The Anti-Inflammatory Effects of Enteral NutritionCentre for Adult & Pediatric Gastroenterology, Institute of Cell and Molecular Science, University of London, United Kingdom Correspondence: Prof. I. R. Sanderson, Adult and Paediatric Gastroenterology, Institute of Cell and Molecular Science, Barts and The London, Turner Street, London. E1 2AD, UK. Electronic mail may be sent to i.r.sanderson{at}qmul.ac.uk. Enteral nutrition has a positive effect on growth in children with active Crohn's disease. The question arises: is this is due only to improved nutrition? Enteral formulas may also directly reduce inflammation, lowering the expression of cytokines like interleukin (IL)-6 that inhibit growth. Four lines of evidence support the hypothesis that enteral formulas directly lessen inflammation: enteral nutrition directly affects the inflamed intestine; changes in inflammatory markers precede repletion of nutrition status; molecular pathways exist linking changes in luminal contents to the expression of class II MHC genes in intestinal epithelium in animal studies; and enteral formulas have a direct effect on cytokine expression by intestinal epithelial cells. Impaired linear growth is common in children with active small bowel Crohn's disease. Growth failure is present in 20%–50% of children at presentation.1 However, it is not clear whether the growth failure is due to a result of an impaired nutrition status or whether it is due to a direct effect of inflammation on growth.2-4
The proinflammatory cytokines interleukin (IL)-6 tumor necrosis factor
(TNF)- Enteral nutrition has positive effects on growth.9 Therefore, the question arises whether the changes seen are a consequence of improved nutrition alone, or whether a direct effect of enteral feeding on inhibiting inflammation is also a significant factor. Our unit has pursued the hypothesis that enteral feeding directly affects inflammation and that this results in improvement of growth.11 This review will examine the possibility that enteral feedings directly alter inflammatory processes. Four lines of evidence will be used to support this hypothesis: enteral feedings directly affect the inflamed intestine, changes in inflammatory markers precede repletion of nutrition status, molecular pathways exist linking changes in luminal contents to the expression of class II major histocompatibility complex (MHC) genes in intestinal epithelium in animal studies, and finally that enteral feedings have a direct effect on cytokine expression by intestinal epithelial cells.
When enteral feedings were initially used in patients with Crohn's disease, they were shown to improve disease activity scores.12,13 However, nutrition status is a component of such scores and the possibility has been raised that the improvements in patients was due only to improved nutrition. Later studies showed that enteral nutrition does directly improve the erythrocyte sedimentation rate (Fig. 1) and the C-reactive protein.14,15 Examining the permeability to nonabsorbable sugar markers also demonstrated a direct effect on the intestine.16 Children admitted for colonoscopy that did not have inflammatory bowel disease excluded lactulose to the same extent as previously tested normal children and adults (Fig. 2). However, children with Crohn's disease had markedly increased permeability to lactulose when compared with absorption of rhamnose (a readily absorbable sugar). Six weeks of enteral feeding caused a dramatic improvement in lactulose/rhamnose permeability ratios in all children tested. The ratios returned to normal in about 50% of the children and improved in the rest. Children whose permeability markers did not return to normal were those with more severe intestinal disease at presentation. More recently, colonoscopy showed that enteral feeding can return the intestine macroscopically to normal.17,18 Furthermore, enteral feeding improves cytokine transcripts in inflamed intestine and return the expression of matrix metalloproteinases to normal.19 Thus, these early studies showed that enteral feeding was a direct treatment for intestinal inflammation and not a superficial restitution of nutrition status. This contrasts markedly with corticosteroids which do not resolve intestinal inflammation.20
Earlier studies clearly showed that inflammation improves markedly over a 6-week period of enteral nutrition.16 However, nutrition restitution is also improved by this time. Measuring early time points after initiating enteral nutrition would detect whether inflammatory changes occurred before nutrition repletion. Studies were therefore designed to compare the finding of nutrition repletion with anti-inflammatory effects.11 If the return to nutrition restitution was important in the improvement of inflammatory processes in the intestine, one would expect changes in inflammatory markers to be delayed until there was an increase in nutrition measures, such as the body mass index. This is not to say that enteral feedings may not contain nutrients that act in a biochemical manner on signal transduction pathways. Such nutrient activities would not constitute nutrition restitution, however. An example of this is the effect of butyrate on gene expression.21 Although butyrate is a known metabolic fuel for colonocytes, its mechanisms of altering gene expression depend on its ability to inhibit histone deacetylases. The mechanisms of this are identical to other agents that inhibit histone deacetylases but deliver no energy to the cell (ie, trichostatin A).
Twelve children who presented consecutively with Crohn's disease were treated with enteral nutrition (AL-110; Nestlé, Vevey, Switzerland). Markers of inflammation and nutrition status were measured from day 3 onwards.11 Markers of inflammation changed rapidly and were virtually normal by day 3 (Figs. 3 and 4), whereas measurements of a nutritional status did not alter significantly until 14 days after treatment commenced. This study therefore showed that changes in inflammation did not occur as a consequence of improved nutrition status (Figs. 5 and 6).
Animal studies enable scientists to examine whether changing the diet has direct consequences on gene expression in the intestine. It is also possible to examine molecular events in the epithelium induced by changes in diet. A useful model is the expression of class II MHC in the epithelium after weaning of mice.22 Class II MHC is responsible for presentation of antigen and its expression on the intestinal epithelium of the mouse occurs after weaning. It is therefore possible to wean mice on to a normal diet (mouse chow) or enteral nutrition to examine the difference between these types of nutrition. The epithelial cells can then be isolated from the mice at varying time points after weaning to study the expression of class II MHC and invariant chain, which is coexpressed. The effects of the enteral feeding (Vivonex; Novartis, Nyon, Switzerland) was dramatically different from normal mouse chow (Fig. 7). Normal mice chow induced an expression of these genes between 20 and 30 days of age whereas during this time period, enteral feeding did not.22 It is known that the class II transactivator is, for cell types so far examined, both necessary and sufficient for class II MHC expression. Experiments were therefore designed to examine whether the diet acted through the class II transactivator (Fig. 8). In the mouse, there are 3 isoforms of class II transactivator factor (CIITA). Interestingly, a normal complex diet increased the expression of class II MHC though CIITA IV (Fig. 9).23 In addition to the dietary regulation of class II MHC, there was a slow time-dependent regulation, which was found to be due to CIITA III. These experiments therefore show that alterations in the diet have recognizable molecular pathways between the intestinal lumen and the signal transduction machinery of the epithelial cell. These effects are difficult to explain based on changes on body mass, and are much more likely direct effects of changes of luminal environment.
Although animal experiments show the linkage between metabolic pathways and enteral feeds, there are of course many cell types in the living intestine and such experiments do not rule out the mediating effect of other cells such as activated T cells. Experiments were therefore designed to study cell lines and their response to pro-inflammatory cytokines in the presence of different enteral feeding formulations. The epithelial cell line Caco 2 produces both IL-6 and IL-8 after stimulation with the pro-inflammatory cytokine IL-1β. A study was therefore designed to examine the effect of different enteral formulas on this stimulation. Enteral formulas that are used in children were diluted with culture media and placed in the culture at the same time as IL-1β. The effects of the IL-1β on IL-6 and IL-8 expression were then measured. Enteral nutrition has a marked effect on inhibiting the production of these 2 cytokines after IL-1β, and further experiments are underway to confirm this observation and to examine the molecular pathways responsible for them.
This review presents evidence from 4 different sources that the effect of enteral nutrition on inflammation is a direct process in the intestine, and does not depend on altering the status of fat and muscle in other parts of the body. It is therefore likely that the improvement in growth, a characteristic feature of enteral nutrition that makes it the preferred treatment over corticosteroids, is in part due to its effect on inflammation. Thus, enteral nutrition acts in 2 ways on improving growth. Firstly, it provides extra nutrients necessary for growth. Secondly, it reduces inflammation through cytokine-mediated pathways. Received for publication December 6, 2004. Accepted for publication February 1, 2005.
DiscussionDr MacDonald: Is there any evidence in infants, the effect in the diet, or weaning, or anything changes things like class II in the humans?Dr Sanderson: I don't think this question has been examined in humans. The MHC class II expression in humans may be different from the mouse epithelial cell. The intestine at birth in the human is more advanced than in a mouse. I would suspect, therefore, if we were to look that the CIITA type III in the human, it would already be active before birth and the MHC class II may already be expressed at birth, but I don't know that, that is just a guess. MHC class II is reported to be absent in the epithelium human fetal intestine, but is expressed in the small intestinal epithelium in late infancy.1 Dr Arnaud-Battandier: After the discussion of this morning concerning the length of nutrition treatment, with your results do you think there is no need to push for 8 weeks of treatment? We can stop at 4 weeks or 3 weeks? Dr Sanderson: You are asking about how long to keep the children on enteral feeds. No study has been done as was mentioned this morning, randomizing children onto different lengths of treatment. It may be a good idea to ask the question directly. If you do a 4-week study vs an 8-week study, would there be differences in inducing remission and would there be differences on the length of remission? Such a study would be expensive because it would need large numbers of children to be involved, but it would be an interesting study. However, I would like to say it is not as simple as just doing the study comparing one group to another. One would have to stratify the effects onto different types of Crohn's disease because our anecdotal experience is that it takes longer for the changes in the large intestine to become normalized than the changes in the small intestine. Children with isolated colonic disease sometimes do go into remission with enteral feeds, but not very rapidly in our experience. In fact, there were no children with isolated colonic diseases in this cohort of children I described.2 I do think that the small intestine needs only a short time. I wouldn't like to speculate whether the macroscopic changes in the large intestine would resolve at the same time as all the changes in the inflammatory markers. It is possible that the inflammation might die away, the activation of T lymphocytes which produce IL-6, may disappear very rapidly, but the healing process of the intestine may take much longer. I can see a situation where the inflammatory markers and the cells responsible for inducing them could rapidly improve, but it would take a longer time for the mucosa to return completely to normal through the mechanism that Tom described this morning. Dr Donnet: It was just in reference to Tom's question and expression in infant intestinal epithelial MHC class II expression. Per Brandtzaeg did a study in the end of the 1980s and he said that in the first week after birth the intestinal epithelial cell was negative. Dr Beattie: Can I make a comment about the speed of the response? If the inflammatory markers are down to normal by day 3 and weight doesn't increase until day 14, that is not necessarily an argument against the nutrition status being a major factor because weight is such a crude marker of nutrition status. It would certainly be presumably possible that by day 3 you could have improved nutrition at the cellular level which was impacting on the inflammatory process. I think that would be a tempting hypothesis to throw in. I would be rather anxious that purely on the basis of the weight not going up until day 14, but the inflammatory markers returning to normal by day 3, that doesn't mean there hasn't been a nutritional effect by day 3. Dr Sanderson: Thanks, Mark. We were careful in choosing our nomenclature to say repletion of nutrition status rather than nutrition itself. I have alluded to experiments where short chain fatty acids could affect the epithelial cell signaling.3 This would be the molecular effect of the fatty acids on the signaling process. Now you could argue that that is a nutritional effect; but it is a semantic argument as to whether you think a molecule that has a nutrient capability when acting on a molecular pathway is an example of a nutritional effect. We can see the semantic argument in the case of short chain fatty acids. I haven't gone into short chain fatty acids in depth.3 However, they are nutrients to the large intestine epithelium. Short chain fatty acids also act through a defined molecular pathway in large intestinal epithelial cells, acting on histone acetylation. They inhibit histone deacetylase, thereby increasing the amount of acetylation, affecting a number of the genes that are important in the immune system. Now that is the effect of a nutrient, but it is not repleting a deficient nutrition status. I think there is a difference between a diet repleting a nutrition status and a molecule acting on the molecular pathways of the inflammation. Dr Beattie: But I think what would be interesting to speculate on is presumably that micronutrient repletion can occur considerably earlier than macronutrient and calorie repletion, and I think an interesting way to speculate on how Modulen works is to not just think whether there is a calorie effect that improves your skin fold thickness, it is another marker, but also the micronutrient repletion effect. So it has the semantic thing, but I think it may give us clues to what we are achieving by giving enteral nutrition such that within 3 days you can downregulate systemic inflammation so much, but we must not discount the fact that it could be the micronutrient or some other nutrient component of the diet that is having that as a direct effect. Dr Sanderson: I didn't mean to give the impression I didn't think that the nutrients in the enteral feed were important. I don't think their importance lies in making children replete in nutrients. If you think about people in Africa who have severe wasting, they can be affected by infections very easily and die, that is due to the fact that they have very poor nutrition status and one needs to replete them, but I don't think that the situation of Crohn's disease is one where the benefit is due to replenishment of the body stores of calories, fat or protein. Dr MacDonald: IL-1β increases defensin expression by epithelial cells, do you think there is a possibility that this is some antimicrobial effect going on here? Another question is, I guess the thrust of your argument was saying that enteral feeds—all of them—could do something, could dampen the epithelial inflammation, and I think it is not in the literature I follow very much but I think people have looked at immune cell function and the effect of these feeds. Is there a knowledge about this, and what is the general consensus of these experiments?
Dr Sanderson: To answer your first question, it is true that
IL-1β upregulates the beta defensins. Whether that is important, I don't
know; but I would suspect that all downstream effects of IL-1β would be
altered in the same way as IL-6 and IL-8. I would expect any NF As to your second question about looking at different immune cells, there is some literature on that. The best literature is where investigators have looked at the particular nutrient and seen its effects; for instance, where they have looked at glutamine and the effect on T cell proliferation.4 But the effects of the formulas as a whole, on these different parameters, is not something that I am very familiar with. Dr MacDonald: You have an inflamed situation, you have an epithelium sitting there that may be contributing to the inflammation, and you have an inflammatory cell there, and I thought that a lot of the discussion is to try to dissect the effects of whether the enteral feed is doing something to the epithelium or doing something to the immune cells, so it is quite easy to grow immune cells in the same way it is quite easy to locate the T cells and look at essential immunology. So I was wondering if someone has taken a T cell or a macrophage for example and showed that if you put it in a milk formula it stops it from making TNF and responds to LPS, but very much like in these experiments that was the macrophage cell lines. Dr Schiffrin: We have used a coculture system in our lab and we have used epithelial cells and different immune cells, and we have been challenging the apical side of the system with bacteria, with some nutrients or some proinflammatory signals. Interestingly, the epithelial cell layer without touching immune cells can send signals, and induce a downregulation of the macrophage proinflammatory status. A blood monocyte is ready to react when it is challenged with bacteria. If you stimulate the epithelial layer, the monocyte in the basolateral compartment will lose the CD14 similarly to the lamina propria macrophage. In IBD, it has been reported that the macrophage remains very active, expressing molecules that enable them to react against bacteria signals. It is true, we have done some work, and the normal epithelial function seems to downregulate macrophage activation. Previous work has shown that PG2 and TGF-β modulate immune to cell reactivity and phenotype as in the basolateral compartment of the cell culture system. Dr Sanderson: I think Tom was wondering that if you don't have epithelial cells in the system at all, you just have immunocytes and put Modulen on immunocytes, what happens? Dr Schiffrin: We haven't done that. We have controlled with and without epithelial cells and the epithelial cell is really having a modulatory activity by signaling to the immune cells within the basolateral compartment under different challenges. Dr Heuschkel: Is there any explanation why the effect that you detected at 24 or 31 days with the downregulation of the MHC is reversed or begins to change despite persistent elemental feed? Dr Sanderson: That is a good question. In our paper in Gastroenterology we actually did try and reverse class II MHC and CIITA, but the problem with that is that the type III is not reversible5 and that is a mechanism that starts late in the mouse and continues to increase, and that results in constant expression of class II MHC thereafter. The question is can you reverse the type IV? We have had problems with that particular experiment, because the difficulty we have is how long to give the enteral feed to try and reverse the type IV. We have done 2 weeks, and we can see some reduction in the type IV, but we could not get it down to the level it was in the mouse before the introduction of the normal diet. So there is a partial reduction of the type IV, but it never goes down exactly to the level that you get before you start the experiments in the first place, which is zero. Now is that a clue to how long one should give enteral feeds in Crohn's disease? I think that would be too big a step because the models are too far apart, this is a normal mouse with a normal expression of class II MHC, and one would have to assume that the molecular pathways that respond to the MHC class II expression were similar to the molecular pathways of Crohn's disease. So therefore you could not take the model and say because you need 10 weeks of elemental diet to reverse the type IV, you need 10 weeks in humans. Dr Heuschkel: But the models are normal animals. Are you planning any work on abnormal animals? Dr Sanderson: We have done the experiments on the IL-10 knockout mouse giving elemental diets. There is some data published from Japan6,7 on enteral diets on various mouse models. The main problem we found is that the severity of the inflammation in the small intestine of the mouse models is very small compared with the large intestine. Most mice models have large intestinal inflammation but the IL-10 knockout does have some small intestinal inflammation; however, it depends on the facility that you have. In some facilities, the small intestinal inflammation is very small indeed. We have had very little small bowel inflammation.
Journal of Parenteral and Enteral Nutrition, Vol. 29, No. 4 suppl,
S134-S140 (2005) This article has been cited by other articles:
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
and interferon (IFN)-
have been linked to growth
suppression.
Top
ENTERAL FEEDINGS IMPROVE THE...
RAPID EFFECTS OF ENTERAL...
B
dependent molecule would be similarly affected, although I haven't tested
that. This question also relates to a concept as to why NOD2 is important.
Because NF
