 I'd like to thank John and the organizers also for the opportunity to be here today and to speak to you about some of the work that we've been doing. We've been interested in gene expression and the role of enhancers in that process for some time now and this interest of course was boosted a few years ago as for many people as a result of the development of chip-chip analysis and all the variations on chip seek analysis that have come out in part due to the great work from the ENCODE consortium and we of course have taken advantage of that technique as well. So I just want to acknowledge people that have been involved in some of the studies I'm going to show you first here and I want to primarily acknowledge Mark Meyer who's done much of the work that I'm going to talk about along with various other individuals in this in this group. I'll talk a little bit about some of the work that we have done with Charles O'Brien and acknowledge some of the analytical work that we've done through Homer with Chris Benner who at that time was actually working with Chris Glass and we're appreciative of that effort. Now what I would like to do first off however is pose this sort of question that is why we study enhancers and of course I don't really need to say an awful lot about that to this group but I just want to point out several things and that is of course gene expression and cell phenotype are primarily governed by enhancers and so a detailed knowledge of mechanisms of regulation are necessary literally for every gene whose misexpression is known to result in human disease and this comes about from the obvious idea that if there's a human root to a gene knockout in mice that results in a phenotype that's an actually an important gene and knowledge of that regulatory capability is very important. Also disease regulatory knowledge regarding genes that play a role in disease progression and there are a number of those as well and finally we've talked a number we talked about this is regulatory knowledge about genes that are linked to RFLP SNPs or indels that increase the risk for disease. All of these are misexpression examples that are very important and therefore we actually think that it's important to actually understand the direct relationships between this regulation and these genes because these details could provide some selective opportunities for therapeutics. So I want to just first of all give you a couple of slides on our entry point into gene regulation and that has to do with the vitamin D hormone and just to remind you that while vitamin D is known to regulate mineral homeostasis and it's involved in certainly in skeletal homeostasis it has a lot of other effects some related to the immune system cardiovascular system and so forth and so it's actually vitamin D is certainly plays a role in in biology that's that's far beyond that that is related to calcium and phosphorus homeostasis. All of these activities of the vitamin D of course are regulated by or mediated by the vitamin D receptor and in fact while we had some understanding very early on that vitamin D might be part of an endocrine system it took the cloning of the vitamin D receptor in the late 80s to to really identify it as as a true member of the steroid nuclear receptor family of genes and in so doing provide a bona fide membership in this family and the vitamin D receptor through domain analysis that was done at that time suggested the amino terminus said the DNA binding domain the business end of the molecule was in fact the C terminal and then the effects of the ligand were to re alter the configuration of this C terminal and to open an act or conform an activation domain we called AF2 that led to activation of gene expression. This is simply a crystallography study that we did with the ligand binding domain to understand how the vitamin D receptor actually worked and other people have done this as well and this is the loudest to understand where the contacts for 125 D3 in this pocket this the activation domain surface with this peptide was identified and so forth so this is allowed us to begin to think about designing analogs although that hasn't actually gone all that well now I want to come back just for a minute and remind you that the primary role of vitamin D has been in calcium and phosphorus homeostasis and what you can see here is that vitamin D plays a really a very complex role in the in the intestine kidney and bone to regulate this and maintain extra cell you have calcium and phosphorus levels and it also does it by regulating PTH levels through the parathyroid gland it also regulates through bone it regulates FGF PTH is primarily involved in in the calcium regulation whereas FGF 23 is involved in phosphate and we've been particularly interested in the bone areas simply because vitamin D actually can mediate not only bone remodeling but of course it also the skeleton is actually the source of calcium and phosphorus and in situations where there's a dietary deficiency of either of these minerals so this is actually a very important area that that needs to be understood and we've focused to some extent on that now when we in the in the 90s we begin to understand three basic principles of the vitamin D receptor and quickly go through these the first one is that the vitamin D receptor interacts on DNA with a partner called RXR RXR plays a role in many other receptors as well but it's a heterodimer that functions it functions on DNA and we were able to identify response elements that were comprised of two hexanucleotide half sites separated by three base pairs so these two principles were important this is a cryo EM structure done by you know morass and his colleagues so this is sort of a contemporary view of it but the third principle was that the vitamin D receptor as many nuclear receptors and other transcription factors simply function to recruit chromatin active co-regulatory proteins at groups of proteins and this is three of them here I won't go through them there are many many of these and we yet don't understand all of them and their individual functions but I think we're we're getting there to understand at least some of them now the problem with this is that this was a very receptor-centric view of life and that is because we were studying the receptor on a couple of genes that we had begun to explore and the problem with that was was we it was very limited to these particular genes and so during the when when the unbiased of methods such as chip chip and chip seek analysis came about we tried to take advantage of that to begin to understand the properties of vitamin D action in on a genome-wide basis sort of an overarching principle again we focused on osteoblasts as bone-forming cells again because they play a large role in regulating through remodeling a bone but also because we could actually look in vitro we could actually look at the differentiation process and compare different stages of osteoblast cells and much like Evan Rosen has has told you about with respect to adipocytes in any event so we did an experiment where we took osteoblasts and just treated them in the absence and presence of 125 and did a chip seek analysis looking at the cystrom for the VDR and here what you can see is that there's about a thousand genes binding sites for the vitamin D receptor in the absence of ligand about 7000 or so in the presence of ligand clearly suggesting that in the case of the vitamin D receptor is that the hormone is very active in promoting DNA binding if we look at RXR we can see the same thing it's a much bigger cystrom probably because it has greater activities and with other receptors but in case when you add 125 D3 you do see that activity and the important aspect of this is if you actually cross these over you actually look at the peak intersex you will see that about 4,000 of these sites have both VDR and RXR certainly these this would support them the idea that RXR is a partner if you do a de novo analysis motif finding analysis in pre osteoblasts as well as osteoblasts these are early cells these are fully differentiated mineralizing cells you can see you see the vitamin D receptor RXR motif which is AGG TCA to AGG TCA motif separated by the three base pairs and in fact you can see that in in both cases there are some additional motifs that are found in these elements in these enhancers as well now this was the most interesting observation and we've talked about this already and that is the discovery in fact that by that the binding sites for the vitamin D receptor as well as many other transcription factors are in fact located distal to the promoter rather than near promoters in fact the bulk of them are either intronic or intergenic and this has had profound effects on our ability to understand the process finally when we look at pre osteoblasts and osteoblasts here's the VDR cystrom what you see is a tremendous contraction of the number of binding sites the cystrom following differentiation and this actually has some important ramifications as well now we were asking then we asked and whether the change in this transcriptome had or this cystrom had an important role to play in gene regulation we did a microarray analysis of pre osteoblasts and then the mature osteoblasts we could see some rather dramatic effects this is a microarray analysis of this as of the gene expression we've done extensive RNA seek since that time but in fact what you can see is both up and down regulated genes are changed dramatically as a result so it's pretty clear that in fact here's the up up regulated genes and here the down regulated genes you can see the contraction in both cases so it's pretty clear that there is a change in gene expression in response to 125 as a function of differentiation now one of the genes that actually goes down the suppressed by 125 because 125 D3 is trophic for the vitamin D receptor and auto regulates it is in fact the vitamin D receptor itself so it suppressed in its expression levels and that has a clear impact on the expression of the gene but the real question here is why all this occurs and in fact one of the things that we noticed was that there's still a number of genes that are targets in the mature osteoblasts and in fact if you look at this is actually quite interesting you see a tremendous change in in the regulatory capability of 125 D3 between early and late cells so for example here in call 2a1 a tremendous down regulation by vitamin D but almost no but a tremendous suppression in the case of the mature osteoblasts and almost no regulation probably a result of the loss of vitamin D receptor binding here and in fact if you look at this chip-seq analysis of this you can see that there's a complete loss of the vitamin D receptor binding activity in this gene however if you look at a couple of other genes this is an interesting one in pp1 and 3 which are right together there is a primary enhancer that we believe regulates both of the expression of both of these genes again you can see there's an incredible increase in responsibly to 125 modest changes in basal level as well and finally if you look at this last gene here you can see almost no expression and then tremendous expression and these don't really correlate with the levels of the vitamin D receptor that are bound here in in each of these particular cases so the real question here is why are these in mature cells much more sensitive to 125 D3 and in a couple of these cases and the that they're able to regulate but even though the the peak height for example is not near as high as it is and the scale over here you can see is actually rather dramatically different so there are many reasons for this and we don't really understand it all but just let me point out a couple of possibilities that first of all is that the we had some clues as to the idea that there were other transcription factors that were present in these in at these enhancers and so we actually did a chip seek of of cebp beta and rungs to now rungs to is a master regulator of the osteoblasts is involved in chromatin opening and cbp betas is involved in chromatin remodeling and what you see here is about 17,000 sites 1700 sites where vdr and rxr actually bound or actually co-occupied by rungs to and cebp beta and all this motif finding analysis reveals pretty much the same thing and what this says when you do a further biome biome informatic analysis is in fact that there's an organization to these motifs in which if you look at on a broad spectrum you can see that there's an organization where the vitamin D receptor is bound between rungs and cbp beta and the the approximation here is quite close there's just a few base pairs in between so we've called this the osteoblasts enhancer complex it's not consistent across the entire genome here but it does suggest that these other players rungs to and cbp beta could have an enormous impact on the regulation mediated by 125 b3 now the second area that I don't have time to really go into is the the genetic and the epigenetic changes that occur during differentiation and just to to summarize these the differentiation and transgift differentiation as we and and the encode group has characterized fully is is that there are significant changes in histone modifications as selected gene loci and those are the ones of course that are changing during the differentiation process the enhancers are highlighted by signature enhancer histone modifications that change as a result of the change of these genes during differentiation these changes in histone marks and the regulatory factors can contribute to the responsibility to 125 b3 and in fact we've also noticed that 125 and other hormones provoke changes in these histone modifications particularly at the level of histone acetylation that again are very histone gene selective for 125 d3 so we think that these are clearly changing the the the epigenome surrounding the genes that are actually changed and influence vitamin D response so this is a list of overarching principles of vitamin D mediated gene regulation that we have identified I just want to reiterate three of these one the distal binding sites the locations of these because they create the complexity that we heard in the last talk and in most of the talks distal regulation where a promoter proximal binding site can be inferred to regulate a gene that's sitting right next to it but the distal regulation makes it very very difficult if 40% of the binding sites for some of these molecules is to greater than 250 kilobases away makes it very difficult to assay epigenetically their epigenetic signature marks largely identified by the end code group and these are really very important and they they mediate a dynamic nature to the genes themselves and finally the vitamin D your systems are highly dynamic based upon the differentiation process and we also believe and I think this is certainly true that disease can alter these effects as well and so these are very important now I want to go through I want to go back to one of the genes that we notice that was regulated and tell you how we look at enhancers and their activities and show that in fact these enhancers are directly linked to the regulation of the genes are many bioinformatic ways but to our way of thinking at least at this point they really are not definitive so we have looked at MMP 13 as another gene we've looked at many genes but this is one I want to give you an example of regulated by vitamin D and differentiation you can see the process over here in the early cells not much regulation here a bit but when you when you differentiate these cells tremendous increase in baseline and then a tremendous increase in vitamin D response now this is a gene called collage an ace MMP 13 a collage an ace 3 the grades extra cellular collagen it is regulated by a whole host of different factors it's certainly aberrantly regulated nearly every cancer or disease with fibrotic complications and importantly it's also regulated and important in atherosclerosis finally it's interesting in people many people here will probably agree previous work on the regulation is focused almost exclusively within the three or four hundred bases near the promoter region of the gene so when we looked at the chip seek analysis that we had had derived from some of our earlier work in early and late osteoblasts here's the gene itself we looked at the vitamin D receptor rungs 2 and c ebp beta in a series of histone modifications and I think it's pretty clear here is the promoter region right here which there is some activity but I think you can see there are three other regions 10 20 and 30 kilobases the 10k region binds the vitamin D receptor 20k region and 30k region bind the c ebp beta and then rungs is bound to the 30k distal region as well and all these are highlighted by marks that contribute to the idea of a histone signature so we were pretty convinced in fact that that it that these were enhancers a question was whether they were mediating the activity of MMP 13 and so what we have done is actually to use the CRISPR Cas9 system to generate a series of daughter cell lines from the from from the from the host cell line and actually look at those and look at the consequence by of deleting the enhancers themselves and here are the ones that we've we've actually created we created a promoter deletion with these that this approximately 400 base pairs we deleted about 200 base pairs around the 10k region that removes the vitamin D responsive elements there's two of them we've identified and then we remove the 30k region again that contains the rungs binding site we also knocked out the vitamin D receptor by that same means and we also knocked out rungs as well to see whether there was a correlation between these measurements and these measurements so this is the data it's a little bit complicated but let me go through it quickly and that is so here's the wild type cells and there's several different clones because there there tends to be some variability and we're trying to be very cautious about this this is basal activity this is the activity in the presence of 125 B3 the first thing you can see is when you knock out the 10k hormonal regulated enhancer you completely lose vitamin D induction but what you gain actually surprisingly is the ability to 125 D3 to actually suppress this gene and I'll come back to this in a little bit if you look at the promoter region you lose a little bit of basal activity but it's still inducible by 125 D3 in the fold is about the same but we're measuring RNA levels MMP 13 RNA levels in this cell and of course slight aside from the genomic modifications we're not adding or subtracting anything to these cells okay so if you look for example at the deletion of the minus 30k region now you see a complete loss of basal activity from that 30k region so and in fact if you then look at but there's some still there's some regulation of the vitamin D itself so that still seems to be a bit operable when you look at a knock rungs knockout or the VDR knockout of these cells those activities generally reflect the activities that's seen with the knockout of the two enhancers so this is the expansion of that just to show you here's the basal level here's the ability 125 to suppress and you can see that this is an expanded version of just these sets you can see a tremendous change so we actually think that they're that these are important observations with respect to MMP 13 and they establish a direct linkage between the enhancers that we've identified here and this gene so I said that vitamin D turns into a suppressor we don't understand it entirely but here's the model that we're looking at here's MMP 13 the direct effects of vitamin D to induce that to induce that gene but vitamin D has very potent down regulatory effects on runks to on the runks to gene itself and that and that also in turn has a direct effect in suppressing Ostrichs which is another key osteoblast transcription factor as well so what we think is happening is when we knock out the vitamin D response itself then the impact of the vitamin D system on down regulating a key basal regulator which is runks to leads to that kind of a suppression and we're following up on this but it has we believe it actually has significant implications for other runks to and vitamin D target gene because you can imagine all sorts of things in genes where there's a different sort of arrangement of regulation so this is a summary of what we of what we think is going on in this chromatin interaction model where this is the the promoter region itself here and we've we've centered the 30k region here because it's a primary player we have the 20 and the 10 and the promoter region interacting and the reason we have this like this is the fact that when you delete the minus 30k region it gets even more complex because what you can do if you follow up with chip analysis of these other sites delete the 30k region if you look now at the ability of the vitamin D receptor interact with the 10k region it's strikingly reduced there's a striking reduction at the 20k for cebp beta and some and and also some striking decrease and even runks activity at the promoter so this seems to be a central organizing mediator of this particular gene and I don't think these conclusions are terribly surprising but at least they do show in fact direct linkage now so where are we going with this we have basically are introducing all these mutations into the mouse and we're trying to look now to see what the consequence of that actually is and I just want to show you one one set we've deleted the minus 10k region in the mouse and when you look at the cells that are derived for in the skeleton that are derived you can see in fact a very similar activity and that is 125d3 in the wild type induces a strong up regulation of mmp13 but surprisingly or perhaps not surprisingly when you delete the minus 10k region you lose the inducibility and you gain this suppressability so we think that this sort of confirms what we've seen in the cell lines and we're hoping to see a similar and perhaps more interesting things in the in the mouse itself with the other deletions now I want to just so in the last few slides I want to go to another gene that we've studied I'm not going to go through this in in detail but this is a gene called rank ligand which we've been studying for a number of years rank ligand is extremely important because it's it's up regulation mediates osteoporosis and almost every osteoporotic disease stayed in humans as well in mice its primary role is in bone but there are many many other activities that have been seen for this particular molecule but the primary consequence of rank ligand knockout is a striking phenotype in the mouse with respect to the skeleton now this just shows you that rank ligand comes from the osteoblast so it's part of the communication that goes on in remodeling it comes and it's regulated from the osteoblasts but it promotes the osteoclast differentiation from hematopoietic cells now we originally did this work with chip chip a number of years ago fact 2004 this is actually a chip a chip seek analysis in which we can see these similar vitamin e responsive regions this is a gene that's regulated by 125 but we can identify a number of binding sites and the important aspect here is in fact that there are four or five of them and in fact that the reason we could never identify a regulatory mechanism for this gene was in fact that the that in this particular case this binding site is 23 kilobases upstream and these sites are in fact about 88 kilobases upstream so there's a whole cluster here but they're they're broadly up regular up up beyond the gene now we've done a lot of work and I'm not going to go through this in detail except to say that we know where there's a lot of binding proteins that are going on here in all of these and we've actually classified these as two types one mesenchymal or osteoblast like enhancers from D1 to D6 and then we've actually identified hematopoietics cells T and B cells that express rank lag and from an even more distal set between 123 and 157 a 53 kb even even further upstream and so these and interestingly D5 seems to be one that mediates both mesenchymal as well as immune cell activity now this is just to show you that when we use that construct or that segment that contains all this information we can fully rescue this run to rank lag and no mouse which has all sorts of skeletal defects and all sorts of in lymph organogenesis there's a whole host of these things but we can fully rescue virtually all of it if we use smaller segments of this gene they are unable to rescue this this mouse so what we've done now is then I'm not going to go through this but we've made four five genomic deletions in the mouse one near the promoter where we deleted about seven kilobases and then some of these other regions that correspond to some of these enhancer regions I've just summarized all the data here I won't go through it in detail to say that we've just finished the phenotyping of literally all of these and each one of these has a basal a regulatory and a cell type specific phenotype that is consistent with what we had identified in the cells themselves so let me summarize this by saying the vitamin D system represents a good paradigm for the regulation of genes by systemic endocrine signals differentiation alters the epigenetic state we think this is actually very very important I've showed you a little bit of this section of the control of of MMP 13 through direct linkage between enhancers and the target gene and then we've done a similar analysis with this rank lag and TNF SF 11 gene that is complicated and we can actually see right distinct regulation in mesenchymal versus a metabolic enhancers and I will stop there thank you very much it sounded from your talk like RXR might be prebound at VDR sites or at least a subset of them yes and at those sites where VDR joins in and our XR is prebound do you see histone modification changes or not so I I would say that we haven't drilled down that deeply but it's absolutely true that RXR appears to be marking these sites we don't know the status of RXR that whether it's a homodimer whether there's another nuclear receptor that's located there or or what the role is but but years ago we began to think that in fact the prebound RXR might be player play a role in perhaps even directing the vitamin D receptor to those sites and that rather than forming in solution in the in the cytoplasm