Tumor progression in Apc(1638N) mice with Exo1 and Fen1 deficiencies.

Flap endonuclease 1 (Fen1) and exonuclease 1 (Exo1) have sequence homology and similar nuclease capabilities. Both function in multiple pathways of DNA metabolism, but appear to have distinct in vivo nucleic acid substrates, and therefore distinct metabolic roles. When combined with Apc(1638N), Fen1 promotes tumor progression. Because of functional similarity to Fen1, and because Exo1 is involved in DNA mismatch repair (MMR) by interaction with Msh2 and Mlh1, genes that cause hereditary nonpolyposis colorectal cancer (HNPCC), we investigated the possibility that Exo1 might also act as a modifier to Apc(1638N). We present evidence that mice with combined mutations in Apc(1638N) and Exo1 and Apc(1638N), Exo1 and Fen1 genes show moderate increased tumor incidence and multiplicity in comparison to Apc(1638N) siblings, implying a low penetrance role for Exo1 in early gastrointestinal (GI) tumorigenesis. Despite a decrease in median survival (10 months) in Apc(1638N) Exo1 mice, their tumors do not progress any more rapidly than those of Apc(1638N). Instead these animals die from infections that are the result of impaired immune response. Apc(1638N) Exo1 Fen1 mice survive longer (18 months), and therefore appear relatively immune competent. They die of invasive GI tumors that display microsatellite instability (MSI). Our results show that Exo1 has a modest tumor suppressor function.

The distinct spectra of tumor-associated Apc mutations in mismatch repair-deficient Apc1638N mice define the roles of MSH3 and MSH6 in DNA repair and intestinal tumorigenesis.

In mammalian cells, mismatch recognition has been attributed to two partially redundant heterodimeric protein complexes of MutS homologues, MSH2-MSH3 and MSH2-MSH6. We have conducted a comparative analysis of Msh3 and Msh6 deficiency in mouse intestinal tumorigenesis by generating Apc1638N mice deficient in Msh3, Msh6 or both. We have found that Apc1638N mice defective in Msh6 show reduced survival and a 6-7-fold increase in intestinal tumor multiplicity. In contrast, Msh3-deficient Apc1638N mice showed no difference in survival and intestinal tumor multiplicity as compared with Apc1638N mice. However, when Msh3 deficiency is combined with Msh6 deficiency (Msh3(-/-)Msh6(-/-)Apc1638N), the survival rate of the mice was further reduced compared to Msh6(-/-)Apc(1638N) mice because of a high multiplicity of intestinal tumors at a younger age. Almost 90% of the intestinal tumors from both Msh6(-/-)Apc1638N and Msh3(-/-)Msh6(-/-)Apc1638N mice contained truncation mutations in the wild-type Apc allele. Apc mutations in Msh6(-/-)Apc1638N mice consisted predominantly of base substitutions (93%) creating stop codons, consistent with a major role for Msh6 in the repair of base-base mismatches. However, in Msh3(-/-)Msh6(-/-)Apc1638N tumors, we observed a mixture of base substitutions (46%) and frameshifts (54%), indicating that in Msh6(-/-)Apc1638N mice frameshift mutations in the Apc gene were suppressed by Msh3. Interestingly, all except one of the Apc mutations detected in mismatch repair-deficient intestinal tumors were located upstream of the third 20-amino acid beta-catenin binding repeat and before all of the Ser-Ala-Met-Pro repeats, suggesting that there is selection for loss of multiple domains involved in beta-catenin regulation. Our analysis therefore has revealed distinct mutational spectra and clarified the roles of Msh3 and Msh6 in DNA repair and intestinal tumorigenesis.

Down-regulation of beta-catenin TCF signaling is linked to colonic epithelial cell differentiation.

The beta-catenin TCF pathway is implicated in the regulation of colonic epithelial cell proliferation, but its role in the regulation of cell differentiation is unknown. The colon carcinoma cell line, Caco-2, spontaneously undergoes G(0)/G(1) cell cycle arrest and differentiates along the absorptive cell lineage over 21 days in culture. In parallel, we show that beta-catenin-TCF activity and complex formation are significantly down-regulated. The down-regulation of beta-catenin-TCF signaling was independent of APC, which we characterized as having a nonsense mutation in codon 1367 in Caco-2 cells, but was associated with a decrease in TCF-4 protein levels. Total beta-catenin levels increased during Caco-2 cell differentiation, although this was attributable to an increase in the membrane, E-cadherin-associated, fraction of beta-catenin. Importantly, down-regulation of beta-catenin-TCF signaling in undifferentiated Caco-2 cells by three different mechanisms, ectopic expression of E-cadherin, wild-type APC, or dominant negative TCF-4, resulted in an increase in the promoter activities of two genes that are well-established markers of cell differentiation, alkaline phosphatase and intestinal fatty acid binding protein. These studies demonstrate, therefore, that in addition to its established role in the regulation of cell proliferation, down-regulation of the beta-catenin-TCF pathway is associated with the promotion of a more-differentiated phenotype in colonic epithelial cells.

Differences in susceptibility to colonic stem cell somatic mutation in three strains of mice.

Different species and different strains of animals commonly show very different sensitivities to carcinogenic regimes, which are often unexplained. A major possible contributory factor is variation in susceptibility to mutation, but this has not been directly demonstrated. This study therefore quantified the colonic stem cell mutation frequency in three strains of mice using two carcinogens. Stem cell mutations were identified using loss of function of glucose 6-phosphate dehydrogenase (G6PD) in individual crypts, a technique validated by several previous studies. The carcinogens dimethylhydrazine (DMH) and ethyl nitrosurea (ENU) were given to Balb/C, C57BL/6J, and C3H mice. In response to DMH, Balb/C mice were most susceptible, with approximately double the stem cell mutation frequency found in C3H and more than ten-fold that found in C57BL/6J (3.3+/-0.71 vs. 1.5+/-0.52 vs. 0.28+/-0.8x10(-4)). In response to ENU, Balb/C mice and C3H mice were equally susceptible, showing a stem cell mutation frequency approximately twice that of C57BL/6J (3.1+/-0.4 vs. 3.1+/-0.65 vs. 1.63+/-0.28x10(-4)). The observed differences among the strains with respect to somatic mutation following DMH treatment are likely to be due to the previously documented differences in metabolic conversion to the active metabolite. However, as ENU is a directly acting, rapidly inactivated mutagen, strain differences in response to ENU are unlikely to be due to strain-dependent metabolism of the mutagen and are likely to reflect differences in DNA repair efficiency, or possibly in stem cell kinetics among the strains studied. Susceptibility to the induction of colonic stem cell mutation is an important factor in susceptibility to carcinogens, whether due to differences in DNA repair or to other factors. Direct quantification of stem cell mutation frequency allows the separate identification of this component of the carcinogenic cascade and shows that it can make a major contribution to the differing susceptibility of different mouse strains.

Tumor-associated Apc mutations in Mlh1-/- Apc1638N mice reveal a mutational signature of Mlh1 deficiency.

Apc1638N mice, which are heterozygous for a germline mutation in Apc, typically develop three to five spontaneous intestinal tumors per animal. In most cases this is associated with allelic loss of wildtype Apc. We have previously reported that the multiplicity of intestinal tumors is increased dramatically by crossing Apc1638N with an Mlh1-deficient mouse strain that represents an animal model of hereditary non-polyposis colorectal cancer (HNPCC). The increased tumor multiplicity in these mice was associated with somatic mutations in the Apc tumor suppressor gene. Here, we have examined the nature and distribution of 91 Apc mutations implicated in the development of intestinal tumors in Mlh1-/- Apc1638N animals. Protein truncation mutations were detected in a majority of tumor samples, indicating that the prevailing mechanism of Apc mutation in tumors is altered from allelic loss to intragenic mutation as a result of Mlh1 deficiency. The observed mutations were a mixture of base substitutions (27%) and frameshifts (73%). Most frameshifts were detected within dinucleotide repeats and there were prominent mutational hotspots within sequences of this sort at codons 927-929, 1209-1211 and 1461-1464. The observed Apc mutations caused protein truncation upstream of the third 20 amino acid beta-catenin binding domain and the first Axin-binding SAMP repeat, yielding Apc proteins that are predicted to be deficient in destabilizing beta-catenin. Our results reveal a characteristic mutational signature in Apc that is attributable to Mlh1 deficiency. This demonstrates a direct effect of Mlh1 deficiency in the mutation of Apc in these tumors, and provides data that clarify the role of Mlh1 in mammalian DNA mismatch repair.

Tumorigenesis in Mlh1 and Mlh1/Apc1638N mutant mice.

An3 1 KAL I MutL homologue 1 (MLH1) is a member of the family of proteins required for DNA mismatch repair. Germ-line mutations in MLH1 lead to the cancer susceptibility syndrome hereditary nonpolyposis colorectal cancer (HNPCC). We generated mice carrying a null mutation in the Mlh1 gene. We showed that mice heterozygous and homozygous for the Mlh1 gene are predisposed to developing tumors of the gastrointestinal (GI) tract, lymphomas, and a number of other tumor types. We also examined the role of adenomatous polyposis coli gene (Apc) gene mutations in the GI tumors of Mlh1 mutant mice by different methods and showed that the GI tumors in Mlh1 mice express little or no adenomatous polyposis coli protein. When an Apc gene mutation was bred into the Mlh1 mutant mice, the GI tumor incidence increased 40-100-fold. The wild-type Apc allele in these tumors was found to contain mutations. Together, these results show that we have developed two mouse models for human HNPCC and that the mechanisms of tumor development in the GI tract of these mice involve loss of Apc gene function in a manner very similar to that seen in the GI tumors of HNPCC.

Somatic mutation of the glucose-6-phosphate dehydrogenase (g6pd) gene in colonic stem cells and crypt restricted loss of G6PD activity.

The study of somatic mutation frequency, particularly stem cell somatic mutation, is important to the understanding of mechanisms of carcinogenesis. The models currently in use for studies in stem cell tissues such as the colon infer the presence of stem cell somatic mutation from alteration in enzyme function, when this has shown to be mutagen dose dependent, restricted to the unit of clonal architecture, and persistent. The present study identifies and characterises somatic mutations in the g6pd gene in individual mouse colonic crypts showing histochemically demonstrable loss of G6PD activity. Microdissection of single crypts, showing either normal or low G6PD activity by histochemistry was performed in mice treated with ethylnitrosourea (ENU), and the presence of point mutations sought by PCR and direct sequencing. Because of the limitation of the small amount of partially degraded (due to fixation) DNA available from each crypt, only about 20% of the coding region of the g6pd gene could be sequenced. Despite this, somatic mutations were identified in 3 of the 9 crypts analysed which showed loss of G6PD activity, but in none of the crypts with normal activity. Each of the mutations identified would be predicted to lead to a decrease in enzyme activity. We conclude that we have confirmed that the crypt restricted loss of G6PD activity is indeed due to stem cell somatic mutation in the g6pd gene, and suggest that the G6PD model can be used as a paradigm for other models where somatic mutation is inferred from a change in histochemically identifiable gene expression.