. The rationale for the data mining approach shown in Fig. 6b was the observation that the Apcl 1 63T/163 a 8T ES cells were shown to be Wnt- and differentiation-proficient by TOP-FLASH reporter and by teratoma assay, respectively 9, this study. Hence, the genes 20 differentially expressed in Apc 163 8T/ i 1

GS

a T when compared with wild type teratomas are likely to result from P-catenin independent effects. The recently characterized chromosomal instability of the Apc-mutant ES cell lines here employed 8 0 may partly account for this differential gene expression. However, it is important to state that loss of this function of Apc in mitosis cannot possibly account for the differentiation 25 defects here reported as Apc168T/1638TES cells were shown to be chromosomally instable but differentiation-proficient. Discussion In this study we have shown that Apc mutations affect the differentiation capacity of 30 mouse ES cells in a quantitative and qualitative fashion depending on the dosage of 3 catenin signaling. This direct correlation between differentiation and Apc/P-catenin signal transduction has implications for the understanding of the cellular mechanism underlying Apc-driven tumorigenesis. Although it is generally accepted that the WO 2004/024901 PCT/NL2003/000640 28 tumor suppressor function of Apc resides in its ability to down-regulate P-catening,15, 81 3, the tissue-specific downstream targets responsible for the broad tumor spectrum observed both in FAP patients 12 and Apc-mutant mouse models 6 are still largely unknown. Our results are in agreement with the role of P-catenin signaling in 5 maintaining stem cell properties in the intestine, as also elegantly illustrated by the failure to form crypt stem cells in mice lacking Tcf4 1 3 . In these mice, the neonatal epithelium is composed entirely of differentiated cells. Thus, the genetic program controlled by P-catenin signaling and executed by Tcf-4 maintains the crypt stem cells of the small intestine and, in view of the data presented here, modulates 10 differentiation. In the colonic epithelium, mitotic cell rates equal terminal differentiation and cell loss rates. Intestinal tumors are the result of an increase in this gain:loss ratio. In view of the present data, the latter increase may be the consequence of a differentiation defect. Activation of P-catenin signaling by Apc mutation will retard and/or inhibit differentiation and ultimately result in an 15 enlargement of the stem cell compartment, e.g. the target cell population that undergoes additional mutations eventually leading to tumor formation. Accordingly, recent computer modeling studies implicate that tumor initiation in the colon is caused by crypt stem cell overproduction, i.e. a differentiation defect 1 4 . In an alternative but analogous scenario, a differentiated colonic epithelial cell undergoes 20 Apc (or -catenin) mutation thus triggering a process of de-differentiation and again the enlargement of the size of the stem cell compartment within the crypt. The latter would be in agreement with the recently postulated 'top-down' model for colorectal tumorigenesis where neoplastic transformation is initiated in a fully differentiated cell that becomes dysplastic, spreads laterally and progressively replace the normal 25 crypt cells3 4 . In general, colorectal tumors do show a high proportion of undifferentiated crypt-like cells with high numbers of dividing cells, although all differentiated cell types are still present 2 2

.

35 . The same correlation between activation of Wnt/P-catenin signaling and stem cell maintenance may also apply to other tumor types as it represents an effective way to sustain tumor growth in a broad spectrum of 30 self-renewing tissues. Accordingly, a large number of tumor types are characterized by P-catenin overexpression and/or nuclear localization. Recently, it was shown that overexpression of Pin1 is responsible for the frequently observed p-catenin WO 2004/024901 PCT/NL2003/000640 29 upregulation in breast cancer, a tumor type where APC or P-catenin mutations are not common 36 . Mutations in different members of the Wnt pathway will result in different signaling dosages associated with specific tumors in susceptible tissues. In conclusion, we report that mutations in Apc and p-catenin affect the capacity of 5 embryonic stem cells to differentiate into the three germ layers in a P-catenin dosage dependent fashion. These results have implications for the understanding of the molecular and cellular basis of tumor initiation by defects in the Wnt pathway. We propose a model where adult somatic stem cell compartments are characterized by tissue-specific P-catenin threshold levels for cell proliferation, differentiation, and 10 apoptosis. Different Apc mutations will result in different levels of intracellular p catenin and will confer different degrees of tumor susceptibility in different tissues. Hence, p-catenin dosage-dependent differentiation may not only explain how a single pathway is involved in the development of different tissues, but also its pleiotrophic role in tumorigenesis. 15 WO 2004/024901 PCT/NL2003/000640 30 References 1. Cadigan, K.M. & Nusse, R. Wnt signaling: a common theme in animal development. Genes Dev, 11, 3286-3305 (1997). 2. Seidensticker, M.J. & Behrens, J. Biochemical interactions in the wnt 5 pathway. Biochim Biophys Acta, 1495, 168-182 (2000). 3. Polakis, P. Wnt signaling and cancer. Genes Dev, 14, 1837-1851 (2000). 4. Kinzler, K.W. & Vogelstein, B. Lessons from hereditary colorectal cancer. Cell, 87, 159-170 (1996). 5. Fodde, R., Smits, R. & Clevers, H. APC, signal transduction and genetic 10 instability in colorectal cancer. Nature Reviews Cancer 1, 55-67 (2001). 6. Fodde, R. & Smits, R. Disease model: familial adenomatous polyposis. Trends in Molecular Medicine 7, 369-373 (2001). 7. Su, L.K. et al. Multiple intestinal neoplasia caused by a mutation in the murine homolog of the APC gene. Science 256, 668-670 (1992). 15 8. Moser, A.R. et al. Homozygosity for the Min allele of Apc results in disruption of mouse development prior to gastrulation. Dev Dyn, 203, 422-433 (1995). 9. Smits, R. et al. Apc1638T: a mouse model delineating critical domains of the adenomatous polyposis coli protein involved in tumorigenesis and development. Genes Dev, 13, 1309-1321 (1999). 20 10. Fodde, R. et al. A targeted chain-termination mutation in the mouse Apc gene results in multiple intestinal tumors. Proc Natl Acad Sci US A, 91, 8969-8973 (1994). 11. Smits, R. et al. Apcl638N: a mouse model for familial adenomatous polyposis associated desmoid tumors and cutaneous cysts. Gastroenterology 114, 275-283 25 (1998). 12. Fodde, R. & Khan, P.M. Genotype-phenotype correlations at the adenomatous polyposis coli (APC) gene. Crit Rev Oncog, 6, 291-303 (1995). 13. Korinek, V. et al. Depletion of epithelial stem-cell compartments in the small intestine of mice lacking Tcf-4. Nat Genet, 19, 379-383. (1998). 30 14. Boman, B.M., Fields, J.Z., Bonham-Carter, O. & Runquist, O.A. Computer modeling implicates stem cell overproduction in colon cancer initiation. Cancer Res, 61, 8408-8411 (2001). 15. Korinek, V. et al. Constitutive transcriptional activation by a beta-catenin-Tcf WO 2004/024901 PCT/NL2003/000640 31 complex in APC-/- colon carcinoma. Science 275, 1784-1787 (1997). 16. Stevens, L.C. The biology of teratomas. Adv Morphog, 6, 1-31 (1967). 17. Hay, E.D. & Zuk, A. Transformations between epithelium and mesenchyme: normal, pathological, and experimentally induced. Am JKidney Dis, 26, 678 5 690 (1995). 18. Friedrich, G. & Soriano, P. Promoter traps in embryonic stem cells: a genetic screen to identify and mutate developmental genes in mice. Genes Dev, 5, 1513-1523 (1991). 19. Tanaka, S., Kunath, T., Hadjantonakis, A.K., Nagy, A. & Rossant, J. 10 Promotion of trophoblast stem cell proliferation by FGF4. Science 282, 2072 2075 (1998). 20. Rudnicki, M.A. & McBurney, M.W. Cell culture methods and induction of differentiation of embryonal carcinoma cell lines, in Teratocarcinomas and embryonic stem cells: apractical approach. (ed. Robertson, E.) 19-50 (IRL Press 15 Limited, Oxford, UK, 1987). 21. Kim, K., Pang, K.M., Evans, M. & Hay, E.D. Overexpression of beta-Catenin Induces Apoptosis Independent of Its Transactivation Function with LEF-1 or the Involvement of Major G1 Cell Cycle Regulators. Mol Biol Cell, 11, 3509 3523 (2000). 20 22. Harada, N. et al. Intestinal polyposis in mice with a dominant stable mutation of the beta-catenin gene. Embo J, 18, 5931-5942 (1999). 23. Urist, M.R. Bone: formation by autoinduction. Science 150, 893-899 (1965). 24. Wozney, J.M. et al. Novel regulators of bone formation: molecular clones and activities. Science 242, 1528-1534 (1988). 25 25. Hollnagel, A., Oehlmann, V., Heymer, J., Ruther, U. & Nordheim, A. Id genes are direct targets of bone morphogenetic protein induction in embryonic stem cells. JBiol Chem, 274, 19838-19845 (1999). 26. Hu, G., Lee, H., Price, S.M., Shen, M.M. & Abate-Shen, C. Msx homeobox genes inhibit differentiation through upregulation of cyclin Dl. Development 30 128, 2373-2384 (2001). 27. Chen, J. et al. Microarray analysis of Tbx2-directed gene expression: a possible role in osteogenesis. Mol Cell Endocrinol, 177, 43-54 (2001). 28. Yamada, M., Revelli, J.P., Eichele, G., Barron, M. & Schwartz, R.J. Expression WO 2004/024901 PCT/NL2003/000640 32 of chick Tbx-2, Tbx-3, and Tbx-5 genes during early heart development: evidence for BMP2 induction of Tbx2. Dev Biol, 228, 95-105 (2000). 29. Monaghan, A.P. et al. Dickkopf genes are co-ordinately expressed in mesodermal lineages. Mech Dev, 87, 45-56 (1999). 5 30. Fodde, R. et al. Mutations in the APC tumour suppressor gene cause chromosomal instability. Nat Cell Biol, 3, 433-438 (2001). 31. Morin, P.J. et al. Activation of beta-catenin-Tcf signaling in colon cancer by mutations in beta-catenin or APC. Science 275, 1787-1790 (1997). 32. Lamlum, H. et al. The type of somatic mutation at APC in familial 10 adenomatous polyposis is determined by the site of the germline mutation: a new facet to Knudson's 'two-hit' hypothesis. Nat Med, 5, 1071-1075 (1999). 33. Smits, R. et al. Somatic Apc mutations are selected upon their capacity to inactivate the beta-catenin downregulating activity. Genes Chromosomes Cancer 29, 229-239 (2000). 15 34. Shih, I.M. et al. Top-down morphogenesis of colorectal tumors, Proc Natl Acad Sci USA, 98, 2640-2645 (2001). 35. Moser, A.R., Dove, W.F., Roth, K.A. & Gordon, J.I. The Min (multiple intestinal neoplasia) mutation: its effect on gut epithelial cell differentiation and interaction with a modifier system. J Cell Biol, 116, 1517-1526 (1992). 20 36. Ryo, A., Nakamura, M., Wulf, G., Liou, Y.C. & Lu, K.P. Pinl regulates turnover and subcellular localization of beta-catenin by inhibiting its interaction with APC. Nat Cell Biol, 3, 793-801 (2001).

WO 2004/024901 PCT/NL2003/000640 33 Table 1. Tissue-specific gene expression profiles in Apc-mutant teratomas. Shaded rows indicate genes whose expression patterns have been validated by immunohistochemistry. The listed entries were selected based on their tissue-specific expression (Unigene) from the unsupervised agglomerative cluster analysis by the 5 error model in Rosetta Resolver v3.0- Gene Expression Data Analysis System where a total of 1484 genes were included for having a fold change > 2 and a p-value less than 0.01 when compared with wild type (Apc'+l+). Note that some genes (marked with an asterisk and listed next to each other in the tables) have independent multiple entries in the Affymetrix microarray here employed. As expected, variations in the 10 corresponding fold change values are observed, although the general trend of differential expression is concordant.

WO 2004/024901 PCT/NL2003/000640 34 a. Neural-0specific genes Fold expression changes when compared with Apc+^ + teratomas Unigene Gene symbol and Title Apc1638NI/ApcI638 Apc1638N/ApcSI72 Apc1638N/Apc1638 Apc1 63 BT/Apcl638 N T T T Mm.1287 Mapt* -100 -100 -2,3 1.1 (microtubule-associated protein tau) Mm.1287 Mapt* -95.5 -100 -2.6 -1.1 (microtubule-associated protein tau) Mm.3304 Nsg2 -74.1 -64.6 -4.5 -1.5 (neuron-specific gene family member 2) Mm.4599 Astnl -53.7 -30.2 -3.4 -1.4 (astrotactin 1) Mm.40615 Kcnq2 (potassium voltage- -52.5 -25.1 -5.1 -2.2 gated channel subfamily Q member 2) Mm.2496 Ina (alpha internrexin neuronal -100 -69.2 -2.7 -1.1 intermediate filament protein) Mm.1419 Aqp4* -45.7 -20.4 -1.3 2.4 (aquaporin 4) Mm.1419 Aqp4* -18.2 -8.1 -1.4 2.2 (aquaporin 4) Mm.34637 Catna2* -39.8 -36.3 -5.6 -1.7 (catenin alpha 2) Mm.34637 Catna2* -8.3 -3.5 -2.6 1.1 (catenin alpha 2) Mm.34637 Catna2* -6.8 -10.5 -2.4 -1 (catenin alpha 2) Mm.5309 Gabrg2 (gamma-aminobutyric -37.2 -24.5 -4.7 -1.3 acid GABA-A receptor subunit WO 2004/024901 PCT/NL2003/000640 35 gamma 2) Mm.5101 Sytli -27.5 -14.5 -12.3 -12.3 (synaptotagmin 1) Mm.2991 Oprl -18.2 -8.5 -7.4 -1.7 (opioid receptor-like) Mm.57194 L1cam -18.2 -11.2 -8.3 -10.2 (L1 cell adhesion molecule) Mm.28562 Syt9 -17.8 -17.8 -3 -1.2 (synaptotagmin 9) Mm.2419 Cdh4 -17.4 -6.6 -2.3 1 (cadherin 4) Mm.1239 Gfap -13.8 -4.8 -1.7 2 (glial fibrillary acidic protein) Mm.20892 Syn2 -8.3 -5.8 -1.7 -1 (synapsin II) Mm.4974 Ncam -7.6 -7.6 -2.5 -1.5 (neural cell adhesion molecule) Mm.4921 Gria2 (glutamate receptor -7.2 -23.4 -4.7 -1.5 ionotropic AMPA2 a2) Mm.18086 Sytl1 -6.6 -6.5 -33 -3.4 8 (synaptotagmin 11) Mm.4920 Grial (glutamate receptor -6.2 -10.7 -3.4 -1.5 ionotropic AMPA1 c xl) Mm.10696 Odz4 -5 -1.2 -1.4 1.2 (odd Ozlten-m homolog 4 Drosophila) WO 2004/024901 PCT/NL2003/000640 36 b. Bone- and cartilage-specific Fold expression changes when compared with Apc* + genes teratomas Unigene Gene symbol and Title Apc163N/Apc1638 Apcim 38 N/Apc1 5 7 2 Apc1 638 NjApcB 3 8 Apc1 63 8 T/Apc1638 N T T T Mm.4987 lbsp -4.4 -19.1 -2 -1.4 (integrin binding sialoprotein) Mm.4778 Cspg3 -10 -4.4 -7.2 -3 (chondroitin sulfate proteoglycan 3) Mm.2759 Agc (aggrecan structural -4.4 -5.2 1.5 -2.9 proteoglycan of cartilage) Mm.5091 Spocki (sparclosteonectin -2.3 -3.7 -2.5 -2 cwcv and kazal-like domains proteoglycan 1) Mm.7964 Cdrap (cartilage derived -10.7 -3.8 -1.3 1 retinoic acid sensitive protein) C. Musce-speciff a genes Fold expression changes when compared with Apcv' teratomas Unigene Gene symbol and Title Apc16 38 N/Apc 1 6 38 Apc1638N/Apc 1 57 2 Apc1638N/ApcI 63 8 Apci 63 BT/Apc 1 638 N T T T Mm.1529 Mylpc (myosin light chain -34.7 -16.2 2.9 1.1 phosphorylatable cardiac ventricles) Mm.46514 Mylc2a -13.2 -9.1 4.7 1.3 (myosin light chain regulatory A) Mm.16528, Myog -1.6 4.2 2.1' 1.8 (myogenin) Mm.3153 Myh11 2.2 1.6 5.1 2.2 WO 2004/024901 PCT/NL2003/000640 37 (myosin heavy chain 11 smooth muscle) Mm.36850 Smtn 19.1 13.5 9.5 -2.2 (smoothelin) d. intestine-specific genes Fold expression changes when compared with Apc* + teratomas Unigene Gene symbol and Title ApcI638N/ApcI 6 3 8 Apc638N/Apcls 5 72 Apc1638NIApc 16 38 Apc 6 3 8 TlApc 638 N T T T Mm.15621 Cftr (cystic fibrosis 5.6 3.8 1.4 2.3 transmembrane conductance regulator homolog) Mm.10805 Coll3al 6.6 6.6 -1.1 1.5 (procollagen type XIII alpha 1) Mm.4010 Vii 8.7 4.2 2.3 1.6 (villin) Mm.27830 Slc7a8 (solute carrier family 8 12.3 2.8 -1.1 1.9 cationic a.a. transporter y system member 7) Mm.57132 Cy i 16.6 4 -1.1 1.3 (Mus musculus 129 cryptdin i gene) Mm.15790 Defcr-rsl 32.4 7.2 1.2 1.3 9 (defensin related sequence cryptdin peptide paneth cells) Mm.14017 Defcr5 47.9 8.7 -1.3 2.7 3 (defensin related cryptdin 5) Mm.14271 Defcr-rs2 (defensin related 60.3 4.4 3 1.8 cryptdin related sequence 2) WO 2004/024901 PCT/NL2003/000640 38 e, Lung-specific genes Fold expression changes when compared with Apc

+

*

+ teratomas Unigene Gene symbol and Title Apc16N/Apc i 638 Apc'638N/Apcl 5 72 Apc' 63 NIApc 1638 Apc1638T/ApcI 6 3 8 N T T T Mm,7420 Tubb4 -16.6 -47.9 -2.6 -1.1 (tubulin beta 4) Mm.1041 MIfl -10 -4.6 -2.4 1.1 4 (myeloid leukemia factor 1) Mm.4225 Tektl -6.8 -3.7 -2 1.3 7 (tektin 1) f. Globin genes Fold expression changes when compared with Apc* + teratomas Unigene Gene symbol and Title ApclB3 8 N/Apc 1 638 Apc16 3 8 N/Apc57 2 Apcl 63 8NIApcl 6 3 8 ApcI638T/ApcI638 N T T T Mm.2308 Hbb-bhl* 4.9 11 4.7 3.2 (hemoglobin Z p3-like embryonic chain) Mm.2308 Hbb-bhl* 26.9 18.6 4.5 -1.7 (hemoglobin Z beta-like embryonic chain) Mm.14175 Hba-x (hemoglobin X ca-like 15.8 28.8 7.2 -1.1 8 embryonic chain in Hba complex) Mm.35830 Hbb-y 100 81.3 16.2 -1.3 (hemoglobin Y beta-like embryonic chain) 5 WO 2004/024901 PCT/NL2003/000640 39 Table 2. Signal transduction gene expression profiles in Apc-mutant teratomas. The listed entries are derived from the unsupervised agglomerative cluster analysis by the error model in Rosetta Resolver v3.0 Gene Expression Data Analysis System (1484 genes included for having a fold change > 2 and a p-value less than 0.01 when 5 compared with wild type) and classified into the different signaling pathways based on the information provided by the Affymetrix Gene Ontology annotations (https://www.affymetrix.com/analysis/index.affx). 10 WO 2004/024901 PCT/NL2003/000640 40 a Wnth pathway genes Fold expression changes when compared with Apc+/* teratomas Unigene Gene symbol and Title Apc638N/Apc1638 ApcI1638N/ApCI 5 72 Apc1638N/Apc1638 Apc1 638 T/Apc1638 N T T T Mm.20355 Wnt4 (wingless-related MMTV 18.6 3.9 5.6 5.2 integration site 4) Mm.22182 Wntil (wingless-related 10.5 5.4 1.9 1.1 MMTV integration site 11) Mm.2438 Wnt6 (wingless-related MMTV 8.1 1.9 -1.3 1.4 integration site 6) Mm.32207 Wnt5a (wingless-related 7.1 3.2 -1 1.1 MMTV integration site 5A) Mm.5130 WntlOa (wingless related 6 1.3 1 1.9 MMTV integration site 10a) Mm.45050 Fzdl 3.3 1.7 -2.2 -3.9 (frizzled homolog 1 Drosophila) Mm.10359 Dkk2 21.9 14.1 1,9 -1.3 3 (dickkopf 2) Mm.7960 Dkkl 11.5 8.3 3.7 1.4 (dickkopf homolog 1 Xenopus laevis) Mm.2029 Leff 2 1.3 -1.8 1.1 (lymphoid enhancer binding factor 1) Mm.5080 Sox17 4.8 4.2 3.5 1.9 (SRY-box containing gene 17) Mm.2580 Sdcl 18.6 6.5 2.2 -1.1 (syndecan 1) Mm.4272 Slugh 6.8 5 2.2 1.2 (slug chicken homolog) Mm.6813 Bmp4 3.9 2.3 1.4 -1.1 WO 2004/024901 PCT/NL2003/000640 41 (bone morphogenetic protein 4) Mm.13828 Wisp2 (WNT1 inducible 5.1 2.8 1.9 1.5 signaling pathway protein 2) Mm.7417 Ccnd3 2.6 3.2 2.1 -1.2 (cyclin D3) Mm.16234 Itga5 (integrin alpha 5 10 6.5 4.8 2.3 fibronectin receptor alpha) Mm.5039 Six2 (sine oculis-related 7.9 8.7 1.9 -1.3 homeobox 2 homolog Drosophila) b. TgfAJ pathway genes Fold expression changes when compared with Apc*' teratomas Unigene Gene symbol and Title Apc1 638 N/ApcI 63 8 Apcl 638 N/Apci 572 Apc1638N/ApcO638 Apc1 638 T/Apc 1 638 N T T T Mm.57216 Bmp2 13.2 3 1.9 -1.2 (bone morphogenetic protein 2) Mm.6813 Bmp4 6.3 4.3 1.4 -1.4 (bone morphogenetic protein 4) Mm.9154 Tgffl1 5.5 2.8 -1.6 -3 (transforming growth factor beta 1) Mm.18213 Tgff2 4.1 2.4 -1.5 -1.4 (transforming growth factor beta 2) Mm.8042 Inhpa 7.8 1.4 -1.7 1.2 (inhibin beta-A) Mm.9404 Nbl1 (neuroblastoma 6 2.6 1.3 2.5 suppression of tumorigenicity 1) WO 2004/024901 PCT/NL2003/000640 42 Mm.19307 Ltbpl (latent transforming 6 1.9 2.2 1.3 growth factor beta binding protein 1) Mm.1763 Msx2 69.2 8.5 2.8 2.3 (homeo box msh-like 2) Mm.870 Msxl 5 3.2 -1.1 1.1 (homeo box msh-like 1) Mm.5194 Dx3 17 4.1 1.4 -1.5 (distal-less homeobox 3) Mm.23467 D5Ertdl89e (DNA segment 11.2 8.9 1.3 1.4 Chr 5 ERATO Doi 189 expressed) Mm.4605 Tbx2 4.8 4.6 1.3 1.6 (T-box 2) Mm.4509 Runx2 8.7 2.6 -1 -1.8 (runt related transcription factor 2) c.Fgfp athway genes Fold expression changes when compared with Apc* teratomas Unigene Gene symbol and Title Apc1638INApc 1638 Apc1638HNIApc1572 Apc1638N/ApcI63 ApcI16 38 T/Apc 1 638 N T T T Mm.4912 Fgfr4 13.8 3.2 -1.3 -1.5 (fibroblast growth factor receptor 4) Mm.16340 Fgfr2 3.5 1.7 1 -6 (fibroblast growth factor receptor 2) Mm.3157 Fgfrl 3.2 1.9 1.2 1.3 (fibroblast growth factor receptor 1) Mm.16340 Fgfr2 2.5 1 -1 -2.5 (fibroblast growth factor receptor 2) WO 2004/024901 PCT/NL2003/000640 43 Mm.2580 Sdcl 18.6 6.5 2.2 -1.1 (syndecan 1) Mm.67919 Mafb (v-maf 11.5 2.3 1.7 1.1 musculoaponeurotic fibrosarcoma oncogene family protein B) d. Retinoic acid pathway genes Fold expression changes when compared with Apc' * / teratomas Unigene Gene symbol and Title Apcl638NIApc 1

G

3 8 Apcl638N/Apc 5 7 2 Apc1638N/Apc 1 638 Apcl 6 3 8 TIApcl 6 3 8 N T T T Mm.1273 Rarg 46.8 10.7 2.8 -2 (retinoic acid receptor gamma) Mm.34797 Crabp l 9.8 6.6 3.3 5.4 (cellular retinoic acid binding protein I) Mm.4757 Crabp2 3.2 3.3 1.4 1.9 (cellular retinoic acid binding protein 11)

Claims (18)

1. A method for modulating differentiation in a population of cells comprising modulating the presence of P-catenin in said population of cells.

2. A method according to claim 1, wherein the presence of P-catenin is modulated in a 5 population comprising stem cells.

3. A method for enhancing differentiation in a population of stem cells, by diminishing the presence of P-catenin in said population of stem cells. 10

4. A method according to any one of claims 1-3, wherein the presence of P-catenin is modulated by modulating the activity of the Wnt signalling pathway in said population.

5. A method according to claim 4, wherein differentiation is enhanced in said 15 population by decreasing the presence of P-catenin in said population of cells.

6. A method according to any one of claims 1-5, wherein the presence of P-catenin in said population of cells is diminished by at least in part enhancing the protein activity of a multiprotein complex comprising GSK3B, axin/conductin and APC. 20

7. A method according to any one of the afore going claims, wherein said population is present in the GI tract of an individual.

8. A method according to any one of claims 1-7, wherein said P-catenin presence is 25 affected below/above a threshold associated with a stem cell/differentiation phenotype in said population of cells.

9. A method according to claim 6 or claim 7, wherein the protein activity is provided through a nucleic acid sequence encoding said at least part of said protein activity. WO 2004/024901 PCT/NL2003/000640 45

10. A method according to claim 9, wherein said nucleic acid sequence is provided in a gene delivery vehicle. 5

11. A Wnt signalling pathway antagonist for use in enhancing differentiation in a population of stem cells.

12. A protein capable of enhancing the protein activity of a multiprotein complex comprising GSK31, axin/conductin and APC for use in enhancing differentiation in a 10 population of stem cells.

13. A protein capable of enhancing the protein activity of a multiprotein complex comprising GSK33, axin/conductin and APC for use in enhancing differentiation in a population of stem cells according to claim 12, which is active APC. 15

14. A gene delivery vehicle comprising a nucleic acid sequence encoding a wnt signalling pathway antagonist for use in enhancing differentiation in a population of stem cells. 20

15. A gene delivery vehicle comprising a nucleic acid sequence encoding at least one protein from a multiprotein complex comprising GSK3B, axin/conductin and APC for use in enhancing differentiation in a population of stem cells.

16. A pharmaceutical formulation suitable for enteral administration comprising 3 25 catenin inhibiting activity.

17. A pharmaceutical formulation according to claim 16, wherein said activity is provided through a gene delivery vehicle comprising a nucleic acid sequence encoding at least one protein from a multiprotein complex comprising GSK3B, axin/conductin 30 and APC, or a nucleic acid sequence encoding a Wnt signalling pathway antagonist. WO 2004/024901 PCT/NL2003/000640 46

18. A pharmaceutical formulation according to claim 16, comprising at least one protein of a multiprotein complex comprising GSK3B, axin/conductin and APC or a Wnt signalling pathway antagonist.