Scientist Stories: Mitchell Lazar, Transcriptional Regulation of Metabolism

Good afternoon, everybody. My name is Bart de Plancken. It's a pleasure to chair this session this afternoon, and it's an even more higher pleasure actually to introduce our second keynote of this conference, Mitch Lazar, Professor Mitch Lazar from the University of Pennsylvania. I think most of you know Mitch. I think he's sort of the father of our conference sitting next to the mother of our conference, and we see both of them actually one after the other actually this afternoon. So that alone is already fantastic, but I think Mitch is much more than that. He's also sort of the father, probably he's father of many things, but—

exactly.

Father, I think, of the nuclear receptor field. I think, you know, having followed actually Mitch as sort of a pioneer in metabolism, introducing functional genomics also, and actually trying to understand how transcriptional regulation plays a pivotal role actually in metabolism, circadian rhythms, rhythms. I think, you know, Mitch, you've been sort of pioneering that field actually tremendously. So it was always a pleasure to actually follow that field. And especially also, of course, your role in actually, yeah, characterizing PPAR gamma, you know, its role actually in fat cell formation. And then beyond that, actually suddenly also actually realizing that, of course, these transcription factors, nuclear receptors don't function by themselves, but also actually, of course, interact with highly complicated complexes. Co-regulators and so forth, which actually add this additional dimension to actually the complexity of transcriptional regulation. And we might actually hear something of that actually this afternoon. So I think that has been very inspiring to many of us. And so as such, it's a fantastic pleasure to actually have you over, Mitch. And not only that, actually, I find out in sort of introducing, you know, or making up this little spiel, that actually it's a homecoming for you because you went to high school in Seosett. Right here. We had an interesting conversation, so this is your home calling. Mitch, thanks so much.

Well, Mort, thanks for the great introduction. Yes, I went to high school a mile from here or so, and it's been great to come back. Until COVID, I pretty much came back almost every year, but it's great to be back after the hiatus. I know a lot of people feel the same way, and I want to add my thanks to Bart, Katie, and Jared for not only inviting me to do this, but There we go. You're all right. How about now? Oh yeah. Well, you probably didn't hear me. I was thanking Jared, Bart, and Katie, not only for inviting me to give this keynote, which I'm very honored by, but for putting together this really fantastic meeting. And as the father, along with Susanna and Dan Kelly, who's the third of our funny family, I guess, that we had. It's actually the 10th anniversary of this meeting. I had not realized this until I— this is actually on the wall of my lab, so it gets some use here. But it's really additional thanks to the organizers because you've kept it relevant. It's just been a fantastic meeting. And actually, I would like to thank I'd like to take the opportunity to have everyone give a round of applause to the organizers. Okay. Three great postdocs in my lab are here at the meeting. I'm going to largely talk about Amy's work today, but Lauren is also here, and Yang, and I hope many of you got to see their posters, but there's still time to have them pull out the poster on the dance floor or something like that, so please interact with them. I know they're having a great time at this meeting. I'm going to take the prerogative of the extra time that I was given rather than to, let's say, add another story, which is very tempting, I must say, but I've decided to be maybe a little provocative and a little philosophical and stimulate some discussion on a more general area and then go back to kind of the usual science. So, let's see how that works out. I wanted to think about where we are in the metabolism field. Of course, it's been 10 years since this meeting started, but it's been around for a long time. And to think about the different things that we all do, and I'll just give examples, but in terms of aspects of our field that are in their infancy and things that are on the other end of the the life spectrum. And these are just meant to be examples, and I might piss some people off, so sorry if I do. But I'll start with, I think this will be relatively safe. I'm putting vitamins at the end here, and the reason I'm doing that is because largely vitamin science, which was huge in the 1920s, 1930s, 1940s, is essentially settled, and I'm saying relatively inactive. There's work on vitamins still, of course, but They got A, B, C, and D, and E, and they were kind of hoping to get to Z and then AA, but they didn't. And that field is sort of settled, but it has a huge medical impact. There's recommended doses of vitamins every day. It affects how we think about so many things, so it was huge, but probably not a lot of people are working in the vitamin field. I'm gonna suggest, and again, this is an incomplete list, these are just examples I picked out, Things on the other end of the spectrum, and I'm listing here transgenerational inheritance of metabolic characteristics, the role of extracellular vesicles in metabolic signaling, and the microbiome, which we've heard great things about, but the actual role aside from treating, let's say, C. difficile in obesity therapy or cancer therapy, I would stipulate is still early. So on this other end of the spectrum, we have Really exciting new mechanisms, lots of potential, but at this time the medical impact is unclear. And as we go across, and maybe you guys could all be thinking as an exercise of what you would have put on here if you were doing this, I put microRNAs here because I think they're a little further but still not a lot of therapies or clear physiology associated. I think the viral 20, 6-year-old of the field is GPCRs. I mean, almost every field we've heard some things today, but almost every talk in one way or another touches on GPCRs. There's another 700 of them without ligands, so there's going to be more of this for a while. So, yeah, I think that's there. But I'm going to talk about nuclear receptors. And I'd like to suggest that nuclear receptors are kind of the middle-aged member of this spectrum, and that's what I'm going to to talk about. The first nuclear receptors were identified and cloned in 1985 by Pierre Chambon and Ron Evans, and we've learned a lot about nuclear receptors since their first identification and characterization. We know that in— we heard earlier from Marion that there are hundreds in worms. In man, there are 48, and we We know they're regulated not only by hormones, which is the way we first got to these original receptors, but by lipophilic metabolites and also drugs. Nuclear receptors are the target of 13% of all the prescriptions for drugs written in the United States, especially when you consider all the estrogenic preparations, tamoxifen, thyroid hormone, and so on. We also know an awful lot about how they work mechanistically. There's the fairly settled, not completely, co-regulator switch mechanism in which co-repressors bind to an unligand receptor and repress transcription at enhancers or silencing elements that get converted to enhancers and activating elements upon the binding of the ligand of the hormone, for example. And This is stuff you find in textbooks. That's why I think it's in the middle-aged kind. There's a lot, and there's also, as I mentioned, a lot of medical interest. And in metabolism, this is a really nice— I'm sure it's incomplete, but it's amazing that Vincent Geyer was able to put this together with his colleagues, just all these things that we're talking about today where one or another or multiple nuclear receptors is playing a role in one or another tissue. There's a lot out there, and there might be a tendency to say, why study nuclear receptors? Why not study some of the cool new things on the left side of the life spectrum? Well, I thought I would raise— I have a list of 5 things. I really could have gone the whole talk on this, but I had to stop. Later at the bar, I can give you my— the also-rans on this list. But questions about nuclear receptors that when I got into this field in the, in the mid to late '80s, I thought would have been answered by now, and certainly along the way. And one of them, we don't talk about this enough I don't think, is the question of how thyroid hormone increases energy expenditure. We have heard at this meeting about a lot of interesting ways to activate, usually people talk about brown fat, and other hormones that can do that., but thyroid hormone is the most powerful hormone in the regulation of basal metabolism that we know of. And this is from 1921, and this is, you know, one— I think we had another— Evan Rosen showed a one-figure paper. This is a one-figure paper. And basically they had someone— and amazingly, because a lot of us had to build cores to measure basal metabolic rate, that's all they could do in 1920 is measure basal metabolic rate.. And when they measured it in people who were hyperthyroid but took out the thyroid, the basal metabolic rate went down dramatically. And if they had people who had low thyroid and they gave them thyroid hormone, it went up dramatically. And so we've actually known that thyroid hormone was a powerful driver of oxidative respiration since before the discovery of insulin, which just celebrated its centennial in 1922. But I would I should say, we still don't know the tissues most responsible for this in people. And again, if people know, then you can point it out at the question and answer period. And despite the identification of thyroid hormone receptors, which I had a little something to do with, we still don't know the underlying mechanism of this. We don't know what genes are activated. It's kind of embarrassing, frankly. I thought we'd know that, but it's not settled science. Here's another one, and this became relevant with Mary Terrell's talk earlier and some of— and and so on. But, you know, she was talking about glucocorticoid effect on fat, but in people, what glucocorticoids do is they affect different fat depots completely differently. They make central fat more and peripheral fat less. And again, love to hear what people think. I would say no one understands the mechanism of this, but this is called Cushing's syndrome. And what's interesting from a circadian point of view, for example, is that if you If you have a Cushing's syndrome because you have a pituitary adenoma, you lose your circadian rhythm and you look like this. But if you are given prednisone once a day, you look like this. So it's not all circadian. And we also don't know what's the difference in these depots. I have confidence that some of the beautiful work— in fact, on the first night we heard from the Dupont lab about IGF-BP2 as a way in omental fat in humans of maybe having a differential effect compared with peripheral, I think that's a great start, but we're not there yet. Number 3 on my list: what's the natural ligand for PPR gamma? In 1994, before 1994, there were no papers on PPR gamma. Now there's 25,000 hits on PubMed. In 1994, there were these two papers from Bruce Spiegelman's lab and my lab identifying it as a fat-specific transcription factor. Then a year later, Steve Cleaver and his colleagues that were at Glaxo at the time realized that drugs that were already on the market for treating diabetes were high affinity ligands. We know that other members of the family have ligands, thyroid hormone, glucocorticoids, etc. That was 1994, 1995. It's 2023. We've had a number of molecules come down the pike and maybe they're all ligands. That's actually a possibility that they're all physiological ligands. I have— I'm kind of agnostic to that, but we certainly don't have a ligand like we do for thyroid receptor or many of the other receptors for this really important receptor, and I thought we would 25 years ago or more. What are these receptors doing on chromatin? Up until the mid-2000s and the explosion of interest in functional genomics, we had about 3 or 4 binding sites for PPR gamma in the genome that were painstakingly discovered by old-fashioned cis/trans assays and so forth. And then I'm highlighting here two back-to-back papers from my lab and from Susanna's lab that demonstrated that there were thousands of binding sites for PPR gamma. And we don't really know how— what it's doing at many of those sites. And you— I hear at many meetings and read papers where people see a ChIP peak and they say it's functional at that site. ChIP-peak doesn't mean it's functional at that site. It just means it binds at that site. And I think there's a tendency to think we already know this stuff among maybe even young new scientists, journal editors, and so on. Oh, this is old stuff. We know everything about nuclear receptors, but we don't. And then finally, because I'm going to address this in the talk, what are the specific roles For components, we've particularly been interested in the co-oppressor complex because we identified it first in 2000, believe it or not. And it has, in addition to nuclear receptor co-oppressor, and there's two flavors of those, NCORP1 and NCORP2. Number 2 is sometimes called SMARt. It has histone deacetylase, a specific one, and other molecules that I'll talk about more in a minute. So we have all these questions still. Yes, the textbooks are replete with telling us how nuclear receptors work, so maybe we don't have to do any more work here. But I've just given you 5, I think, pretty good examples of why we need to do more work here. So what do you do? What do you do if you're worried that people who are reviewing grants are going to say, why do we have another nuclear receptor grant here? We know everything about nuclear receptors, or journals, or what have you. And so Will Rogers, who was the famous American humorist, said, If you find yourself in a hole, the first thing to do is to stop digging. So that could be what you do. You just say, "I'm out of here." I think what a lot of people do, and there's nothing wrong with this, I think science has to be a combination of this, is to move to another place and start digging a new hole. I think if that happens too often, you're going to get a lot of superficiality. And it's just a style thing. Some scientists like to do that more than others. But I personally, and that's why I said this is provocative and philosophical. I think we need both. I think we need people who— and labs and groups that dig deeper as well as the ones digging new holes. Now, as we dig, sometimes you get a few little morsels of stuff, and when that happens, that's called incremental. Incremental doesn't look like it, but it's become a four-letter word in science these days. It's a word that, oh, that's incremental. Now, 10 years ago it was descriptive. People would say, that's descriptive, and that was a way of saying forget this crap. And now that's incremental. But I would highlight, and I hope everyone has read or should read, especially the trainees, this beautiful article. It's really like one page by Bill Kaelin, Nobel laureate, but also a brilliant and deep thinker. And what Bill says here is that what we really need instead Science is more carefully done things, and if you're working on an important project, just learning one new thing that's going to help you get to the next new thing, but you're confident about it and it's robust, is a good thing. But again, that's not many of the— what we'd like to see today is a story that goes from A to B to C to D to E, but what Bill argues, I think, pretty Nicely here is that that will lead to non-reproducibility because you can't possibly do every one of those steps as carefully as you need to do an individual step. We're in a crisis of non-reproducibility right now in science. I mostly get information on that from Twitter these days, but it's everywhere. The crisis of non-reproducibility in science is spilling over to a loss of faith in science in the general public. I've had several conversations just on the beach the other day and so on about this. And I think it's all, in my opinion, highly related to this issue. So I would ask people to really think about this as they evaluate other people's science and think about what your science is going to do. And of course, if you're really lucky and you keep doing incremental science, you could eventually hit the pot of gold. And that's what we— that's what we all want to do, a new discovery And, you know, these were the clues. Okay, so that was my soapbox, and now I'm going to segue into this question, one of the top ones I've wondered about for a long time. You have this whole complex delivered, the co-repressors are binding the nuclear receptors, and Then they're bringing in, and many people including me have referred to them as boats that are docking on the nuclear receptors that are recognizing the specific genes based on sequence-specific DNA binding transcription factors, and they bring in all these other things. But, you know, I hate why questions because we can never answer them, but there's a bunch of things here. And do they all do the same thing? You might think they do. They're coming together. Well, I'm going to tell you one very short story about something that's published, but I think it fits this theme, and then unpublished work. The first thing is about the relationship between histone deacetylase 3 and N-CORP. When we first found histone deacetylase 3, and not 1 or 2, et cetera, it's HDAC3 specifically, in the N-CORP complex, it was pleasing, if you will, because it was right at the time when HATS had just been discovered and recognized to be regulating activation, and it was then thought, well, if they regulate activation, then there must be a deacetylase that regulates, that inactivates genes. And Stuart Schreiber discovered the first HDAC. That was HDAC1. They were just named numerically. And HDAC3 is highly related to HDAC1, but it's not the same, and HDAC1 cannot replace HDAC3 in this complex. So it made a lot of sense that this co-repressor complex should have a stoichiometric relationship and belonging, if you will, of a histone deacetylase. Matt Gunther, who actually purified this complex, then when he was studying it further made this discovery a long time ago now, which is that the HSTAC3 polypeptide by itself had essentially no detectable enzyme activity. This is an E. coli, but it could have been a reticulocyte or a baculovirus. But it turned out that he actually tried every combination, but when he mixed it with SMART is another name for N-CORP2. When he mixed it with N-CORP2, he now could activate HDAC activity. And furthermore, we mapped this to a region that we refer to as the deacetylase activating domain, which is a small region of N-CORP2, it's about 70 amino acids, it includes a SAM domain and then some other very specific amino acids that are not in other SAM domains specific to N-CORP2 and SMART. Its structure was crystallized by John Schwabe who provided this beautiful images of how the DAD and HDAC3 come together. Again, I'm not going to go too much into that, but it's really quite beautiful. And since we know how it works, we can make catalytic site mutants, and I'm going to show you what happens with the catalytic site mutant, this tyrosine to phenylalanine mutation at Y298F, but you can see it's dead. So as we were trying to get beyond the biochemistry to the physiology and try to understand what's happening in the living organism, Zheng Sun, when he was a postdoc in the lab, he's now at Baylor with his own associate professor in his own lab, knocked out HDAC3 in the liver. And we like to use a technique that was pioneered at the Penn Vector Core at our Penn Diabetes Center, and many people have used it. I think I saw others in this, and I'm glad people are using this because it's a way to knock out genes specifically in hepatocytes, not just in the liver, but in hepatocytes, and to do it in the adults so you're not worrying about developmental things that you might have to worry about if you use, for example, albumin Cre. And this is the AAV-TBG human thyroxine binding globulin Cre, and you, Zhang knocked out HDAC3. When he did that, you can see the knockout is quite good. And what you can see is we have, when we grind up these livers, there's a lot of HDAC activity. Sorry, this is actually a pull-down of HDAC3, of N-Corps rather. There's a lot of HDAC activity and there's nothing here. And look what happens to liver triglycerides. They massively increase. This is Oil Red O staining for neutral lipids. This is the actual amount of triglycerides. It goes from 5 to 10-fold. Again, most of us, and certainly in my lab, we are attracted to— I think someone else said this earlier in the meeting too— robust findings. It's fine to have a 10% change, 20% change, but it might not be biologically significant and it's extremely hard to dissect the mechanism. But here we had something we could play with. So what Cheng did first was to make sure it was really due to the absence of HDAC3 because these are complicated. They've got, you know, floxed genes and you put in a virus and et cetera, et cetera. But he could replace HDAC3 with a flag tag, so it's a little a little bit bigger here, and he picked an amount that was roughly similar to the endogenous amount of HDAC3. And you can see that it restored HDAC activity, and importantly, it rescued the phenotype. But the surprising result is that when he made— he had made the catalytic mutant, that Y298F that I showed you earlier, and when he put that in, and actually it didn't even express to the level of the wild type, he was able to largely rescue— remember rescue here would mean reduced triglycerides and increased oil red O staining. It largely, not completely, but largely rescued the phenotype of the HDAC3. So that told us that HDAC3 didn't need HDAC activity to do this very important function, the major function we found for it in the liver. And work I'm not going to show you, but we've done the reciprocal, which is to make mutants of N-COR in that dead domain that cannot interact or activate HDAC3, and we have the same thing. They don't get fatty liver even though there's no detectable HDAC3 activity. So what's in a name? And again, I think we've seen other examples of this in this conference already too, the repurposing, if you will, or the fact that different proteins may have different functions. But here's the histone deacetylase 3, but we, I I've shown you that it has functions that do not require deacetylase activity. And so now we're trying to revisit the idea of how the co-repressor complex represses and figure out what else HDAC3 might be doing as a deacetylase as well as what else it's doing that doesn't require its deacetylase activity. Okay. And the last part of the talk I want to talk about is Another newer unpublished set of experiments addressing the questions of what nuclear receptor co-repressor complexes components actually do. I already introduced why we might be interested in that and why it's not settled science. I just showed you that the HSEC3 knockout has a markedly fatty liver. Eckhart Treuter at the Karolinska knocked out GPS2, and actually they're protected from fatty liver, those mice. Stefan Herzog's group knocked out TBL1, and they found a fatty liver, but it was modest. It was sort of a 50% increase, not this 5-fold massive effect. And a couple of groups have knocked out the NCORs separately, and they didn't do much. NCO2, or SMART, has no overt phenotype, and NCO2 has the same direction for fatty liver, but again, very mild, nothing like the HDAC3 knockout. So that raises the question of whether there's redundancy of Ncor1 and 2. Here's what they look like. They're big proteins, 270 kilodaltons. They're very, very similar. And they certainly have some unique features because the germline knockout of Ncor is lethal even though Ncor2 is still there, and the germline knockout of Ncor2 is lethal even though is lethal even though Ncor1 is still there. So, there's been a lot of thinking that they're not redundant, and here the question was, was there redundancy? I do want to take one moment to point out that this figure here is taken from a really nice review by Marty Provolsky. Marty Provolsky was a terrific scientist in the nuclear receptor field who unfortunately passed away from myeloma during COVID And because it was COVID, because it was the COVID time, a lot of people do not know that maybe. But he's really a terrific person. He was a self-described nerd, but he had a very excellent sense of humor. And I just thought in his memory I would tell you one thing. He was a very careful scientist, but one time he was giving a talk at a Keystone meeting, and in the question and answer period I asked him about whether he did a certain control. And Marty said, "Controls? We don't do those in my lab anymore. We've had some rather unpleasant experiences with them in the past." So goodbye, Marty. Rest in peace. Bad segue, Amy, if you're here, but I hadn't really thought about that. But I'm going to tell you about Amy's work. She had a poster the other night. I hope many people got to talk to her. There was a lot more data on her poster than I'm going to show you here. I'm sure you're saying, "Thank goodness," but here we go. So Amy used that same system but this time to knock out Ncor1 and 2 in suitably floxed mice. And when she did this and compared the transcriptomes of the Ncor1/2 double knockout to HDAC3, Ncor1, controls, including in this case you can see that the Ncore2 knockout and the controls all group together in this PCA plot. Ncore1 and HDAC3 are very close, HDAC3 farther away, and that's, we think, and I'll show you data that is consistent with, that's because they're both causing a fatty liver. It's just that one is much more powerful than the other, possibly because Ncore2 is not there.. And NCORP1 and NCORP2 are in a different zone. And when you drill down on these genes, and you can do this many different ways and we debate how to do this in our lab, but I'm gonna point out that the genes sort of in common, even with GPS2 which has this opposite phenotype, are genes involved in fatty acid oxidation. And that we think could explain why it has its own separate, 'cause it's activating those genes to repress or activating those genes. And that would fight against, if you will, the de novo lipogenesis and so on that's at the core of what NCORR does and what HDAC3 does. But what came out of the NCORR 1,2 double knockout that we did not expect were a lot of genes involved in glucose metabolism. We saw no real hint of that from any of the other subunits. And many of these were downregulated, which made us Remember, this is, of course, I've been saying it's a co-repressor complex, so we don't expect when you knock it out for things to go down. We expect them to go up because it was normally repressing them. So that's surprising. Now, what are these genes? Everything in blue here, and these are just different ways to do gluconeogenesis. Many people here know a lot more about this than I do. We've heard a lot about this at the meeting. But all of these pathways, you can see all these things in blue, they're they're all downregulated when you take away Ncor1 and 2. And that's shown here. The fonts are a little bit small, I'm sorry, but gluconeogenesis, amino acid catabolism, and these were specific. These don't happen in the HDAC3 knockout. If you look at the HDAC3 knockout and the Ncor knockout, look what they have in common. Again, it's fatty acid metabolism. So yes, the repressor complex represses genes involved in fatty acid metabolism, Ncor, and HDAC3. HDAC3 do that? It's acting as a repressor, so that's the part that's in the textbook that's absolutely or relatively correct. But what's going on with these downregulated genes? And it's not subtle. This is, again, this is not a z-score thing. This is major robust changes on the order of 3 to 5 fold reduction. And consistent with this, When Amy fasted these mice, they have a marked, they also have a reduction in glycogen stores, which I'm not showing you, but they have a marked hypoglycemia almost immediately with the fast 'cause they don't have glycogen stores, and then it continues as they do other gluconeogenic or attempt to do other gluconeogenesis. So what's going on there? How is Ncore 1 and 2 normally activating these genes? To get at this, Amy performed ATAC-seq to look at just what chromatin is differentially opened when you take away Ncor1 and 2. And that's what this is. So this is ATAC-seq with a differential signal to get on this heat map. It's a differential signal, and only Ncor1 and 2 binding sites are shown here in the first place. And you can see there's many more sites that produced ATAC-seq signal, maybe somewhat surprising, although to me And this is something we think about and I don't think there's a settled answer. Opening of chromatin could happen by a repressor just as much as an activator since it still has to bind to DNA and so on. But anyway, a lot more reduced ATAC-seq signal. And here's what it looks like on two arguably of the most important gluconeogenic genes, PGC-1 and PEPCK. And you can see that this is the RNA-seq and this is just showing you the massive reduction in the double knockout of both PEPCK and PGC-1. And then ATAC-seq, we see reduction in these peaks that correlate with N-core binding sites. So that's built into this, but these are often easier to look at the actual screenshots. So then Amy looked at genes that had the reduced ATAC-seq signal. Near genes that are downregulated in the double knockout liver. So this is one flavor of functional genomics, of course. And then we did that, we took advantage of a terrific website run by Shirley Liu, formerly at Dana-Farber, but it still exists. I'm not sure who's running it now, actually. It might be Miles Brown. It's called CistromeDB, and what this does is rather than look at a whole bunch of genes and then bioinformatically look at the sequences and guess what transcription factors might be there just by the sequences, which is not a good way to do it even though a lot of people do it. This actually takes advantage of all existing cistromes that are out there. It grades them in an unbiased way for quality, and there's lots of them for the liver. And then it says of all the things that there are already cistromes out for, which one correlates best with whatever you're interrogating. Here we're interrogating these sites where the ATAC-seq signal went down. What you can see is it was, you see Ncore, that's nice, but here's NR3C1. NR3C1 is the glucocorticoid receptor. We've heard a lot about the glucocorticoid receptor in liver from Claudio and fat today from Mary. But I wanna remind everyone that glucocorticoids, even though we heard other things they do and they're very pleiotropic in their effects, but they're the major regulators of hepatic glucose production. So much so that their name, which was given to them by Kendall when he purified them, is based on glucose cortex steroid, the fact that they are involved in glucose metabolism. So people sometimes actually forget that, but that is what they were named for. So what Amy did was— to now say, okay, I'm going to see if these mice are responsive to glucocorticoids. So she treated them with dexamethasone acutely, and she found that there was blunted dexamethasone response in a number of the same gluconeogenic genes that I mentioned earlier. But there were other genes where they were not blunted. There was still the effect, but these had other functions, which again, you might expect that. So this robust and selective reduction in dexactivation of glucose metabolic genes in these N-core double knockout livers. And as a result, or related to that, when she ChIP'd for GR on these sites, and it's complicated here, but I'll just tell you that you're seeing it's chipping and it goes away in the double knockout. So it's another name badge. This is the nuclear receptor co-repressor complex, and it definitely is a co-repressor complex and definitely uses HDAC3 and its enzyme activity for some of that repression. However, what I've shown you here is that they also have activating functions. What we know of it so far is that this is due to an unexpected role in keeping chromatin open in the liver. It plays a major role in the physiology of glucocorticoid action as well as fasting response. I'm not sure what happened there, but not all co-repressor complex proteins have the same function in metabolism, which was not actually clear before. And in some ways it's counterintuitive. You think they're in this complex, it's stoichiometric, you'd think they'd evolve to do the same thing, but I hope I've convinced you they definitely don't. But there's much more digging to do. What's recruiting NCoRs to the DNA? How do they open chromatin? Development, and et cetera, et cetera. So finally, I just want to put it back in the context of how I began this lecture. It's easy to walk away after you've dug enough, especially if you already found some stuff here. But there's often more digging to do and more things to find. And one of my heroes, scientific heroes, Claude Bernard, the first experimental biologist and the great French physiologist, had several great quotes, but this one I think is pertinent here. It's what we think we know already that prevents us from learning. And I think we've already seen a lot of that, not just in this talk, but throughout the conference, and I think that's great. So keep on digging. And I'd just like to thank my lab group, I mentioned Amy. These are other postdocs. Actually, Yang is here. She contributed to this work. Past members of the lab, we got the NCORP2 mice from Ron Cohen. One last plug for those posters that you can still look at the abstract and you might buy a beer for any of these folks and they'll pull out their posters for you. Our work is funded by NIH and JPB Foundation and ADA. And I'll stop there. Thank you for your time and attention.

Happy to take any questions. Thanks so much, Mitch, for this great philosophical overview. Great story. And just a note to all the speakers, there's 10 seconds left, right?

As a keynote.

Darn, I tried to—

Questions?

Bart, over here. Oh, yeah, sorry. Thanks, Mitch, I appreciate that. I also appreciated the philosophical treatise. My question is, and this is hard, I realize how hard it is to answer this, how frequently are these co-repressor complexes really stoichiometric? You know, is every time that Ncor1 is there, is HDAC3 there, or is it possible that there are scenarios where HDAC3 simply isn't there, and that may be some of the complexes that are activating instead of repressing.

Yeah, I mean, that is a fantastic question. We've had a hard time— when we've done, you know, old-fashioned column exclusion chromatography, it's always in the void volume. In fact, if you add up the— by the way, there are other things that are not stoichiometric that you don't capture when you just do a, you know, typical purification. But if you add up the molecular weights of everything there, it doesn't come close to this greater than 2 million or 3 million that we see. And to make it even worse, it's not just a question of could there be substoichiometric amounts, but TBL1 in particular is a tetramer. And the region that it binds N-core with is not the region that tetramerizes. Tetramerizes with. So this could, you know, you can imagine a polymer if you really took that out to the limit. So I'm sure it's a really big complex. Actually, we're trying using epitope-tagged versions that are endogenously knocked in with better biochemistry to address this, and also with different tags on different proteins to get a sense of that. But it's a great question, but it could go either way, sub-stoichiometric and just massive.

In your cistrome analysis, you showed, you know, a couple of candidates, and amongst them were CBP, which is, you know, a coactivator. So is there a poised state of some of these NQR1 corepressor complexes that are being partitioned amongst the, you know, the SLLAs and the HAT complexes? Have you guys done any work to delineate that?

Yeah, it's a great question. I think there's two parts to the answer. One, I've talked— spoken to Susanna about this actually at this meeting. There are a lot of examples of this kind of thing. But the other thing is that remember when they do those cistrones, they're going to be having situations where probably in some cases there was less glucocorticoid or more glucocorticoid, and so it might be active or not active. And I think that's also part of— in this way of looking at this, because that's not based on the same experiment that that we did. But it definitely could be what you're saying, more of a poised situation. Absolutely, I agree. Mitch, that's great as always. So I'm just trying to understand. So if N-Corps is clearing the way for GR to bind, right, that's the model? It's opening chromatin. Doesn't mean it's binding where N-Corps is. But N-Corps also binds to the GR and acts as a co-repressor. No, it doesn't. It does not? No, sorry. So I was just trying to understand how it— Sorry. No one has really shown, to my— someone has, let me know, but I don't think that's been shown for glucocorticoid receptor. It's really for more the RXR and reverb and receptors like that. Sorry. Great talk. So I just wondering, because in addition to the uncle, like the nuclear receptor co- repressor, there are also a group of like transcription factors such as Synth3a, Synth3b. They can form— they are the key components of the silencing complex. So what do you think the relationship between those like transcription factor groups like silencing complex and also like the group of UNCO and these like nuclear receptor-associated co-repressors, and how do they coordinate with each other like to regulate the transcription of genes? Thank you. Yeah, it's a great question. I mean, there are many co-repressor complexes, and Sin3 and HDAC1, for example, are in the NERD complex as the most famous, but there are others. It is true, and I guess I'll go out on a slight limb here, that I told you some of the history of how people were so excited about HDAC being in the N-Corps complex because that was the model.. But at the time that there was only one HDAC and it was HDAC1, several groups— and you can PubMed this to find out who they were— found in biochemical or sometimes microinjection antibody experiments that HDAC1, and in some cases SID3, was in this complex. But it isn't. If it is, it's in much, very substoichiometric, tiny trace amounts, and that's been shown by many, many labs over the years since then. I view that as, and this is again philosophical, it was sort of the right answer. There was an HDAC in there. There was only one HDAC known, and a lot of things that are not reproducible, and this is completely opinion, are when people kind of think they know the answer and they get that answer. And I think that is what that's an example of. But if you purify the complex and show me a lot of HDAC1 in there, I'll change my mind.

Maybe I take the privilege of asking the last question. I think there's nobody else out there. No? I mean, I'm teasing you a little bit, but I think one of your questions is, of course, very valid in the transcription regulation field in general. Why do you think there are so many binding sites?

You probably have a perspective on it. Yeah, you're actually asking me that question, huh? I think that some of them are for looping and hub formation, and that someday we're going to find— and we actually have one example of this in our own lab at the moment that we think is real where they individually as enhancers, they weren't particularly strong, but together they're forming a loop that seems to stabilize, in our case, repressor binding. So I think that's one possibility. But I think that, you know, we probably needed to— there's probably other things that are happening in chromatin. I'm sure you've thought about this too, so I'll turn it around in a second. But I think there are other things happening in chromatin and that we'll be learning from the work of many good labs, and eventually it'll make sense to us. But I don't have— a lot of times these sites don't have enhancer marks. They don't have increased H3K27 acetylation. They don't have RNAs that are proportional to what's happening to the gene. So I don't think they're functional in the sense of— except for what I just said about the looping— in the sense of transcription.

What do you think? I think that that's right. I think there's also a big component in terms of context, right? Having sort of the right partners in a sense that then actually that leads to functionalization of that particular site. And often these nuclear receptors are just hanging around and waiting for the right partners to actually come in. And I wonder also whether that actually also has something to do with the complex you see, and then it sort of turns it from maybe a co-repressor and a co-activator depending on other partners in crime, so to speak. Yeah, no, it's a great point.

Last question, yeah. Sure. I was wondering if you were thinking of it maybe as a pioneer factor. We don't, you know, the sort of a formal definition of a pioneer factor meaning you can show that it opens chromatin, we've in no way have we shown that, but a bit related to that, I mentioned that the DAD region, the region that activates HDAC3, has a so-called SANTh domain. SANTh domain is this really interesting domain that is found only in transcriptional regulators. Actually, the N in SANTh stands for NCORP. It was a Trends paper by someone who named that domain, but NCORP got in there. Outside of this DAD domain that I'm showing you here, usually STAT domains are histone-binding domains. And actually we showed many years ago, we haven't figured out what it's doing, but we showed many years ago that N-Corps and SMARt, the other STAT domain, can bind histones. So one idea we have that we're trying to test now is that that's how it's getting attached to at those sites. It's something about its ability intrinsically to bind histones maybe doesn't even need a transcription factor, a sequence-specific factor at some sites, but that's just speculation.

All right, let's really give a round of applause for Mitch. Thanks so much.