So I'm pleased to be here this morning and to be able to tell you a little bit about our work in the area of how the microbiome might be an integrator of gene-environment interactions in production of the metabolic syndrome, and particularly insulin resistance. And I have, these are my disclosures, but none of the work I will present is related to any of the pharmaceutical or biotech companies. It's all NIH-funded research. So as a starting point, I think most people in this room are very familiar with the concept of metabolic syndrome. But I think it's important to keep in mind that the metabolic syndrome is much bigger than just diabetes and obesity, that the metabolic syndrome includes not only central obesity and glucose intolerance or diabetes, but it includes hypertension, dyslipidemia is an accelerated atherosclerosis, sort of the classic metabolic syndrome, as well as things that are now added to this metabolic syndrome phenotype, including hepatic steatosis and cholesterol gallstone formation, reproductive disorders like polycystic ovarian disease. We know that these individuals also have increased risk of cancer and increased risk of Alzheimer's disease. And when they get Alzheimer's, it tends to be more rapidly progressive. So that's another important risk factor. And ultimately, all of this impairs longevity. So this is really all a complex syndrome. And from my perspective, what brings it all together as a single syndrome in some way is that in all of these, a prominent pathophysiologic feature is insulin resistance. And so when we think of this syndrome and we think of insulin resistance, we usually think about it as some interaction between a group of polygenes and a group of environmental factors. And what I'd like to explore with you in the next 15, 30 minutes or so is, you know, where in this does the microbiome fit? Is this a third factor? And if so, is it reflecting genes, environment, or some combination of those? And what in this cluster of metabolic syndrome could it be accounting for? Now, most of the work I'm going to show you is based on animal studies. But I like to show some human physiology just to put this in perspective. And before, I thought we had a shorter time. I had a couple of slides. But here's one that I think is particularly revealing about the complexity of metabolic syndrome. And it comes from the work, the classic work by the late Jerry Reven in Stanford, who was interested in the relationship between obesity and insulin resistance in a non-diabetic population. And he gathered together some 300 to 400 individuals who were not diabetic. And in these individuals, he not only measured a bunch of physiological measurements, including their BMI, but he also measured their insulin sensitivity. And the way he measured insulin sensitivity was a measure called steady state plasma glucose or SSPG. And for those of you who don't know this test, basically, you infuse glucose and insulin at a fixed dose simultaneously. After a couple of hours, people reach a steady state of glucose. That's the SSPG part. And of course, this is a balance between how much insulin action there is to get rid of the glucose that's being infused. So it's a measure of insulin resistance in a rather pure but simple form. And what this slide shows is a relationship which we all understand and are taught, which is that with increasing BMI, there is an increasing level of insulin resistance as measured by SSPG. But what I like to point out when you look at this slide is look at the range of variation among these individuals. And at any given BMI, you can have a hugely variable level of insulin resistance so that there are some individuals here with a BMI of 30 who are just as insulin sensitive as the most insulin sensitive individuals in the population. And there are others who are just as insulin resistant as the most insulin resistant individuals in the population. And so the question is, what accounts for this? Is this genes? Is this environment? Is this perhaps the microbiome? And how would we sort this out? Well, to do this, what we have tried to do is, in a way, model some of these features using mouse models where we can control some of the variables. And the mice models that we've employed are three inbred strains of mice, which are common laboratory mice. The C57 Black 6 mouse from Jackson Laboratories. Another mouse from Jackson Laboratory, which is called the 129S1 strain, which is the Jackson strain of 129. And a close relative of that 129, the 129S6, which is actually sold by Taconic. Actually, both these 129 mice came out of a single founder about 50 years ago. They've been bred apart for some 50 years. So they've accumulated some variation drift in their genome, but they're very highly genetically similar. And what's interesting about these mice is that they, in a way, capture the heterogeneity of metabolic syndrome and obesity in these three mice. So here is the response of weight gain in these three mice when challenged with a high fat diet, a 60% high fat diet. And it's pretty clear that the B6 mouse shown in red and the 129 mouse from Taconic gain equal amounts of weight, and much more weight than the 129 mouse from Jackson, even though it's closely related to the other 129 mouse, which is somewhat considered obesity resistant because it's eating the same high fat diet. If you look beyond just the weight gain, however, what you see is also heterogeneity with response to glucose tolerance. So the B6 mouse, the mouse that we use for most studies of metabolic syndrome in mice, becomes very glucose intolerant, very hyperglycemic in this glucose tolerance test. Note that the 129 from Jackson, which didn't gain much weight, has the next worst glucose tolerance. And the 129T from Taconic, which gained as much weight as the B6, has almost completely normal glucose tolerance test. Now is this insulin secretion or insulin resistance? Well, if you measure insulin resistance, in this case, as illustrated by the HOMA-IR, which is a simple measurement of insulin resistance, you can see that the B6 mouse has the highest HOMA-IR, is the most insulin resistant. And both 129 mice are actually quite insulin sensitive, despite their difference in glucose tolerance. So here we have three mice, the B6, which is obesity prone, diabetes prone, and very insulin resistant. The 129 from Jackson, which is obesity resistant, diabetes resistant, and insulin sensitive. And then the 129 from Taconic, which is obesity prone, but nonetheless is diabetes resistant and insulin sensitive. So is this genes or is this environment? Or in the context of today's session, could this be the gut microbiome? So to begin to address this question, what we decided to do was to get these three strains of mice from their respective vendors and bring them to the Jocelyn Diabetes Center where I do my research and environmentally normalize them. So we bred them for three generations at Jocelyn in the same mouse facility with the same food, the same animal handlers, not in the same cage, but sharing the same environment. We did this for three generations. So we called these Jocelynized mice. So we had then still, we had our B6, our 129 from Taconic, our 129 from Jackson. And they were Jocelynized. And for all three of these, we did not only metabolic phenotyping, but we analyzed what happened to the gut microbiome. So to start off with, what does the gut microbiome look like in these inbred mice from a commercial vendor? So this is the mice as they arrive from the vendor, and this is based on 16S sequencing. So this was done a number of years ago. It's not metagenomics, but you can get the picture very clearly. If you look up here at the bacteroides in this heat map, the dark red means a high presence. The dark blue means essentially no presence of a bacterial species. So you can see that the two mice that come from Jackson share actually the two bacteroides species that they come with, whereas the one from Taconic has a different bacteroides, and those are the only bacteroides in these mice. And then they each have a range of different firmicutes and other microbiota. So they have quite different microbiomes. But what happens after they're bred at Jocelyn? Do they become the same? Well, the first thing is they become different. That's for sure. The Jocelyn is not Taconic, is not Jackson Laboratories where they have super facilities. We just have an average old mouse facility, and you can see already they have a much more diverse microbiome in our animal facility. With many species of bacteroides and many species of firmicutes and other microbiota, but even in the same environment with the same dyad and the same animal handlers, et cetera, they still come to a different new steady state with different microbiota. And so they have the B6 and the 129 both from Jackson compared to the 129 from Taconic are each unique in their bacterial species with different bacteroides and different firmicutes and other bacterial species as well. So they readjust their microbiome, but it is not the same. And what is the implication of this in terms of metabolic health? Well, I won't have time to show you all of the data, or I didn't think I was going to have time to show you all the data, but I can say that things change. One of the things that most obviously changes is their glucose tolerance test. So this was the glucose tolerance test I showed you in the first slide, which are glucose tolerance tests done within the first few days of arrival of the mice from the external vendor to the Jocelyn, and you can see this is the B6 showing the worst glucose tolerance, the 129, the two 129 strains from showing the different glucose tolerance. And now you can see after being Jocelyn bred for three generations, the B6 is still the most glucose intolerant, but the two 129 strains now have almost identical glucose tolerance. So changing the microbiome has really significantly changed this metabolic profile, and in fact it's also changed to some extent the weight gain profile and other features which I won't have time to review for you today. So is this still really the microbiome that's doing it, or could it be some other feature? And to do that we wanted a different or second way to manipulate the microbiome, and for that we decided to use antibiotic treatment. And so again we took the same three strains of mice, we had them at the Jocelyn so that they were in our facility, and we took the mice and divided them into groups and put one group on placebo drinking water, just plain drinking water, and the other two groups on two different antibiotics, vancomycin, an antibiotic which kills a lot of gram-positive organisms in the gut and is not absorbed, and metronidazole, an antibiotic which kills a lot of gram-negative organisms and is absorbed, and then we subjected them to high-fat diet challenge and asked the question, how does this change their behavior in terms of metabolic health? And what we found, and I'll just show you the results here for the B6 mouse because it was the one that was most metabolically abnormal, and I'm just going to show you the fasting glucose as one measure, but everything sort of follows this, is that when this high-fat fed B6 mouse with this metabolic syndrome is given systemic antibiotics in its drinking water and the microbiome of course is changed by this, there is a new steady state of blood glucose in the fed state, both with metronidazole and with vancomycin. So this mouse has improved its metabolic state by changing the microbiome through the use of antibiotic treatment. This still doesn't prove, however, that it is the microbiome. Those of you who do microbiome research will know that this is usually not regarded as sufficient proof. The way you want to test if it is the microbiome is actually ask, if you transfer that microbiome to a germ-free mouse, can you transfer the phenotype even without the antibiotic treatment, even without the strain difference, et cetera? And so what we did in this particular paradigm is we took these B6 mice. We gathered the microbiome from the control mice, the high-fat diet, the high-fat plus antibiotic treated mice. We took new B6 mice, which were germ-free, so they had no gut microbiota, and then we transferred the microbiome from each of these animals into the germ-free mice. And after this, during this transfer, some of the mice were kept on normal chow, and others were given a high-fat diet, but no further antibiotics. And what is very remarkable about this experiment is that you can transfer some of this metabolic improvement due to antibiotic treatment without giving any antibiotics, simply by transferring the gut microbiome. So shown in this panel in the upper right, you see the glucose tolerance now in the mice who are recipients. These are germ-free mice, but on a normal chow diet who were given normal chow microbiota from a normal chow-fed animal. They have the best glucose tolerance. The germ-free mice who are on high-fat diet and given microbiota from a high-fat-fed animal have the worst glucose tolerance. And in between are the glucose tolerance tests in the animals on high-fat diet who were given gut microbiota from other high-fat-fed animals who had been given antibiotics, but who themselves had not received antibiotics. So we've transferred the glucose tolerance phenotype, or at least a large part of the glucose tolerance phenotype, by simply transferring the microbiome. And this actually is due, this improvement in glucose tolerance is due to an improvement of insulin resistance, and I can say that this is true even at the signaling level. So if you look at the signaling level, this lower right panel shows you insulin signaling in muscle. In this case, all these animals are high-fat-fed. I'm just comparing the high-fat diet to the two animals that got the microbiota from the antibiotic-treated mice, but were not antibiotic-treated themselves. And if you look at insulin receptor, insulin-stimulated receptor phosphorylation, you can see that that's increased in the mice receiving the vancomycin-treated microbiome or the metronidazole mouse microbiome. And that's true also at the level of AKT and ERK phosphorylation. So we've transferred not only the improved glucose tolerance, we've actually transferred improved molecular insulin signaling. And I won't have time to show you all the data, but this is true also in the liver, and it's true in other organs of the body that are not even involved in metabolism. For example, we're very interested also in how the microbiome might affect the brain and behavior. And so in these mice, we do behavioral tests, and I'm just going to show you one behavioral test here. This is kind of my favorite behavioral test. Margie Mark has other favorite behavioral tests, but this is called the marble-burying test. It's a simple test. Even endocrinologists can do this test. You put mice in a cage. You put in this array of 15 marbles in the cage. So that's the starting point over here. And you give the mice five or 10 minutes, and you ask, how many marbles do they bury? Because it's sort of a nesting behavior. So a normal mouse will bury a certain number of marbles. And then you can see that here's the high-fat-fed mice. These are the mice, actually, the original mice from this cohort, not the gut microbiome transfer. You can see they bury more marbles. They are a little bit more obsessive. But if they've been antibiotic-treated, they don't bury more marbles. And this is quantitated for you here. So marble-burying behavior sort of correlates with sort of this anxiety or compulsive behavior. I think it's obsessive-compulsive, but who knows? I'm not a psychiatrist. But they bury more marbles, and that's reversed with antibiotic treatment. But what's really cool is if you do this microbiome transfer into mice that have never been seen these antibiotics, they do exactly the same thing. So the ones getting the chow diet bury fewer marbles than the ones getting the high-fat diet and getting the microbiome from a high-fat-diet animal. And if they get the microbiome from an animal that was high-fat-fed but given antibiotics, even if they're not getting antibiotics, they bury less marbles. And there's other behaviors like this. So this is a systemic effect. This is going to have effects potentially on every organ of the body, including the brain and behavior. So how does all this occur? And we're probably going to hear a couple of other thoughts about this. But the effects of the microbiome, as we know, are very complicated. First of all, we know that there are integrations of many effects. There are integrations of the effects of genomes and the environment. So here's our individual subjected to a high-fat diet who comes with his or her own genetics. The gut response to this is going to be different, not only in terms of the changes in the microbiota, but how these affect the gut barrier. It will affect, as a result of changing the gut barrier, the immune response of that gut barrier. It will also affect endotoxins and cytokine release. And it will certainly affect metabolism because of the metabolic breakdown products, and you'll get a change in metabolites, which could affect signaling. And then these will cause metabolic regulation, immune regulation, and all this will come back together when we look at our metabolic syndrome phenotype. So being interested in metabolism, we've actually focused primarily on the metabolic aspects of this rather than the immune aspects, and I just want to show you how powerful metabolism is as an intermediate regulator of many of these microbiome effects. So to do this, we use the technique of metabolomics, which is based on LC-MS-MS technology. For those of you who don't do it, it's a very powerful technology to measure literally thousands of metabolites. When you do this type of assay, we do it in collaboration with Clary Klisch at the Broad Institute who runs a core that does this. You literally initially measure over 25,000 peaks of metabolites in the blood, most of which are unknown, and I'll come back to that at the very end. But of these, about 500 or 600 can be considered identified metabolites, and you can actually give them names based on what we know, and you can ask, does the metabolome differ in these three inbred strains of mice? And here in this heat map, you see how different mice, inbred strains of mice are in terms of their metabolome, their metabolic profile. So at the top, I just say the color code is that blue means low abundance, red means high abundance, and so this first cluster of metabolites are very abundant in 129 from Jackson, but not in the other strains. Here's metabolites that are present in both 129 strains, but not in the B6. Here are metabolites that are present, particularly high in the 129 from Taconic, but also a little bit in the B6, and then here are metabolites that are almost exclusively high in B6 mice. And there are lots of interesting metabolites right here. I won't go through all of them for you, but I will just point out a couple. So up at the top, for example, is taricolic acid, a bile acid that's very high in 129 mice. Taricolic acid is known as an anti-inflammatory bile acid, so they have the higher level of this anti-inflammatory bile acid, which could potentially blunt the immune response or the inflammatory response to high-fat diet, maybe accounting in part for their insulin sensitivity. Down in this lower group is in TMAO, the metabolite that's been associated with cardiovascular disease and metabolic syndrome. That's high in the B6 mouse. And then there are other metabolites, such as this is the pair of beta-hydroxybutyrate at the top and alpha-hydroxybutyrate at the bottom. So that's 3-hydroxybutyrate and 2-hydroxybutyrate, and you can see that they're high in different animals. And this will get important in the end, because I'm going to show you that these metabolites have their own metabolic effects. They actually can modify things like insulin signaling and insulin responsiveness. So actually knowing what the metabolites are and the metabolomic profile is really important. Now I want to also give you a couple of caveats to the way we all do this, and just allow you to think about a third dimension or a fourth dimension in this. And that is within the body, there are also differences in different compartments. So we almost always, and I think most of us almost always, think about systemic blood flow. And the metabolite profiles I show you were done in cardiac blood, so that's systemic blood. But of course, the gut is being drained preferentially by the portal vein, which is going through first the liver, where there's an element of clearance. Some things are cleared, some things may be released. So there might be different levels in portal blood versus systemic blood. And many lipids, particularly large ones with large unsaturated chains, fatty acids, are actually not being absorbed by the portal vein, but are being absorbed by lymphatics as chylomicrons, and are being drained directly. And of course, then there's what's in the gut itself, which is still different. So you can have metabolites at different levels in different places in the same animal, same individual. And this is a PCA plot, so this is a distribution of the known metabolites, just sampling portal blood from a bunch of animals on high-fat diet, chow, and antibiotics, and cardiac blood. And I think that it's very clear that the first separation of these is not by the high-fat diet, not by the obesity, not by the antibiotics, but it's by which blood you're sampling. So the upper circles are in the portal blood. There's 111 metabolites that are enriched in the portal blood over the peripheral blood by a factor of at least 1.25. And conversely, there are about 75 metabolites, which are actually more enriched in the cardiac blood than the portal blood, because they are actually being taken up by the liver or released. And they are also going to be affecting systemically more than the liver. And I'm going to give you an example of just a couple of these. This is from the TCA cycle, and I don't want to go through all the details of the TCA cycle here, but this TCA cycle is illustrated for you in this slide. And shown in these box plots at the bottom are the relative levels. On the left, let me just see, on the left is the portal blood, on the right's the cardiac blood. You have B6 mice on chow, high fat, high fat plus vancomycin, and high fat plus metronidazole. So you're seeing all four conditions, and you're comparing the portal blood to the cardiac blood. So you can see that for these metabolites, these TCA cycle intermediates, in every case the portal blood levels are higher, significantly higher, than the cardiac blood levels. And that some of these, as they change with either diet or antibiotics, you get reflection of this in the systemic circulation, but it's not always the same. Sometimes it's more, and sometimes it's less. And this dissociation can be seen as you get further maybe into other branches of this. So for example, I'm showing you the glutamate metabolism branch here, and you can see that some of these metabolites, like glutamate itself, show a very similar pattern, but look how much the level in cardiac blood changes as you go from chow to high fat, to high fat plus vancomycin or metronidazole, whereas the level in the portal blood doesn't change. So different pools can be changing under different circumstances. And some of these metabolites, like citrulline, have also been related to changes in hepatic insulin sensitivity. In this regard, it's important to keep in mind that the bacteria themselves, the microbiota themselves, also have metabolism. So we have a TCA cycle, and they have their own TCA cycle, but not every microbe has the full TCA cycle. So some microbes can do half the first half of the TCA cycle, some can do the last half, and some can do both halves. So we're exchanging metabolites with them, and they're exchanging metabolites with us and with each other at the same time, and so I won't go through all of this, but you have to think about the complexity of this exchange from bacterial and host metabolites, and the fact that microbes, different categories of microbes, may have an incomplete TCA cycle or have some metabolic enzymes which are not present in other microbes. Finally, I'm going to come back to the insulin sensitivity part, and just say that we still haven't figured this out, but one way we're trying to approach this is to look across all of these models, are there metabolites that are either positively or negatively correlated with insulin resistance or insulin sensitivity, and across all of them, conditions, so that we could use these as a guide to figure out which metabolites should we pursue as potential mediators of these microbiome effects. So I'm showing you a list of four metabolites here. These are metabolites that positively correlate with insulin resistance. So this is a scale of zero to four for insulin resistance, combining all the different kinds of treatments we've done, and you can see that for these four metabolites, there's a positive correlation. The more insulin resistant the animal, the higher the level of these four metabolites. They include amino adipate, alpha-hydroxybutyrate or 2-hydroxybutyrate, et cetera, also carnitines and acylglycines. And then here are metabolites that negatively correlate, going in just the opposite direction, so you can see the list at the bottom there, which includes adipate and several other fatty acids or several other lipid species. The one I just want to illustrate for you as an example is 2-hydroxybutyrate. That's alpha-hydroxybutyrate, not beta-hydroxybutyrate, which we think of as a ketone body, but this is alpha-hydroxybutyrate, which you can see the correlation. Alpha-hydroxybutyrate itself can modify insulin signaling. Here's an experiment where we've taken FAO cells, these are liver cell line in culture, we've exposed them to the metabolite at increasing concentrations or time, and we've measured insulin signaling. And so you can see the effects of either six hours of treatment or 48 hours of treatment at two different concentrations, both within the physiologic range. And here's basal and insulin stimulated IRS, one phosphorylation, which is suppressed in a dose and time dependent manner, as well as AKT phosphorylation, which is also suppressed in a dose and time dependent manner. And if you quantitate this, increasing this metabolite across its physiologic range can cause more than a 50% decrease in insulin stimulated signaling in vitro in this liver culture system. So what I've tried to show you in the last 30 minutes or so is that figuring out this interaction between genes, environment, diet, and the gut microbiome, and metabolic syndrome is complex. That each of these is talking to and communicating with the others. Somehow, this conversation, however, is increasing the risk of diet-related obesity, diabetes, and metabolic syndrome. We think that a major driver of this are critical metabolites that are being modified by both the diet intake and the gut microbiota, as well as by the host. And these will be at different levels in different compartments, including portal and peripheral blood, and will affect overall metabolism in different ways, including, actually, metabolism in the brain, mood, and behavior. Finally, my last point to make is that this is just the tip of an iceberg. And to do that, I'm showing you now a heat map of the unknown metabolites. So every one of these bars in this heat map represents a peak in the HPLC mass spec, mass spec analysis, which is going from, in some cases, very high levels in B6 to virtually undetectable in 129, and vice versa, and changing with diet and antibiotics. There's 1,000 of these. Some of them change by 100-fold under these conditions, and we haven't a clue as to what they are or what they do. So for those of you who are still looking for something to do, this is something to do, because I think in here is the clue to many aspects of the crosstalk between the gut microbiome and metabolic health. And with that, I'll end, and I just want to acknowledge the people who've done this work. This project was started many years ago by Siegfried Usser, when he was a fellow in the lab in Shiho Fujisaka. But we've had great other fellows who've joined in, Carly Sederquist, Marian Soto, who did the work on the brain, Amral Tindes, whose work I didn't have a chance to talk about today, and the current person who's looking at the peripheral levels, Francois Moreau. Thank you very much for your attention. Thank you, Dr. Kahn. We're now open for questions. Thanks. Dave Schumann, Cleveland Clinic. I'm sort of curious to focus on the TMAO that was the bad guy in Stan Hazen's work, and the fact that you list carnitine, or a metabolite, was it 25-carnitine? Is it just carnitine, is one of the things that had a big effect on insulin resistance? Yeah, it was, let's just go back, because I'm not, let's see if I can go backwards with this. I can't go backwards, so, and I don't remember if to tell you the truth, but I think, go on and finish your question. You had a plot where it sort of, it went up from left to right, which seemed to be different levels of insulin resistance across species. Right, those were across insulin resistance and the level of the metabolite, so that was at a higher level, was associated with more insulin resistance. Because the carnitine would have then been associated with much higher levels of TMAO, maybe? Yes, and it was. Okay, and then, so carnitine, and then TMAO, which is then thought to be very, very atherogenic, is then, mediates insulin resistance, or is associated with it? Well, it could. We haven't tested TMAO specifically in the in vitro, and so I think that one of the things that you're pointing out- I mean, I realize it's one in 10,000 hypotheses or something. So that's one of the things I was going to say, is that so many of these metabolites correlate with each other, and so the only way to know, in a way, if a single metabolite is causal, or has a potential causal effect, is either to give it as an isolated, the way we did, for example, with the alpha-hydroxybutyrate, or, obviously, to do something to knock out that conversion, and then see which part of that, and we haven't done that for the TMAO. We haven't really pursued TMAO as a metabolite of interest to us. I mean, you've done a great job with this. It's always fun to see something that starts to split it apart into something that you might play with, but what are you going to do now? What are we going to do now? Well, if I had- How do we figure this out? Yeah, well, if I had the infinite amount of money that it would take to do it, actually, I would bet on that next to the last slide with the heat map. I think that we're looking under, you know, we're like the drunk who looks for his keys under the light post, even that's not where he lost them, because the light's best, you know, that joke. But I think we're like that, and because we can see it the most, but I think that, you know, there's so much richness in these unknowns, and I've tried very hard, I've tried a little bit to get some funding for this, because it's a huge project, but so far, not. But I think that, personally, if I had to bet, I would do that. But other things that we're actually really trying to better understand is this difference between what the liver's exposed to and peripheral tissues, because we know in metabolic syndrome that you can have a dissociation between hepatic insulin resistance and peripheral insulin resistance, say, in muscle, so you can get fatty liver, which is sort of, you know, lipid metabolism in the liver is continued, but glucose metabolism is not. So those are the two areas we're pursuing, but there's many other questions that could be pursued. Thank you. Sure. So for those of you who have come in a little bit late, we actually only have two speakers today, so we are lucky that we can sort of go into a little bit more detail with the questions. So if you do have questions, please do come up and ask them. Yeah. Gordon Cutler. I really just mostly have comments. I thought that was a spectacular talk, Ron. I'm almost as old as you, and over my career, I, which is unlike yours, really I'm in the observer state now rather than the doer. There's certain people, certain lists that, whether they're in my field or not, I always go to their talk if I see them, and you're one of those, and I advise the young people to do the same. And the other thing that I loved about this is that it's something that took me a while to realize, but that I always told fellows, and that is that there is so much known, both in science and medicine, that it seems overwhelming. Of course, nobody can encompass it all, and yet the amount that's unknown is vastly greater. And sort of like what Rumsfeld, when he was secretary, he called them the unknown unknowns. The other thing that's, to me, amazing is Jeff Gordon was one of my fellow interns. He was just an intern like anybody else, and I look at his career and how he started this field of the microbiome, and it was sitting there, and everybody knew we had a microbiome. But when we were being trained, nobody did anything about it or thought anything about it other than, you know, the colon is full of bacteria. Thank you for the nice comments, Gordon, and good to see you. Maybe we'll talk afterward. Yeah. I also would say, as you find out whether the humans have some of these same correlations, you know, I'd love to know in time to fix my metabolism. All right, but we'll work on that. Mark Tuttle, Wellesley College. Thanks so much, Ron, for a wonderful talk. That was really exciting. I really found your metabolomics data really compelling and I'm sorry if I miss this, but have you looked for any sex differences in the metabolomics and also your gut brain, your Marble Baron experiments? So I'm glad you asked the question. All of our studies were done in male mice, but I do think that this is another dimension which I didn't get a chance to talk about, so I'll just make a plug for it, is that in many of the things we measure, both at the whole body level and even at the cellular level, there is sexual dimorphism. We've gotten, we actually, one of my fellows had a poster yesterday about sexual dimorphism in IPS cells in tissue culture, human IPS cells in tissue culture, where they show different, males and females show different signaling in the absence of adding any hormones, any sex hormones and so forth. So sexual dimorphism is another dimension. It will certainly, work by you and others have shown that it certainly can affect the microbiome. It will certainly affect the way the metabolites are processed to give different products as well, not just sex steroids, but other metabolites because bile acid metabolism is different between males and females, etc. And so I think that that has to be another variable for people doing work to keep in mind and we need to understand it for sure, but I don't think we have it yet. You take sex differences and overlay it with your thousand unknown metabolites, you've got some nice questions. Yeah, we should do that. I like that. So I'm gonna actually indulge myself and ask you a question as well. There is a ton of literature in the veterinary world and in the animal husbandry world looking at low-dose chronic antibiotic therapy to increase the bulk of young animals, poultry, pigs, as well as cows. Makes them heavier, makes the farmers get more bang for their buck. You show something a little bit different with your metronidazole and your vancomycin. Can you discuss that and what your thoughts are there? Yeah, so there's two questions in here in a certain way, which I think we sometimes conflate a little bit. So it is known, as you say, that if you give animals, most animals supplement antibiotics into their feed, they will grow larger. And this is used by farmers for pigs and cows. I don't think for other animals, but for those two for sure. And the mechanism that I think is not really completely clear. I, you know, we assume it's somehow related to the microbiome, but how it's related I don't I don't think we know. At least I don't know for sure. And I don't know how it relates to what we see here. And so I think that is something. And I think that a lot of times, I will say since there's some endocrine people here who probably study growth hormone and growth, and of course a lot of people assume that in pediatrics, my daughter's a pediatrician, when I asked her why babies are bigger and kids are taller, she always says, oh, you know, it's a healthier diet, right? Well maybe it's a healthier diet, but I think there's some more things left to be discovered there too. I think that they're, you know, not only do we get the nutrients that we measure, but we get other small molecules in our diet. Other, and it could be that the levels of growth are also mediated by things that are not these organic molecules that we're following, which is what we do when we do follow metabolomics, but also inorganic molecules. And I think that that could be another factor. So I think we have to keep an open mind. I don't know the answer, but I think it's a terrific question. All right, last two questions and then we'll let Dr. Khan have a break. Hi, thanks again for a great talk. Sean Seltzer from Montreal. I just found it interesting that the nude mice seem to have like the best glucose tolerance, and then, and with the antibiotics, it seemed to improve as well and from the other, from the original mouse. But my question is, do you think that there's a component of inducing malabsorption with the administration of antibiotics, and have there been any studies where they've used maybe a parenteral approach to giving like a glucose load and looking at insulin signaling and sensitivity? Yeah, yeah. So two things. I would say in our studies, we have measured fecal caloric loss, and it doesn't seem to account for the differences in weight gain, but I must say that fecal caloric estimations are not very precise. And the amount of calories that it takes to change a mouse to get a little heavier or a little lighter or so small, they're probably within the noise of that measurement. So I don't think we can a hundred percent exclude fecal caloric differences, but we've tried and it doesn't seem to be that, but I agree with that. And the second part of the question was? Well, just if you know if there have been maybe a parenteral nutrition. Yeah, and with, we've never done full parenteral nutrition. The glucose tolerance tests that I showed were actually done as IP, intraperitoneal glucose tolerance tests. So the glucose delivery is in a way kind of systemic rather than gut drive. You see differences if you do oral glucose tolerance too, but you can see them with IP. So it isn't simply, say, an incretin effect, but incretins also change in this context. Thanks. Last question, thank you. This maybe just reflects my own ignorance, but is it possible or do you know if anybody is looking at this sort of overwhelming heat map that you showed at the end and using AI to try and understand the relationships there? Yeah, we have not found a partner who wants to do that, but if you know somebody who's good at this, we're open because we have lots of data. For every, I will say for every species that's identified by mass spec, you know the mass charge ratio. You have a range of things that it could be because mass spec gives you the mass charge ratio, at least of the ions that come out of it. So you have some clues, and I think if somebody were maybe really good at this, they could help deconvolute it. Eventually you have to test it, but they would certainly, that would be a way to go. Thank you. Thank you. All right, our next speaker today is Dr. Zheng Kuang from Carnegie Mellon, and the talk will be on microbiota, circadian rhythms, and lipid metabolism. Okay, so good morning everyone. So first of all, I want to thank the organizer for the opportunity to speak at this meeting. So my lab started just a few months ago, and we are mainly interested in the data microbiome and how it regulates the circadian rhythms in the host. So today I'm going to talk about a story about how the microbiota regulates the circadian rhythms of lipid metabolism, and it's mainly done when I was a postdoc in Laura Hooper's lab. And I have no financial relationship to this course, and this is a QR code to access. So as you may know, there are trillions of different microorganisms living in our gut, and these microbeads are collectively called the gut microbiota. And people found that the microbiota functions just like another organ, and has very extensive crosstalks with different systems in our body. So all that was mainly interested in understanding how the microbiota regulates metabolism and immunity in the intestine. And we basically focused on a single layer of intestinal epithelial cells. And it's mainly because, first of all, the epithelial cells, the main function of the epithelial cells is to uptake nutrients, so it has a huge impact to host metabolism. And also the epithelial cells form the first line of cells that separate the host from the microbeads. So it has a very important role to protect the host from different pathogens. But because metabolism and immunity are very different functions, right, so we're wondering how these two processes are coordinated in the same type of epithelial cells. And so to understand this question, we introduced the concept of circadian rhythms. So circadian rhythms are synchronized gene expression and physiology to help coordinate the function of the host in response to the 24-hour daylight-night cycle. So it's well known that the circadian rhythms are very important for metabolism and immunity, but recently people noticed that actually there are very extensive crosstalks between the circadian rhythms and the gut microbiota. So for example, people found that the gut microbiota also shows circadian rhythms in abundance, so those bacteria are actually cycling in the gut. And people also show that the germ-free mice, which do not have any bacteria, they have very different circadian rhythms compared to the conventional raised mice. So our question is, how does the bacteria in the gut regulate the circadian rhythm in the host since they do not sense the light signal directly, right? So we know that the circadian rhythms are driven by the circadian clock, which is composed of several transcription factors and form multiple feedback loops to drive the rhythm of gene expression and also physiology. But if you really look at how the circadian clock works, you may see one of the main mechanisms is through histone modification. So there are many different modifications on the histone proteins, such as acetylation, mycelation, or phosphorylation, and those modifications can regulate the transcription by either opening the chromatin, closing the chromatin, or recruiting other transcription factors. And the circadian clock is able to modify the histones and regulate transcription. So for example, the clock, BMO1, are the central component of the circadian clock, and it can acetylate histones and activate transcription. So there are also histone deacetylase, such as HX3, which can be recruited by the circadian clock and remove histone acetylation and repress transcription. So you can see that those proteins work together to drive the rhythm of histone acetylation and the rhythm of gene expression. So based on this model, we hypothesize that the gut microbiota may regulate host circadian rhythms through histone modification. And so to test this idea, we use the germ-free mice, which do not have any bacteria, and we compare them with the conventional race of mice. And we collected multiple time points across one daylight cycle. So for those who are not familiar with circadian biology, we use the Zeigert time here. So Zeigert time zero, or ZT0, means the beginning of the day, and ZT12 means the beginning of the night. And we collected the small intestinal epithelial cells and did the RNA-seq and histone acetylation chip-seq to look at how transcription and chromatin states are affected by the microbiota. So the first question is, how does histone acetylation look like in the intestine? So here shows the six time points from the conventional mice. So you can see we ordered the six time points from top to bottom at ZT0 to ZT20. So each row or each track represents one time point. So you can see there are multiple histone acetylation peaks in this example region, and each region represents one active promoter or enhancer. So if we focus on one signal, you can see the signal is actually dynamic across the six time points. So it's low at the beginning of the day at ZT0, and it increases at a peak at ZT8, and it decreases again and becomes low at ZT20. So you can see this signal has a circadian region. But actually you can see all the signals here are synchronized, and they consistently peak at ZT8 and become low at ZT20. So this data tells us that histone acetylation has a circadian region in the conventional mice. But how about the germ-free mice? So here shows the four time points from the germ-free mice of the same exact genomic region. So first of all, you can see the same peaks at the same locations from the germ-free mice. But it's very obvious that most of the signals, they are not cycling anymore, and they are constantly high. So this means that the circadian rhythm of histone acetylation is microbiota-dependent. So we also look at the genes that are regulated by the cycling peaks, and we found many different pathways enriched by the cycling peaks. And in particular, we noticed many metabolic and elution uptake pathways, such as lipid metabolic process or amino acid transport genes. So so far, the data suggests that the gut microbiota is essential for the circadian rhythm of histone acetylation in the intestinal epithelial cells, which potentially regulate energy uptake and metabolism. But now it raises several interesting and important questions. So for example, how does the microbiota regulate the rhythm of histone acetylation? And how does it impact our energy uptake and metabolism in the gut? And how does it affect our metabolic health? So to address all these questions, a first important step is to identify the enzyme, the chromatin modifier, that is regulated by the microbiota and drive the rhythm, right? Because if we know the enzyme, then we can use mouse genetics to answer all the questions. So we know that there are three major classes of chromatin associated proteins. And the first class is called the writer, so they can put the modification on the histone proteins. The second class is called the eraser, so they can remove a modification from histones. And the third class is called the reader, so they can recognize a specific modification and produce downstream effects. So as I was showing you earlier, we observed the histone acetylation is cycling in the conventional mice, but it's not cycling, and it's constantly high in the germ-free mice. So this pattern suggests that maybe there's an eraser which can remove histone acetylation at some time point of the day in the conventional mice, but this eraser is not active in the germ-free mice. And in our specific case, we are looking for a histone deacetylase or HDAC. So our model is that there is one HDAC that is cycling in the conventional mice, but it's not active in the germ-free mice. So we are trying to identify this enzyme. So first of all, we look at the expression of all the histone deacetylase in the conventional and germ-free mice. And we found that there is one HDAC, HDAC3, which shows reduced expression in the germ-free mice. So this data fits our first criterion, right? So this enzyme is activated by the microbiota. But we have another criterion, which is this enzyme should also be rhythmic in the conventional mice. So we first look at the expression of HDAC3. Now we didn't see a very robust cycling expression pattern in the conventional mice. So then we were wondering if it's not an expression, maybe the activity is rhythmic. So we know that HDAC3 functions by binding to the target genes and deacetylating histones. So we are wondering whether the DNA binding activity is rhythmic or not. So here shows one example gene, SLC25A45. We are not going to talk about the function of this gene, but we just want to use this as an example. So we did a cheap qPCR against the HDAC3 in both conventional and germ-free mice. So you can see the DNA binding activity of HDAC3 is rhythmic in the conventional mice, as shown by the black line. However, the binding is dramatically reduced in the germ-free mice and the circadian rhythm is also dampened in the germ-free mice. So this suggests that the microbiota is able to promote the rhythmic DNA binding activity of HDAC3. So we actually look at many example genes and we consistently see the same pattern. But now the question is, how does the microbiota regulate this DNA binding activity, right? So we know that HDAC3 functions by first forming this co-repressor complex with nCoA1, and then this complex can be recruited by the circadian clock to the target genes and deacetylate histones and repress gene expression. So we are wondering whether the microbiota can regulate this protein-protein interaction and complex formation. So to test this idea, we did a co-IP experiment in the mouse intestinal epithelium. So you can see we IP HDAC3 and brought against nCoA1 and we clearly see an interaction signal in the conventional mice. However, the signal is dramatically reduced in the germ-free mice. So this data tells us that the microbiota can promote the interaction between HDAC3 and nCoA1 and potentially regulate this DNA binding activity. So so far all the evidence suggests that HDAC3 is the enzyme that is activated by the microbiota and drives the circadian rhythm in histone isolation. And to firmly test this idea, we generated these conditional lockout mice where HDAC3 is specifically deleted in the intestinal epithelium cells. And then we repeat the histone isolation chip-seq experiment. So as you can see here, histone isolation is still cycling in the white-type mice. You can see the signal is high at ZT8 and low at ZT20. So this is expected. However, if you look at the lockout mice, you can see most of the signals, they are not cycling anymore and they are constantly high. So this data confirms that HDAC3 is the enzyme that drives the circadian rhythm of histone isolation. And because we know that histone isolation regulates the gene suppression, right, so we are wondering how HDAC3 affects the circadian rhythm of gene suppression in the intestine. So we also collected multiple time points across a daylight cycle from both white-type and lockout mice and did an RNA-seq experiment. So we found that there are a lot of genes that are robustly cycling in the wetter mice, as you can see from this heat map. However, the circadian rhythms are dramatically dampened in the H3-localized mice. And we also found that many metabolic and lutein uptake genes are enriched in this group of genes. So together, those data suggest that the gut microbiota is able to activate a histone deacetylase, H3, in the intestinal epithelial cells, which drives the circadian rhythm of histone isolation and the rhythm of energy uptake and metabolic gene expression. But now the question is, how does this signaling pathway affect our metabolic health? So we first did a metabolomic experiment, right? And we found that actually many nutrient absorption is actually affected by deleting the enzyme in the intestinal epithelial cells. But in particular, we noticed that there is a long list of lipid uptake and metabolic genes. So you can see from this heat map, those genes show very robust cycling pattern in the wetter mice, but the circadian rhythms are dampened in the H3-localized mice. So we are wondering whether H3 could regulate the circadian rhythm of lipid uptake. So to test this idea, we collected a serum from both white-type and the localized mice across a daylight cycle. And we measured the triglyceride level from those serum samples. So you can see, triglycerides are actually cycling in the white-type serum, as shown by the black line. However, the level is reduced and the circadian rhythms are dampened in the H3-localized mice. So this data give us a first idea or evidence that epithelial cell H3 could regulate the absorption of lipids. And we also know that, because lipid absorption is important for diet-induced obesity and fat accumulation. So we're wondering whether epithelial cell H3 is involved in this obesity process. So to test this idea, we fed mice with high-fat diet. I will look at the body weight and the body fat. So if the mice are not fed on high-fat diet, only on chow diet, we didn't see any difference between the white-type and the localized mice, in terms of body fat and body weight. But after we fed the mice with high-fat diet for 10 weeks, we observed that the white-type mice gained significantly more weight and fat compared to the localized mice. So this data suggests that the H3-localized mice are resistant to high-fat diet-induced obesity. And we also know that this process, this lipid accumulation process, is very sensitive to the environment. So we're wondering how does the environmental factors affect this H3-dependent obesity process. So in particular, we are interested in two factors. The first one is the 24-hour daylight cycle, and the second one is the bacteria in the gut. So to study the first factor, we used the jet lag system. And for the second factor, we used a cocktail of antibiotics. So jet lag is very common in the modern society, and it's associated with many metabolic disorders. So to study jet lag, instead of flying the mice internationally, we keep the mice in light boxes, and we maintain this 24-hour daylight cycle. So as you can see from this plot, we first turn on the light at 6 a.m., and turn off the light at 6 p.m., and we keep this schedule for three days. And then we shift the schedule by turning on the light eight hours earlier, and keep the new schedule for another three days, and then shift the schedule again. So we keep this shift every three days, and probably you can imagine that the mice in those light boxes are very similar to the mice that are flying, for example, from Atlanta to Paris, and stay in Paris for three days, and fly from Paris to Beijing, and stay in Beijing for another three days, and then fly back to Atlanta. So to look at how the jet lag impacts this fat accumulation, we fed the mice with a high-fat diet, and we look at how much weight they gain. So first of all, we focused on the water mice, where you can see. We observe that the water mice and jet lag condition gain significantly more weight compared to the water mice and the normal light. So this is expected, because we know jet lag can induce obesity. But what is really interesting is that when we delete HX3 in the epithelial cells, we observe that the mice are no longer responding to the jet lag interruption. So you can see the lung colon mice accumulate very similar weights without jet lag interruption. So this data suggests that maybe epithelial cell HX3 is essential for jet lag-induced obesity. And then we look at a second factor, the bacteria. So we treat mice with a cocktail of antibiotics to remove all the bacteria in the gut. And then we also fed the mice with high-fat diet. So now you can see, after we treat the mice with antibiotics, the water mice gain significantly less weight and fat compared to the water mice without antibiotics treatment. So this is expected, because we know that the gut microbiota is essential for fat accumulation. But we also saw that the lung colon mice, after we treat with antibiotics, they show no difference compared to the lung colon mice without antibiotics. So this means that HX3 in the epithelial cells is essential for microbiota-dependent fat accumulation. But now the question is, how does epithelial cell HX3 regulate its fat accumulation process? So we just did a standard metabolic cage experiment, right? We're looking at the food intake and energy expenditure, but we didn't see any difference in both food intake and energy expenditure. So we're wondering, maybe there's something happening in the absorption process. So we focus on the epithelial cells in the small intestine. We first measured the lipid level in the epithelial cells from both the white-tap and lung colon mice. And we saw that there are much less lipids in the HX3 lung heart epithelial cells. So we're wondering, where does the lipid go? Because we know they eat the same amount of lipid or the same amount of food. And we actually found that the mice basically pooped them out. So we also measured the lipids from the feces of white-tap and lung heart mice. And you can see that there are more lipids in the HX3 lung heart feces. So this suggests that maybe there's a defect in lipid absorption. But how does HX3 regulate lipid absorption? So just like energy transportation, right? So you can increase the lipid absorption by either increase the intake amount or you can reduce the cost during transportation. So that's the way we are wondering which of these processes are regulated by HX3. So to understand this question, we go back to the genes that are regulated by this enzyme. And as you can see from this heat map, it's obviously there are two class of genes. So for the first class, basically the genes are out-regulated in the HX3 lung heart mice. And because we know that HX3 can only function as a transcriptional repressor, so it can remove histone isolation and repress transcription. So we're wondering whether this is the mechanism for the first class of genes. So here is one example, CPT1A. So CPT1A is an enzyme which can bring lipid from the cytosol into the mitochondria for fatty acid oxidation. So we observe that histone isolation is cycling in the wartime mice. You can see the signal is higher at 88 and lower at 80-20 at the promoter of this gene. However, the histone isolation is not cycling and is constantly high in the HX3 lung heart mice. So this suggests that HX3 can remove histone isolation and repress the expression of CPT1A. So this kind of suggests that HX3 can repress fatty acid oxidation in the intestine. But how about the second class of genes? So why do they show reduced expression in the lung heart mice? And what's the biological function for those genes? So here we notice a very interesting gene called CD36. And we're going to focus on this gene in the rest of the talk. So CD36 is actually a membrane protein which can transport lipids from the lumen into the epithelial cells for downstream transportation. So consistent with the gene expression pattern, we also observe that the protein of CD36 is reduced in the HX3 lung heart mice. And the level is also reduced in the germ-free mice. So this suggests that maybe the microbiota is able to activate CD36 potentially through a HX3-dependent mechanism. So it can promote lipid uptake. But we know that HX3 is a repressor, so how can it activate the expression of CD36? So to answer this question, so we're first wondering, why does HX3 directly regulate CD36 expression? So we did a QPCR experiment at the promoter of CD36. And you can see that in the conventional mice, as shown by the black line, so HX3 does bind to CD36, and the binding is also rhythmic. However, the binding is dramatically reduced in the germ-free mice, and the circadian rhythm is also dampened. So this suggests that HX3 does bind to CD36, and it's microbiota-dependent. So we're wondering whether this binding could affect the level of histone isolation or not. So we also look at two histone isolation marks at the promoter of CD36. But we didn't see a significant difference between the white type and black-colored mice. So this means that there's a direct regulation at this promoter, but it's not through histone isolation. So what could be the alternative mechanisms? And to answer this question, we noticed that there is another paper showing that HX3 can actually deacetylate a transcription cofactor called a PDC1-alpha. And this deacetylation can increase the transcription activity of PDC1-alpha and ER-alpha, and it can activate the expression of UCP1 in the adipose tissue. So we're wondering whether this could be the same mechanism for CD36 expression in the intestine. So to test this model, we first did a co-IP experiment. And so we basically IP PDC1-alpha and brought against the HX3 in the mouse intestinal epithelium. And we found that these two proteins do interact with each other. And then we also used a QPCR to show that all the three transcriptional cofactors, they are co-localized at the promoter of CD36. And finally, we used a Lucifer's assay, and we showed that HX3 can promote the transcription activity of PDC1-alpha and ER-alpha for the promoter of CD36. So basically, we confirmed that this could be the mechanism for HX3 to activate the expression of CD36. So to summarize the story, I've shown you that the gut microbiota is essential for the circadian rhythms of histone isolation in the intestinal epithelium. So on one hand, it can promote the basically binding activity of HX3, which remove histone isolation and regulate the circadian rhythm of lutein uptake and fatty acid oxidation. And on the other hand, we also found that HX3 can deacidate long histone proteins, such as PDC1-alpha. And this will increase the expression of lipid uptake genes, such as CD36, and increase lipid uptake. So right now, our lab is trying to identify the microbiome factors that are driving this reason. So for example, whether there are any specific bacteria that is required for HX3 activation and for the circadian rhythm of histone isolation. And whether there are any specific molecules that can be produced by the bacteria and drive this kind of rhythm. So hopefully, we will make some progress and I can show the data in the next meeting. So with that, I want to thank everyone in my postdoc lab and also my postdoc mentor, Laura Hooper. And I also want to thank the people in my current lab, which are currently carrying out the experiment, as I mentioned. And I also want to thank our collaborators, who helped us at different steps of the project. And also my funding sources. So with that, I'm happy to take any questions. Thank you. Thank you. All right, go ahead, please. Okay, thank you. Great talk. I really enjoyed it. This is Hong Xiaoran from Indiana University. I think you have demonstrated the remarkable effect of HX3 in the gut epithelium cell knockout and resistant to diet-induced weight gain. I just wonder, since you observed that the phenotype is mostly attributed to the lipid absorption, so you think, I just wonder if the HDAC histone modification is mainly rely on the enterocytes in the epithelium. There are also important cell types in the epithelium, like enteroendocrine cells, important for weight regulation, and appendic cells, as you mentioned, immunity, and LGR5 stem cells. Any thoughts about that? Yeah, that's a great question. So actually, so here we only focus, we mainly focus on the enterocytes, but we did observe that other cell types could potentially be regulated by HX3 too. So here, the RNA-seq and the ChIP-seq data are all done by the bulk epithelium cells, so we cannot distinguish specific cell types. But I think the last step is to figure out, to focus on single cell types and see how HX3 could function differently in different cells. But so far, we don't know the answer, yeah. Gary Winton from Adelaide. Thank you for very elegant and interesting data. There are two components to circadian rhythmicity. One is light-induced, which is the dominant effect, but there's also diet-induced, a rhythm which operates independent of light. Do you have any data as to whether the effect you're seeing is an entirely light-dependent effect, or whether it's dependent, at least to some extent, on the diet-induced rhythm? Yeah, that's a great question. So I think, the short answer is, I think both light and diet are involved in regulating this kind of circadian rhythms. So one evidence is that we use light box. So when we switch the light schedule for two weeks and make sure the mice are adjusting to the new schedule, and we can switch the circadian rhythms. But we were also wondering whether diet is very important for this kind of regulation. So because people have already shown that diet can switch the rhythm of the microbiome, right? So if you keep the light schedule while you switch the feeding cycles, you can basically switch the rhythm of the bacteria in the gut. So we're wondering whether the diet can also switch the rhythm in the intestinal epithelium. So the experiment we did is, we keep the light schedule same, but we switched the feeding cycle. And we didn't switch the circadian rhythm of histone isolation. But what we observed is that the circadian rhythm is lost. So we feel that the epithelial cells got confused. So they got signal from the light and also the signal from the bacteria or from the food. And it's kind of two signals that kind of cannot be really interpreted in the epithelial cells. That's why we didn't see either a switch or at the same circadian rhythms in the epithelial cells. So are you suggesting that they're interdependent? I think definitely these two signals converge in the epithelial cells. And we are still trying to figure out how these two signals crosstalk in the epithelial cells. Thank you. Maybe the circadian clock is involved, right? Like the alpha is able to recruit HDX3. But maybe there are some molecules from the bacteria which also influence this process. So that's the two ideas we are thinking about right now. So I have a question that sort of builds on that. When you had your jet lag wild type and you showed that there were changes that occurred, did you put them back on a circadian cycle to see whether those changes reversed? That's a great question. So we didn't do that. So we just keep the shift for eight or 10 weeks and we saw the difference. I'd be very interested to see. Yeah, that would be very interesting to see whether they can restore afterwards and then back in the normal schedule, yeah. Thank you. Last question. Thank you very much for the nice talk. Alessia Perino from EPFL Switzerland. So I was wondering if you looked at bile acids and because of course, I mean, they regulate lipid absorption and they are gut microbiome products. So I was wondering if you had a look at them. Yeah, that's a great question. So, but unfortunately, we haven't looked at the bile acid right now. Yeah, but I think it definitely play a very important role in lipid absorption. So that's something we could definitely look at in the future, yeah. Could collaborate on that. That's great. Yeah, thank you. All right, there's no other questions and we'll adjourn the session. Thank you so much for joining us today. Thank you.

gut microbiome

gene-environment interactions

metabolic syndrome

insulin resistance

metabolic health

insulin signaling

circadian rhythms

lipid metabolism

histone deacetylase HDAC3

energy uptake and metabolism

intestinal epithelium