Oncogene (2006) 25, 7512–7521 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc REVIEW Wnt signalling in the mouse intestine AR Clarke Cardiff School of Biosciences, Museum Avenue, Cardiff University, Cardiff, UK The ApcMin/ þ mouse has emerged as a powerful model of adenomatous polyposis (FAP), a human syndrome human intestinal tumour predisposition. As such, it has characterized by the development of multiple polyps in provided a platform for studying genetic and epigenetic the colon. In the mouse, genetic modelling of the modifiers of adenoma predisposition, and for assessing the involvement of the Wnt pathway was first made possible chemotherapeutic potential of a plethora of different by the creation of the ApcMin/ þ mouse, which emerged agents. The development of new conditional and hypo- out of an ethylnitrosourea mutagenesis programme and morphic Apc alleles, together with models carrying was subsequently characterized to carry a point muta- mutations in other Wnt pathway components, has greatly tion in the murine Apc gene (Su et al., 1992). The ApcMin/ þ extended the scope of experimentation. Together these mouse has proven to be a relatively good model of the approaches are being used to identify and validate key human FAP syndrome, and has been used extensively to critical targets of the Wnt pathway, such as Mash2, identify genetic and epigenetic modifiers of disease, as Tiam1 and the Eph/Ephrins. They have also established a well as to assess therapeutic regimens such as exposure fundamental role for Wnt in the development and to a range of different chemopreventatives. In recent maintenance of normal intestinal physiology, and in years this model has been joined by other Apc mutant particular control of the stem cell niche. These activities mice, both constitutive and conditional, and also by a are now being dissected at the level of individual Wnt cohort of models targeting different elements of the Wnt components, with some surprising dependencies revealed. pathway. These models have increasingly rationalized In terms of adenoma development, these models also our understanding of the role played by the Wnt support a ‘just right’ notion for tightly controlled b-catenin pathway in neoplasia, and are beginning to reveal the activity both in normal physiology and neoplastic develop- complex interplay that underlies normal intestinal ment. They also indicate a two-stage dependency for some physiology mechanisms as well as identifying putative Wnt pathway targets, with an initial requirement that novel disease markers and therapeutic targets. is subsequently overcome to permit progression. Finally, these models establish that the Wnt pathway does The ApcMin/ þ mouse not operate in isolation, and that both normal and diseased physiology develops in a dynamic interplay with Initially identified through the development of overt ApcMin/ þ other pathways such as the Notch, Hedgehog and BMP clinical symptoms of intestinal disease, the pathways. The comprehensive understanding arising from mouse carries a mutation at codon 850 and develops these studies should lead the identification of novel multiple small intestinal adenomas, in addition to a prognostic markers and therapeutic targets, and also open smaller number of colonic polyps. The prevalence of the possibility of tissue engineering in the intestine. small intestinal lesions represents one of the major Oncogene (2006) 25, 7512–7521. doi:10.1038/sj.onc.1210065 differences between the human syndrome and the mouse model, the explanation for which remains unclear. Keywords: Wnt;Apc;intestine The predisposition to intestinal polyposis also changes on different genetic backgrounds, a phenomenon which led to the identification of the first modifier of neoplasia, Mom-1 (Dietrich et al., 1993), subsequently identified as secretory phospholipase A2 (Pla2g2a, MacPhee et al., 1995). More recently other modifiers have been Introduction identified, including Mom-2 (Silverman et al., 2002) and a region on the distal arm of chromosome 18 (Haines Deregulation of the Wnt pathway is well recognized to et al., 2005). underpin the early stages of colorectal neoplasia in both As a relatively accurate genetic model of human the human and the mouse. This is perhaps most clearly disease, the ApcMin/ þ mouse has been used extensively demonstrated by the recognition that mutations in the to characterize the response to potential therapeutic adenomatous polyposis coli gene (APC) underlie familial strategies, and a comprehensive list of the effects of 269 chemopreventative agents has been assem- Correspondence: Professor AR Clarke, Cardiff School of Biosciences, bled at http://www.inra.fr/reseau-nacre/sci-memb/corpet/ Museum Avenue, Cardiff University, Cardiff, CF10 3US, UK. indexan.html (Corpet and Pierre, 2003). The pertinent E-mail: [email protected] question arising from these studies is the extent to which Wnt signalling in the mouse intestine AR Clarke 7513 these results relate to human disease. By virtue of the downregulation of b-catenin. The ‘just right’ model sheer number of studies performed in the ApcMin/ þ implies that different Apc mutations result in different mouse, it is becoming possible to address this question, levels of pathway activation and thereby differing although there remain immense technical difficulties in degrees of tumour susceptibility. This strongly argues direct mouse/man comparisons. It is therefore encoura- that not all Apc mutations are equivalent, a conclusion ging that a meta-analysis of chemopreventative agents which also follows from the multifunctional nature of Apc, used in mouse and human has shown relatively good and which is supported by the phenotypes of mice bearing correspondence for a number of agents, such as sulindac different mutations in Apc. Compared to the ApcMin/ þ (Corpet and Pierre, 2005). Further encouragement that mutation, mice bearing an earlier truncation (at codon the ApcMin/ þ mouse is truly relevant to human disease 716) develop higher numbers of polyps (Oshima et al., derives from the observation that expression of the 1995), whereas mice heterozygous for the later 1638N modifier Pla2g2a significantly correlates with survival chain terminating mutation have an attenuated phenotype for patients with gastric adenocarcinoma (Leung et al., relative to the ApcMin/ þ mouse (Smits et al., 1997). The 2002). Microarray studies also argue in favour of direct 1638N mice develop five to six adenomas of the small human relevance of the ApcMin/ þ mouse, with trans- intestine during the first 6 months of life, and have also criptional profiles derived from the mouse successfully been reported to develop aberrant crypt foci in the colon identifying novel gene targets in human colorectal that retain Apc protein, suggesting heterozygosity for this tumours (Reichling et al., 2005). mutation is sufficient to initiate lesion development Somewhat fewer studies have been reported assessing (Pretlow et al., 2003). Finally, mice have been constructed genetic modification of the ApcMin phenotype, however bearing hypomorphic alleles wherein expression of Apc is this still represents a substantial and growing database reduced to 20 and 10% of wild-type levels and polyp of genetic backgrounds known to accelerate or suppress formation is reduced compared to the 716 mutation (Li tumorigenesis. For example, deficiency of many DNA et al., 2005). These hypomorphic alleles are characterized repair proteins accelerates polyp formation including by levels of b-catenin expression that inversely correlate the mismatch repair proteins (e.g. Reitmair et al., 1996), with Apc levels, and have been used to argue for a the thymidine glycosylase Mbd4 (Millar et al., 2002), threshold level of Apc (15% of wild type) at which a single and the base excision repair gene Myh (Sieber et al., polyp will develop per mouse (Li et al., 2005). 2004). In terms of the Wnt pathway, both heterozygosity and complete deficiency of the proposed Wnt target gene cyclin D1 have been shown to suppress adenoma The primary consequences of inactivation of Apc: formation, although adenomas do still develop in the conditional models absence of cyclin D1 (Wilding et al., 2002;Hulit et al., The development of the different constitutive Apc 2004). Indeed, the status of cyclin D1 as a Wnt target is mutants has greatly advanced our understanding of somewhat debatable, as it appears not to be an the subtleties of altered Wnt signalling upon adenoma immediate target of the Wnt pathway in vivo, becoming formation. However, these models have shed rela- deregulated in a delayed manner in a subset of tively little light on the precise mechanisms through Wnt-activated cells (Sansom et al., 2005). For another which Wnt activation leads to adenoma initiation. This proposed target, peroxisome proliferator-activated difficulty has been at least partially circumvented by receptor (PPAR)-delta, the story is also rather unclear. the development of strategies to conditionally in- In this case, both activation and inactivation of this activate Apc within the adult intestinal epithelium. pathway have been shown to accelerate intestinal Two approaches have been used to achieve this, based tumour formation in the ApcMin/ þ mice, albeit to on the use of a loxP-flanked Apc allele in conjunction differing extents (Gupta et al., 2004, Harman et al., with either an inducible Cyp1A-Cre transgene (Sansom 2004, Reed et al., 2004). The interpretation here must be et al., 2003) or a tamoxifen regulable, intestinal-specific that activation of PPAR-delta accelerates neoplasia, Villin-CreER transgene (Andreu et al., 2005). The but that this is not a relevant mechanism to normal phenotypes reported by both systems are remarkably adenoma formation in the ApcMin/ þ mouse. These rather similar. Inactivation of Apc leads to the rapid reloca- complex results for both cyclin D1 and PPAR-delta lization of b-catenin to the nucleus and rapid changes in underline the necessity of for validating proposed Wnt both the appearance of enterocytes and in the gross targets and Wnt-dependent mechanisms in an appro- histology of the crypt. Cell movement within the crypt is priate in vivo setting.