Qualitative and Quantitative Relationship between Dysplastic Aberrant Crypt Foci and Tumorigenesis in the Min/+ Mouse Colon

FAP,² an inherited form of human colon cancer, is caused by germ-line mutations in the tumor suppressor gene, APC. FAP is characterized by the development of hundreds to thousands of adenomas in the large intestine. If not removed, some of these lesions inevitably progress to carcinomas (1). In addition, somatic mutations in the APC gene are found in 85% of sporadic colorectal cancers (1, 2). Various germ-line mutations in the murine Apc, which is homologous to the human APC, cause multiple intestinal neoplasia in mice (3). Min/+ mice are heterozygous for a nonsense mutation in the Apc gene at codon 850 (4). In both FAP and Min/+ mice the intestinal tumorigenesis seems to be dependent on somatic genetic events that lead to the inactivation of both alleles (5, 6, 7). Loss of APC function precludes the post-translational down-regulation of β-catenin, which consequently accumulates in cytoplasm and translocates into the nucleus where it complexes with the transcription factor Tcf-4 and activates specific genes such as c-MYC and cyclinD1. Increased levels and biological activity of free β-catenin may also result from a mutated β-catenin gene (3).

ACF, putative preneoplastic lesions in the colon of carcinogen-treated rodents, have been used as a short-term bioassay to evaluate the role of nutritional components and chemopreventive agents at an early stage of colon carcinogenesis (8). They are scored by light microscopic examination of the mucosal surface and have also been identified in patients with sporadic colorectal cancer and patients with FAP (9). Although ACF share many morphological, genetic, and biochemical features with colonic tumors (7, 10), there is not a simple association between ACF formation and tumorigenesis. In carcinogen-treated rodents, a negative correlation between ACF formation and tumor formation (11) and a discrepancy between distribution of ACF and distribution of tumors (12) is even reported. In carcinogen-treated rodents and patients with sporadic colorectal cancer, the number of tumors is minuscule compared with the large number of ACF, indicating that only a very small fraction of ACF progresses to the stage of a tumor (9, 11). This is consistent with the observation that a large fraction of ACF is hyperplastic whereas only a small fraction of ACF shows dysplasia, a hallmark of malignant potential (9, 13). In contrast, most ACF in FAP patients are dysplastic, and their histopathology resembled that of the numerous adenomas developing (9). It has been proposed that only the dysplastic ACF progress to adenomas and adenocarcinomas (7) and that these lesions are closely related to APC mutations. In AOM-treated rats, small dysplastic lesions undetectable as ACF by surface examination and with β-catenin gene mutations have been described in histopathological sections (14). Such mutations were also observed in rat and mouse colonic tumors (15, 16). Increased expression of β-catenin has been reported in colonic tumors of AOM-treated mice and in spontaneous intestinal tumors of Min/+ mice (17).

Contrary to the hypothesis that ACF are preneoplastic lesions, we did not observe spontaneous formation of such classical ACF in the colon of Min/+ mice (18), although we demonstrated preliminarily that classical ACF may be formed in AOM-treated Min/+ mice (19). However, in untreated Min/+ mice we discovered small lesions of a different type, which we denoted ACF_Min. In contrast to classical ACF, these lesions were not elevated above the surrounding mucosa, and their detection was totally dependent on methylene blue staining and transillumination. ACF_Min exhibited dysplastic crypts similar to those found in adenomas. Also, the C57Bl/6J background of Min/+ mice is susceptible to AOM-induced ACF formation (20).

The aim of the present work was to monitor the development of classical ACF and ACF_Min in the colon of AOM-treated Min/+ mice to evaluate their role as potential precursors of adenomas. To ensure a large tumor outcome, AOM was given at weeks 1 and 2 after birth, when the Min/+ mice are particularly susceptible to carcinogens (21). The morphological relationship between the lesions was examined by histopathology and expression of β-catenin as determined by immunohistochemistry.

The mice were bred at the National Institute of Public Health, Oslo, Norway, from inbred mice originally purchased from The Jackson Laboratory (Bar Harbor, ME). The Min/+ pedigree was maintained by mating C57BL/6J-+/+ (wild-type) females with C57BL/6J-Min/+ males, and procedures to secure inbreeding were followed. The Min/+ mice were identified by allele-specific PCR on DNA isolated from blood (21). Water and diet were given ad libitum. The mice were given a breeding diet, SDS RM3 (E), from Special Diets Service (Witham, United Kingdom), during gestation and until 5 weeks of age. Thereafter they were given a standard maintenance diet from B&K Universal (Grimston, United Kingdom).

AOM from Sigma Chemical Co. (St. Louis, MO) was dissolved in 0.9% NaCl. Min/+ mice or their +/+ littermates were given 5 mg/kg body weight AOM s.c. at weeks 1 and 2 after birth, when the animals were particularly susceptible to carcinogens (21). Other groups of mice received similar volumes of the vehicle 0.9% NaCl. The mice were killed by cervical dislocation at 7 or 11 weeks of age.

The colons were removed, rinsed in ice-cold PBS, slit open longitudinally, and fixed flat between wet (PBS) filter papers for 48 h in 10% neutral buffered formalin prior to 30-s staining with 0.2% methylene blue (George T. Gurr Ltd., London, United Kingdom) dissolved in the same formalin solution. At least 24 h after staining, the mucosa was examined by transillumination in an inverse LM (18). Criteria used to identify AC constituting classical ACF were their increased size, bright blue staining, slightly elevated appearance from the surrounding mucosa, and often a more elongated shape of the luminal opening. The size of classical ACF was scored as focal crypt multiplicity (AC/classical ACF). Criteria used to identify aberrant crypts constituting ACF_Min were their increased size, bright blue staining, and flat appearance hidden in the surrounding mucosa. Because the ACF_Min were not observed as elevated structures, their bright blue staining observed with transillumination was essential for identification. The size of ACF_Min was scored as AC/ACF_Min. Lesions with more than 12 aberrant crypts were scored as tumors, and their diameters were scored with an eyepiece graticule. The relationship between tumor diameter (d) and crypt multiplicity (c) was empirically determined to be: c = 0.5 × d².

Areas with mucosal lesions, identified by LM surface examination of whole-mount colon preparations, were dissected, washed in PBS, and dehydrated in a graded series of ethanol from 70–100%, before critical point drying (Balzers Union, Balzers, Liechtenstein) from CO₂. The dried specimens were oriented and mounted on stubs under a stereomicroscope and then sputter-coated with platinum. SE was carried out using a JSM 840 SEM [Japan Electron Optical Laboratory (JEOL), Tokyo, Japan] operated at 15 keV.

Areas with mucosal lesions, identified by LM surface examinations of whole-mount colon preparations, were dissected, embedded in paraffin wax, cut in parallel with the mucosal surface, and stained with H&E.

Paraffin-embedded formalin fixed sections were prepared; deparaffinized; and rehydrated in xylene, graded alcohol, and water. The endogenous peroxidase activity was blocked by incubating the sections in 0.44% hydrogen peroxide for 30 min at room temperature. After the sections were blocked in normal goat serum for 20 min, they were incubated over night at 4°C with mouse monoclonal anti-β-catenin antibody (Transduction Laboratories, Lexington, KY). Then the sections were incubated with goat antimouse antibody and ABC-AP reagent according to the manufacturer’s instructions (Vectorstain ABC-AP kit; Vector Laboratories, Burlingame, CA). Applying 3,3′-diaminobenzidine tetrahydrochloride containing 0.075% of hydrogen peroxide for 10 min showed peroxidase activity. The sections were stained with hematoxylin for 10 s and mounted with coverslips.

To reduce unspecific coloring we also tried an avidin/biotin blocking kit (Vector Laboratories) to quench the endogenous biotin/avidin and mouse-on-mouse blocking from Vector M.O.M. immunodetection kit (Vector Laboratories). None of these treatments improved the β-catenin staining or reduced background coloring and were therefore not included in the protocol.

We used a two-way ANOVA (SigmaStat software; Jandel Scientific, Erkrath, Germany), after rank transformation of the data attributable to their non-normal distribution, in order to test the following null hypotheses: (a) there is no change from week 7 to 11; and (b) there is no effect of gender. The Bonferroni all pairwise multiple comparison procedure was used to isolate the groups that were different. The Mann-Whitney rank order test was used to compare two groups. A P < 0.05 was considered significant.

By transillumination and surface examination of whole-mount colon preparations, we identified two types of ACF. In the AOM-treated Min/+ mice and +/+ littermates, we detected classical ACF as well as ACF_Min. In vehicle-treated Min/+ mice controls, we only detected ACF_Min, and in vehicle-treated +/+ mice controls, no aberrant crypts of either type were seen. Histopathological and immunohistochemical examination of lesions identified as classical ACF (Fig. 1,A) revealed normal or hyperplastic crypts (10/10 classical ACF examined; Fig. 1,B) and no cytoplasmic overexpression of β-catenin (0/10; Fig. 1,C). The β-catenin expression in these lesions was, as in normal adjacent crypts, restricted to the membrane of the cell-cell borders. In contrast, lesions identified as ACF_Min (Fig. 1, D and G) were uniformly (15/15 ACF_Min examined) characterized by moderate to severe dysplasia (Fig. 1, E and H) and cytoplasmic overexpression of β-catenin (Fig. 1, F and I) and displayed the same feature in Min/+ mice and +/+ mice. The characteristic picture of dysplasia and cytoplasmic β-catenin accumulation in ACF_Min was also seen in tumors (Fig. 1, J–L) of Min/+ mice (6/6 tumors examined) and +/+ mice (2/2 tumors examined). The dysplastic crypts of ACF_Min and tumors (Fig. 1, E, H, and K) were characterized by increased crypt diameter, crowding of immature crypt cells, pseudostratification, loss of goblet cells, and a general increase in the size of the nuclei, as we reported previously in untreated Min/+ mice (18).

In the lesions with dysplasia and cytoplasmic β-catenin accumulation described above, nuclear translocation of β-catenin was not prominent. To test whether this was attributable to our routine long-term formalin fixation (and storage) of the whole-mount preparations, we treated two Min/+ mice with AOM in a second experiment and fixed their preparations ≤24 h in formalin. Surprisingly, the dysplastic lesions in these preparations displayed both cytoplasmic and nuclear accumulation of β-catenin (data not shown), indicating that nuclear immunostaining of β-catenin was dependent on the duration of formalin fixation.

In the SEM, classical ACF were easily observed as elevated lesions (Fig. 2,A) with similar ultrastructure as the surrounding epithelium, including the presence of goblet cells. In contrast, ACF_Min with few crypts displayed enlarged crypts with a flat appearance (Fig. 2,B). Large ACF_Min had enlarged and swollen crypts still with a relative flat appearance (Fig. 2,C). ACF_Min had less goblet cells compared with the background epithelium (Fig. 2 C).

The number of AOM-induced ACF was significantly higher in Min/+ mice than in +/+ mice at week 7 (P < 0.01) but not at week 11 (Table 1). In +/+ mice, the number of ACF was significantly increased from week 7 to 11. The focal crypt multiplicity was significantly increased from week 7 to 11 in both Min/+ mice and +/+ mice (P < 0.001, data not shown).

In Min/+ mice treated with AOM, a maximal 13–16-fold increase in ACF_Min number at week 7 (P < 0.002) and a maximal 18–20-fold increase in tumor number at week 11 (P < 0.001) was induced compared with vehicle-treated Min/+ controls (Table 1). Apparently there was an association between induction of ACF_Min and formation of tumors, because the decrease in ACF_Min number from week 7 to 11 corresponded with the increase in tumor number in the same period. Because the number of ACF_Min + tumors was virtually constant, it was likely that a large number of ACF_Min at week 7 grew to the size of a tumor at week 11. Such a recruitment of newly formed tumors was supported by the smaller mean tumor size found in AOM-treated Min/+ mice at week 11 compared with that found in tumors in vehicle-treated Min/+ mice (P < 0.001; data not shown). To visualize this hypothesis we transformed tumor diameter to crypt multiplicity (see “Materials and Methods”), pooled ACF_Min and tumors from both males and females in a size distribution plot, and subtracted the size distribution of the Min/+ control background (Fig. 3). On the basis of the areas under the curves for the size distributions of AOM-induced lesions at weeks 7 and 11 (Fig. 3,B), ∼56% of the ACF_Min at week 7 had grown to the size of a tumor at week 11. With this presumed association between ACF_Min growth and tumor formation, the ACF_Min (Fig. 3,B) grew considerably faster than the classical ACF (Fig. 3 A). This was illustrated by the significantly larger mean crypt multiplicity of pooled ACF_Min and tumors at week 11 compared with the corresponding mean crypt multiplicity of classical ACF (33.4 ± 2.6, n = 418 versus 6.2 ± 0.3, n = 157, P < 0.001).

Also in +/+ mice, the AOM treatment resulted in the induction of ACF_Min and tumors. However, the number of these lesions was at the same low level as in control Min/+ mice.

By transillumination and surface examination of whole-mount colon preparations of AOM-treated Min/+ mice, we were able for the first time to identify specific lesions that show a continuous development from the monocryptal stage to adenoma. We called these lesions ACF_Min because we first discovered them in untreated Min/+ mice (18). The uniform picture of dysplasia and cytoplasmic overexpression of β-catenin in monocryptal ACF_Min, large ACF_Min, and adenomas indicated a qualitative relationship between ACF_Min and tumorigenesis. A quantitative relationship was also observed because the dramatic decrease in ACF_Min number from week 7 to 11 was paralleled by a reciprocal increase in tumor number. By calculating the crypt multiplicity of ACF_Min and tumors pooled, we determined a high growth rate for ACF_Min. In all of the AOM-treated Min/+ mice, it was possible to distinguish ACF_Min from classical ACF, which were not dysplastic and did not display cytoplasmic accumulation of β-catenin. The number of classical ACF was virtually constant between week 7 and 11, the focal crypt multiplication of classical ACF was modest and they were not found in untreated Min/+ mice. Therefore, in contrast to ACF_Min, the classical ACF did not seem to be related to tumorigenesis.

Contrary to what was reported for tumors of Min/+ mice (17), we did not see significant nuclear immunostaining of β-catenin in the dysplastic lesions. This was probably related to the long-term formalin fixation applied in the present study, because nuclear immunostaining of β-catenin was indeed confirmed in dysplastic lesions from an additional experiment where the preparations were fixed no longer than 24 h. This is in concurrence with a previous study demonstrating that the immunohistochemical detection of nuclear β-catenin in colorectal tumors may be dependent on the method of fixation (22).

The fact that AOM treatment led to the formation of ACF_Min in both Min/+ mice and in +/+ mice indicated that these lesions are general precursors of adenomas in the colon. Although the ACF_Min and tumors in both Min/+ mice and in +/+ mice were characterized by altered β-catenin accumulation, the initiating mechanisms induced by AOM might be different in the two genotypes. In Min/+ mice, which harbor only one intact Apc allele, it is likely that the formation of ACF_Min is initiated by the inactivation of the remaining wild-type Apc allele, leaving the crypt cell with lost Apc function. Such a mechanism, which implies that the Min/+ mice are more vulnerable than +/+ mice to AOM treatment, is consistent with the 50–100-fold higher frequency of ACF_Min observed in Min/+ mice than in +/+ mice at week 7. Accordingly, the tumor frequency at week 11 was ∼100-fold higher in Min/+ mice than in +/+ mice. Furthermore, genetic analyses of lesions from untreated Min/+ mice (5, 6, 7) and Min/+ mice treated with N-ethyl-N-nitrosourea (23) suggest that intestinal tumorigenesis in these animals are dependent on somatic events that lead to the inactivation of both Apc alleles. On the other hand, in AOM-treated +/+ mice, which harbor two intact Apc alleles, it is more likely that β-catenin gene mutation rather than “two hits” at the Apc locus is the main cause of lost β-catenin control. This speculation accords with results from other studies in wild-type rodents. Interestingly, the morphological feature of ACF_Min appears identical to the histologically altered crypts with macroscopically normal-like appearance recently identified in early rat colon carcinogenesis induced by AOM (14). Histologically altered crypts with macroscopically normal-like appearance, which could not be distinguished from normal adjacent crypts in whole-mount preparations examined by surface microscopy, were characterized by altered β-catenin control and frequent β-catenin gene mutations. Such alterations in β-catenin were also observed in colonic tumors from rat and mouse (15, 16, 17). Additional genetic analyses of ACF_Min in Min/+ mice and +/+ mice are required to elucidate the mechanisms for their altered β-catenin accumulation.

We described previously ACF_Min in the colon of untreated Min/+ mice as flat structures that could not be recognized in the SEM (21). In the present study, we could identify large ACF_Min by scanning electron microscopy probably because the lesions had developed large, swollen crypts that were slightly elevated upon expansion in the mucosa. Disrupted Apc control of the β-catenin/Tcf pathway, a signal that is essential for the maintenance of the proliferative compartments of the intestine during embryogenesis (24), may explain the immature (dysplastic) appearance of crypt cells constituting ACF_Min. The relative flat appearance of ACF_Min could be a result of disrupted Apc control of the cell cycle (25) and cell anchoring (26), as well as disrupted Apc-driven migration (27) and apoptosis (28). Furthermore, ACF_Min resemble a subset of ACF, the dysplastic ACF, that are described previously as potential precursors of adenomas in rodent and human colon carcinogenesis (7, 10, 13, 29). Interestingly, these dysplastic ACF, which are characterized by their larger size (13, 29), accord with the fast growing ACF_Min in our study. If these dysplastic ACF initially possess a flat morphology like small ACF_Min, they might not be identified by conventional ACF scoring. However, at later stages they might be scored, as slightly elevated structures with dysplasia, like large ACF_Min.

The induction of classical ACF by AOM was apparently not affected by Apc status because the number of these lesions was approximately the same in Min/+ and +/+ mice, and because they displayed normal β-catenin expression. The classical ACF in our experiment resemble the previously described hyperplastic ACF, which are not assumed to be precursors of adenomas (9). These hyperplastic lesions, which constitute the majority of ACF, are characterized by mutations in ras (7, 9, 30) rather than in Apc (7) or β-catenin (14). These lesions are frequent in tumor resistant mice treated with AOM (31) and in non-neoplastic colonic disease (27).

In conclusion, we identified two distinct populations of altered crypts, ACF_Min and classical ACF, in the colon of Min/+ mice after AOM treatment. Apparently, these populations represent separate directions of development already from the monocryptal stage. The ACF_Min, which resemble the dysplastic ACF described previously, demonstrated a continuous development from the monocryptal stage to adenoma with fast crypt multiplication and altered control of β-catenin. At the initial stage, the dysplastic lesions were flat structures hidden in the surrounding mucosa and therefore not detectable as ACF by conventional methods. In contrast, the classical ACF, which resemble the hyperplastic ACF as described in the literature, were characterized by slow growing crypts with normal β-catenin expression, and they were probably not related to tumorigenesis. Additional studies are warranted to characterize the phenotypic and genotypic features of both types of ACF and their relationship to colonic carcinogenesis in FAP as well as in sporadic colorectal cancer.

We thank Marit Hindrum and Ingeborg Løstegaard Goverud for excellent technical assistance.