Complete Deletion of Apc Results in Severe Polyposis in Mice

Home , P21

Oncogene (2010) 29, 1857–1864 & 2010 Macmillan Publishers Limited All rights reserved 0950-9232/10 $32.00 www.nature.com/onc SHORT COMMUNICATION Complete deletion of Apc results in severe polyposis in mice

AF Cheung1, AM Carter1, KK Kostova1, JF Woodruff1, D Crowley1,2, RT Bronson3, KM Haigis4 and T Jacks1,2

1Koch Institute and Department of Biology, MIT, Cambridge, MA, USA; 2Howard Hughes Medical Institute, MIT, Cambridge, MA, USA; 3Department of Pathology, Tufts University School of Medicine and Veterinary Medicine, Boston, MA, USA and 4Masschusetts General Hospital Cancer Center, Harvard Medical School Department of Pathology, Charlestown, MA, USA

The adenomatous polyposis coli (APC) gene product is region of APC termed the mutation cluster region and mutated in the vast majority of human colorectal cancers. result in retained expression of an N-terminal fragment APC negatively regulates the WNT pathway by aiding in of the APC protein (Kinzler and Vogelstein, 1996). the degradation of b-catenin, which is the transcription Genotype–phenotype correlations involving germline factor activated downstream of WNT signaling. APC APC mutations suggest that different lengths and levels mutations result in b-catenin stabilization and constitutive of APC expression can influence the number of polyps WNT pathway activation, leading to aberrant cellular in the gut, the distribution of polyps and extra-colonic proliferation. APC mutations associated with colorectal manifestations of the disease (Soravia et al., 1998; cancer commonly fall in a region of the gene termed the Nieuwenhuis and Vasen, 2007). Specifically, patients mutation cluster region and result in expression of an that present clinically with attenuated FAP have N-terminal fragment of the APC protein. Biochemical and reduced polyp counts and a later age of colon cancer molecular studies have revealed localization of APC/Apc to development. Interestingly, germline APC mutations different sub-cellular compartments and various proteins associated with attenuated FAP fall 50 and 30 of the outside of the WNT pathway that associate with truncated mutation cluster region, presumably resulting in APC APC/Apc. These observations and genotype–phenotype truncations bearing different protein domains than in correlations have led to the suggestion that truncated classical FAP and/or hypomorphic APC expression. APC bears neomorphic and/or dominant-negative function These observations suggest a gain-of-function or domi- that support tumor development. To analyze this possibility, nant-negative role for APC truncated at the mutation we have generated a novel allele of Apc in the mouse that cluster region. Paradoxically, patients with whole-gene yields complete loss of Apc protein. Our studies reveal that APC deletions, although rare, have been found to present whole-gene deletion of Apc results in more rapid tumor with classical FAP (Herrera et al., 1986; Sieber et al., development than the APC multiple intestinal neoplasia 2002). More thorough analysis of the phenotypic con- (ApcMin) truncation. Furthermore, we found that adenomas sequences of different heritable APC alleles is hampered by bearing truncated Apc had increased b-catenin activity when the low frequency of FAP and attenuated FAP as well as compared with tumors lacking Apc protein, which could the existence of numerous genetic and environmental lead to context-dependent inhibition of tumorigenesis. modifiers of the disease (de la Chapelle, 2004). Thus, we Oncogene (2010) 29, 1857–1864; doi:10.1038/onc.2009.457; and others have turned to mouse models that faithfully published online 14 December 2009 recapitulate human disease as a tool for examining the consequences of various Apc mutations. Keywords: Apc truncation; Apc null; colon cancer; The APC multiple intestinal neoplasia (ApcMin) was mouse model the first heritable mutant allele of Apc developed in the mouse (Su et al., 1992). ApcMin mice develop numerous intestinal lesions resembling human FAP. The ApcMin Colorectal cancer is one of the leading causes of cancer allele, which was discovered in an ethylnitrosourea death in the Western world. More than 80% of mutagenesis screen, comprises a nonsense mutation in colorectal cancers have mutations in the adenomatous codon 850 of Apc and gives rise to a stable, truncated polyposis coli (APC) gene, which is known to occur early Apc protein similar to those found in human colon in tumor development. Most of these APC gene cancers. Since the characterization of the ApcMin mouse, alterations, which are found both in the germline in additional alleles of Apc have been generated, involving individuals with hereditary familial adenomatous poly- truncations both 50 and 30 to the Min mutation (Oshima posis (FAP) and in sporadic colon cancers, fall in a et al., 1995; Smits et al., 1999; Colnot et al., 2004; Gaspar et al., 2009; Pollard et al., 2009). As observed Correspondence: Dr T Jacks, Koch Institute and Department of in human patients bearing mutations in APC, these Biology, MIT, 77 Massachusetts Avenue, E17-517, Cambridge, alleles have yielded different phenotypic consequences MA, USA. relating to intestinal tumorigenesis (Fodde and Smits, E-mail: [email protected] 2001). The basis for these genotype–phenotype correla- Received 21 August 2009; revised 9 November 2009; accepted 12 November 2009; published online 14 December 2009 tions remains largely unknown. Null Apc yields more severe polyposis than ApcMin AF Cheung et al 1858 The APC/Apc gene product is a critical negative ing novel, conditional and constitutive null alleles of Apc regulator of the canonical Wnt signaling pathway. Apc in mice. We performed sequential gene targetings in forms a complex with Axin and glycogen synthase C57BL/6 embryonic stem cells, inserting LoxP sites kinase-3b that phosphorylates b-catenin on N-terminal 90 kb apart in the Apc locus flanking all 15 protein- residues, allowing for subsequent ubiquitination and coding exons of Apc to generate a Cre-regulatable null proteosomal degradation. When the ability of Apc to allele, Apcfle1–15 (Figure 1a). We then tested our allele by interact with this destruction complex is compromised, crossing Apcfle1–15 mice to Villin-Cre transgenic mice, b-catenin accumulates in the cytoplasm and translocates which express Cre in a mosaic manner throughout the into the nucleus in which it binds to Tcf/Lef transcrip- intestinal epithelium (el Marjou et al., 2004). Apc þ /fle1–15; tion factors to regulate target genes, such as c-myc and Villin-Cre mice developed several lesions in their small Axin2 (Clevers, 2006). Moreover, mutations in b-catenin intestines and colons by 4–5 months of age, whereas that lead to protein stabilization are observed in human Apc þ /fle1–15 mice without Cre never did, even when aged colon cancer and comparable mutations can initiate for >1 year (Figures 1b, d and f; data not shown). An intestinal tumorigenesis in mice (Morin et al., 1997; existing conditional Apc allele developed by Colnot et al. Harada et al., 1999). (2004), Apcfle14, encodes a truncated Apc protein after Beyond b-catenin deregulation, mutant APC has been deletion of exon 14 by exposure to Cre recombinase suggested to contribute to colon tumorigenesis through activity. By histological examination of tumor morphol- other mechanisms. First, the vast majority of sporadic ogy, grade and stromal recruitment, we found that colon cancers in humans have mutations in APC, adenomas from Apc þ /fle1–15;Villin-Cre mice closely re- whereas other cancers with deregulated WNT signaling sembled those found in Apc þ /fle14;Villin-Cre littermates have an equal or greater proportion of mutations in (Figures 1b–g). These data show that conditional b-Catenin or other WNT pathway components (Giles, deletion of Apc grossly phenocopies conditional trunca- 2003). This observation points to a special role for APC tion of Apc. mutation specifically in colon cancer. Second, Apc is a Next, we crossed Apc þ /fle1–15 mice to Meox-Cre relatively large protein of approximately 312 kDa and transgenic animals, which express Cre in the embryo, has a number of protein-binding motifs. In fact, Apc has for generating a heritable null allele of Apc, ApcDe1–15 been shown to interact with a variety of proteins outside (Tallquist and Soriano, 2000; Figure 1a). To confirm the of the Wnt pathway and has been found in various sub- loss of Apc expression upon genomic deletion of the cellular locations (Fodde et al., 2001; Aoki and Taketo, locus, we performed quantitative PCR to detect Apc 2007; Mccartney and Nathke, 2008). Furthermore, the message with Taqman probes spanning exons 9 and 10 majority of APC mutations result in a stable, truncated and 13 and 14. Analysis of complementary DNA polypeptide, lacking most b-catenin and Axin-binding generated from whole colons of 1-month-old animals domains that lie in the middle portion of the protein. In revealed the expected 50% reduction of Apc mRNA these mutant proteins, the microtubule-binding domain in Apc þ /De1–15 mice when compared with a wild-type and EB1-binding site at the C-terminus of APC are also littermate (Figure 1h). Apc þ /De1–15 mice were aged to lost, but the N-terminal armadillo repeats and oligo- allow for spontaneous tumor formation. Histological merization domains are retained (Nathke, 2004). comparison revealed no significant phenotypic differ- We took a genetic approach to dissect the role of ence between tumors derived from Apc þ /De1–15 and truncated Apc in intestinal tumorigenesis by construct- Apc þ /Min mice on a C57BL/6 background (data not

Figure 1 Generation of Apcfle1–15 and ApcDe1–15 alleles. (a) Schematic of the wild-type Apc genomic locus showing non-coding exon 0 and coding exons 1, 14 and 15. A floxed Puromycin resistance (PuroR) gene was targeted to intron 1 of Apc in C57BL/6 embryonic stem (ES) cells. ES cells were selected on Puromycin and screened by Southern blot for homologous recombination. Targeted clones were then transfected with a Cre expression plasmid to excise the PuroR cassette leaving behind a single loxP site. After confirming successful removal of PuroR, the intergenic region 30 to Apc was targeted with a construct containing a Neomycin resistance gene flanked proximally by frt sites on both sides and a distal loxP site on one side (loxP-frt-NeoR-frt cassette). ES cells were selected on Neomycin and again screened by Southern blot for homologous recombination. Correctly targeted ES cells were injected into FVB blastocystes to generate chimeric mice. Chimeras were mated to wild-type C57BL/6 mice and pups were screened by coat color and PCR for germline transmission of the targeted allele. The NeoR cassette was removed in vivo by germline FlpE expression leaving a loxP and a frt site 30 to Apc. This strategy yielded the conditional null allele Apcfle1–15. Germline expression of Cre by C57BL/6 Meox- Cre mice resulted in excision of the genomic region flanked by targeted loxP sites yielding a constitutive null allele ApcDe1–15. ApcMin mice, to which ApcDe1–15 mice were compared, were purchased from Jackson Laboratories (Bar Harbor, ME, USA). Both alleles were subsequently maintained by breeding to C57BL/6 mice originally from Jackson Laboratories. (b–g) Hematoxylin and eosin (H&E)-stained sections of small intestinal (b, c) and colonic (d–g) adenomas from Apc þ /fle1–15;Villin-Cre (b, d, f) and Apc þ /fle14;Villin-Cre (c, e, g) mice. Villin-Cre transgenic mice were a gift from Dr S Robine. Scale bar, 100 mm. (h) Quantitative real-time PCR of complementary DNA (cDNA) from whole colons of 1-month-old wild-type (black bar) and Apc þ /De1–15 (hatched bar) and colon tumors from aged Apc þ /De1–15 (open bar) and Apc þ /Min (gray bar) mice. Taqman probes (ABI) spanning exons 9 and 10 (top) and exons 14 and 15 (bottom) of Apc were used. Expression was normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) for each sample. Bars, mean þ s.e.m. Animal studies were approved by the committee for animal care of the Massachusetts Institute of Technology and were conducted in compliance with Animal Welfare Act Regulations and other federal statutes relating to animals and experiments involving animals and adheres to the principles set forth in the 1996 National Research Council Guide for Care and Use of Laboratory Animals (institutional animal welfare assurance number, A-3125-01).

Oncogene Null Apc yields more severe polyposis than ApcMin AF Cheung et al 1859 shown). Individual colonic tumors isolated from aged As Apc þ /De1–15 and Apc þ /Min mice were isogenic on a Apc þ /De1–15 mice showed nearly complete loss of Apc C57BL/6 background, we sought to compare their expression (Figure 1h). This expression pattern could phenotypes more closely for determining the influence reflect Apc promoter hypermethylation or loss of of truncated Apc on tumor development. We first heterozygosity, yielding tumors with ApcDe1–15/De1–15 geno- examined the survival of Apc-mutant cohorts by aging type, as has been described for Apc þ /Min mice on C57BL/ animals until they became moribund. Strikingly, 6 background (Haigis et al., 2004). Because the residual Apc þ /Min mice had a 30% increase in median survival Apc expression detected is likely derived from stromal compared with Apc þ /De1–15 mice within both female and cells within the lesions, our expression analyses support male subgroups (Figures 2a and b). Morbidity of aged the successful generation of a null allele in mice. These animals was supported by reduced hematocrits, which is data indicate that truncated Apc is not required for indicative of anemia that occurs commonly in mice with intestinal tumor formation and that the presence of extensive intestinal tumor burdens (Figure 2c). truncated Apc does not significantly affect the tumor We went on to analyze the cause of reduced survival phenotype in the gut. in Apc þ /De1–15 mice. Aside from a single mouse that

Oncogene Null Apc yields more severe polyposis than ApcMin AF Cheung et al 1860

Figure 2 Apc þ /De1–15 mice have reduced survival compared with Apc þ /Min mice. Cohorts of mice on C57BL/6 background were progeny of Apc-mutant fathers and wild-type Apc mothers. Females in survival study were kept virgin as they aged and animals were checked for moribundity on a daily basis. To reduce the inﬂuence of environmental modiﬁers and spontaneous mutation in our colony, cohorts were progeny of multiple parent pairs and data collection for disparate groups was performed concurrently. (a) Survival curves for Apc þ /De1–15 (solid line) and Apc þ /Min (dashed line) female mice. Median survival values were 119 (n ¼ 11) and 154 (n ¼ 10) days, respectively. (b) Survival curves for male mice of the same genotypes in (a). Median survival values were 106.5 (n ¼ 10) and 140.5 (n ¼ 10) days, respectively. The P-values by chi-square statistic are indicated for each comparison. (c) The percentage of hematocrits for wild-type and Apc mutants when moribund. Bars, mean.

developed a mammary tumor, no additional extra- by cleaved-caspase 3 immunohistochemistry of adeno- intestinal lesions was apparent (data not shown). Thus, mas revealed no apparent difference between mutant the simplest explanation to account for altered survival Apc genotypes (data not shown). On the other hand, between mutant Apc genotypes would be a difference upon staining for phospho-histone H3, a marker of in tumor development. We killed additional cohorts mitotic cells, we detected consistently more positive cells of animals at 3 months of age to examine tumor in small intestinal adenomas of Apc þ /De1–15 mice when distribution, multiplicity and size. The distribution of compared with those in Apc þ /Min mice (Figure 3i). Thus, lesions through the intestinal tract did not differ our data suggest that decreased survival of Apc þ /De1–15 between Apc þ /De1–15 and Apc þ /Min animals (Figures 3a mice compared with Apc þ /Min mice can be attributed to and b). However, Apc þ /De1–15 mice had on average more rapid tumor development in the absence of more tumors when compared with Apc þ /Min mice truncated Apc. (Figures 3c–f). This result was more pronounced in Our data indicate that truncated Apc may be less females in which the small intestinal tumor count was efﬁcient at promoting the early stages of intestinal increased by 460% and the colonic tumor count by tumor development than the complete loss of Apc 4115%. In males, although the trend was apparent, the protein. This conclusion falls at odds with the com- differences were not statistically signiﬁcant because of monly held view that truncated APC favors tumor the variability in tumor number. development, which is suggested by the high prevalence Colonic tumors were dissected from these mice and of APC truncations in human colorectal cancer and by measured with calipers for determining tumor volumes. studies analyzing truncated APC/Apc in model systems. In female Apc þ /De1–15 mice, tumors were more than twice However, the fact that APC truncation is frequently the volume of those in Apc þ /Min females (Figure 3g). observed may simply be due to a higher likelihood of There was no distinction, however, among colonic point/frameshift mutations occurring in APC when tumor volumes in males (Figure 3h). We reasoned that compared with focal deletions of the locus, particularly the difference in tumor size could be due to either as an early event in tumorigenesis. As mentioned above, increased cell death or reduced proliferation in the whole-gene deletions have recently been detected presence of truncated Apc. Measurement of apoptosis in patients showing classical FAP, indicating that

Oncogene Null Apc yields more severe polyposis than ApcMin AF Cheung et al 1861

Figure 3 Apc þ /De1–15 mice show faster tumor development compared with Apc þ /Min mice. C57BL/6 Apc-mutant females (a, c, e, g) and males (b, d, f, h) were killed at 3 months of age for analysis. To reduce the influence of environmental modifiers and spontaneous mutation in our colony, cohorts were progeny of multiple parent pairs and data collection for disparate groups was performed concurrently. Intestines were flushed with phosphate-buffered saline (PBS), cut open longitudinally and fixed flat on bibulous paper in four segments. Fixed intestine were stained with 1:10 dilution in water of 1% methylene blue in 95% ethanol and destained in 70% ethanol before counting tumors by dissection microscope. (a, b) Distribution of tumors along intestinal tracts of Apc þ /De1–15 (n ¼ 12 for each gender) and Apc þ /Min (n ¼ 12 for each gender) mice. (c–f) Total tumor counts in small intestines (c, d) and colons (e, f)of Apc-mutant mice in (a, b). Bars, mean þ s.e.m., *Po0.05 by rank sum test. (g, h) Volumes of colonic tumors from Apc þ /De1–15 (n ¼ 28 for females and n ¼ 32 for males) and Apc þ /Min (n ¼ 13 for females and n ¼ 24 for males) mice in (a, b). Colonic tumor volumes were determined by plucking lesions from the surface of intestine and measuring tumor masses by calipers. Volumes were calculated based on an ellipsoid shape: volume ¼ (4/3)*pi*x*y*z, with x, y and z being the widths of each perpendicular axis. Bars, mean, **Po0.005 by rank sum test. (i) Box and whisker plots of phospho-histone H3 þ cells quantified per nm2 small intestine tumor area from Apc þ /De1–15 (open bar, n ¼ 27) and Apc þ /Min (filled bar, n ¼ 13) female mice. aÀb-catenin (cat no. 9582), a-cleaved caspase-3 (cat no. 9661) and a-phospho-histone H3 (cat no. 9701) antibodies were purchased from Cell Signaling Technology (Danvers, MA, USA). Immunohistochemistry was performed on paraffin sections as previously described (Haigis et al., 2008). truncated APC is not required for human disease (Sieber directed at addressing the role of truncated APC/Apc et al., 2002). This inference concurs with our findings as in tumorigenesis have mainly used immortalized cell Apc þ /De1–15 mice had accelerated disease, rather than an lines, which undoubtedly possess secondary changes. attenuated phenotype, compared with Apc þ /Min mice. In advanced lesions, with additional genetic/epigenetic Furthermore, in vivo model systems used to study Apc alterations, the tumorigenic effects of truncated Apc have involved non-mammalian organisms and have could be different. focused primarily on the role of truncated Apc in Recently, truncated APC/Apc was shown to posi- organismal development. On the other hand, studies tively regulate WNT/Wnt signaling by enhancing

Oncogene Null Apc yields more severe polyposis than ApcMin AF Cheung et al 1862

Figure 4 Colon tumors from Apc þ /Min mice seem to have higher levels of Wnt pathway activation compared with Apc þ /De1–15 mice. Quantitative real-time PCR for (a) Axin2, (b) b-catenin and (c) c-Jun in colon tumors from aged Apc þ /De1–15 and Apc þ /Min mice. Bars, mean, *Po0.05 using the Student’s t-test. Samples were prepared by ﬂushing fresh colons with cold phosphate-buffered saline (PBS) and plucking colonic tumors from intestinal surface. Tissues were then minced with a razor blade before Trizol extraction for RNA. RNA was quantiﬁed using spectrophotometer, and RNA quality was assessed using agarose gel electrophoresis and visualization of 28S and 18S ribosomal RNA bands. Non-degraded samples were subjected to complementary DNA (cDNA) synthesis using random primers and Superscript III reverse transcriptase. Samples were treated with RNase H to degrade template RNA after cDNA synthesis was complete. Taqman probes and reagents were purchased from Applied Biosystems (Foster City, CA, USA). Relative expression of transcripts was determined by normalization to glyceraldehyde-3-phosphate dehydrogenase (GAPDH).

b-catenin stability (Takacs et al., 2008). We performed researchers have found a divergence from Knudson’s western blot and immunohistochemistry on colon ‘two-hit’ hypothesis regarding inactivation of tumor tumors from Apc þ /Min and Apc þ /De1–15 mice, but failed suppressor genes. Specifically, the nature of ‘second-hit’ to observe a consistent difference in b-catenin protein mutations seemed to be dependent on the location of the levels (Supplementary Figure 1). However, increased ‘first-hit’ in APC, indicating an active mechanism b-catenin could be obscured by the abundance of the beyond simple loss of function (Lamlum et al., 1999; protein in epithelia and by the already high levels of Albuquerque et al., 2002). The result of these APC b-catenin as a consequence of Wnt pathway deregula- mutations was hypothesized to yield APC gene products tion. Therefore, we used quantitative PCR to determine with some ability to downregulate b-catenin activity. whether there was evidence for increased Wnt signaling Although supported by in vitro studies, this idea was in the presence of truncated Apc. Through analysis of only recently substantiated with the development of complementary DNA from Apc-mutant colon tumors, an Apc1322T allele in mice (Pollard et al., 2009). Similar we discovered that tumors from Apc þ /Min mice had more to our findings, C57BL/6 Apc þ /1322T mice developed than twice the expression of Axin2 than tumors from more severe polyposis and their tumors showed dimin- Apc þ /De1–15 mice (Figure 4a). Although many b-catenin ished nuclear b-catenin expression when compared with target genes have been difficult to verify because Apc þ /Min mice. To our knowledge, our report is the first multiple upstream pathways converge on them, Axin2 to connect the lack of Apc protein expression to has been deemed an accurate reporter of b-catenin submaximal b-catenin activity in the context of cancer transcriptional activity (Jho et al., 2002). In addition, we development within a whole organism. found increased b-catenin message, which is consistent One way to reconcile the observed reduction in tumor with a report that b-catenin mRNA is stabilized in a phenotype despite apparently enhanced b-catenin activ- feed-forward manner in cells with active Wnt signaling ity in Apc þ /Min mice is to postulate that excess b-catenin (Gherzi et al., 2007) (Figure 4b). Moreover, c-Jun, a activity might engage oncogene-induced tumor suppres- proto-oncogene transcriptionally induced by b-catenin, sive pathways (Courtois-Cox et al., 2008). Oncogenic increases nearly threefold in expression in tumors from b-catenin ectopically expressed in mouse embryonic Apc þ /Min mice when compared with those in Apc þ /De1–15 fibroblasts has been shown to induce expression of mice (Figure 4c). Other reported b-catenin target genes, p19Arf to cause p53 upregulation and cell-cycle arrest such as c-Myc and cyclin D1, were not changed between (Damalas et al., 2001). As tumors from Apc þ /Min mice tumors of the different genotypes (Supplementary show greater b-catenin activity when compared with Figure 2). These data indicate increased b-catenin those in Apc þ /De1–15 mice, the reduced tumor develop- activity in colon adenomas expressing truncated ApcMin. ment in Apc þ /Min mice could be explained by increased The association between reduced b-catenin activity activation of tumor suppressors due to enhanced and more severe intestinal polyposis that we have oncogenic activity. Western blot analyses of colon identified in this study has precedent in both human tumors from Apc þ /Min and Apc þ /Min mice, however, and mouse studies. Through analyses of APC mutations yielded no appreciable differences in the induction of in familial and sporadic colon cancer in humans, p19Arf, p53 or p21 (data not shown). Nonetheless, this

Oncogene Null Apc yields more severe polyposis than ApcMin AF Cheung et al 1863 notion of tumor suppressor activation in the context of oncogenic K-ras and/or p53 deficiency, may be further high b-catenin activity requires further examination, as revealing. Our novel conditional and constitutive null the lesions that we interrogated were large adenomas. alleles of Apc will be invaluable in these future studies. Our studies show that whole-gene deletion of Apc yields more rapid tumor development than Apc truncation in mice, a result that calls into question the pro- tumorigenic effects of truncated Apc. Mice with Conflict of interest haploinsufficiency for Apc show decreased survival, more severe intestinal polyposis and increased colonic The authors declare no conflict of interest. tumor progression when compared with mice bearing the ApcMin truncation. These observations support a gain-of-function role for truncated Apc, although not one that is required for tumorigenesis or is advanta- Acknowledgements geous for tumor development. Further genetic and We thank members of the Jacks labs, Keara Lane in molecular studies will be required to elucidate the particular, for experimental advice and assistance. This work mechanism of action. Moreover, as our studies have was supported by the Howard Hughes Medical Institute and focused on the role of truncated Apc in adenomas, partially by the Cancer Center Support (core) Grant P30- the comparisons of Apc þ /Min and Apc þ /De1–15 animals in CA14051 from the National Cancer Institute. TJ is a Howard the context of additional genetic alterations, such as Hughes Investigator and a Daniel K Ludwig Scholar.

References

Albuquerque C, Breukel C, van der Luijt R, Fidalgo P, Lage P, Haigis K, Kendall K, Wang Y, Cheung A, Haigis M, Glickman J et al. Slors FJM et al. (2002). The ‘just-right’ signaling model: APC (2008). Differential effects of oncogenic K-Ras and N-Ras on somatic mutations are selected based on a specific level of activation proliferation, differentiation and tumor progression in the colon. of the beta-catenin signaling cascade. Hum Mol Genet 11: Nat Genet 40: 600–608. 1549–1560. Harada N, Tamai Y, Ishikawa T, Sauer B, Takaku K, Oshima M et al. Aoki K, Taketo M. (2007). Adenomatous polyposis coli (APC): a (1999). Intestinal polyposis in mice with a dominant stable mutation multi-functional tumor suppressor gene. J Cell Sci 120: 3327–3335. of the beta-catenin gene. EMBOJ 18: 5931–5942. Clevers H. (2006). Wnt/beta-catenin signaling in development and Herrera L, Kakati S, Gibas L, Pietrzak E, Sandberg AA. (1986). disease. Cell 127: 469–480. Gardner syndrome in a man with an interstitial deletion of 5q. Am J Colnot S, Niwa-Kawakita M, Hamard G, Godard C, Le Plenier S, Med Genet 25: 473–476. Houbron C et al. (2004). Colorectal cancers in a new mouse model Jho EH, Zhang T, Domon C, Joo CK, Freund JN, Costantini F. of familial adenomatous polyposis: influence of genetic and (2002). Wnt/beta-catenin/Tcf signaling induces the transcription of environmental modifiers. Lab Invest 84: 1619–1630. Axin2, a negative regulator of the signaling pathway. Mol Cell Biol Courtois-Cox S, Jones SL, Cichowski K. (2008). Many roads lead to 22: 1172–1183. oncogene-induced senescence. Oncogene 27: 2801–2809. Kinzler KW, Vogelstein B. (1996). Lessons from hereditary colorectal Damalas A, Kahan S, Shtutman M, Ben-Ze’ev A, Oren M. (2001). cancer. Cell 87: 159–170. Deregulated beta-catenin induces a p53- and ARF-dependent Lamlum H, Ilyas M, Rowan A, Clark S, Johnson V, Bell J et al. growth arrest and cooperates with Ras in transformation. EMBOJ (1999). The type of somatic mutation at APC in familial 20: 4912–4922. adenomatous polyposis is determined by the site of the germline de la Chapelle A. (2004). Genetic predisposition to colorectal cancer. mutation: a new facet to Knudson’s ‘two-hit’ hypothesis. Nat Med Nat Rev Cancer 4: 769–780. 5: 1071–1075. el Marjou F, Janssen KP, Chang BH, Li M, Hindie V, Chan L et al. Mccartney B, Nathke I. (2008). Cell regulation by the Apc protein (2004). Tissue-specific and inducible Cre-mediated recombination in Apc as master regulator of epithelia. Curr Opin Cell Biol 20: the gut epithelium. Genesis 39: 186–193. 186–193. Fodde R, Smits R. (2001). Disease model: familial adenomatous Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, Vogelstein B polyposis. Trends Mol Med 7: 369–373. et al. (1997). Activation of beta-catenin-Tcf signaling in colon cancer Fodde R, Smits R, Clevers H. (2001). APC, signal transduction by mutations in beta-catenin or APC. Science 275: 1787–1790. and genetic instability in colorectal cancer. Nat Rev Cancer 1: Nathke I. (2004). APC at a glance. J Cell Sci. 117: 4873–4875. 55–67. Nieuwenhuis MH, Vasen HF. (2007). Correlations between mutation Gaspar C, Franken P, Molenaar L, Breukel C, Van Der Valk M, Smits site in APC and phenotype of familial adenomatous polyposis R et al. (2009). A targeted constitutive mutation in the APC tumor (FAP): a review of the literature. Crit Rev Oncol Hematol 61: suppressor gene underlies mammary but not intestinal tumori- 153–161. genesis. PLoS Genet 5: e1000547. Oshima M, Oshima H, Kitagawa K, Kobayashi M, Itakura C, Taketo Gherzi R, Trabucchi M, Ponassi M, Ruggiero T, Corte G, Moroni C M. (1995). Loss of Apc heterozygosity and abnormal tissue building et al. (2007). The RNA-binding protein KSRP promotes decay of in nascent intestinal polyps in mice carrying a truncated Apc gene. beta-catenin mRNA and is inactivated by PI3K-AKT signaling. Proc Natl Acad Sci USA 92: 4482–4486. PLoS Biol 5. Pollard P, Deheragoda M, Segditsas S, Lewis A, Rowan A, Howarth Giles R. (2003). Caught up in a Wnt storm: Wnt signaling in cancer. K et al. (2009). The Apc 1322T mouse develops severe polyposis Biochim Biophys Acta (BBA) Rev Cancer 1653: 1–24. associated with submaximal nuclear beta-catenin expression. Haigis K, Hoff PD, White A, Shoemaker AR, Halberg RB, Gastroenterology 136: 2204–2213 e1–13. Dove W. (2004). Tumor regionality in the mouse intestine reflects Sieber OM, Lamlum H, Crabtree MD, Rowan AJ, Barclay E, the mechanism of loss of Apc function. Proc Natl Acad Sci USA 101: Lipton L et al. (2002). Whole-gene APC deletions cause classical 9769–9773. familial adenomatous polyposis, but not attenuated polyposis

Oncogene Null Apc yields more severe polyposis than ApcMin AF Cheung et al 1864 or ‘multiple’ colorectal adenomas. Proc Natl Acad Sci USA 99: Su LK, Kinzler KW, Vogelstein B, Preisinger AC, Moser AR, Luongo 2954–2958. C et al. (1992). Multiple intestinal neoplasia caused by a mutation in Smits R, Kielman MF, Breukel C, Zurcher C, Neufeld K, Jagmohan- the murine homolog of the APC gene. Science 256: 668–670. Changur S et al. (1999). Apc1638T: a mouse model delineating Takacs C, Baird J, Hughes E, Kent S, Benchabane H, Paik R et al. critical domains of the adenomatous polyposis coli protein involved (2008). Dual positive and negative regulation of wingless signaling in tumorigenesis and development. Genes Dev 13: 1309–1321. by adenomatous polyposis coli. Science 319: 333–336. Soravia C, Berk T, Madlensky L, Mitri A, Cheng H, Gallinger S et al. Tallquist MD, Soriano P. (2000). Epiblast-restricted Cre expression in (1998). Genotype-phenotype correlations in attenuated adenoma- MORE mice: a tool to distinguish embryonic vs. extra-embryonic tous polyposis coli. Am J Hum Genet 62: 1290–1301. gene function. Genesis 26: 113–115.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene