In this paper, the first aim we had with Dr. Amandine Everard, was to investigate whether some key molecules involved in the innate immune system, mainly MyD88 [myeloid differentiation primary response gene 88], may contribute to the development of obesity, diabetes and low grade inflammation.

This is not something novel, because we knew that MyD88 or Toll-like receptors are involved in the crosstalk between microbes and the host in the context of obesity resistance, but here, more specifically what we did was to delete more precisely MyD88 only in the intestinal epithelial cells, and via a system which was inducible. In other words, we were able to have the normal mice with a normal immune system, a normal gut microbiota, and then we can induce the deletion by using tamoxifen.

The power of the model was this key inducible system. Some people have demonstrated that, in absence of MyD88 in the whole body, mice developed obesity. Some others said they are protected. Some people said that in absence of MyD88 you have a different gut microbiota, and some others have shown that deleting MyD88 on the epithelial cells but via a non-inducible system is associated with changes in gut microbes and colitis. So there was no clear explanation on what the real contribution of MyD88 was in this crosstalk between microbes and host in the intestinal epithelial cells.

After we prepared and generated this kind of mice, we first validated that MyD88 can be deleted on the epithelial cells… and we saw indeed that we can delete MyD88 from the epithelial cells without affecting the different other organs. Then we took mice who were born normally and developed a normal microbiota, normal immune system; we fed the mice with a high-fat diet after having induced the deletion in adult mice. We found that they were gaining less weight. The body weight gain was lower in absence of MyD88 in the epithelial cells of the intestine, and the fat mass development was also lower.

We were surprised by the fact that we didn’t find any changes in food intake. So mice were not eating less food and there was no fat malabsorption or no energy loss in the feces. So we put the mice in the metabolic cages, [and] we did find an increased energy expenditure. So that was the first discovery of the paper: that in absence of this specific innate immune protein MyD88, mice were able to increase their energy expenditure, although they were fed with a high-fat diet. Because we know that a high-fat diet feeding decreases energy expenditure, and that may also contribute to body weight gain, of course.

Following this finding, we also decided to analyze and investigate glucose metabolism. We found here, again, that besides the lower body fat mass, mice exhibit a lower inflammatory tone and a lower plasma glucose. So they had lower insulin resistance and an improved profile of inflammatory markers, suggesting indeed that there are probably impacts of intestinal MyD88 upon all these different parameters that we know are involved in the development of insulin resistance.

Interestingly, when we measured the hepatic lipids – so steatosis – in absence of MyD88 in the intestine, mice were completely protected against the diet-induced hepatic steatosis… maybe due to the increase in energy expenditure or an increased lipid oxidation, but they were completely protected against this hepatic lipid accumulation. We also found that plasma LPS [lipopolysaccharide] (the metabolic endotoxemia) was completely reversed. We know that a high-fat diet increases plasma LPS, and in absence of MyD88 in the intestinal epithelial cells, mice were completely resistant, again, against this diet-induced metabolic endotoxemia.

If we take into account all these data, we can suggest that MyD88 at the level of the intestinal epithelial cells contributes to a reinforcement of the gut barrier, thereby reducing metabolic endotoxemia, although the mice are fed with the high-fat diet. This effect may also contribute to the reduction in hepatic steatosis and an improved insulin resistance. That’s one of the key points.

Then of course, knowing that we are changing the immune system at the level of the epithelial cells, we investigated if the abundance of immune cells, and immune markers, were modified in the intestine. We did find a decrease in Treg cells [regulatory T cells] following a high-fat diet. But the decreased Treg cells was not observed in the knockout mice. So again, if we delete MyD88 in the intestinal epithelial cells, mice were protected against this high-fat-diet-induced modification of the immunity at the level of the intestine, suggesting that there are crosstalks between the fats and the intestinal epithelial cells, or maybe the gut microbiota, because we know that the gut microbiota can change the situation of immune cells in the intestinal epithelial layer.

This is one of the key reasons why we decided to investigate the gut microbiota composition. First, because we found a decrease in plasma LPS, and second, because we found that markers of antimicrobial peptide production were changed, were improved, in the knockout, and also the immune cells.

When we analyzed the gut microbiota composition, we indeed found modification of the gut microbiota that was induced by the deletion. And 72 different OTUs [operational taxonomic units] were different between high-fat diet knockouts and high-fat control mice. So deleting the MyD88 in the intestinal epithelial cells, although it was performed in adult mice, changed the gut microbiota composition.

We transferred this microbiota into germ-free mice just to see if the modification of the gut microbiota contributes to the improvement of the gut barrier and also the fat mass development, and when we transferred the gut microbes into [non-knockout] germ-free mice and we fed the germ-free with a high-fat diet we found that transferring the gut microbiota partially protects against diet-induced obesity and also reduces low-grade inflammatory tone in the germ-free mice, suggesting that there are changes in gut microbes, or specific gut microbes, that contribute to this phenotype. Thus, the reduction in obesity and inflammation is likely due to modification of the gut microbiota.

Finally, one of the questions that we had also was: You have a specific model where you can induce the deletion of MyD88 under high-fat [diet conditions]. You observed a reduction in body weight gain and fat mass development. But is there any place for a putative therapeutic role of this deletion? So what we did was to put the mice on a high-fat diet for a couple of weeks, for six weeks exactly. They gained weight, they became obese. And then thanks to this inducible model we were able to induce the deletion in obese mice. Then, if we induced the deletion in obese mice, they lost weight and they had improved glucose tolerance and a reduction in inflammatory tone also, suggesting and showing indeed that, if we delete MyD88 in obese mice, we can observe a decrease in body weight [and] fat mass, [and] improved glucose tolerance also.

Below, I answer some questions from the GMFH editors:

How does this new research fit in with your previous research on prebiotics?

We found that the prebiotics were able to change the immune markers at the level of the intestine. So I had in mind that microbes are modified when using prebiotics. We found, for instance, a change in REG3γ [regenerating islet-derived protein 3 gamma] and different markers of the antimicrobial peptide production at the level of the gut, which means that microbes dialogue with the host, and we know that microbes dialogue with the host via for instance the Toll-like receptors. We knew that the microbes were modified by the prebiotics; we also have different data in hand suggesting that indeed the prebiotics increase gut hormones… but the link with the innate immune system has never been, to my knowledge, investigated; only at the level of the epithelial cells in the context of obesity.

By using prebiotics in our previous studies we found these different links. So the parallel with the prebiotics study was, first, when we change gut microbes, we can change the epithelial cells and the gut barrier function. But why, and via which mechanism? The idea here was to verify the effect of deleting specific proteins of this crosstalk, because we know that almost all the Toll-like receptors will need MyD88 to transmit the signal. In other words, by using prebiotics to change gut microbiota, and then by changing gut microbiota, we change host metabolism; here we were able to change the crosstalks the another way around: I mean, by changing the host.

In the philosophy of the lab, and what I’m trying to develop as a researcher, is this idea of crosstalks and dialogue between microbes and host, host and microbes. We want to understand what’s the contribution of one or other system to obesity, diabetes and inflammation.

What individual differences will you need to take into account when making parallels to humans in this line of research?

First, we have to keep in mind that we are not all equal in terms of response when we digest fat, in terms of body weight gain. If you have one obese patient, and you put this guy on a low-calorie diet, he will lose weight, let’s say 10 or 15 kilograms. You can give the same low-calorie diet to another person, with the same physical activity, and he will lose less weight. So it means that there is something somewhere, a physiological response of the body, that is different between these two guys.

One of the ideas I have — of course we have to prove it, but I think this is something that may exist according to recent data from colleagues — is that the microbiota might be different between both subjects. They have a microbiota different from me, suggesting there are probably some signatures that are different from one to the second. Then I think that the response of the immune system itself, and the innate immune system in our case, at the level of the gut, might be different between obese subjects also. Not only if you’re slim; but also between [individual] obese and diabetic subjects. It means that we can argue that we are different not only because of the microbes, but also because of the way we are able to dialogue with microbes.

There are also data suggesting that fat requires Toll-like receptor 4 to induce the increase in low-grade inflammatory tone, and in this context the fact that MyD88 might be differently expressed may also suggest that the obese patients may respond differently to a high-fat diet because of these responses. So we have data in hand, and other colleagues have suggested, that the microbial response or the intestinal immune system might be different between these subjects.

What are the potential clinical applications of this research?

I think that from a therapeutic point of view, although we have no clear tool in hand to target specifically MyD88 in the intestine and then induce body weight loss, what I think might be possible to do is to understand how the microbes are discussing with the epithelial cells, or with the immune system. Then maybe this also suggests that if we are able to reduce the overstimulation of the different Toll-like receptors related to MyD88, it might be one way to reduce the body weight gain and fat mass development and improve the gut barrier function. This is something that we will try to investigate. If we can target only in the epithelial cells the expression of MyD88 — I cannot say how so far — maybe we will be able to reproduce the same effects.

Then in the simple understanding of the physiology of obesity and diabetes, this study suggests indeed the existence of a crosstalk between host and microbes and vice versa. This also clearly demonstrates that we cannot say that all the different organs and all the different cells are equivalent in terms of response
against diet-induced obesity. In other words, here we found that epithelial cells are really crucial to detecting the fat coming from the diet, and that the role of the epithelial cells is not only the absorption of the nutrients, but also maybe the role of sensing fat. Not only sensing fat via specific receptors, but maybe
via the innate immune system. Can we [find a] similar relationship in humans? I have no data in hand to say yes or not. But I know that there are tools, and I hope that we will be able to demonstrate that there are some tools that are beyond this MyD88; some of these proteins [are] probably inhibited by specific inhibitors.

Targeting the MyD88 in the whole body is probably not the solution; it’s too risky because we need the immune response. But targeting the response only in the epithelial cells is something worth pursuing.


Everard A, Geurts L, et al. (2014) Intestinal epithelial MyD88 is a sensor switching host metabolism towards obesity according to nutritional status. Nature Communications DOI: 10.1038/ncomms6648