Ionizing radiation, inflammation, and their interactions in colon carcinogenesis in Mlh1-deficient mice.

Genetic, physiological and environmental factors are implicated in colorectal carcinogenesis. Mutations in the mutL homolog 1 (MLH1) gene, one of the DNA mismatch repair genes, are a main cause of hereditary colon cancer syndromes such as Lynch syndrome. Long-term chronic inflammation is also a key risk factor, responsible for colitis-associated colorectal cancer; radiation exposure is also known to increase colorectal cancer risk. Here, we studied the effects of radiation exposure on inflammation-induced colon carcinogenesis in DNA mismatch repair-proficient and repair-deficient mice. Male and female Mlh1(-/-) and Mlh1(+/+) mice were irradiated with 2 Gy X-rays when aged 2 weeks or 7 weeks and/or were treated with 1% dextran sodium sulfate (DSS) in drinking water for 7 days at 10 weeks old to induce mild inflammatory colitis. No colon tumors developed after X-rays and/or DSS treatment in Mlh1(+/+) mice. Colon tumors developed after DSS treatment alone in Mlh1(-/-) mice, and exposure to radiation prior to DSS treatment increased the number of tumors. Histologically, colon tumors in the mice resembled the subtype of well-to-moderately differentiated adenocarcinomas with tumor-infiltrating lymphocytes of human Lynch syndrome. Immunohistochemistry revealed that expression of both p53 and β-catenin and loss of p21 and adenomatosis polyposis coli proteins were observed at the later stages of carcinogenesis, suggesting a course of molecular pathogenesis distinct from typical sporadic or colitis-associated colon cancer in humans. In conclusion, radiation exposure could further increase the risk of colorectal carcinogenesis induced by inflammation under the conditions of Mlh1 deficiency.

Humans are exposed to multiple environmental carcinogens including genotoxic and non-genotoxic chemicals, heavy metals, inflammatory agents, and both UV and ionizing radiation. The leading factors contributing to human cancer are carcinogens in tobacco smoke followed by those in the diet, and inflammation. Heritable factors are estimated to account for 0–42% of cancers, depending on the organ,1 in addition to the known genetic influence on the varying susceptibility of individuals to carcinogen exposures, including for radiation-induced cancer.2 Combined exposures to environmental carcinogens in susceptible individuals in the population thus represent a high-risk scenario that warrants further investigation. Although interactions of ionizing radiation and genotoxin or hormone exposures have been studied,3 little is known about interactions between ionizing radiation and inflammation in cancer-prone individuals.

Colorectal cancer (CRC) is the third most frequent malignancy worldwide4 and is mostly sporadic. Yet heredity, mostly inherited defects in DNA mismatch repair (MMR), is a significant factor in 20–25% of cases.5–7 The MMR pathway is responsible for repair of base mismatches and small insertions/deletions arising during DNA replication8,9 and is important for maintaining genomic stability and preventing tumorigenesis.10 Mutations in the MMR gene mutL homolog 1 (MLH1) are involved in hereditary non-polyposis colorectal cancer, or Lynch syndrome.11

Inflammation is an important risk factor for gastrointestinal diseases and tumors,12 with colitis-associated CRC the most serious long-term complication of inflammatory bowel disease.13 Mild inflammation accelerates colon carcinogenesis in MMR-deficient mice (Mlh1−/−, Msh2),14,15 and can itself inactivate MMR through Mlh1 promoter methylation.16 Microsatellite instability (MSI) is also known to occur in colitis-associated CRC and dysplastic lesions in humans,17 with one study revealing that 46% of colon neoplasms/lesions with a high degree of MSI showed hypermethylation of the MLH1 gene.18

Colon cancer risk also increases following exposure to ionizing radiation,19 now a consideration when evaluating adjuvant radiotherapy for abdominal cancers,20,21 and justifying computed tomography colonography, a first-line screening modality to detect CRC. In addition to repairing endogenous damage, MMR is active in the response to radiation-induced DNA damage.22,23 We previously showed a small but significant acceleration of intestinal tumor development after exposure to 2 Gy X-rays in 10-week-old Mlh1−/− mice, with no sensitivity observed in Mlh1+/− heterozygotes.15,24 Yet, scarce data are available on the interaction between ionizing radiation and inflammation on colon carcinogenesis, particularly in conjunction with inherited MMR deficiency. Lynch syndrome patients are a subpopulation that may be at high risk for radiation-induced colon cancer, particularly in the context of chronic inflammation, warranting investigation for radiation protection purposes. Furthermore, the sensitivity of Lynch syndrome patients and mouse models of Mlh1 deficiency provides an opportunity to observe interactions that are difficult to study in the wider population, but may have significant impact given the prevalence of both CRC and chronic inflammation.

Here, we examined the combined effect of ionizing radiation and induced inflammatory colitis on colon carcinogenesis in Mlh1−/− and control Mlh1+/+ mice, and examined molecular pathogenesis through the expression of the CRC-associated proteins adenomatosis polyposis coli (Apc), β-catenin, p53, and p21.

Materials and Methods

Mice

Mice carrying a disrupted exon 4 of Mlh1, whose genetic background was >95% C57BL/6, were provided by Prof. Liskay (Oregon Health and Science University, Portland, OR, USA).25 The Mlh1 genotype was determined by PCR from mouse ear punch DNA as previously described.24 Male and female mice were housed separately in stainless steel cages (4–5 mice per cage) with water and a standard laboratory diet (MB-1; Funabashi Farm, Chiba, Japan) provided ad libitum. They were maintained at 23 ± 2°C and 50 ± 10% humidity, on a 12:12 h light : dark cycle. All experiments were reviewed and approved by our institution's animal welfare committee.

Induction of inflammatory colitis

Treatment with dextran sodium sulfate (DSS) (molecular weight, 36 000–50 000; MP Biomedicals, Aurora, OH, USA) was used to induce colitis.26,27 For induction of colitis, mice received 1% DSS in drinking water for 1 week when 10 weeks old. One cycle of the 1% DSS regimen, unlike repeated cycles of higher DSS doses, rarely leads to colon carcinogenesis and death in wild type (WT) mice.26,28 Intake of distilled water (DW) or 1% DSS solution per cage (4–5 mice) was measured throughout the treatment periods. DW and DSS load (mL/day/mouse) was calculated.

Experimental design

A total of 226 Mlh1+/+ and Mlh1−/− mice were divided into 12 experimental groups as shown in Figure1. A dose of 2 Gy has been shown to accelerate colon carcinogenesis in susceptible mice models, yet the most sensitive time for radiation exposure differs between studies.24,29 Therefore, the single whole-body irradiation of 2 Gy was given to mice at either 2 or 7 weeks old, to cover both of the observed peaks. X-ray irradiation was carried out using a Pantak X-ray generator (Pantak, East Haven, CT, USA) as previously described.24 Most irradiated Mlh1−/− mice develop thymic lymphomas before colon tumors arise, which makes it difficult to quantitatively assess the role of Mlh1 deficiency in colon carcinogenesis. Therefore, in this experiment, thymectomy was carried out at 5 weeks old to exclude the confounding factor of thymic lymphoma in mice of all groups. All mice were weighed before being killed under terminal isoflurane anesthesia when 25 weeks old.

Figure 1

Overview of the experimental design to ascertain the effects of radiation exposure on inflammation-induced colon carcinogenesis in DNA mismatch repair-proficient (Mlh1+/+) and repair-deficient (Mlh1−/−) mice. DSS, dextran sodium sulfate; F, female; M, male.

Histopathological examination

After death, the colons were harvested and the mucosal layer spread out over filter paper to expose any lesions protruding into the lumen. The flattened tissues were then fixed with 10% neutral-buffered formalin for 24 h and then examined closely for lesions. Colon lesions were then stained with H&E for histological analysis, with mucosal dysplasia and neoplasms diagnosed according to previously established criteria.31,30,32

Immunohistochemical staining

Serial 4-μm-thick paraffin sections were used for immunohistochemical examination to determine the expression of Apc, β-catenin, p53, and p21 proteins. These sections were stained with primary antibodies at the following dilutions: 1:50 for Apc (polyclonal; Abcam, Tokyo, Japan), 1:100 for β-catenin (monoclonal; BD Transduction Laboratories, Lexington, KY, USA), 1:500 for p53 (polyclonal; Novocastra Laboratories, Newcastle, UK), and 1:500 for p21 (polyclonal; Santa Cruz Biotechnology, Santa Cruz, CA, USA). Standard protocols for immunohistochemistry were followed as described previously.15

Harvesting DNA from colon lesions by laser capture microdissection

β-catenin- and p53-positive focal areas and adjacent normal mucosa were isolated from the immunostained tissues by laser capture microdissection (MMI CellCut; Molecular Machines and Industries, Eching, Germany) and collected onto the adhesive cap of the provided 500-μL microcentrifuge tubes. DNA was extracted using the QIAmp DNA Micro kit (Qiagen, Hilden, Germany) according to their specific protocol for extraction of genomic DNA from laser capture microdissection tissues.

Mutation analysis of β-catenin and p53 genes

The PCR primers were designed to amplify exon 3 of β-catenin, which encodes the GSK-3β phosphorylation sites, a known mutation hotspot. The sequences of the primers were as follows: for exon 3, ctnnb1-F (5′-GGA GTT GGA CAT GGC CAT GG-3′) and ctnnb1-R (5′-TCA ACA TCT TCT TCC TCA GG-3′). The PCR primers to amplify p53 exons 6, 7, 8, and 9 were designed, giving coverage of the coding region for the DNA-binding domain. The sequences of the primers were as follows: for exon 6, E6-1F (5′-TCC CGG CTT CTG ACT TAT TC-3′) and E6-1R (5′-TGC CTG TCT TCC AGA TAC TCG-3′), E6-2F (5′-CGG GTG GAA GGA AAT TTG TA-3′) and E6-2R (5′-AAG ACG CAC AAA CCA AAA CA-3′); for exon 7, E7-1F (5′-GTA GGG AGC GAC TTC ACC TG-3′) and E7-1R (5′-CCC CAT GCA GGA GCT ATT AC-3′), E7-2F (5′-TAC CAC CAT CCA CTA CAA GTA CA-3′) and E7-2R (5′-CTA CCA CGC GCC TTC CTA C-3′); for exon 8, E8-F2 (5′-TGC TGG TCC TTT TCT TGT CC-3′) and E8-R2 (5′-TGA AGC TCA ACA GGC TCC TC-3′); for exon 9, E9-F (5′-GGA GGA GCC TGT TGA GCT TC-3′) and E9-R (5′-ATG CGA GAG ACA GAG GCA AT-3′). The PCR products were directly sequenced using a BigDye Terminator Version 3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) and ABI PRISM 3100 Genetic Analyzer (Applied Biosystems) following the manufacturer's protocol.

Statistical evaluation of data

All statistical analyses were carried out using R version 2.15.0 software function nlm (http:www.r-project.org). An unpaired Student's t-test was used to analyze average body weight (BW) and for intake of DW and 1% DSS. Fisher's exact probability test was used to analyze the incidence of colon lesions; the Kruscal–Wallis test with Steel–Dwass multiple comparisons method and the Mann–Whitney U-test were used to analyze the multiplicity of colon lesions. The results were considered significant at P < 0.05.

Results

Changes in BW

Although BW was significantly lower in all treated male mice on the Mlh1 background compared to treatment-matched WT males, only the combined exposures and X-rays at 2 weeks old caused a significant reduction below the weight of the Mlh1 control males (Table1). Only DSS combined with irradiation at 7 weeks caused a significant reduction in BW for WT males compared to the WT control males. In contrast, female mice on the Mlh1 background showed no changes in BW, whereas the controls showed the lowest BW for WT females, with the three single treatments each causing a significant increase in weight at 25 weeks old.

Table 1

Effect of X-ray irradiation (2 Gy) and/or dextran sodium sulfate (DSS) on body weight (BW) in Mlh1+/+ and Mlh1−/− mice

Group no.	Treatment	Genotype	BW, g
			Male (no. of mice)	Female (no. of mice)
1	Untreated control	+/+	30.8 ± 1.3 (10)	22.3 ± 1.5 (10)
2		−/−	30.1 ± 2.1 (9)	22.4 ± 3.9 (8)
3	X-rays (2 weeks)	+/+	30.1 ± 2.4 (10)	24.2 ± 1.4** (10)
4		−/−	26.5 ± 3.3*,# (8)	22.2 ± 1.9# (9)
5	X-rays (7 weeks)	+/+	31.0 ± 1.2 (10)	24.4 ± 2.5* (10)
6		−/−	28.4 ± 1.8## (9)	21.9 ± 3.9 (8)
7	DSS (10 weeks)	+/+	31.5 ± 2.5 (10)	24.2 ± 1.7* (10)
8		−/−	28.6 ± 2.3# (9)	22.6 ± 2.8 (10)
9	X-rays (2 weeks)/DSS	+/+	29.9 ± 2.3 (10)	24.3 ± 2.7 (10)
10		−/−	27.2 ± 1.5**,## (9)	22.2 ± 2.7 (9)
11	X-rays (7 weeks)/DSS	+/+	28.9 ± 2.3* (10)	24.6 ± 3.8 (10)
12		−/−	26.2 ± 2.3**,# (10)	21.8 ± 2.1 (8)

The number of Mlh1−/− mice per group ranges from 8 to 10 for each sex because a small number of mice died due to thymectomy or DSS treatment. Data are shown as the mean ± SD.

P < 0.05,

P < 0.01, significant difference from no treatment group of same genotype by an unpaired Student's t-test.

P < 0.05

P < 0.01, significant difference from Mlh1+/+ mice treated with same condition by an unpaired Student t-test.

Intake of DSS and DW

The mean daily intake of DSS and DW for mice of each Mlh1 genotype is shown in Table2. Although female mice drank less water than males, males and females of both genotypes drank less when DSS was present, and the drop in intake was more pronounced in females, leading to a lower relative intake of DSS in females compared to males.

Table 2

Intake of dextran sodium sulfate (DSS) and distilled water (DW) in Mlh1+/+ and Mlh1−/− mice

	Genotype	Male	Female
DSS	+/+	3.20 ± 0.09*	2.70 ± 0.04
	−/−	3.33 ± 0.16*	2.59 ± 0.03
DW	+/+	3.63 ± 0.07*	3.30 ± 0.05
	−/−	3.63 ± 0.07*	3.34 ± 0.03

Data are shown as the mean ± SE (mL).

P < 0.05, significant difference from female mice of same genotype by an unpaired Student's t-test.

Macroscopic view of colon lesions induced by single and combined exposure to X-rays and DSS

The typical gross appearance of colon lesions, which varied in size from 0.5 to 6 mm in diameter, is shown in Figure2. All lesions were sessile types and protruded into the lumen as a polyp (Fig.2b–d) or plaque (Fig.2a). The large lesions, but no small lesions, had hemorrhage on part of the upper surface (Fig.2c,d). No other tumors developed during the experimental period.

Figure 2

Macroscopic view of representative colon lesions from Mlh1−/− mice treated with X-ray irradiation and dextran sodium sulfate (DSS) alone or in combination. (a) X-rays (7 weeks) + DSS, female (arrow, single small protruded lesion). (b) X-rays (7 weeks) + DSS, male (arrow, single sessile polyp without hemorrhage). (c) DSS only, female (arrow, single large sessile polyp with hemorrhage). (d) X-rays (2 weeks) + DSS, male (arrow, single large sessile polyp with hemorrhage).

Combination of X-rays and DSS enhanced colon lesions in Mlh1−/− mice

We first evaluated the incidence and multiplicity of colon lesions induced by the combinations of radiation exposure and DSS treatment, as shown in Figure3. No colon lesions were observed in WT mice with any treatment. A single colon lesion was observed in an untreated male and one in a female exposed to X-rays at 2 weeks old, an indication of the low baseline frequency without DSS induction. All the remaining lesions arose following DSS alone or in combination with radiation at 2 or 7 weeks old, with an incidence in males of 55.6%, 77.8%, and 90%, respectively. The increase in the incidence in males receiving the combined X-rays and DSS treatments above untreated controls reached statistical significance, but not in those given DSS alone. In females, both DSS alone and in combination with X-ray irradiation at 2 weeks old significantly increased incidence above that observed in untreated females, with a higher frequency when the DSS followed irradiation at 2 weeks old. Unlike the males, no additional lesions compared to DSS alone were observed in females after irradiation at 7 weeks old. Although the number of mice available per group did not permit a formal test of synergy between the two treatments, the near complete penetrance in males given DSS combined with irradiation at 7 weeks old was remarkable, and the putative additional colon lesions observed across the four combined-exposure groups exceeded what might have been expected from even an additive effect from the X-ray-induced incidence.

Figure 3

Incidence and multiplicity of colon lesions in Mlh1+/+ and Mlh1−/− mice treated with X-ray irradiation or dextran sodium sulfate (DSS) alone or in combination. Incidence of colon lesions (number of mice with at least one colon lesion compared to total number of mice) is shown as a percentage for each group (a, b) and the multiplicity (number of colon lesions per mouse) is shown as the mean ± SE for each group (c, d). *P < 0.05, **P < 0.01 versus control, Fisher's exact probability test. #P < 0.05, ##P < 0.01 versus control, Kruscal–Wallis test with Steel–Dwass multiple comparisons method. $P < 0.05 versus treatment-matched female mice, Fisher's exact probability test. †P < 0.05 versus treatment-matched female mice, Mann–Whitney U-test.

Multiplicity of colon lesions largely mirrored the trends observed in the incidence data, and was consistent with the various treatments influencing the probability of colon lesions for each mouse rather than increasing the number of colon lesions in the susceptible mice. The difference in multiplicity between males and females was due to a consistently lower number of colon lesions found per tumor-bearing mouse in females, despite a similar incidence of colon lesions between the sexes.

Histopathological findings of colon lesions induced by single and combined exposure to X-rays and DSS

Mucosal hyperplasia was slightly elevated above the mucosal surface and consisted of glandular hyperplasia. The glands were lined with hyperplastic basophilic epithelium with goblet cell differentiation; there was no cellular atypia (Fig.4a). Dysplasia consisted of one or more glands lined with hyperchromatic atypical columnar epithelium with closely packed nuclei (Fig.4b). Dysplastic cells showed loss of normal progression from germinal to fully differentiated cells, and goblet cells were not seen. Tubular adenoma (Fig.4c) and tubular adenocarcinoma (Fig.4d) were seen as endophytic lesions in the mucosa. These tumors consisted of irregular-sized glands lined with dysplastic glandular epithelium with many tumor-infiltrating lymphocytes (TIL) (Fig.4e). In adenocarcinoma, there was obvious invasion of the muscularis mucosa by neoplastic glands (Fig.4d).

Figure 4

Histopathology of colon lesions in Mlh1−/− mice (H&E) treated with X-ray irradiation or dextran sodium sulfate alone or in combination. (a) Hyperplasia (group 10, female). (b) Dysplasia (group 12, male). (c) Tubular adenoma (group 10, female). (d) Well-differentiated tubular adenocarcinoma (group 12, male). (e) Magnified view of the boxed area in (d). Numerous tumor-infiltrating lymphocytes are present in the neoplastic epithelium and stroma (arrows).

Histological examination of colon lesions indicated that combined exposure of male Mlh1−/− mice to X-rays and DSS increased the number of low-grade adenomas, dysplasia, and hyperplasia compared with DSS treatment alone (Fig.5). In females, compared with DSS treatment alone, combined exposure to X-rays at 2 weeks old also increased the number of high-grade adenomas and hyperplasia. In contrast, the failure to observe an enhanced incidence of colon lesions in females with combined exposure to DSS and X-rays at 7 weeks old was due to an unexplained decrease in low- and high-grade adenomas and carcinomas.

Figure 5

Histological classification of colon lesions in male (a) and female (b) Mlh1−/− mice treated with X-ray irradiation or dextran sodium sulfate (DSS) alone or in combination.

Figure6 indicates the distribution of colon lesions in the distal, middle, and proximal areas of the colon. Colon lesions were distributed in all three segments at similar rates in males, whereas in females colon lesions were mainly observed in the middle area of the colon. Low- and high-grade adenomas and adenocarcinomas were mainly distributed in the distal and middle areas in both male and female Mlh1−/− mice. However, low-grade lesions such as hyperplasia and dysplasia were predominantly distributed in the proximal area in male Mlh1−/− mice.

Figure 6

Distribution of lesions in the colon in male (a) and female (b) Mlh1−/− mice treated with X-ray irradiation or dextran sodium sulfate alone or in combination.

Immunohistochemical staining of Apc, β-catenin, p53, and p21 in colon lesions of Mlh1−/− mice

We analyzed the expression of Apc, p53, β-catenin, and p21 proteins in colon lesions of Mlh1−/− mice by immunohistochemistry (Fig.7, Table3). Apc, which was strongly expressed in the nuclei of normal mucosal epithelium (Fig.7a), became faint in adenomas and adenocarcinomas (Fig.7b), suggesting that inactivation of Apc protein is a late event in colon carcinogenesis in this mouse model. The levels of β-catenin protein were concentrated in membranes of normal epithelium (Fig.7c) but could be found in the nuclei and cytoplasm of adenomas and adenocarcinoma cells (Fig.7d). Positivity of strong nuclear β-catenin staining was 84% and 75% of adenomas and adenocarcinomas from mice given DSS combined with irradiation at 2 and 7 weeks old, respectively, higher than in tumors from mice given DSS alone (42%). This also indicated that nuclear localization of β-catenin is a late phenomenon. Weak, diffuse expression of p53 was frequently observed in adenomas (57.7%) and adenocarcinomas (60.9%), with strong expression of p53 in some atypical glands only rarely observed in adenomas (1/10, 10%) and adenocarcinomas (1/6, 16.7%) (Fig.7f) and no expression observed in normal epithelium (Fig.7e). The expression level of p21 protein was reduced in most colon tumors compared to normal colon epithelium (Fig.7g,h). Taken together, these results indicate that dysregulation of Apc, p53, and β-catenin were all late events in colon carcinogenesis in Mlh1−/− mice treated with X-rays and DSS exposure.

Figure 7

Immunohistochemical analysis of adenomatosis polyposis coli (Apc), β-catenin, p53, and p21 expression in normal epithelium (a, c, e, and g) and adenocarcinoma (b, d, f, and h) from Mlh1−/− mice treated with X-ray irradiation and dextran sodium sulfate (group 10 or 12). Arrows indicate strong expression of p53 in atypical glands.

Table 3

Immunohistochemical staining of adenomatosis polyposis coli (Apc), β-catenin, p53, and p21 in colon lesions in Mlh1−/− mice treated with X-ray irradiation and/or dextran sodium sulfate (DSS)

Treatment Colon lesions	Apc			β-catenin			p53			p21
	Nucleus			Membrane	Cytoplasm/nucleus		Nucleus			Nucleus
	−	+	++	+	+	++	−	+	++	−	+	++
Untreated control
Normal epithelium	0/1 (0)	0/1 (0)	1/1 (100)	1/1 (100)	0/1 (0)	0/1 (0)	1/1 (100)	0/1 (0)	0/1 (0)	0/1 (0)	0/1 (0)	1/1 (100)
Hyperplasia	–	–	–	–	–	–	–	–	–	–	–	–
Dysplasia	–	–	–	–	–	–	–	–	–	–	–	–
Adenoma	–	–	–	–	–	–	–	–	–	–	–	–
Adenocarcinoma	0/1 (0)	0/1 (0)	1/1 (100)	0/1 (0)	0/1 (0)	1/1 (100)	1/1 (100)	0/1 (0)	0/1 (0)	0/1 (0)	0/1 (0)	1/1 (100)
X-rays (2 weeks)
Normal epithelium	0/1 (0)	0/1 (0)	1/1 (100)	1/1 (100)	0/1 (0)	0/1 (0)	1/1 (100)	0/1 (0)	0/1 (0)	0/1 (0)	0/1 (0)	1/1 (100)
Hyperplasia	–	–	–	–	–	–	–	–	–	–	–	–
Dysplasia	–	–	–	–	–	–	–	–	–	–	–	–
Adenoma	–	–	–	–	–	–	–	–	–	–	–	–
Adenocarcinoma	0/1 (0)	0/1 (0)	1/1 (100)	0/1 (0)	0/1 (0)	1/1 (100)	0/1 (0)	1/1 (100)	0/1 (0)	0/1 (0)	1/1 (100)	0/1 (0)
DSS
Normal epithelium	0/10 (0)	0/10 (0)	10/10 (100)	14/14 (100)	0/14 (0)	0/14 (0)	14/14 (100)	0/14 (0)	0/14 (0)	0/14 (0)	0/14 (0)	14/14 (100)
Hyperplasia	0/1 (0)	0/1 (0)	1/1 (100)	2/2 (100)	0/2 (0)	0/2 (0)	2/2 (100)	0/2 (0)	0/2 (0)	0/2 (0)	2/2 (100)	0/2 (0)
Dysplasia	–	–	–	–	–	–	–	–	–	–	–	–
Adenoma	0/3 (0)	0/3 (0)	3/3 (100)	2/5 (40)	2/5 (40)	1/5 (10)	2/5 (40)	3/5 (60)	0/5 (0)	1/5 (20)	4/5 (80)	0/5 (0)
Adenocarcinoma	1/6 (16.7)	0/6 (0)	5/6 (88.3)	0/7 (0)	3/7 (42.9)	4/7 (57.1)	1/7 (14.3)	6/7 (85.7)	0/7 (0)	1/7 (14.3)	6/7 (85.7)	0/7 (0)
X-rays (2 weeks)/DSS
Normal epithelium	0/20 (0)	0/20 (0)	20/20 (100)	25/25 (100)	0/25 (0)	0/25 (0)	25/25 (100)	0/25 (0)	0/25 (0)	0/22 (0)	0/22 (0)	22/22 (100)
Hyperplasia	–	–	–	2/2 (100)	0/2 (0)	0/2 (0)	2/2 (100)	0/2 (0)	0/2 (0)	0/1 (0)	1/1 (100)	0/1 (0)
Dysplasia	0/4 (0)	0/4 (0)	4/4 (100)	4/4 (100)	0/4 (0)	0/4 (0)	4/4 (100)	0/4 (0)	0/4 (0)	1/4 (25)	3/4 (75)	0/4 (0)
Adenoma	0/8 (0)	3/8 (37.5)	5/8 (62.5)	3/11 (27.3)	0/11 (0)	8/11 (72.7)	6/11 (54.5)	5/11 (45.5)	0/11 (0)	3/9 (33.3)	6/9 (66.7)	0/9 (0)
Adenocarcinoma	0/8 (0)	6/8 (75)	2/8 (25)	0/8 (0)	0/8 (0)	8/8 (100)	4/8 (50)	4/8 (50)	0/8 (0)	3/8 (37.5)	5/8 (62.5)	0/8 (0)
X-rays (7 weeks)/DSS
Normal epithelium	0/15 (0)	0/15 (0)	15/15 (100)	23/23 (100)	0/23 (0)	0/23 (0)	23/23 (100)	0/23 (0)	0/23 (0)	0/23 (0)	0/23 (0)	23/23 (100)
Hyperplasia	–	–	–	3/3 (100)	0/3 (0)	0/3 (0)	3/3 (100)	0/3 (0)	0/3 (0)	0/3 (0)	3/3 (100)	0/3 (0)
Dysplasia	0/1 (0)	1/1 (100)	0/1 (0)	2/4 (50)	1/4 (25)	1/4 (25)	1/4 (25)	3/4 (75)	0/4 (0)	1/4 (25)	3/4 (75)	0/4 (0)
Adenoma	1/8 (12.5)	5/8 (62.5)	2/8 (25)	2/10 (20)	2/10 (20)	6/10 (60)	2/10 (20)	7/10 (70)	1/10 (10)	1/10 (10)	9/10 (90)	0/10 (0)
Adenocarcinoma	0/6 (0)	2/6 (33.3)	4/6 (66.7)	0/6 (0)	0/6 (0)	6/6 (100)	2/6 (33.3)	3/6 (50)	1/6 (16.7)	3/6 (50)	3/6 (50)	0/6 (0)

Staining graded as: −, negative; +, positive; ++, strongly positive. Percentages are shown in parentheses. –, Without colon lesions.

Exon 3 of the β-catenin gene was sequenced from a total 36 β-catenin-positive lesions (1 dysplastic lesion, 15 adenomas, and 20 adenocarcinomas), yet no mutations were observed in this mutation hotspot region (data not shown). In addition to the colon lesions with positive staining for p53, there were two mice with colon tumors showing regions of strong p53-positivity (one with a strong p53-positive adenocarcinoma and the other with a strong p53-positive adenoma). Within these two tumors, DNA was extracted from various regions showing strong p53 immunostaining (Fig.8a–e). In the adenocarcinoma, three separate regions of strong p53 staining were sampled (Fig.8f–h) each showing independent p53 mutations (Fig.8k–m). The mutations were all single-nucleotide substitutions: the first leading to a Q201K amino acid substitution (homologous to the human Q204K mutation observed in Hodgkin's disease;33 a GT>AT change at the splice donor site of intron 6, which has been shown at the equivalent site in a human case to create a six amino acid insertion;34 and a seemingly silent change at the start of exon 6. In the adenoma, three regions of strong p53 staining were sampled (Fig.8i,j) with two of them harboring single-nucleotide substitutions in the analyzed area of the DNA binding domain (Fig.8n,o). Interestingly, the first was also at the start of exon 6, but this time an AG>AA change at the splice acceptor site. As mutations at this acceptor site (and at the position preceding it) have been shown in human cancer to result in omission of exon 6, yielding a premature stop codon,35,36 it is likely that both this mutation and the otherwise silent change in the adenocarcinoma produce aberrantly spliced p53 protein. The second mutation in the adenoma is predicted to result in an R303K substitution (equivalent to R306K in human p53). Aberrant p53 DNA sequences were not detected in two moderately p53 stained areas adjacent to the strong p53 staining tumor regions, four p53-postive areas from other moderately staining lesions, nor normal non-staining tissue adjacent to each of the five lesions studied in total (data not shown).

Figure 8

Distinct strong positive p53 staining areas were observed in one adenocarcinoma (a–c) and one adenoma (d,e) induced by combined X-ray and dextran sodium sulfate (DSS) exposure in Mlh1−/− mice (group 12). Re-imaging after microdissection shows the isolated cells from the same field above (f–j). For each region, p53 sequencing results are shown corresponding to the fields above (k–o) in the lower trace, compared to the upper trace sequenced from adjacent normal tissue. Red arrows indicate mutated sites and red letters show amino acid substitutions.

Discussion

We have shown that combined exposure to radiation and mild inflammation could potentiate colon carcinogenesis under MMR-deficient conditions, although the trend was more pronounced in male mice. Dextran sodium sulfate (DSS)-induced cancer in WT mice typically requires long exposure periods or multiple cycles of DSS treatment, yet in Mlh1−/− mice, 1% DSS exposure for 1 week induces colon neoplasms with short latency (by 25 weeks old).15 DSS treatment is able to promote tumors in WT mice when combined with initiating chemical treatments,37–39 but did not promote tumors in irradiated WT mice, perhaps due to insufficient dose.40 In contrast, even a single 2 Gy dose combined with DSS induced colon tumors in MMR-deficient mice, consistent with other evidence supporting synergistic mutagenicity of radiation in DNA repair-deficient cells.23,41

A subset of human CRC with MMR pathway defects are characterized by high levels of genome-wide MSI.42 It is estimated that 15% of sporadic CRC exhibits MSI whereas the remaining cases are primarily attributed to an early loss of APC by chromosomal instability and/or mutation leading to adenoma by upregulation of β-catenin,43 with rare progression to carcinoma (usually marked by eventual inactivation of P53),44 or colitis-associated CRC, marked by early loss of P53 in mucosa subjected to prolonged inflammation,45 leading to increasing dysplasia before ultimate loss of APC and progression to carcinoma.46

Tumors with inherited (Lynch syndrome) and acquired inactivation (usually through promoter methylation) of the MMR pathway (sporadic CRC) are postulated to follow different tumorigenic pathways.47 In Lynch syndrome, CRCs are thought to arise by stochastic mutations in a constellation of genes as a result of disrupted DNA repair, with a number of genes with unstable short nucleotide repeats thought to drive transformation.46 Lynch syndrome CRCs typically progress from adenomas, which morphologically tend to show features associated with increased risk of progression48–50 and, unlike sporadic CRC, show a very high probability of transition to carcinoma.44

Lynch syndrome carcinomas are characterized by the presence of TIL and Crohn's-like lymphocytic reactions.51–54 They fall into three main groups, the most common of which is moderate to well-differentiated, well-circumscribed mucinous carcinoma often associated with adjacent signet ring cells, mucinous cells with extracellular mucin, and tubulovillous or villous adenomatous remnants.55,56 The second type is poorly differentiated adenocarcinoma with a medullary growth pattern that is well circumscribed and lacks abundant desmoplastic stroma.56 The third type is well to moderately differentiated adenocarcinoma with occasional serrated polyps.55,57

The principal pathological characteristics of colon lesions induced here by combined exposure to X-rays and DSS in Mlh1−/− mice were dysplasia, adenoma with low- and high-grade dysplasia, and well-differentiated adenocarcinoma with TIL, akin to the well-differentiated adenocarcinoma with TIL seen in human Lynch syndrome. Aberrant APC, β-catenin, and P53 were identified mainly in the more advanced lesions, suggesting that these were late events, distinct from the early causal roles of APC or P53 inactivation in the two canonical pathways in human sporadic CRC or colitis-associated cancer. Histologically, the tumors induced by the combination of X-rays and DSS were similar to those in mice treated with DSS alone, although they tended to show higher expression of β-catenin and p53. The emergence of multifocal areas of inactivated p53 due to independent mutation events within tumors induced by the combined treatment may be a sign of increased tumor progression. The lack of β-catenin mutations might suggest that the increased levels of β-catenin staining are secondary to mutations in other genes, such as Apc. It is possible that differences in these and other biological markers in tumors that appear histologically similar are indications of the mutational history of the tumor and the pathway by which they arose, rather than defining a difference in tumor type.

The higher tumor incidence and multiplicity in male Mlh1−/− mice parallels the increased CRC risk in men with Lynch syndrome58 with both environmental and genetic causes proposed.59 Sensitivity to DSS-induced inflammation and radiation-induced intestinal cell apoptosis have also been shown to differ by sex,60–62 although the mechanisms are not well understood.

In conclusion, our results indicate a potential for exposure to ionizing radiation to further increase the progression of preneoplastic lesions in Mlh1−/− mice with inflammatory colitis. Given the coalition of radiogenic mutation, MMR-deficiency, and inflammatory processes, the tumors arising following such combined exposures are likely to exhibit a complex series of traits rather than recapitulating any single human CRC subtype. Comprehensive genetic analysis in the future may be able reveal more information about the specific interactions between the various exposures. In the meantime, these results are in agreement with the principle that signs of inflammation should be closely monitored in MMR-deficient patients, such as those with Lynch syndrome, before exposure to radiation for therapy, diagnostic, or screening purposes.