In this study we have shown that a protocol of 3 intra-rectal TNBS injections successfully causes colitis in pre-pubertal rats, characterized by inflammation and fibrosis on histology, and MR-colonography, although the rats did not exhibit growth delay.
The TNBS chronic colitis model in mice developed by Lawrance et al. showed interesting properties mimicking natural history of IBD in humans, with a progressive reduction in the proportion of mice having severe inflammation between the 2nd and 6th dose of TNBS, and a progressive increase of the incidence of severe fibrosis (50% of mice at the 6th dose) (10). Lawrance et al. also showed the dynamic nature of fibrosis, persisting 2 to 4 weeks after the last dose of TNBS.
There have been very few published studies on chronic TNBS colitis in pre-pubertal rats (11, 12). Hence our first objective was to adapt the TNBS chronic colitis model to young rats, with low mortality and significant fibrosis before reaching adult age. The window of opportunity was narrow, since in Sprague-Dawley rats the peri-natal period lasts until Postnatal Day 21 (PND 21), the pre-pubertal period until PND 42 (6 weeks) and the pubertal period until PND 63 (9 weeks) (13). We chose to induce colitis starting at 5 weeks of age and all rats were analyzed at 8 weeks of age, before the end of the pubertal period.
Among the signs of colitis expected in a TNBS model, weight loss and linear growth failure are classical features (11, 12). In our model there was no significant growth failure based on weight and tail length in the TNBS group, although significant histological inflammation indicates that the dose of TNBS was sufficient. A possible explanation results from the fact that the rats were fed a high energy high protein (25.2% crude protein) rat breeding diet during the whole study instead of transitioning to normal diet as they increased in age, in order to allow between group comparisons with the MODULEN IBD® fed rats. In comparison, rats in other chronic TNBS models are fed a standard rat chow with normal protein intake (generally 17–19% crude protein). In this case, the rat breeding diet might have had the effect of nutritional supplementation that could have counterbalanced the detrimental nutritional effect of colitis. Ballinger et al. showed, in a model of acute TNBS colitis in pre-pubertal Wistar rats with pair fed controls that colitis caused impaired linear growth caused in part by inflammation and in part by undernutrition (12). Nutritional supplementation in the colitis group increased weight almost to the value of controls (12).
The other possible explanation for the absence of growth failure in the TNBS group could be that the rats with the most severe colitis died during the protocol and that those that survived had a better nutritional status. Indeed there was a significant mortality rate of 5/10 in the TNBS group vs. 2/10 in the control group. We adapted and lowered the doses of TNBS compared to those previously reported, but maybe insufficiently for these young animals. In the absence of comparable studies in this age group it might be necessary in future studies to lower the doses 2 and 3 of the protocol.
Choosing the correct dose of TNBS is a challenge when putting into balance colonic inflammation, fibrosis and morbidity. Among the animal models of intestinal fibrosis (14), the chronic TNBS colitis requires repeated doses of TNBS to induce fibrosis. The effect of TNBS also has a high variability between species and between animals in an experimental group (10, 15). Previous studies have shown that TNBS models based on 2 to 3 doses of TNBS cause severe but transient colitis. Fibrosis, as confirmed by elevated expression of collagen and components of the extra-cellular matrix, becomes particularly evident after 4 doses (11), 6 doses (6) and even 8(10) repeated administrations of TNBS.
Our study is the first to assess pre-pubertal colitis in rats using MR-colonography. The MR-colonography characteristics of our model were similar to our previously reported adult rat chronic TNBS colitis model, with a significant increase of colon wall thickness and stenosis with no spontaneous enhancement of T1w wall signal (6). There is persisting need for imaging techniques that could quantify the amount of fibrotic tissue in strictures and assist the decision of medical or surgical management. Diffusion MR techniques are promising, but could not be applied to our model due to the small size of the animal colons and the “noise” due to the diffusion technique. Other MRI techniques applied in mice or rodents have ambitioned to better identify the fibrotic nature of a stricture, such as dynamic contrast enhanced MRI studies (with injection of Gd-DTPA) (16) or non-contrast MRI magnetization transfer and T2-weighted signal intensity ratio (comparing bowel wall intensity at the area of greatest thickening, to paraspinous muscle signal intensity) (17).
Moving on to the second objective of our study, feeding rats exclusively with MODULEN IBD® during study protocol did not have an anti-inflammatory effect based on histological scores and MRI images.
Compared to the TNBS group, rats in the MODULEN group had significant growth failure starting within 6 days after first dose of TNBS, without subsequent catch-up growth. This growth failure could be explained by the colitis in itself, although nutritional factors could have amplified the weight loss. Indeed, although the rat breeding diet fed to the TNBS group has a high protein content, MODULEN IBD® is on the contrary low in protein and high in fat, with 42% energy from fat, 44% from carbohydrates and 14% from protein. The growth curve of the MODULEN IBD® group is very similar to that of a group of rats fed a protein-restricted diet (12% crude protein), in a model studying compensatory growth (18).
Protein restriction alone does not seem to explain why the severity of colitis in the MODULEN group was not less than in TNBS group. The majority of rats from the MODULEN group had moderate to severe inflammation and coexisting fibrosis on histology. MRI showed signs of intense inflammation with increased thickness and signs such as ulcers and a mucosal flap that were not seen in the TNBS group, but did not show significant stenosis. This would tend to suggest a pro-inflammatory effect in the MODULEN group.
Once again, nutritional aspects of the diet could explain in part these observations. Xue Li et al. showed that a high fat diet (60% energy from fat compared to 10% in controls) given for 4 weeks before induction of TNBS colitis in C57BL/6 mice increased severity of weight loss and histological damage, and increased pro inflammatory cytokines (19). They showed that high fat diet caused an increase in oxidative stress, tight junction dysfunction and increased mucosal permeability.
We chose to use MODULEN IBD® in its original form and not as a supplementation to the rodents’ standard diet as other teams have done (20, 21) in order to best reproduce the effect of an exclusive enteral nutrition by MODULEN IBD® as recommended in pediatric IBD induction therapy. We also chose not to administer the MODULEN IBD® by oral gavage since the model was to be the least invasive possible. We had observed in a previous study of MODULEN treatment in acute TNBS colitis that the rats ate the powder and did not develop growth failure (tail length) over a period of 10 days.
Admitting the alternative hypothesis that difference in macronutrient composition is not significant in explaining the increased severity of colitis in the MODULEN group, another explanation would be that a component of the diet had a pro-inflammatory effect. A characteristic of MODULEN IBD® is its enrichment in TGF-beta 2 (> 24 ppm) (22). TGF- β has several isoforms. TGF-β1 is the prominent cytokine involved in fibrotic procedures in multiple organs, through a signaling cascade leading to differentiation of fibroblasts to a myofibroblast phenotype which in turn causes collagen deposition in the extracellular matrix. TGF-β2 is classically associated with an anti-inflammatory effect. However there have been recent reports of increased TGF-β2 levels in colon strictures in a rat model of 2,4,6-trinitrobenzene sulfonic acid (TNBS) colitis (23). In our study, not only were there fewer strictures in the MODULEN group, but also a decrease in all fibrosis markers compared to controls, although only alpha-SMA decrease was significant.
As in other animal models of intestinal fibrosis, we used histology as the gold standard for confirming fibrosis and therefore validity of the chronic model (24, 25). One reason we did not observe an increase in “classical” fibrosis markers such as CTGF, alpha-SMA and COL1A1 could be the location of the sample. Indeed, in patients with stricturing ileal Crohn’s disease there has been shown an increase in alpha-SMA and COL I – III expression only in stenotic tract, and not in non-involved ileum (26).
However, considerations such as timing of analyses are also paramount, due to the dynamic nature of the fibrotic process. Briefly, fibrosis in the intestine is based on increased extracellular matrix production by activated myofibroblasts, and decreased degradation of these extracellular matrix components (4, 27, 28). Myofibroblasts can originate from fibroblasts and from multiple other cellular sources, via epithelial to mesenchymal transition (EMT- and endothelial to mesenchymal transition (EndoMT). Many growth factors such as TGF-beta and CTGF are implicated in activation of mesenchymal cells, either alone or by cooperative interaction (29). Upon activation, myofibroblasts express alpha-SMA, which is commonly used in studies as a marker of activated myofibroblast presence. Once activated, myofibrobasts produce extracellular matrix components according to a certain temporality. For example, Zhu et al., in a chronic (5-doses of TNBS) colitis model in adult Sprague-Dawley rats showed increased expression of COL1A and COL3 mRNA as soon as the 2nd week of colitis, and collagen deposition starting at 3rd week (30). During the final resolution of fibrosis phase, myofibroblasts are reduced by apoptosis (31).
When applying these theoretical considerations to our model, we can hypothesize that the timing at which the colons were harvested coincided with the beginning of resolution of colon fibrosis. Indeed, histology showed submucosal fibrosis, but the markers of myofibroblast activation (CTGF and COL1A1) were not increased compared to controls suggesting the myofibroblasts were no longer activated. On the contrary, the decrease of alpha-SMA in the MODULEN group could be interpreted as a sign of myofibroblast apoptosis and possibly, as an anti-fibrotic effect of MODULEN IBD®. This would need to be confirmed by serial analyses, with a longer follow-up after last dose of TNBS in order to confirm accelerated resolution of fibrosis in the MODULEN group compared to the TNBS group.