C57BL/6J mice carrying the Min allele of Adenomatous polyposis coli (Apc) develop numerous adenomas along the entire length of the intestine and consequently die at an early age. This short lifespan would prevent the accumulation of somatic genetic mutations or epigenetic alterations necessary for tumor progression. To overcome this limitation, we generated F1ApcMin/+ hybrids by crossing C57BR/cdcJ and SWR/J females to C57BL/6J ApcMin/+ males. These hybrids developed few intestinal tumors and often lived longer than 1 year. Many of the tumors (24-87%) were invasive adenocarcinomas, in which neoplastic tissue penetrated through the muscle wall into the mesentery. In a few cases (3%), lesions metastasized by extension to regional lymph nodes. The development of these familial cancers does not require chromosomal gains or losses, a high level of microsatellite instability, or the presence of Helicobacter. To test whether genetic instability might accelerate tumor progression, we generated ApcMin/+ mice homozygous for the hypomorphic allele of the Nijmegen breakage syndrome gene (Nbs1ΔB) and also treated ApcMin/+ mice with a strong somatic mutagen. These imposed genetic instabilities did not reduce the time required for cancers to form nor increase the percentage of cancers nor drive progression to the point of distant metastasis. In summary, we have found that the ApcMin/+ mouse model for familial intestinal cancer can develop frequent invasive cancers in the absence of overt genomic instability. Possible factors that promote invasion include age-dependent epigenetic changes, conservative somatic recombination, or direct effects of alleles in the F1 hybrid genetic background. [Cancer Res 2009;69(14):5768–75]

Human colorectal cancer is a leading cause of cancer death in the world. In the United States alone, 148,810 new cases and 49,960 deaths were expected in 2008 (1). Once a colorectal tumor has metastasized to other sites, the prognosis is quite poor, because no effective treatment of disseminated colon cancer has yet been identified.

Three models have been proposed to explain the acquisition of metastatic potential (reviewed in ref. 2). The predisposition model states that the genetic composition of the host determines whether a primary tumor metastasizes. By contrast, the initiation model states that neoplastic cells within the primary tumor have acquired mutations or other stable changes during the early stages of tumorigenesis that eventually lead to metastasis. Finally, the progression model states that a subpopulation of neoplastic cells within a primary tumor sequentially acquires genetic mutations, epigenetic alterations, or both that drive progression, culminating in metastasis. Note that these models are not mutually exclusive (e.g., a combination of constitutional allelic variants and acquired somatic mutations might be required for metastasis). Also, each scenario likely requires time to unfold.

Mouse models of hereditary human colorectal cancer have not yet fully addressed the issue of metastatic potential because the existing models do not consistently develop advanced cancers. C57BL/6J (B6) mice carrying one of several different adenomatous polyposis coli (Apc) mutations (ApcMin, Apc580S, and ApcΔ716) develop numerous adenomatous polyps (reviewed in ref. 3). Plausibly, these tumors do not progress because the mice have a shortened lifespan, owing to the heavy tumor load especially in the small intestine. This barrier can be partially overcome by introducing additional mutations. B6 ApcMin/+ mice lacking EphB2, EphB3, or p53 activity as well as cis-compound ApcΔ716Smad4 heterozygotes develop invasive adenocarcinomas (reviewed in ref. 4). The frequency of invasive tumors in these models ranged from 5% to 55% when reported. Other models with low tumor multiplicity, including Apc1638N/+, Muc2−/−, ApcloxP/+; CDX2P-NLS Cre, Smad3−/−, and Tgfβ1−/−; Rag2−/− mice, also develop advanced cancers (59). ApcloxP/+; CDX2P-NLS Cre mice carry a loxP-targeted Apc allele and the CDX2P-NLS Cre recombinase transgene; expression of Cre recombinase in the colon results in the loss of Apc activity and consequently the formation of tumors (5). Again, the frequency of invasive tumors varies widely, ranging from 3% to 50% when reported. In Apc1638N/+ and Smad3−/− mice, the cancers were reported to have metastasized, although the penetrance of this phenotype in both cases was quite low (8, 10). Here, we report a high frequency of invasive adenocarcinomas and local metastases in long-lived ApcMin/+ mice. In contrast to the multiply mutated mouse models, we have investigated the state of the genome in adenocarcinomas that arise in the familial Min setting. Finding no evidence for spontaneous genomic instability, we have gone on to explore whether imposed genetic instability or Helicobacter infection can accelerate the age-dependent tumor progression in this model.

Mice. All mice were investigated in the McArdle Laboratory for Cancer Research under protocols approved by the Institutional Animal Care and Use Committee at the University of Wisconsin following Association for Assessment and Accreditation of Laboratory Animal Care guidelines. Two types of F1ApcMin/+ hybrids were produced by breeding C57BR/cdcJ (BR) and SWR/J (SWR) females to B6 ApcMin/+ males. To test whether increased genetic instability affected tumor progression, females carrying the Nijmegen breakage syndrome gene (Nbs1ΔB) allele were bred to B6 ApcMin/+ males and the resulting progeny with a mixed genetic background of 129 and B6 alleles were intercrossed to generate ApcMin/+ mice with 0, 1, or 2 copies of the Nbs1ΔB allele. Mice were genotyped at Apc and Nbs1 using a PCR assay as described previously (11, 12).

Ethylnitrosourea treatment. Mice ages between 45 and 75 days were given either a single intraperitoneal injection of ethylnitrosourea or two injections separated by 14 days. Each injection of ethylnitrosourea was 50 μg/g body weight.

Helicobacter testing. Fecal pellets were collected and DNA was isolated with spin columns using the Blood and Body Fluid protocol provided by Qiagen. Samples were analyzed using the following PCR assay. The reaction included 1 μmol/L each of Helicobacter genus-specific primers 5′-TATGACGGGTATCCGGC-3′ and 5′-ATTCCACCTACCTCTCCCA-3′, designed from conserved regions of the 16S rRNA gene (13), 200 μmol/L of each dNTP (dATP, dCTP, dGTP, and dTTP), 1.5 mmol/L MgCl2, template DNA, 1.25 units Taq polymerase, and distilled water in a total volume of 25 μL. Reactions were heated to 94°C for 30 s once followed by 45 cycles of denaturation at 94°C for 2 s, primer annealing at 53°C for 2 s, and extension at 72°C for 30 s in a thermocycler (MJ Research PTC-100). Each PCR was analyzed by electrophoresis on an agarose gel containing ethidium bromide and photographed under UV light. The following controls were included with each PCR run: no sample, a sample from a Helicobacter-infected mouse, and a sample from a Helicobacter-free mouse. Positive samples contained a 375 bp band as confirmed with DNA molecular weight standards.

Tumor scoring. All mice were sacrificed when moribund. The number of desmoid fibromas on the abdominal wall was scored and the intestinal tract was removed, opened longitudinally, and laid out as described previously (14). Samples were fixed in 10% buffered formalin for 16 h and then placed in 70% ethanol for long-term storage. The number of intestinal tumors was scored with a dissecting microscope by a single observer (R.B.H.) blind to the genotype of the mice.

Histologic analysis. Intestinal tumors were embedded in paraffin, sectioned, and stained with either H&E or β-catenin as described previously (15). Sections were analyzed by three pathologists (H.C.P., M.K.W., and R.S.), each blind to the genotype of the mice.

Genetic integrity. To determine the Apc status of intestinal tumors, DNA was isolated from samples scraped from 20 to 40 paraffin sections using neighboring H&E-stained sections as a guide. The Apc status was ascertained using the Apc heterozygosity index as described previously (16).

To test for microsatellite instability (MSI), intestinal tumors and adjacent normal tissue were microdissected during necropsy. Each tumor was cut in half; one piece was snap frozen in liquid nitrogen, whereas the other piece was fixed overnight in 10% buffered formalin for histologic analysis. DNA was isolated from each frozen sample using the Magnesil Genomic Fixed Tissue System from Promega and analyzed for MSI as described previously (17).

Intestinal tumors, adjacent normal tissue, and kidneys were collected for comparative genomic hybridization. Again, each tumor was divided so that histologic analysis could be done. DNA from tissue that had been snap frozen in liquid nitrogen was analyzed using FVB DNA as the reference as described previously (18). Although tumors contain both neoplastic and normal cells, our analysis of the Apc status in tumors by quantitative PCR indicates that the normal cell admixture is <30%, so the comparative genomic hybridization platform should permit any significant change in copy number to be detected.

B6 ApcMin/+ mice fail to develop invasive intestinal cancers consistently, presumably owing to their short lifespan of ∼100 days. This limited time likely restricts the execution of each conceivable pathway to metastasis. To test this possibility, we generated Min mice with an extended lifespan. BR females were crossed with B6 ApcMin/+ males and the resulting (BR × B6)F1ApcMin/+ hybrids were held until moribund. The average lifespan of these mice was 232 ± 66 days, considerably longer than the average lifespan of B6 ApcMin/+ controls (Fig. 1; Table 1A). The increase in longevity was related to a dramatic reduction in tumor multiplicity: (BR × B6)F1ApcMin/+ hybrids developed, on average, 19 ± 10 intestinal tumors, whereas B6 ApcMin/+ controls developed, on average, 122 ± 45 (Table 1A). Suppression of tumorigenesis was evident along the entire length of the intestinal tract from the duodenum to the descending colon (Fig. 2). This observation indicates that the BR strain carries at least one dominant modifying allele that suppresses tumorigenesis. The majority of intestinal tumors (56%) from (BR × B6)F1ApcMin/+ hybrids were invasive adenocarcinomas (Fig. 3A; Table 1A). Longevity also affects extracolonic lesions: (BR × B6)F1ApcMin/+ hybrids developed significantly more desmoid fibromas than B6 ApcMin/+ controls (data not shown).

Figure 1.

F1ApcMin/+ hybrids live significantly longer than B6 ApcMin/+ controls. Two hundred (SWR × B6)F1ApcMin/+, 184 (BR × B6)F1ApcMin/+, and 97 B6 ApcMin/+ mice were allowed to age until moribund. The percent surviving versus age is plotted.

Figure 1.

F1ApcMin/+ hybrids live significantly longer than B6 ApcMin/+ controls. Two hundred (SWR × B6)F1ApcMin/+, 184 (BR × B6)F1ApcMin/+, and 97 B6 ApcMin/+ mice were allowed to age until moribund. The percent surviving versus age is plotted.

Close modal
Table 1.

Effect of background/age, Nbs1 genotype, ethylnitrosourea treatment, and Helicobacter status on tumor multiplicity and progression

Mouse typenLifespan (d), mean ± SDIntestinal tumor count, mean ± SDInvasive adenocarcinomas/total tumors (%)*
(A) Phenotype of F1 and controls     
    (BR × B6)F1 72 232 ± 66 19 ± 10 104/185 (56) 
    (SWR × B6)F1 79 370 ± 99 15 ± 10 39/45 (87) 
    B6 97 100 ± 18 122 ± 45 0/74 (0) 
(B) Effect of Nbs1     
    B6 Nbs1+/+ 17 144 ± 48 23 ± 23 6/53 (11) 
    B6 Nbs1+/ΔB 28 175 ± 79 24 ± 14 ND 
    B6 Nbs1ΔB/ ΔB 22 142 ± 42 31 ± 27 10/47 (21) 
    B6 Apc+/+Nbs1ΔB/ΔB 13 ND 0 ± 0 NA 
(C) Effect of ethylnitrosourea treatment     
    (BR × B6)F1, one ethylnitrosourea treatment 202 ± 47 44 ± 22 8/94 (9) 
    (BR × B6)F1, two ethylnitrosourea treatments 180 ± 8 58 ± 20  
    (SWR × B6)F1, one ethylnitrosourea treatment 289 ± 86 19 ± 14 13/35 (38) 
    (SWR × B6)F1, two ethylnitrosourea treatments 190 ± 21 40 ± 17  
(D) Effect of Helicobacter     
    (BR × B6)F1, Helicobacter negative 61 312 ± 82 10 ± 4 9/30 (30) 
    (BR × B6)F1, Helicobacter positive 20 319 ± 100 12 ± 7 13/41 (32) 
    (SWR × B6)F1, Helicobacter negative 58 396 ± 144 13 ± 10 12/38 (32)§ 
    (SWR × B6)F1, Helicobacter positive 12 498 ± 75 8 ± 4 7/29 (24)§ 
Mouse typenLifespan (d), mean ± SDIntestinal tumor count, mean ± SDInvasive adenocarcinomas/total tumors (%)*
(A) Phenotype of F1 and controls     
    (BR × B6)F1 72 232 ± 66 19 ± 10 104/185 (56) 
    (SWR × B6)F1 79 370 ± 99 15 ± 10 39/45 (87) 
    B6 97 100 ± 18 122 ± 45 0/74 (0) 
(B) Effect of Nbs1     
    B6 Nbs1+/+ 17 144 ± 48 23 ± 23 6/53 (11) 
    B6 Nbs1+/ΔB 28 175 ± 79 24 ± 14 ND 
    B6 Nbs1ΔB/ ΔB 22 142 ± 42 31 ± 27 10/47 (21) 
    B6 Apc+/+Nbs1ΔB/ΔB 13 ND 0 ± 0 NA 
(C) Effect of ethylnitrosourea treatment     
    (BR × B6)F1, one ethylnitrosourea treatment 202 ± 47 44 ± 22 8/94 (9) 
    (BR × B6)F1, two ethylnitrosourea treatments 180 ± 8 58 ± 20  
    (SWR × B6)F1, one ethylnitrosourea treatment 289 ± 86 19 ± 14 13/35 (38) 
    (SWR × B6)F1, two ethylnitrosourea treatments 190 ± 21 40 ± 17  
(D) Effect of Helicobacter     
    (BR × B6)F1, Helicobacter negative 61 312 ± 82 10 ± 4 9/30 (30) 
    (BR × B6)F1, Helicobacter positive 20 319 ± 100 12 ± 7 13/41 (32) 
    (SWR × B6)F1, Helicobacter negative 58 396 ± 144 13 ± 10 12/38 (32)§ 
    (SWR × B6)F1, Helicobacter positive 12 498 ± 75 8 ± 4 7/29 (24)§ 

NOTE: All mice are ApcMin/+ unless otherwise specified.

*

Intestinal tumors were isolated from a randomly selected subset of mice of each strain.

Wilcoxon rank-sum P values of the three pair-wise comparisons ≥ 0.3. ND, not done; NA, not applicable.

One or two ethylnitrosourea treatments.

§

One tumor in this set showed metastasis by direct extension to a regional lymph node.

Figure 2.

Tumors develop along the entire length of the intestinal tract in F1ApcMin/+ hybrids. Mice were allowed to age until moribund and then sacrificed. The small intestine was divided into four equal segments, with region 1 being closest to the stomach and region 4 being closest to the cecum. The number of tumors in these segments and the colon was scored and is plotted as the percentage of the total. Interestingly, the distribution of tumors skewed toward region 1 if F1ApcMin/+ hybrids are treated with ethylnitrosourea (ENU; P < 0.001 each by χ2 test for ethylnitrosourea-treated BR and SWR hybrids compared with their untreated counterparts).

Figure 2.

Tumors develop along the entire length of the intestinal tract in F1ApcMin/+ hybrids. Mice were allowed to age until moribund and then sacrificed. The small intestine was divided into four equal segments, with region 1 being closest to the stomach and region 4 being closest to the cecum. The number of tumors in these segments and the colon was scored and is plotted as the percentage of the total. Interestingly, the distribution of tumors skewed toward region 1 if F1ApcMin/+ hybrids are treated with ethylnitrosourea (ENU; P < 0.001 each by χ2 test for ethylnitrosourea-treated BR and SWR hybrids compared with their untreated counterparts).

Close modal
Figure 3.

Tumors in long-lived F1ApcMin/+ hybrids often become invasive adenocarcinomas and sometimes even metastasize by direct extension to regional lymph nodes. F1ApcMin/+ hybrids ranging in age from 224 to 387 days became moribund and were sacrificed. A, H&E-stained sections of four representative invasive adenocarcinomas from these mice. The top two photographs are of (SWR × B6)F1 and the bottom two are of (BR × B6)F1. Tumors often exhibit ulceration of the mucosal surface and extensive invasion of neoplastic cells through the muscle wall into the serosa and mesentery. Bar, 1 mm. B to D, in a 367-day-old (SWR × B6)F1ApcMin/+ mouse, a tumor appeared to have spread by direct extension into a regional lymph node. The tumor, mesentery, and enlarged lymph node were removed, fixed in 10% buffered formalin, embedded in paraffin, and sectioned. Sections were assessed by immunohistochemistry for overexpression of β-catenin (left) or stained with H&E [middle; each area outlined with a black rectangle is shown enlarged (right)]. In this tumor, neoplastic cells have invaded into the muscle wall of the gastrointestinal tract (B and C) and largely effaced and replaced normal tissue in the lymph node leaving only residual intact areas (D). The cytoplasm and nucleus of each tumor cell stained intensely for β-catenin (brown) unlike normal tissue. Bar, 1 mm for β-catenin photographs.

Figure 3.

Tumors in long-lived F1ApcMin/+ hybrids often become invasive adenocarcinomas and sometimes even metastasize by direct extension to regional lymph nodes. F1ApcMin/+ hybrids ranging in age from 224 to 387 days became moribund and were sacrificed. A, H&E-stained sections of four representative invasive adenocarcinomas from these mice. The top two photographs are of (SWR × B6)F1 and the bottom two are of (BR × B6)F1. Tumors often exhibit ulceration of the mucosal surface and extensive invasion of neoplastic cells through the muscle wall into the serosa and mesentery. Bar, 1 mm. B to D, in a 367-day-old (SWR × B6)F1ApcMin/+ mouse, a tumor appeared to have spread by direct extension into a regional lymph node. The tumor, mesentery, and enlarged lymph node were removed, fixed in 10% buffered formalin, embedded in paraffin, and sectioned. Sections were assessed by immunohistochemistry for overexpression of β-catenin (left) or stained with H&E [middle; each area outlined with a black rectangle is shown enlarged (right)]. In this tumor, neoplastic cells have invaded into the muscle wall of the gastrointestinal tract (B and C) and largely effaced and replaced normal tissue in the lymph node leaving only residual intact areas (D). The cytoplasm and nucleus of each tumor cell stained intensely for β-catenin (brown) unlike normal tissue. Bar, 1 mm for β-catenin photographs.

Close modal

The predisposition model states that metastatic potential is a heritable trait influenced by genetic polymorphisms of the host. To test whether the development of intestinal cancers was peculiar to (BR × B6)F1ApcMin/+ hybrids, we generated another group of F1 hybrids by crossing SWR females to B6 ApcMin/+ males. The resulting (SWR × B6)F1ApcMin/+ hybrids had an average lifespan of 370 ± 99 and developed, on average, 15 ± 10 tumors along the entire length of the intestinal tract (Fig. 1; Table 1A). This reduction in tumor multiplicity was expected because SWR carries the resistance allele of the Mom1 locus, which encodes the secretory phopholipase A2 (19). Here too, the majority of tumors in (SWR × B6)F1ApcMin/+ hybrids were invasive adenocarcinomas (Fig. 3A; Table 1A). Moreover, 2 tumors of 67 metastasized by direct extension to regional lymph nodes (Fig. 3B). This frequency is likely an underestimate because the metastatic lesions were identified owing to the greatly enlarged lymph nodes, which were harvested at necropsy. Thus, the development of invasive cancers in Min mice is not peculiar to a particular strain. Rather, it appears to depend on either the elapsed time of tumor development or the age of the host—two intertwined variables in these experiments.

We analyzed the status of the Apc wild-type allele in 10 advanced tumors from F1ApcMin/+ hybrids. The majority (8 of 10) exhibited loss, most likely by somatic recombination as in B6 ApcMin/+ mice. We did not analyze the status of Mom1. BR is closely related to B6 and carries the sensitive allele of Mom1. Therefore, the reduction in tumor multiplicity and increase in longevity exhibited by (BR × B6)F1ApcMin/+ hybrids cannot be explained by Mom1 status.

In the progression model, tumors grow and metastasize as a consequence of accumulated genetic mutation, epigenetic alterations, or both throughout the genome. Tumors from F1ApcMin/+ hybrids ranging in age from 232 to 609 days were tested for MSI using a multiplex assay described previously (17). One F1 adenocarcinoma was MSI-high exhibiting mutations in 2 of 8 markers, whereas 13 F1 adenocarcinomas were MSI-low exhibiting mutations only in 1 of 8 markers (Table 2). The majority of the mutations were found in mBAT-67, a very long mononucleotide repeat marker that is likely to be less stable than shorter mononucleotide repeat markers. Notably, the low level of MSI observed in invasive adenocarcinomas from long-lived F1ApcMin/+ animals does exceed that observed in short-lived B6 ApcMin/+ controls. Only 1 of 20 adenomas in B6 animals ranging in age from 90 to 100 days was MSI-low. Thus, a low level of MSI may participate in familial tumorigenesis, but it is not measurably enhanced during progression to the adenocarcinoma.

Table 2.

MSI analysis of invasive intestinal adenocarcinomas with mononucleotide repeats

GenotypeMouse IDAge (d)Sample pairChange in allele size (bp)
D1Mit79mBat-24mBat-26mBat-27*mBat-37mBat-59mBat-64mBat-67
B6 Mlh1−/− 4934 269 −2 −3 −2 −3, −7 −7, −11 −6, −12 
(BR × B6)F1ApcMin/+ 1851 362 
 2012 297 ND 
 2056 232 
(SWR × B6)F1ApcMin/+ 2058 530 
 2079 536 −24 
   ND 
 2217 520 −25, −34, −40 
   
 2232 517 10 ND 
   11 ND 
 1840 601 12 −22, −38 
 1888 591 13 −30 
 1902 595 14 −21 
 1904 595 15 ND 
 2000 571 16 −38 
 2001 571 17 
 2221 533 18 −25 −15, −26, −35 
   19 ND ND 
 2223 533 20 −27, −37 
   21 −12, −20 
 1986 579 22 −12, −24, −33 
 1885 604 23 −18 
 1898 609 24 −26 
 2041 578 25 
 2047 577 26 ND −21 
GenotypeMouse IDAge (d)Sample pairChange in allele size (bp)
D1Mit79mBat-24mBat-26mBat-27*mBat-37mBat-59mBat-64mBat-67
B6 Mlh1−/− 4934 269 −2 −3 −2 −3, −7 −7, −11 −6, −12 
(BR × B6)F1ApcMin/+ 1851 362 
 2012 297 ND 
 2056 232 
(SWR × B6)F1ApcMin/+ 2058 530 
 2079 536 −24 
   ND 
 2217 520 −25, −34, −40 
   
 2232 517 10 ND 
   11 ND 
 1840 601 12 −22, −38 
 1888 591 13 −30 
 1902 595 14 −21 
 1904 595 15 ND 
 2000 571 16 −38 
 2001 571 17 
 2221 533 18 −25 −15, −26, −35 
   19 ND ND 
 2223 533 20 −27, −37 
   21 −12, −20 
 1986 579 22 −12, −24, −33 
 1885 604 23 −18 
 1898 609 24 −26 
 2041 578 25 
 2047 577 26 ND −21 
*

mBat-27 (accession no. L24372) primer sequences were GGGAAGACTGCTTAGGGAAGA and ATTTGGCTTTCAAGCATCCATA (48). The other microsatellite markers were described previously (17).

Tumors from F1ApcMin/+ hybrids were also tested for chromosomal deletions and gains by comparative genomic hybridization. We prepared genomic DNA from invasive adenocarcinomas, normal intestinal epithelium, and kidney. A few clones on the arrays consistently exhibited a change in copy number among all samples including normal tissue, indicating that they were polymorphic between B6 and FVB (Fig. 4). However, no tumor-associated aberrations were detected in the DNA from these tumors. Thus, large chromosomal deletions or duplications are not necessary for the progression of invasive adenocarcinomas in this particular model of familial cancer.

Figure 4.

Invasive adenocarcinomas from long-lived F1ApcMin/+ hybrids do not exhibit chromosomal instability. Two invasive adenocarcinomas (A and B; bar, 2 mm), adjacent normal tissue, and kidneys were removed from a (BR × B6)F1ApcMin/+ male that lived for 312 days. DNA was prepared and analyzed by comparative genomic hybridization (C). The log2 ratio is plotted for each locus. None of the autosomal chromosomes exhibited any gains or losses when comparing results obtained with DNA from tumors (red square and red triangle) with those obtained with DNA from adjacent normal tissue (black diamond). By contrast, the X chromosome appears to be lost in all samples. This change merely reflects the fact that test DNA for this example was isolated from a F1ApcMin/+ male and the control DNA was isolated from a female. A total of seven invasive adenocarcinomas from three F1ApcMin/+ hybrids were analyzed. Similar results were observed in all cases.

Figure 4.

Invasive adenocarcinomas from long-lived F1ApcMin/+ hybrids do not exhibit chromosomal instability. Two invasive adenocarcinomas (A and B; bar, 2 mm), adjacent normal tissue, and kidneys were removed from a (BR × B6)F1ApcMin/+ male that lived for 312 days. DNA was prepared and analyzed by comparative genomic hybridization (C). The log2 ratio is plotted for each locus. None of the autosomal chromosomes exhibited any gains or losses when comparing results obtained with DNA from tumors (red square and red triangle) with those obtained with DNA from adjacent normal tissue (black diamond). By contrast, the X chromosome appears to be lost in all samples. This change merely reflects the fact that test DNA for this example was isolated from a F1ApcMin/+ male and the control DNA was isolated from a female. A total of seven invasive adenocarcinomas from three F1ApcMin/+ hybrids were analyzed. Similar results were observed in all cases.

Close modal

Although apparently not necessary, chromosomal instability might be sufficient to drive tumor progression in Min mice. Cells homozygous for a hypomorphic allele of Nbs1, Nbs1ΔB, exhibit defects in the intra-S and G2-M checkpoints and marked chromosomal instability (20). Nbs1ΔB/ΔBp53−/− mice develop lymphomas much earlier than Nbs1+/+p53−/− mice (12). We generated Min mice carrying 0, 1, and 2 copies of the mutant Nbs1ΔB allele and held them until moribund. The significant reduction in Nbs1 activity did not affect tumor progression: 21% of the tumors in ApcMin/+Nbs1ΔB/ΔB mice and 12% in ApcMin/+Nbs1+/+ mice were invasive adenocarcinomas (Table 1B). Further, no difference was observed in lifespan, tumor multiplicity (Table 1B), or tumor distribution (data not shown). Thus, a reduction in Nbs1 activity does not strongly affect intestinal tumorigenesis in the Min mouse. One explanation is that reduced Nbs1 activity might not cause genetic instability in the intestinal epithelium; this possibility has not been specifically tested.

Is genetic mutation sufficient to drive tumor progression in Min mice? We treated F1ApcMin/+ hybrids with one or two doses of a strong somatic point mutagen. Mice received 50 μg ethylnitrosourea/g body weight per intraperitoneal injection at ages between 45 and 75 days and were held until moribund. We choose a low dose of ethylnitrosourea in young adults rather than a high dose because we wanted to avoid a dramatic increase in tumor multiplicity and consequently reduced lifespan. Surprisingly, treatment of (BR × B6)F1ApcMin/+ hybrids appeared to impede tumor progression, with only 9% of the tumors being invasive adenocarcinomas (Table 1C). This effect might simply reflect the significant decrease in lifespan owing to the increase in adenoma multiplicity. Treated mice developed, on average, 50 ± 22 tumors, whereas contemporaneous controls developed 19 ± 10 tumors. Ethylnitrosourea had its strongest effect in the proximal half of the small intestine (Fig. 2). In addition, treatment of (SWR × B6)F1ApcMin/+ hybrids with ethylnitrosourea did not increase tumor progression. Thus, this situation, designed to increase the levels of somatic mutation during tumorigenesis, clearly did not enhance tumor progression in the Min mouse.

The risk of developing intestinal cancer is clearly influenced by environmental factors, including diet and bacterial flora. Several epidemiologic studies indicate that risk of colon cancer is higher for individuals infected with Helicobacter pylori. This link is also apparent in some animal models: the development of intestinal cancers in Smad3-deficient mice depends on the presence of Helicobacter. To test whether age-dependent tumor progression in Min mice was affected by Helicobacter status, we analyzed tumors from F1ApcMin/+ hybrids for which the status of Helicobacter was known. No effect on the frequency of invasive adenocarcinomas was observed between Helicobacter-negative mice and contemporaneous Helicobacter-positive controls (Table 1D).8

8

The frequency observed in this experiment was noticeably lower than that in the initial experiment (Table 1D versus Table 1A). During the 6 years that elapsed between these two experiments, our colony underwent significant environmental changes including a switch in diet and the implementation of a microisolator system.

We have identified a homogenous genetic and environmental context in which ApcMin/+ mice develop intestinal tumors that spontaneously progress and metastasize locally. A key barrier to the development of advanced cancers in the Min mouse has been the short lifespan on the sensitive B6 genetic background. F1ApcMin/+ hybrids were generated by crossing BR or SWR females to B6 ApcMin/+ males. These mice developed one-tenth the number of tumors observed in controls and often lived longer than 1 year. Many of the tumors in these hybrids were invasive adenocarcinomas and a few even metastasized by extension into regional lymph nodes (Fig. 3; Table 1). The development of these familial cancers did not require a high level of MSI, chromosomal gains or losses, or the presence of Helicobacter. Several classes of genomically conservative processes can be considered to drive progression in this setting. Prominent candidates can be suggested from the known epigenetic changes observed by the analysis of the tumor transcriptome. We have previously shown that advanced cancers from long-lived Min mice exhibit core molecular signatures of the epithelial-to-mesenchymal transition. In particular, vimentin expression is elevated not only in neoplastic cells of the primary tumor but also in those invading the submucosa and musculature (21). A full characterization of the epigenome during tumor progression in the Min model may identify other candidate genes whose deregulation mediates progression.

The genomically conservative pathway to invasive intestinal adenocarcinomas that we have observed in the familial Min model does not rule out a role for genomic instability in other contexts such as sporadic colon cancer. Indeed, mutations in several genes have been shown to enhance the tumor phenotype of Min and other mouse models. We have investigated whether a generalized increase in somatic mutation can increase the number of invasive adenocarcinomas and metastatic lesions as well as decrease the time required for these types of tumors to form in the Min model. We treated adult F1ApcMin/+ hybrids with a low dose of ethylnitrosourea and found that this treatment shortened the lifespan, increased tumor multiplicity, and reduced progression.9

9

The increase in tumor multiplicity was unexpected, because our previous study had found that treatment of young B6 ApcMin/+ adults with ethylnitrosourea failed to affect tumor multiplicity (22). Key differences between this study and that of Shoemaker and colleagues are genetic background (F1 versus B6) and the age at which the treated mice were sacrificed (≤100 versus 188 ± 32 days).

Thus, any molecular changes that occur spontaneously during tumor progression cannot be fully replaced by treating mice with a strong point mutagen. Perhaps, the salient changes are not point mutations. Other possibilities will be discussed below.

Our results do not yet support a particular model of metastatic potential. Consider the predisposition model. Both (BR × B6)F1 and (SWR × B6)F1ApcMin/+ hybrids develop a significant number of invasive adenocarcinomas. BR and SWR are quite distinct and were derived from separate stocks in the 1930s and 1940s (23). Nonetheless, these strains could share one or more allelic variants that affect metastatic potential. The propensity for metastasis must be mapped to particular host loci and tested by transplantation for autonomy versus nonautonomy to investigate this hypothesis. Lifsted and colleagues recently showed that the development of mammary cancers and extensive metastasis in mice carrying the (MMTV-PyMT)634Mul transgene was profoundly affected by genetic background (24). In addition, Park and colleagues showed that resistant and susceptible strains can be stratified based on the expression of 17 genes in normal mammary tissue (25). Together, these findings have lent support to the host predisposition model.

Constitutional polymorphic variants are known to affect other aspects of intestinal tumorigenesis in the ApcMin/+ mouse model. B6 ApcMin/+ mice develop, on average, 100 tumors, whereas AKR ApcMin/+ mice develop, on average, <1 tumor (16). This difference is partially due to Mom1, encoding Pla2g2a and at least one other modifier of intestinal tumorigenesis (19). The reduction in tumor multiplicity does not necessarily lead to increased longevity because AKR mice frequently develop lymphomas. Other polymorphic modifiers of the Min phenotype have been identified including Mom2, Mom3, Mom6, and Mom7 (4). Mom2 is a spontaneous mutation that maps to distal chromosome 18 and disrupts Atp5a1, encoding the α subunit of the ATP synthase (26). Mom7 maps to the pericentromeric region of chromosome 18 and may directly regulate the loss of heterozygosity of distal elements (27). The genome-wide identification of polymorphic modifier loci for the Min phenotype is just beginning.

Somatic mutations also can affect intestinal tumorigenesis. Armitage and Doll estimated that five to six changes are necessary for a cancer to form (28). Analyses of human tumors at different stages have revealed that the activation of KRAS promotes growth and the loss of p53 favors metastasis. These alterations appear not to occur spontaneously during the development of intestinal adenomas and invasive adenocarcinomas in studies of two mouse models (29, 30). Possibly, the context, for example, sporadic versus familial disease or even the specific mutation in Apc, dictates which pathway is favored. The possible role of somatic mutations in progression in mouse models is suggested only by studies of these models carrying germ-line mutations in loci other than Apc. Janssen and colleagues showed that constitutional expression of an activated form of KRAS does affect tumor multiplicity and progression in Apc1638N/+ mice (31). Nearly half of the tumors from KRASV12G; Apc1638N/+ mice progressed to invasive adenocarcinomas by age 105 days (16 of 34) unlike Apc1638N/+ controls (1 of 8). Similarly, we have shown that a lack of p53 activity leads to tumor progression in the intestines of young Min mice (14). B6 ApcMin/+p53−/− mice developed a small proportion of invasive adenocarcinomas by age 90 days (2 of 42) in contrast to B6 ApcMin/+p53+/+ controls (0 of 78). Tumor progression in mouse models is also affected by mutations in several other genes including Atp5a1 (Mom2), Ephrin A1, EphB2, EphB3, ERα, Fen1, Netrin-1, and Tsp1 (reviewed in ref. 4). The genome-wide identification of mutant alleles that modify the Min phenotype is also just beginning.

Two pathways lead to genetic instability in colorectal cancer (32): MSI and chromosomal instability. Alberici and colleagues reported that adenomas in Apc1638N/+ mice display numerous chromosomal aberrations (33). Further, several investigators have found that sporadic human colorectal tumors display numerous chromosomal aberrations. For example, Martin and colleagues reported that those chromosomal regions that had undergone amplification and deletion carried known cancer genes including EGFR and Myc and many of affected regions were also altered in lung cancers, glioblastomas, or multiple myelomas (34). However, it remains a challenge to determine whether the genetic and genomic instabilities observed in these advanced cancers are triggers rather than bystanders acquired during growth under the relaxed constraints of the autonomous neoplasm. Thus, a chromosomal region that is commonly altered may either carry a critical gene affecting progression or else simply be inherently unstable in its structure.

Selection for genetic instability may be much stronger in sporadic disease than familial disease. Some intestinal adenomas from ApcMin/+ mice and familial adenomatous polyposis patients have been reported to form with relatively stable karyotypes (15, 35). Conservative somatic recombination at the Apc/APC locus plays a role in the formation of these tumors. Two independent studies have shown that a reduction in activity of the Bloom's syndrome gene (Blm) enhances the Min phenotype (36, 37). In both cases, the analysis of markers on chromosome 18 revealed that the wild-type allele of Apc was lost by somatic recombination in several tumors. Haigis and Dove showed that the wild-type allele of Apc is almost always lost by somatic recombination to initiate tumorigenesis in B6 ApcMin/+ mice (38). They found the Robertsonian translocation Rb(7.18)9Lub reduces the multiplicity of intestinal adenomas in this mouse model by disrupting the pairing of the chromosome 18 homologues during mitosis. The extent of somatic recombination throughout the entire genome is unknown. Interestingly, McMurray and Gottschling showed that wild-type yeast cells switch to a hyperrecombinational state as they age (39). Loss of heterozygosity occurs by somatic recombination in young cells and by break-induced replication in old cells. These forms of genetic instability would not be detected by comparative genomic hybridization because copy number is unaltered. Thus, somatic recombination in tumors of F1ApcMin/+ hybrids could generate homozygosity for allelic variants that drive progression, culminating in metastasis. We did find loss of heterozygosity at the Apc locus in the majority of advanced tumors from F1ApcMin/+ hybrids.

By contrast, epigenetic silencing may play a role in the formation or progression of these genetically stable neoplasms. Silencing may be reflected by the CpG island methylator phenotype (reviewed in ref. 40). Intriguingly, DNA hypermethylation is an age-dependent process in the intestinal epithelium (41, 42).

In summary, we have tested invasive adenocarcinomas from long-lived F1ApcMin/+ hybrids for chromosomal instability and MSI. Overall, the tumors in this particular model of familial disease display a stable karyotype and only a low level of MSI. These invasive tumors can metastasize by extension into regional lymph nodes. Bodmer and Tomlinson have argued that genetic instability is not required for tumor development because normal mutation rates coupled with selection are sufficient to account for the molecular changes observed (43, 44). Thus, although genetic instability may be able to drive progression in some biological contexts, it is not the only route to the locally invasive adenocarcinoma. Rusan and Peifer have reviewed the possible positive and negative effects of genomic instability on cancer (45). Indeed, genetic instability can inhibit tumorigenesis in certain other biological contexts by triggering apoptosis (46). Rao and colleagues have reported that the lack of mitotic aberrations in Min mice reduces tumor multiplicity in the small intestine but increases it in the colon (47).

The genetically conservative pathway to invasion described in this report must now be analyzed more deeply by newly emerging methods. Does the development of invasive potential reflect the elapsed time of tumor development or the age of the host? Longitudinal studies of individual tumors arising at different times in an animal may enable one to disentangle these two factors. Does it reflect an accumulation of conservative somatic recombination events that render particular constitutional alleles homozygous in F1ApcMin/+ animals? Does it reflect an age-dependent epigenetic silencing of one or both alleles at particular anti-progression loci? The capacity of the Min mouse model to display a high frequency of locally metastatic behavior represents another step in the development of better animal models for human colon cancer.

No potential conflicts of interest were disclosed.

Note: Current address for J.H.J. Petrini: Program in Molecular Biology, Memorial Sloan-Kettering Cancer Center, New York, NY. Current address for R.B. Halberg: Department of Medicine, University of Wisconsin School of Medicine and Public Health, Madison, WI.

Grant support: NIH grants P30 CA014520 (University of Wisconsin Paul C. Carbone Comprehensive Cancer Center), R37 CA63677 and U01 CA84227 (W.F. Dove), and U01 CA84118 (D.G. Albertson) as well as National Aeronautics and Space Administration grant NAG 9-1525 (J.W. Bacher).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dr. Alexandra Shedlovsky for critical reading of the article and Jane Weeks, Harlene Edwards, and the other members of the McArdle Histotechnology Facility for expert support.

1
Cancer facts & figures 2008. Atlanta (GA): American Cancer Society; 2008.
2
Threadgill DW. Metastatic potential as a heritable trait.
Nat Genet
2005
;
37
:
1026
–7.
3
McCart AE, Vickaryous NK, Silver A. Apc mice: models, modifiers and mutants.
Pathol Res Pract
2008
;
204
:
479
–90.
4
Kwong LN, Dove WF. APC and its modifiers in colon cancer. In: Nathke I, McCartney B, editors. APC proteins. Austin: Landes Bioscience; 2008.
5
Hinoi T, Akyol A, Theisen BK, et al. Mouse model of colonic adenoma-carcinoma progression based on somatic Apc inactivation.
Cancer Res
2007
;
67
:
9721
–30.
6
Kucherlapati M, Nguyen A, Kuraguchi M, et al. Tumor progression in Apc(1638N) mice with Exo1 and Fen1 deficiencies.
Oncogene
2007
;
26
:
6297
–306.
7
Yang W, Velcich A, Lozonschi I, et al. Inactivation of p21WAF1/cip1 enhances intestinal tumor formation in Muc2−/− mice.
Am J Pathol
2005
;
166
:
1239
–46.
8
Zhu Y, Richardson JA, Parada LF, Graff JM. Smad3 mutant mice develop metastatic colorectal cancer.
Cell
1998
;
94
:
703
–14.
9
Engle SJ, Hoying JB, Boivin GP, et al. Transforming growth factor β1 suppresses nonmetastatic colon cancer at an early stage of tumorigenesis.
Cancer Res
1999
;
59
:
3379
–86.
10
Fodde R, Edelmann W, Yang K, et al. A targeted chain-termination mutation in the mouse Apc gene results in multiple intestinal tumors.
Proc Natl Acad Sci U S A
1994
;
91
:
8969
–73.
11
Dietrich WF, Lander ES, Smith JS, et al. Genetic identification of Mom-1, a major modifier locus affecting Min-induced intestinal neoplasia in the mouse.
Cell
1993
;
75
:
631
–9.
12
Williams BR, Mirzoeva OK, Morgan WF, et al. A murine model of Nijmegen breakage syndrome.
Curr Biol
2002
;
12
:
648
–53.
13
Riley LK, Franklin CL, Hook RR, Jr., Besch-Williford C. Identification of murine Helicobacter by PCR and restriction enzyme analyses.
J Clin Microbiol
1996
;
34
:
942
–6.
14
Halberg RB, Katzung DS, Hoff PD, et al. Tumorigenesis in the multiple intestinal neoplasia mouse: redundancy of negative regulators and specificity of modifiers.
Proc Natl Acad Sci U S A
2000
;
97
:
3461
–6.
15
Haigis KM, Caya JG, Reichelderfer M, Dove WF. Intestinal adenomas can develop with a stable karyotype and stable microsatellites.
Proc Natl Acad Sci U S A
2002
;
99
:
8927
–31.
16
Shoemaker AR, Moser AR, Midgley CA, et al. A resistant genetic background leading to incomplete penetrance of intestinal neoplasia and reduced loss of heterozygosity in ApcMin/+ mice.
Proc Natl Acad Sci U S A
1998
;
95
:
10826
–31.
17
Bacher JW, Abdel Megid WM, Kent-First MG, Halberg RB. Use of mononucleotide repeat markers for detection of microsatellite instability in mouse tumors.
Mol Carcinog
2005
;
44
:
285
–92.
18
Snijders AM, Nowak NJ, Huey B, et al. Mapping segmental and sequence variations among laboratory mice using BAC array CGH.
Genome Res
2005
;
15
:
302
–11.
19
Cormier RT, Hong KH, Halberg RB, et al. Secretory phospholipase Pla2g2a confers resistance to intestinal tumorigenesis.
Nat Genet
1997
;
17
:
88
–91.
20
Petrini JH, Theunissen JW. Double strand break metabolism and cancer susceptibility: lessons from the mre11 complex.
Cell Cycle
2004
;
3
:
541
–2.
21
Chen X, Halberg RB, Burch RP, Dove WF. Intestinal adenomagenesis involves core molecular signatures of the epithelial-mesenchymal transition.
J Mol Histol
2008
;
39
:
283
–94.
22
Shoemaker AR, Moser AR, Dove WF. N-ethyl-N-nitrosourea treatment of multiple intestinal neoplasia (Min) mice: age-related effects on the formation of intestinal adenomas, cystic crypts, and epidermoid cysts.
Cancer Res
1995
;
55
:
4479
–85.
23
Beck JA, Lloyd S, Hafezparast M, et al. Genealogies of mouse inbred strains.
Nat Genet
2000
;
24
:
23
–5.
24
Lifsted T, Le Voyer T, Williams M, et al. Identification of inbred mouse strains harboring genetic modifiers of mammary tumor age of onset and metastatic progression.
Int J Cancer
1998
;
77
:
640
–4.
25
Park YG, Zhao X, Lesueur F, et al. Sipa1 is a candidate for underlying the metastasis efficiency modifier locus Mtes1.
Nat Genet
2005
;
37
:
1055
–62.
26
Baran AA, Silverman KA, Zeskand J, et al. The modifier of Min 2 (Mom2) locus: embryonic lethality of a mutation in the Atp5a1 gene suggests a novel mechanism of polyp suppression.
Genome Res
2007
;
17
:
566
–76.
27
Kwong LN, Shedlovsky A, Biehl BS, et al. Identification of Mom7, a novel modifier of ApcMin/+ on mouse chromosome 18.
Genetics
2007
;
176
:
1237
–44.
28
Armitage P, Doll R. The age distribution of cancer and a multi-stage theory of carcinogenesis.
Br J Cancer
1954
;
8
:
1
–12.
29
Smits R, Kartheuser A, Jagmohan-Changur S, et al. Loss of Apc and the entire chromosome 18 but absence of mutations at the Ras and Tp53 genes in intestinal tumors from Apc1638N, a mouse model for Apc-driven carcinogenesis.
Carcinogenesis
1997
;
18
:
321
–7.
30
Shoemaker AR, Luongo C, Moser AR, Marton LJ, Dove WF. Somatic mutational mechanisms involved in intestinal tumor formation in Min mice.
Cancer Res
1997
;
57
:
1999
–2006.
31
Janssen KP, Alberici P, Fsihi H, et al. APC and oncogenic KRAS are synergistic in enhancing Wnt signaling in intestinal tumor formation and progression.
Gastroenterology
2006
;
131
:
1096
–109.
32
Lengauer C, Kinzler KW, Vogelstein B. Genetic instabilities in human cancers.
Nature
1998
;
396
:
643
–9.
33
Alberici P, De Pater E, Cardoso J, et al. Aneuploidy arises at early stages of Apc-driven intestinal tumorigenesis and pinpoints conserved chromosomal loci of allelic imbalance between mouse and human.
Am J Pathol
2007
;
170
:
377
–87.
34
Martin ES, Tonon G, Sinha R, et al. Common and distinct genomic events in sporadic colorectal cancer and diverse cancer types.
Cancer Res
2007
;
67
:
10736
–43.
35
Sieber OM, Lamlum H, Crabtree MD, et al. Whole-gene APC deletions cause classical familial adenomatous polyposis, but not attenuated polyposis or “multiple” colorectal adenomas.
Proc Natl Acad Sci U S A
2002
;
99
:
2954
–8.
36
Luo G, Santoro IM, McDaniel LD, et al. Cancer predisposition caused by elevated mitotic recombination in Bloom mice.
Nat Genet
2000
;
26
:
424
–9.
37
Goss KH, Risinger MA, Kordich JJ, et al. Enhanced tumor formation in mice heterozygous for Blm mutation.
Science
2002
;
297
:
2051
–3.
38
Haigis KM, Dove WF. A Robertsonian translocation suppresses a somatic recombination pathway to loss of heterozygosity.
Nat Genet
2003
;
33
:
33
–9.
39
McMurray MA, Gottschling DE. An age-induced switch to a hyper-recombinational state.
Science
2003
;
301
:
1908
–11.
40
Issa JP. Colon cancer: it's CIN or CIMP.
Clin Cancer Res
2008
;
14
:
5939
–40.
41
Yatabe Y, Tavare S, Shibata D. Investigating stem cells in human colon by using methylation patterns.
Proc Natl Acad Sci U S A
2001
;
98
:
10839
–44.
42
Kim JY, Siegmund KD, Tavare S, Shibata D. Age-related human small intestine methylation: evidence for stem cell niches.
BMC Med
2005
;
3
:
10
.
43
Tomlinson IP, Novelli MR, Bodmer WF. The mutation rate and cancer.
Proc Natl Acad Sci U S A
1996
;
93
:
14800
–3.
44
Bodmer W. Genetic instability is not a requirement for tumor development.
Cancer Res
2008
;
68
:
3558
–60.
45
Rusan NM, Peifer M. Original CIN: reviewing roles for APC in chromosome instability.
J Cell Biol
2008
;
181
:
719
–26.
46
Weaver BA, Cleveland DW. Aneuploidy: instigator and inhibitor of tumorigenesis.
Cancer Res
2007
;
67
:
10103
–5.
47
Rao CV, Yang YM, Swamy MV, et al. Colonic tumorigenesis in BubR1+/−ApcMin/+ compound mutant mice is linked to premature separation of sister chromatids and enhanced genomic instability.
Proc Natl Acad Sci U S A
2005
;
102
:
4365
–70.
48
Yao X, Buermeyer AB, Narayanan L, et al. Different mutator phenotypes in Mlh1- versus Pms2-deficient mice.
Proc Natl Acad Sci U S A
1999
;
96
:
6850
–5.