Establishing the colitis-associated cancer progression mouse models

Abstract

Inflammatory bowel disease (IBD) has been reported as an important inducer of colorectal cancer (CRC). The most malignant IBD-associated CRC type has been highlighted as colitis-associated cancer (CAC). However, lack of CAC cases and difficulties of the long follow-up research have challenged researchers in molecular mechanism probing. Here, we established pre-CAC mouse models (dextran sulfate sodium [DSS] group and azoxymethane [AOM] group) and CAC mouse model (DSS/AOM group) to mimic human CAC development through singly or combinational treatment with DSS and AOM followed by disease activity index analysis. We found that these CAC mice showed much more severe disease phenotype, including serious diarrhea, body weight loss, rectal prolapse and bleeding, bloody stool, tumor burden, and bad survival. By detecting expression patterns of several therapeutic targets—Apc, p53, Kras, and TNF-α—in these mouse models through western blot, histology analysis, qRT-PCR, and ELISA methods, we found that the oncogene Kras expression remained unchanged, while the tumor suppressors–Apc and p53 expression were both significantly downregulated with malignancy progression from pre-CAC to CAC, and TNF-α level was elevated the most in CAC mice blood which is of potential clinical use. These data indicated the successful establishment of CAC development mouse models, which mimics human CAC well both in disease phenotype and molecular level, and highlighted the promoting role of inflammation in CAC progression. This useful tool will facilitate the further study in CAC molecular mechanism.

More and more supporting evidence from genetic, pharmacological, and epidemiological studies have figured out that inflammation plays important regulatory roles both in the initiation and the progression stages of cancer development.¹ Among inflammation-associated colorectal cancer (CRC), the most well-known type has been highlighted as colitis-associated cancer (CAC).² Previous studies have indicated that more than 50% of the inflammation-CRC patients will die from CAC.³ Such a high mortality of CAC argues the need to probe into the molecular mechanism.⁴ However, lack of CAC cases and difficulties of follow-up research are the biggest obstacles for molecular mechanism probing.

To address this question, we established the pre-CAC (dextran sulfate sodium [DSS] group and azoxymethane [AOM] group) and CAC mouse model (DSS/AOM group) to mimic human CAC development through the DSS/AOM method, followed by analyzing the disease activity index as previously reported.⁵ We found that DSS treatment resulted in serious diarrhea (Table 1), diarrhea-relevant fluctuation and loss of body weight (Figure 1a), serious rectal prolapse and bleeding (Figure 1b), and lymphocytes infiltration in the colon/rectal of mice (Figure 1c, arrow) in the DSS group and DSS/AOM group compared with the AOM group and control group; while the mice of the DSS/AOM group showed more severe disease phenotypes, including more body weight loss (Figure 1a) and much earlier (the first cycle) and more possibilities of getting serious rectal prolapse and bleeding (5/10) compared with the DSS group (2/10) (Table 1; Figure 1b). These findings and how DSS treatment featured in accumulating effect, recovering time of mice from diarrhea, and prolonged time of diarrhea (Table 1) were consistent with previous studies that DSS can be administered in drinking water to mice in multiple cycles to create a chronic inflammatory state by exerting its toxic effects on the colonic epithelium directly.⁶ We also found that the lymphocyte infiltration only accelerated the CAC tumor development in the DSS/AOM group mice (10/10) (Table 1; Figure 1c, star), which were mainly distributed in the distal colon/rectum (Figure 1d and e), while mice of the AOM group and DSS group only showed inflammatory swelling (Figure 1f, star). These data indicated that the DSS-induced inflammation strongly advanced the AOM-induced carcinogenesis in the DSS/AOM group mice because the dosage and prolonged time of DSS or AOM treatment we performed here were insufficient to induce obvious CAC tumors alone (Figure 1f). It should be noted that these CAC tumors displayed the neoplasm extruded to the enteric cavity both in the early and late stages of tumor progression (Figure 1g, left), and were featured with highly non-typical or primary tumors phenotype but without invading the submucosa (Figure 1g, right). Furthermore, mice death was observed in the DSS group (2/10), and tumor burden exacerbated the survival of the DSS/AOM group mice as expected (much earlier, 5/10) (Table 1; Figure 1h). These consequences indicated the successful establishment of the pre-CAC and CAC mouse models and highlighted the importance of inflammation in CAC development.

Table 1.

The disease activity index of the mice.

		Control group			AOM group			DSS group			AOM/DSS group
		Cycle 1	Cycle 2	Cycle 3	Cycle 1	Cycle 2	Cycle 3	Cycle 1	Cycle 2	Cycle 3	Cycle 1	Cycle 2	Cycle 3
Diarrhea	First emerged at day X/cycle	/	/	/	/	/	/	3–4	4–5	4–7	3–4	3–5	4–6
	Duration time/cycle	/	/	/	/	/	/	11–15	10–13	10–12	11–16	11–13	10–13
	Mice (n)	0/10	0/10	0/10	0/10	0/10	0/10	10/10	10/10	9/9	10/10	9/9	8/8
Prolapse	Mice (n)	0/10	0/10	0/10	0/10	0/10	0/10	0/10	1/10	1/9	1/10	2/9	3/8
Death	First emerged at day X/cycle	0	0	0	0	0	0	/	7	6	10	7	6–8
	Mice (n)	0/10	0/10	0/10	0/10	0/10	0/10	0/10	1/10	1/9	1/10	1/9	3/8

Figure 1.

DSS/AOM combination treatment efficiently induce CAC in mice. (a) Measurement of mice body weight every day each week during the DSS and/or AOM administration showed the significant body weight loss of the DSS/AOM group mice. All data are mean ± SD. Statistics: Student’s t-test, *P <0.05. (b) Brightfield imaging showed the diarrhea of the DSS group mice (arrow) and the serious bloody stool (arrow) of the DSS/AOM group mice. (c) Hematoxylin and eosin (H&E) staining indicated that the DSS group with normal colon/rectum (star) and DSS/AOM group with colon/rectum tumor tissue (star) presented the lymphocytes infiltration (arrow) (scale bar, 200 μm). (d) Brightfield imaging of the tumors (arrow) burdened colon (left) of the mice in the DSS/AOM group and the inset shows an amplified view of distal colon with numerous tumors of varying sizes (right). (e) Calculation of the tumors distribution in the DSS/AOM group mice indicated that the majority of tumors are located in the distal colon and are <2 mm in size. (f) Brightfield imaging of the colon/rectum of mice indicated of the inflammatory swelling in the colon/rectum of the DSS group and AOM group. (g) H&E staining of the normal colon/rectum tissues of the control group mice (left) and the early and late stage tumors of the DSS/AOM group mice (right) (scale bars, 150 μm); all with enlarged insects (scale bars, 50 μm) which displayed the neoplasm extruded to the enteric cavity and non-typical and non-invading features. (h) Kaplan–Meier analysis of the four mice groups showed the worse survival of the DSS/AOM group. The P value is subscribed into log-rank test.

Then we compared CAC progression here with human CAC by checking the expression of several tumor-driving genes, such as Kras, Apc, p53, and TNF-α. Previous studies have indicated that these therapeutic targets genes play decisive roles in CRC development and their expression are intricately regulated by gene mutation, depletion, and signaling pathways.⁷ Through analysis of the colon/rectum of the four mice groups, we found that Kras only showed a humble change in mRNA levels (qRT-PCR; Figure 2a and c) and protein expression (western blot and immunohistochemistry [IHC]; Figure 2b, d, and e) between the four groups. This reminded us of the stabilized tumor-promoting Kras-GTP mutation rather than expression changes in CRC.⁸ However, Apc did not respond to DSS but was downregulated by AOM in the AOM group and even more in the DSS/AOM group; and p53 was downregulated in response to the AOM or DSS treatment and were reduced even further in the DSS/AOM group, which indicated the tumor suppressor function of Apc and p53. These consequences are in accordance with the previous findings that both DSS and AOM can cause genes mutations and genes expression alteration as presented here and reported previously.^6,9,10 We also found that the TNF-α level in blood was accelerated with malignancy which displayed a higher expression level in CAC mice and a lower expression in pre-CAC mice while it remained very low in control normal mice through ELISA analysis (Figure 2f). Furthermore, the high level of blood TNF-α showed worse survival possibly because of the CAC progression and tumor burden (Figure 2g). These results indicated the positive correlation of TNF-α and CAC development which was in accordance with the relevant findings in CAC patients and highlighted the clinical use of TNF-α in predicting CAC progression.¹¹ Taken together, these results enhanced the impression that the DSS/AOM mouse model potently monitors the molecular changes in human CAC and that inflammation can lead to gene mutation and gene expression alterations (Kras, Apc, p53, and TNF-α) to confer the transformation of cancer cells.^2,12

Figure 2.

Apc, p53, Kras, and TNF-α are differently expressed with CAC development in mice. (a) qRT-PCR analysis and (b) western blot analysis of Apc, p53, and Kras in the colon/rectum of the mice in the control group and AOM group. (c) qRT-PCR analysis and (d) western blot analysis of Apc, p53, and Kras in the colon/rectum of the mice in the DSS group and DSS/AOM group. (e) Paraffin-embedded sections of the colon/rectum from mice were performed IHC analysis of Apc, p53, and Kras. Scale bar, 50 μm. (f) ELISA analysis of the TNF-α protein level in mice blood at week 9 showed the positive correlation with malignancy. (g) Kaplan–Meier analysis of survival of the TNF-α high group (TNF-α ⩾27 ng/L) and the TNF-α low group (TNF-α <27 ng/L) mice. The P value is subscribed into log-rank test. The mean ± SD of three separate experiments was plotted. Statistics: Student’s t-test; *P <0.05, **P <0.01 (a–f).

In summary, we successfully established pre-CAC and CAC mouse models through the DSS/AOM method. These CAC mice showed much more severe disease phenotypes (such as body weight loss, bloody stool, inflammation infiltration, tumor burden, and worse survival) compared with pre-CAC mice. Several tumor-driving genes, including Kras, Apc, p53, and TNF-α, were analyzed and their different expression patterns with malignancy (Kras expression remained unchanged, Apc and p53 expression were downregulated, and TNF-α protein was elevated with CAC progression) conferred the promoting roles of inflammation in CAC progression and was in line with the human CAC development.¹² These CAC development mouse models can mimic human CAC progression well and provide a useful tool for further research.