WO2000018913A1

WO2000018913A1 - Apc-2

Info

Publication number: WO2000018913A1
Application number: PCT/NL1999/000595
Authority: WO
Inventors: Johannes Hendrikus Van Es; Mark Alan Pelfer; Johannes Carolus Clevers
Original assignee: Universiteit Utrecht
Priority date: 1998-09-25
Filing date: 1999-09-24
Publication date: 2000-04-06
Also published as: AU6010699A

Abstract

The invention relates to the field of human or veterinary medicine, more specifically to the field of cancer, its diagnosis, treatment and prevention and to the field of drug discovery against cancer. The invention provides an isolated and/or recombinant nucleic acid or a specific fragment, homologue or derivative thereof, corresponding to a gene located in humans on the p arm of chromosome 19 at around position 19p13.3 and comprising a tumour suppresser gene or specific fragments thereof.

Description

Title : APC-2 .

The invention relates to the field of human or veterinary medicine, more specifically to the field of cancer, its diagnosis, treatment and prevention and to the field of drug discovery against cancer. Despite extensive knowledge relating to the multitude of cancer forms (varying in appearance from solid tumours and related metastases in distinct parts of the body to leukaemia's of blood cells that circulate throughout the body, and varying from being totally benign to being aggressively malignant) effective therapy of cancer is difficult and in general restricted to three types; treatment via radiation, via chemotherapy and via surgery.

Possibilities for a more specific therapy, directed against the underlying cause of a specific cancer or group of cancers are currently virtually non-existing. Extensive efforts are directed at providing such specific drugs through drug discovery attempts that try to identify candidate drugs for specific cancer therapy.

Development of cancer often starts with changes in a cell that lead to the unrestricted development and division of that first cell into an ever dividing population of cells. These changes are often an accumulation of mutations or other alterations in key genes that occur chronologically, whereby the mutated cell population looses its original, often specialised character and acquires more and more of a cancerous nature. Normal processes of growth regulation of cells are dysfunctioning in the altered cells. Transcription of genes that are normally only little expressed in said cell type is in cancerous cells no longer controlled. Activation of transcription of genes by transcription factors that would otherwise be dormant in the specific cell type can for example lead to the so typical unrestricted growth and neoplastic nature of cancer. An example are mutations in suppresser genes that function normally by generating proteins that are suppressing transcriptional pathways which are no longer of use in a specialised cell. Mutated suppresser genes do not longer help keep the growth of a cell at bay. Drugs directed against or intervening with the specific protein-protein or protein-DNA interactions in transcriptional pathways controlling cell growth or development can be considered typical candidate drugs for later use in specific cancer therapy, especially when such pathways have gone awry and lead to unrestricted growth cells.

A typical example of a transcriptional pathway gone wrong and leading to development of cancer, can be found with adenomatous polyposis coli (APC) . Mutations in this gene are among the most common disease-causing events in humans, approximately 50% of the population will, during a normal life span, develop colorectal polyps initiated by APC mutations. Individuals who inherit APC mutations develop thousands of colorectal tumours . APC protein interacts with at least six other proteins, β-catenin, γ-catenin, tubulin, EB1, hDLG and Z 3/GSKβ kinase that may be involved with communicating APC related growth control . Colon carcinoma cells with mutant APC contain large amounts of monomeric, cytoplasmic β-catenin. Reintroduction of wild-type APC reduces the overall amount of β-catenin.

Especially during the last decade the molecular genetic analysis of colorectal cancer has revealed that the adenomatous polyposis coli {APC) tumor suppresser gene, originally identified as the gene responsible for familial adenomatous polyposis (FAP) , plays a rate-limiting role in colorectal tumour formation: it is mutated in the majority of sporadic colorectal tumours and inactivation of both APC alleles occurs at early stages of tumour development in mouse and man. Moreover, although the colorectal tumours represent the hallmark of FAP, germline APC mutations often result in a broad spectrum of lesions of ecto-, meso- and endodermal origin. In fact, FAP patients are at high risk for the development of desmoids (fibromas) , duodenal and gastric tumors, congenital hypertrophies of the retinal epithelium (CHRPEs) , epidermal cysts, osteomas, CNS tumours, and others. The latter observation clearly indicates that the APC gene plays a critical role in the maintenance of tissue homeostasis at many different sites throughout the organism. The importance of the role played by the APC gene in homeostasis and tumourigenesis has been confirmed by several functional studies: the 312 kDa APC protein is involved in a variety of cellular processes including E-cadherin-mediated cell adhesion, Wnt signaling, apoptosis, microtubule assembly, cell motility and differentiation. In particular, the β-catenin regulating activity by APC seems to be crucial for its developmental and tumour-suppressing function both in relation with its role in cadherin-mediated cell adhesion and in Wnt signaling.

APC competes with E-cadheπn for the interaction with the cytoplasmic proteins termed β- and γ-catenin (also known as armadillo and plakoglobin) with which they form mutually exclusive complexes. The competition between APC and E- cadherin is likely to be regulated by phosphorylation: β- catenin phosphorylation by receptor tyrosine kinases such as v-src, c-erbB-2, EGF, and HGF correlates with the dissociation of cadherin-catenin complexes from the actin cytoskeleton, decreased cell-cell adhesion and increased cell motility. Protein tyrosine phosphatases are often bound to the cadherin-catenin complex and may counteract tyrosine kinases in regulating cell adhesion.

The interaction with β-catenin makes of APC also a key member of a very important signalling cascade, namely the Wingless (in Drosophila) or Wnt (in mammals) signal transduction pathway. The Wnt/Wingless pathway is now recognised to function in very critical biological processes such as embryonic induction, the generation of cell polarity, and the specification of cell fate. In general, the secreted Wnt/Wingless glycoproteins interact with receptors of the frizzled gene family, thus activating the cytoplasmatic phosphoprotein dishevelled (dsh) . Dsh inhibits the function of the serine/threonine kinase GSK-3β or of its Drosophila homolog zeste whi te 3 (Zw3) . Inhibition of GSK-3β results in the accumulation of hypophosphorylated β-catenin in the cytoplasma and its translocation into the nucleus where it forms nuclear heteromeric complexes with members of the Tcf/Lef/pangolin family of HMG transcription factors.

These complexes can activate target genes (for example ul trabi thorax and engrailed in Drosophila, and siamoi s in

Xenopus) the identity of which is still largely unknown in mammalian cells. In the absence of the Wnt signal, GSK-3β forms a multiprotein complex together with conductin (and/or axin) , APC, and β-catenin thereby promoting the phosphorylation of the latter two proteins. This interaction leads to the rapid β-catenin degradation through the ubiquitin-proteasome pathway. Hence, loss of APC in mammalian cells results in a critical loss over β-catenin/Arm control, leading to constitutive signalling to the nucleus and subsequent cell transformation, most likely because of the uncontrolled activation of the downstream target genes. Indeed, recent studies show that the c-myc oncogene is a target of the

APC/ β-catenin/ TCF- 4 signalling pathway.

As the Wnt pathway plays an essential role in a broad range of cell types, tissue specificity seems to be conferred by the two extremes of the signalling cascade, namely the extracellularly secreted Wnt glycoproteins and their frizzled receptors, and the downstream Wnt-effectors, namely the Tcf/Lef family of HMG transcription factors. An example of the tissue-specificity of some components of the Wnt signalling cascade has been recently provided: Tcf4^''^~ (knockout-) mice die shortly after birth due to a single abnormality, i.e. the absence of the proliferative compartment in the intervillus region of the small intestine. hTcf-4 was found to be the only member of the Tcf family expressed in colonic epithelium and in the genetically predisposed colorectal polyps found in APC and it was suggested that wild type APC regulates the interaction between β-catenin with hTcf4 whereas mutant APC leads to accumulation of β-catenin. In colon cells, APC therefore regulates the formation of transcriptionally competent β- catenin/Tcf-4 complexes, and loss of APC function results in uncontrolled activation of Tcf-4 target genes, thereby contributing to colon tumourigenesis. This observation was supported by the finding that the constitutive transcriptional activity of Tcf-4-responsive reporter genes was strongly increased in APC ^"/^" colon carcinoma cells but inactive in non-colonic cells or cell lines of tumour or non- tumour origin. Loss of APC function can thus be seen as a crucial event in the early transformation of colonic epithelium. The Tcf^"4 gene is expressed in a variety of sites, including the developing CNS and the gut epithelium throughout life. Therefore, it seems likely that the function of Tcf4 in maintaining the crypt stem cells is unique for the small intestinal epithelium, whereas functional redundancy with other members of the Tcf family may compensate the primary defect in other tissues. On the other hand, APC and β-catenin appear to represent more general members of the pathway, as suggested by the fact that germline APC mutations lead to tumorigenesis in tissues of endo-, ecto- and mesodermal origin both in man¹⁷ and mouse

Above finding of APC and APC related hTcf-4 deregulation is, albeit enhancing insight in tumourigenesis, restricted to the narrow field of colon carcinoma. However, there is an obvious need for gaining understanding and access to diagnosis and treatment in a much broader field. The invention provides among others access to and insight in protein-protein or protein-DNA interactions in a transcriptional pathway controlling cell growth or development throughout a wide range of cells and tissues of the body, and provides means, such as nucleic acid, protein, cells and experimental animals and methods to identify candidate drugs, for example for use in cancer therapy. In a first embodiment, the invention provides an isolated and/or recombinant nucleic acid or a specific fragment, homologue or derivative thereof, corresponding to a novel tumour suppresser gene (herein also called APC-2) located in humans on the p arm of chromosome 19 at around position 19pl3.3 and comprising a nucleic acid sequence as shown in figure 1. Where in this application the definition "nucleic acid" is used, both RNA and DNA, in single or double-stranded fashion, and nucleic acid hybridising thereto is meant. Examples of specific hybridisation are given in the experimental part. "Specific fragment" herein meaning a nucleic acid or part thereof that is functionally or structurally related to or hybridising with a distinct nucleic acid or fragment thereof. Typical examples of such a specific fragment as provided by the invention are fragments related to or hybridising with an isolated and/or recombinant nucleic acid or a specific fragment, homologue or derivative thereof, corresponding to a gene located in humans on the p arm of chromosome 19 at around position 19pl3.3 and comprising a nucleic acid sequence or specific fragment thereof as shown in figure 4. "Homologue" herein meaning a related nucleic acid that can be found with another species. "Derivative" herein meaning a nucleic acid that has been derived by genetic modifications, such as deletions,, insertions, and mutations from a distinct nucleic acid or fragment thereof. "Corresponding" herein meaning having a nucleic acid sequence homology of at least 50%, preferably of at least 70, more preferably of at least 80%, most preferably of at least 90%. The invention also provides cDNA, or a specific fragment, homologue or derivative thereof, corresponding to an APC-2 gene, as for example in the experimental part of the description where cloned cDNA corresponding to said gene and its expression patterns are determined. APC-2 stands at the basis of a transcriptional pathway employed in cell -development much as APC does, and the APC-2 gene plays a critical role in the maintenance of tissue homeostasis at many different sites throughout the organism. In the experimental part, the expression pattern of APC2 was studied in relation to APC expression. APC and APC2 are both ubiquitously expressed with particularly high levels in the central nervous system, we noticed preferential expression of APC2 in lung and kidney (figure 2) , indicating that that germline and/or somatic APC-2 mutations lead to tumourigenesis in tissues of endo-, ecto- and mesodermal origin, for example in lung and/or kidney. Loss of APC-2 (function) will result in cancer, particularly in lung and kidney cancer, but likely also in any of the other tissues that generally express APC2. APC-2 controls the Wnt/ Beta-catenin/Tcf pathway in much the same way as APC, i.e. by interacting with GSK-3beta, Axin/conductin, and beta-catenin. Loss of APC2 leads for example to constitutive formation of nuclear complexes between beta-catenin and members of the Tcf family. APC2 also interacts with APC since the N-terminal dimerization domain is virtually identical. Mutations in one (APC-2 or APC) therefore affect function of the other.

The invention for example provides an isolated and/or recombinant nucleic acid or a specific fragment, homologue or derivative thereof, corresponding to a novel tumour suppresser gene (APC-2) comprising cloned human, mouse or Drosophila genomic fragments of APC-2.

The invention also provides an expression vector comprising a nucleic acid as provided by the invention. For example, such a vector comprising a nucleic acid corresponding to a human or mouse APC-2 gene or fragment thereof has been provided in the experimental part of the description. It is within the skills of the artisan to provide vectors that have been provided with single or multiple nucleic acid changes, deletions and/or insertions or other mutations of a nucleic acid sequence.

The invention also provides a cell comprising a genome in which a nucleic acid sequence corresponding to a nucleic acid according to the invention has been modified. Such a modification can comprise a (for example site-directed or transposon-directed or chemically induced) mutation of a nucleic acid encoding a (fragment) of a APC-2 gene, be it in the intronic or exonic sequences of said gene. The invention also provides a cell comprising a nucleic acid according to the invention that has been introduced via recombinant means known to the skilled artisan, for example via homologous recombination techniques or by using a vector according to the invention.

The invention also provides a cell capable of expressing an APC-2 protein or derivative or fragment thereof. Such a cell as provided by the invention is for example derived of a cell comprising a genome in which a nucleic acid sequence corresponding to a nucleic acid according to the invention has been modified or is derived of a cell comprising a nucleic acid according to the invention. Expression of proteins in recombinant cells is in itself a technique available to the average artisan.

The invention also provides a conventional cell capable of expressing APC-2 mRNA or protein or derivative or fragment thereof. Cells expressing a specific mRNA can easily be detected by a skilled artisan, for example by applying hybridisation techniques with nucleic acid probes. Similarly, cells expressing a specific protein can be detected by a skilled artisan by applying techniques such as electrophoreses and/or immunological detection. Such a conventional cell as provided by the invention is for example present in a mouse embryo, a preferred example is a stem cell, but is more preferably derived of a cell line, such as an embryonic stem cell line.

The invention also provides a cell capable of expressing APC-2 mRNA or protein or derivative or fragment thereof, wherein said cell is additionally provided with a nucleic acid encoding a repressible gene, such as a β-catenin gene, Wnt gene or other gene of the Wnt pathway or derivative or fragment thereof. At least six Wnt family genes (i.e. Wnt-1, -2, -3a, -4, -5a, -5b) are known, which are expressed in largely overlapping regions within the embryo. Null mutants in at least three Wnt genes have been described. In a cell provided by the invention which is additionally comprising a nucleic acid encoding a repressible gene or derivative or fragment thereof, Wnt induced β-catenin complexing and subsequent target gene trans-activation, and the influence of a variety of compounds thereof, are studied.

The invention also provides a cell capable of expressing a APC-2 mRNA or protein or derivative or fragment thereof, wherein said cell is additionally provided with a nucleic acid encoding a reporter gene such as a Tcf- responsive reporter gene. Tcf -responsive reporter (Tcf- reporter) genes are those constructs which comprise a readily detectable or assayable gene (such a luciferase, β- galactosidase, chloramphenicol acetyltransferase) linked in cis to a Tcf-responsive element. Such responsive elements are known in the art and any such elements can be used. An optimal Tcf motif contains the sequence CCTTTGATC . From one to twenty copies, and preferably from three to six copies, of the motif may be used. Mutation of the sequence to CCTTTGGCC abrogates responsiveness. Another necessary part of such constructs is a minimal promotor, such as the c-Fos or the herpesvirus thymidine kinase promotor. Transcription of the reporter gene may be performed by any means known in the art , usually by assaying for the activity of the encoded gene, although immunological detection methods can also be used. In addition, transcription can be monitored by measuring the transcribed mRNA directly, typically using oligonucleotide probes. In a cell provided by the invention which is additionally comprising a nucleic acid encoding a reporter gene, target gene activation, and the influence of a variety of compounds thereof, are studied by studying the effect of compounds on the transcription of the reporter gene.

The invention also provides an animal comprising a genome in which a nucleic acid sequence corresponding to a nucleic acid according to the invention has been modified or introduced. Such an animal is for example a transgenic animal obtained by modifying an embryonic stem cell. Stem cell modifications, as well as modifications of other cells, are known to the average skilled artisan. Such an animal is for example a homozygous knock-out animal (APC-2^"7") or a heterozygous (APC-2^+/") animal, for example a cross between a knock-out animal with a wild type animal or is otherwise modified in a nucleic acid fragment in its genome that is corresponding to a nucleic acid provided by the invention. A preferred embodiment of the invention is wherein said animal is a mouse.

The invention also provides an APC-2 protein or derivative or fragment thereof. Such a protein or derivative or fragment thereof according to the invention is for example encoded by a nucleic acid according to the invention or produced by a cell according to the invention. Proteins, be it natural proteins or recombinant versions thereof can easily be isolated by a skilled artisan, for example when at least a part of the amino acid sequence, or a specific antibody directed against the protein, is provided. The invention provides such an isolated or recombinant protein or (poly) peptide or derivatives or fragments thereof.

In addition, the invention provides an antibody (anti- APC-2) specifically directed against an APC-2 protein or (poly) peptide or derivative or fragment thereof according the invention. It is within the skills of the average skilled artisan to provide a (synthetic) antibody directed against a protein or fragment thereof, once the amino acid sequence or the isolated protein is provided.

The invention also provides a method for identifying a candidate drug comprising use of a cell or an animal or a protein or an antibody according to the invention. Drugs directed against or intervening with the specific protein^" protein or protein-DNA interactions in APC-2 related transcriptional pathways controlling cell growth or development can be considered typical candidate drugs for later use in specific cancer therapy, especially when such pathways have gone awry and lead to unrestricted growth of cells. Candidate drugs, often first selected or generated via combinatorial chemistry, can now be tested and identified using a method provided by the invention. Such a drug can be a classically derived chemical compound, or can be a compound having been derived from an APC-2 nucleic acid or protein as provided by the invention, such as reacting with or derived from the N-terminal dimerisation domain, the β-catenin binding domain, the axin/conductin binding domain or other specific fragments of the protein. Such a candidate drug or compound can for example be tested on and selected for their effect on APC-2 regulated transcription, as for example measured by using a reporter gene, such as reporter plasmid pTOPFLASH or the mutant negative control plasmid pFOPFLASH (Korinek et al . , 1997). For example, a test compound or drug which inhibits the transcription of said reporter gene is a candidate drug for cancer therapy. A preferred embodiment of the invention is a method for identifying a candidate drug wherein said drug is tested on β-catenin/APC-2 regulated transcription, for example when said transcription is part of a Wnt/Wingless signalling cascade. For example, a test compound or drug which inhibits the formation of a β- catenin/complex is a candidate drug for cancer therapy. In a most preferred embodiment of the invention, said testing is performed under conditions wherein Wnt/Wingless signalling can be repressed, thereby for example studying the interaction between Wnt/Wingless signalling and a test compound. A method provided by the invention for identifying a candidate drug can be accomplished in vitro or in cells. If the method is to be accomplished in cells, then a reporter gene such as a Tcf-responsive reporter gene must be introduced into the cell. Any means for introducing genetic material into cells can be used, including but not limited to infection, transfection, electroporation. Suitable cells provided by the invention are cells or cell lines recombinantly or conventionally expressing APC-2 mRNA or APC- 2 protein.

A preferred embodiment of the invention is a method for identifying a candidate drug wherein said drug is for use in a cancer patient. In the above it is illustrated that deregulation of transcription is an important cause of cancer. Now that it is possible to study deregulation of transcription-activation by allelic variants of APC-2 in a wide range of cells under many conditions, for example under various conditions of Wnt-signalling, it is possible to test for drugs that intervene in transcription-activation by APC- 2. Such intervening drugs are extremely useful to treat cancers wherein APC-2 regulated transcription and re- programming gene expression in cells is an underlying cause of the cancer. An example of such a method provided by the invention is wherein said cancer is related to foetal cancer, cancers with a stem cell phenotype, brain, lung, kidney or intestinal cancer.

The invention also provides a method for diagnosing cancer comprising use of a nucleic acid or a cell or an animal or a protein or an antibody provided by the invention. For example cells that express APC-2 gene products or malignant variants (that for example are changed in expression pattern or structure, or truncated protein) thereof can now easily be detected in samples taken from a patient suspected of having a cancer. The invention is described more in detail in the figures and the experimental part of the description without limiting the invention thereto.

Experimental part

The APC tumor suppressor protein controls the Wnt pathway by the formation of a complex with GSK-3b, axin/conductin and b- catenin, inducing the rapid degradation of the latter. In colon cancer, loss of APC leads to accumulation of β-catenin in the nucleus, where it binds and activates the Tcf-4 transcription factor reviewed in ¹. Here, we report the identification of APC relatives in mammals (APC2) and flies (dAPC2). APC2, resembling APC in size and overall domain structure, was further analysed. It contains 20-aminoacid repeats and SAMP repeats, which in APC constitute interaction domains for β-catenin and for axin/conductin respectively. Like APC, APC2 regulates the formation of active β- catenin/TCF complexes, as demonstrated in the axis determination assay in Xenopus . Human APC2 maps to chromosome 19pl3. APC and APC-2 perform comparable functions in development and cancer.

The tumour suppressor gene APC is conserved from man ² to fly ³. A search of mammalian EST databases with the human APC sequence revealed the existence of an EST clone (Genbank Accession number H50183), potentially encoding a fragment of a protein distinct from APC, but with distant similarity to armadillo repeats 2 to 6 of APC proteins. The EST clone was derived from a human fetal brain cDNA library. Subsequent sequencing of the corresponding 1200 bp EST cDNA clone allowed us to make several corrections in the submitted sequence and extended it further to the 5 ' of the cDNA fragment. We termed the gene APC2. This finding prompted us to perform a similar database screen for Drosophila . Again, the search revealed the existence of an EST clone, distantly related but not identical to the known Drosophila APC (dAPC) . The sequencing of the nearly full-length cDNA confirmed the existence of a second APC relative in the fruit fly (Figure 1) •

To identify human tissues from which further APC2 sequences could be cloned, tissue-specific expression of APC2 was compared to that of APC by probing normalised poly-A mRNA dot blots with APC2- and APC-specific cDNA probes ⁴. Like APC, APC-2 was broadly expressed with highest levels throughout the central nervous system (Figure 2) . We then used this fragment to screen human fetal kidney and brain cDNA libraries ⁵. In addition, we obtained a mouse genomic Pi clone which contained the complete APC2 gene ⁵. Compilation of sequences allowed us to identify a single long open reading frame of 2247 amino acids for APC2. In Figure 1, the predicted domain structure of APC-2 is compared to that of APC, dAPC and dAPC2. Highest levels of identity were found in the N-terminal 728 amino acids, within the dimerization domain and the armadillo repeat region. C-terminal to the armadillo repeat region, the similarity dropped steeply. Despite this, protein sequence motifs that allow APC to interact with β-catenin, the so called 20-aminoacid repeats ^1,D, could be identified (Figure 1) . In APC, SAMP repeats ⁷ are interspersed between the 20-amino acid repeats and mediate binding to Axin/conductin ⁷'⁸. The 20 amino acid/SAMP repeat region of APC has been shown to harbour the functional regions to control β-catenin in colorectal cancer cells ⁹ and in Xenopus axis induction¹⁰. Like dAPC, dAPC2 contained 15- and 20- amino acid β-catenin interaction domains, but lacked obvious SAMP repeats (Figure 1) .

Having established APC2 (for example in human, mouse and fly) as an candidate disease gene, we determined its chromosomal map by FISH analysis on metaphases of human leukocytes. The gene could be unambiguously assigned to chromosome 19pl3 (see Fig.3). As a prelude to gene disruption, we obtained a murine PI genomic clone. This clone comprised the complete APC2 gene. In addition, it contained the proprotein convertase-4 gene, which has been mapped to chromosome 10 in mouse and chromosome 19 in man ⁿ. Consultation of the Online Mendelian Inheritance Map-database did not reveal any immediate candidate disorders resulting from inherited or sporadic APC2 mutations.

APC plays a key control function in the Wnt signaling pathway by its ability to complex with axin/conductin and GSK-3β. In the currently held scenario ¹'⁸, β-catenin is believed to be phosphorylated by GSK-3β when it participates in this complex. This modification of β-catenin leads to its rapid ubiquitination and subsequent destruction by the proteasome ¹². Signaling through the Wnt cascade decreases the activity of GSK-3β. The consequent alteration of the phosphorylation status of β-catenin rescues it from destruction and allows it to travel to the nucleus to associate with Tcf factors ^{13, 14}. This in turn results in the activate transcription of specific Tcf target genes. In colon cancer and melanoma, mutations in APC or β-catenin result in the constitutive formation of nuclear β-catenin/Tcf-4/Lef, resulting in constitutive transcription of target genes ¹⁵ , such as c-Myc ¹⁶. The activity of Tcf-4 has been shown to be essential for maintenance of stem cells in the intestine ¹⁷.

The presence in APC2 of conserved interaction domains for β-catenin and for axin/conductin predicted its involvement in the control of β-catenin-mediated Tcf target gene transcription. To test this directly, we applied a well- established, functional assay: the Xenopus axis duplication assay. We chose to express the minimal fragment of APC-2 containing the 20-amino acid and the two SAMP repeats. We injected RNA encoding from the same regulatory region of APC2 (aa 705 to 1902) into early Xenopus embryos to assess its potential to control β-catenin signaling in vivo. Injections into the dorsal blastomeres of 4-cell stage embryos led to a potent, dose-dependent ventralisation (Table I) . In contrast, when injection were performed in the ventral blastomeres we did not notice any effects, particularly in terms of dorsalization/axis duplication.

These data obtained for APC2 agree well with the proposed downregulatory activity of the 20 aminoacid/SAMP repeat region of APC on the halflife of β-catenin as described above. It would therefore appear that APC and APC2 exert very similar controlling effects on β-catenin signaling. However, Gu biner and colleagues have recently performed Xenopus embryo injections with RNAs encoding full length APC as well as fragments of APC that are comparable to the APC2 fragments utilised by us ¹⁰. In contrast to our findings, no effects were seen upon dorsal injections whereas ventral injections potently induced secondary axes. These observations do not simply fit into the current theoretical framework of APC ' s biochemical activities. One conclusion may be that APC and APC2, although structurally and biochemically similar, may exert very different or even opposite biological effects. It should be born in mind, however, that experimentation in the Xenopus system always relies on overexpression. Since APC and APC2 likely function in a complex with at least two other negative regulators of β-catenin (i.e. axin/conductin and GSK-3 ) their overexpression may have unpredictable effects on the functional integrity of this multi-protein complex, and thus on the halflife of β-catenin.

We determined APC2 to be a tumor suppressor protein. It is not directly obvious which tumour types could not be caused by loss of APC2. Possibly, the 5-10% of colorectal tumors that have intact APC and β-catenin genes ¹⁵ have undergone mutational changes in APC2. It has always been puzzling why the overwhelming majority of neoplasms in Familial Adenomatous Polyposis patients occur specifically in the intestine, while APC, β-catenin and Tcf's are much more broadly expressed. Possibly, APC and APC2 perform a redundant tumor suppressor function in many tissues. The identification of APC2 and dAPC2 will allow the creation of mouse and fly strains mutant in these genes. Such animal models will shed more light on the unique functions of these genes and on their interactions with APC and other components of the Wingless/Wnt pathway.

Methods

Reporter gene assays were performed as in (7). In brief, 2 x 10⁶ cells were transfected with plasmids by electroporation . After 24 hours, cells were harvested and lysed in 1 mM DTT, 1 % Triton X-100, 15 % glycerol, 25 mM Tris pH 7.8 and 8 mM MgCl₂. cDNAs encoding Myc-tagged versions of b-catenin and hTcf-4 were inserted into the mammalian expression vector pCDNA (Invitrogen) . Sequences of pTOPCAT, pFOPCAT, pTOPFLASH and pFOPFLASH are available upon request. pCATCONTROL, encoding the CAT enzyme under the control of the SV40 promoter, was purchased from Promega. FISH was performed on metaphases derived from phytohaemagglutinin-stimulated human blood cells using standard procedures . The human APC2 probe was nick-labeled with biotin and competed for repititive sequences with 1 mg COtl DNA. Hybridization was performed overnight at 37°C in the presence or absence of a digoxigenin-labeled chromosome 19 painting probe. Biotin was detected with avidin-Cy3 and amplified with a biotinylated anti-avidin antibody and a second layer of avidin-Cy3. The painting probe was detected with a FITC conjugated anti- digoxygenin antibody. Chromosomes were counterstained with DAPI . Gel retardation assays were performed as described elsewhere (7). Extracts were prepared from intact nuclei that were washed four times to avoid contamination with cytoplasmic b-catenin. As the optimal Tcf/Lef probe, we used a double-stranded 15-mer CCCTTTGATCTTACC; the control probe was CCCTTTGGCCTTACC. All oligonucleotides were from Isogen (Maarssen, Holland) . The b-catenin antibody was purchased from Transduction Laboratories (Lexington, KY) . A typical binding reaction contained 3 mg nuclear protein, 0.1 ng radiolabeled probe, 100 ng of dldC, in 25 ml of binding buffer (60 mM KC1, 1 mM EDTA, 1 mM DTT, 10% glycerol). Samples were incubated for 20 min at room temperature, antibody was added, and the samples incubated 20 min further.

Table 1. Dorsal injection of APC-2 suppresses endogenous axis formation

mRNA injected n DAI Percent

1 ng APC-2 32 5 63 4 31 3 3 2 3 1 0

3 ng APC-2 32 56

22

3 16 2 6 1 0

9 ng APC-2 39 5 15 4 26 3 23 2 26

1 10

Xenopus laevis embryos were injected with APC-2 mRNA at the four-cell stage, in each dorsal blastomere with the amount indicated. The embryos were scored at stages 31, 32 according to the standard dorso-anterior index (DAI⁹) . A normal embryo is assigned DAI 5, whereas embryos lacking dorso-anterior structures are assigned DAI 0. LEGENDS

Fig. 1. Comparison of APC, APC2, dAPC and dAPC2 Domain structures conserved between these APC relatives are indicated in various boxes.

Fig. 2. Differential expression of APC and APC2 in fetal and adult human tissues. Dot blots of 50 normalized polyA mRNA samples isolated from a large variety of tissues and several controls were hybridized with APC2 and APC specific probes according to the manufacturer's instructions (Genome Systems). While all tissue RNA samples yielded positive signals with both probes when compared to control RNA samples supplied on the blot, high levels of expression for both genes were observed throughout the central nervous system. Relative differences between APC and APC2 expression in a selected set of tissues are visualized. To that end, signals on individual dots were quantified by phosphorimaging; the signal obtained on the lung mRNA samples was arbitrarily set at 100% for each of the two individual probes . The background obtained with human C₀tl DNA at 0%. The original dotblot data are available upon request

Fig. 3. Human APC2 maps to chromosome 19pl3.3 Metaphase spreads of human leukocytes were hybridized with a 6 kb APC2-specific cDNA probe. The APC2 signal appears in red. The inserted figure combines the APC2 signal with a chromosome 19 paint in green.

Fig. 4. Human and mouse APC2 nucleotide sequences and amino acid sequences. REFERENCES AND NOTES

1 K.M. Cadigan, R. Nusse. Genes Dev. 11: 3286 (1997); P. Polakis. Biochim. Biophys Acta 1332 F127 (1997) 2. K.W. Kinzler et al. Science 253, 661 (1991); G. Joslyn et al. Cell 66, 601 (1991) . 3. S. Hayashi et al. Proc. Natl. Acad. Sci. USA 94, 242 (1997); Y. Ahmed, S. Hayashi, A. Levine, E. Wieschaus. Cell 93, 1171 (1998)

4. M&M dotblot

5. M&M cloning 6. J. Groden et al. Cell, 66, 589 (1991); B. Rubinfeld et al. Science, 262, 1731 (1993); L.-K. Su, B. Vogelstein, K.W. Kinzler, ibid. 262, 1734 (1993); B. Rubinfeld et al. Science, 212, 1023 (1996) . 7. J. Behrens et al. Science 280, 596 (1998) 8. L.Zeng et al. Cell 90, 181 (1997); C. Sakanaka, J.B. Weiss, L.T. Williams. Proc. Natl. Acad. Sci USA 95 3020 (1998); S. Ikeda et al. EMBO J. 17, 1371 (1998); M.J. Hart et al. Current Biol.8, 573 (1998); H. Yamamoto et al. Mol Cell Biol.18, 2867 (1998) 9. S. Mune itsa et al., Proc. Natl. Acad. Sci. U.S.A. 92, 3046 (1995)

10 K. Vleminckx et al. J. Cell Biol. 136: 411 (1997) 11. M. Mbihay, N.G.Seidah, M. Chretien, E.M. Simpson. Genomics 26, 123 (1995) . 12. H. Aberle et al. EMBO J. 16: 3797 (1997).

13. M. Molenaar et al., Cell 86, 396 (1996); M. van de Wetering et al. Cell 88, 789 (1997)

14. J. Behrens et al., Nature 382, 638 (1996); 0. Huber et al., Mech. Dev. 59, 3 (1996); E. Brunner et al. Nature 385, 829 (1997); J. Riese et al. Cell 88, 777 (1997) 15. V. Korinek et al., Science 275, 1784 (1997); P. Morin et al., Science 275, 1787 (1997); B. Rubinfeld et al., Science 275, 1790 (1997)

16. T.-C. He et al. Science 281, 1509 (1998)

17 V. Korinek et al. Nature Genetics 19, 379 (1998)

18 V. Korinek et al. Mol Cell Biol. 18, 1248 (1998

Claims

1. An isolated and/or recombinant nucleic acid or a specific fragment, homologue or derivative thereof, corresponding to a gene located in humans on the p arm of chromosome 19 at around position 19pl3.3 and comprising a nucleic acid sequence or specific fragment thereof as shown in figure 4.

2. A nucleic acid according to claim 1 which is a cDNA molecule .

3. A nucleic acid according to claim 1 or 2 which is of human origin.

4. An expression vector comprising a nucleic acid according to anyone of claims 1 to 3.

5. A cell comprising a nucleic acid according to anyone of claims 1 to 3 or a vector according to claim 4.

6. A cell according to claim 5, additionally provided with a nucleic acid encoding a repressible gene, preferably a gene found in the Wnt pathway or derivative or specific fragment of said gene and/or provided with a nucleic acid encoding a reporter gene, preferably a Tcf-responsive reporter gene.

7. An animal comprising a cell according to claim 5 or 6.

8. A protein or derivative or fragment thereof encoded by a nucleic acid according to anyone of claims 1 to 3 or produced by a cell according to claim 5 or 6.

9. An antibody specifically directed against a protein or derivative or fragment thereof according to claim 8.

10. A method for identifying a candidate drug comprising use of a cell according to any of claims 5 or 6 or an animal according to claim 7 or a protein according to claim 8 or an antibody according to claim 9.

11. A method according to claim 10 wherein said drug is for use in a cancer patient.

12. A method according to claim 11 wherein said cancer is related to foetal cancer, cancers with a stem cell phenotype, brain, lung, kidney or intestinal cancer.

13. A method for diagnosing cancer comprising use of a nucleic acid according to anyone of claims 1 to 3 , a cell according to anyone of claims 5 or 6 or an animal according to claim 7 or a protein according to claim 8 or an antibody according to claim 9.