A novel murine model of inflammatory bowel disease and inflammation-associated colon cancer with ulcerative colitis-like features

Abstract

Mutations that increase susceptibility to inflammatory bowel disease (IBD) have been identified in a number of genes in both humans and mice, but the factors that govern how these mutations contribute to IBD pathogenesis and result in phenotypic presentation as ulcerative colitis (UC) or Crohn disease (CD) are not well understood. In this study, mice deficient in both TNF and IL-10 (T/I mice) were found to spontaneously develop severe colitis soon after weaning, without the need for exogenous triggers. Colitis in T/I mice had clinical and histologic features similar to human UC, including a markedly increased risk of developing inflammation-associated colon cancer. Importantly, development of spontaneous colitis in these mice was prevented by antibiotic treatment. Consistent with the known role of Th17-driven inflammation in response to bacteria, T/I mice had elevated serumTh17-type cytokines when they developed spontaneous colitis and after systemic bacterial challenge via NSAID-induced degradation of the mucosal barrier. Although TNF production has been widely considered to be be pathogenic in IBD, these data indicate that the ability to produce normal levels of TNF actually protects against the spontaneous development of colitis in response to intestinal colonization by bacteria. The T/I mouse model will be useful for developing new rationally-based therapies to prevent and/or treat IBD and inflammation-associated colon cancer and may further provide important insights into the pathogenesis of UC in humans.

Figure 1. Body weights of T/I mice begin to lag soon after weaning.

Mean ± SEM body weights are shown for wild type and T/I male mice between 2 and 9 wks of age. The number of litters and mice represented are 8 litters and 2–24 mice/point for wild type (WT) and 14 litters and 3–12 mice/point for T/I mice. Similar data was also obtained for the following strains that did not differ from wild type values at any of the 8 time points tested: Il10 ^−/− (IL10 KO), 5 litters, 3–21 mice/point; Tnf ^−/− (TNF KO), 5 litters, 7–13 mice/point; T-het/I, 14 litters, 2–6 mice/point. The SEMs for WT mice ranged from 0.2 to 0.6 g and thus are hidden by the markers chosen. *indicates a significant difference between WT and T/I mice; p values ranged from 0.0001 to 0.03. Similar trends were seen in female mice (not shown).

Figure 2. Spontaneous colitis in T/I and T-het/I mice.

The colon histologic score is shown as a function of age for T/I (panel A) and T-het/I (panel B) mice. Each dot represents a single mouse studied. Dashed lines divide mice into groups with no colitis (histologic score ≤12), mild colitis (histologic scores between 13 and 24), and moderate to severe colitis (histologic score ≥25). The percentage of mice in each of these severity categories is shown at the right of each figure. Significantly fewer T-het/I mice had moderate to severe colitis compared with T/I mice (p = 0.0004; relative risk 0.39, 95% confidence interval 0.22–0.68).

Figure 3. Kaplan-Meier survival curves for T/I and T-het/I mice.

The probability of survival as a function of time is shown for T-het/I vs. T/I mice. Censored data for mice euthanized prior to achieving humane endpoints is marked with a small vertical line. The number of mice at risk is provided beneath the graph. T/I mice were significantly more likely to die or to require euthanasia for humane reasons than T-het/I mice (p = 0.0032; log rank test).

Figure 4. Colitis and inflammation-associated colon cancer in T/I and T-het/I mice.

A. Spontaneous colitis in T/I mice (rectum shown) is characterized by epithelial hyperplasia and infiltration of the lamina propria with inflammatory cells, including large numbers of neutrophils. Crypt abcesses are common (arrow). Inflammation typically involves the mucosa only, with little to no inflammation present in the submucosa (Sub M). B. Piroxicam-triggered colitis (shown) and spontaneous colitis (not shown) in T-het/I mice has a similar histologic apprearance to colitis in T/I mice (rectum shown; arrow indicates crypt abcess). C. In contrast, colitis (rectum shown) in Il10 ^−/− mice is typically transmural as indicated by arrowheads, with inflammation present in the submucosa and/or muscularis propria (MP). D, E. Colon cancers that develop in T/I mice (and T-het/I mice, not shown) are typically mucinous adenocarcinomas that may show extensive local invasion. The tumor shown in D is from proximal colon; the tumor in E is from rectum. M indicates mucin lakes. Ink on tumor (arrowheads) indicates invasion completely through the bowel wall (the arrows in E point to the serosal surface). F. T/I (and T-het/I, not shown) mice with colitis often exhibited squamous metaplasia in the rectum, that could give rise to carcinomas with both glandular and sqamous features (arrows). G. Higher magnification view of mucinous adenocarcinoma in a T/I mouse. H, I. T/I mice given placebo food spontaneously developed severe colitis by 8 wks of age (I), whereas T/I mice given matched food containing antibiotics showed no evidence of inflammation at this time point (H). Arrow indicates a crypt abcess, shown at higher magnification in J. Scale bar  = 100 µm in A - C, F - J and 1 mm in D and E.

Figure 5. Distribution of colonic inflammation in Il10 ^−/− (IL10 KO) vs. T/I mice.

The % of colonic area that exhibited inflammatory changes in each segment of the colon (an assessment required for calculation of histologic scores [12]) is shown for a cohort of Il10 ^−/− mice with colitis triggered by exposure to piroxicam (n = 14) and an age-matched cohort of 11–12 wk T/I mice with spontaneous colitis (n = 14). The dashed line indicates the median area involved in each colon segment. The p values shown represent comparison of disease extent (involving 0–60% vs. >60% of the segment area) in the 2 genotypes using Fisher’s exact test.