 Okay, well thanks very much. Really appreciate the honor of being invited to talk to you. Meet some old friends, meet new people, see the facilities. It's been a great visit so far, so I don't have any disclosures. So it is Grand Rounds, so I thought I'd start with a case presentation. This is a 56-year-old woman from the Philippines who presented with a complaint of exertional lightheadedness and fatigue over the past month. No chest pain, no palpitations, no syncope has passed history of hypertension and hyperurusemia, and on exam really only noticeable finding was a pulse rate of 40. Labs were not too revealing, a mildly elevated ANA, but the rest of the labs were quite unremarkable. On chest x-ray there was mild cardiomegaly, and on electrocardiogram, however, atrial flutter with high-grade AV block and some secondary ST and T-wave changes. Cardiac MRI was performed, which was consistent with the chest x-ray, showing mild cardiomegaly with a slight increase in systolic and diastolic volumes, and a normal to borderline LV dysfunction with an ejection fraction in the low 50s. And there was late gadolinium enhancement mild, but in the mid and basal septum, and I think you can, if the mouse works, see a little bit of that over on this image here. On further questioning, the patient recalled that actually there were several additional family members who had a slow heart. Her sister, her mother, maternal uncle, and paternal grandmother, and that's illustrated on the pedigree here, and that prompted the team to recommend genetic testing. And so a standard commercial panel for 84 genes was sent off that looks for known causes of cardiomyopathies, inherited arrhythmias, cardiac and skeletomyopathies. And as usual, there were a couple of variants of unknown significance in a variety of genes, in calcium channel subunit, plaqueophyllin, tyrosine phosphatase inhibitor. But there was a pathogenic mutation in the LAMIN gene, LAMIN-A, which encodes the LAMIN-AC nuclear LAMIN-A protein. And the clinical course and the genetic testing actually were consistent with LAMIN-AC cardiomyopathy with progressive conduction system disease. The patient initially had a dual chamber pacemaker, developed a decline in ejection traction, episodes of non-sustained VT, and because of the high risk of lethal arrhythmias, with LAMIN mutations, the pacemaker was upgraded to a bi-V implantable defibrillator. It was interesting to see this case recently because it reminded me of a paper we published quite a few years back with the Seidman group in which they introduced a mutation into the mouse LAMIN gene, and the mice developed apoptosis and decay of the AV node. The wild-type AV node is shown up here, mostly in blue. But down here in the mutant mice, there was loss of a lot of cells in the AV node, and if you phenotype them and did EKGs, you could see the wild-type had normal sinus rhythm, but the mutants had various degrees of AV block, primary and first-degree and second-degree AV block. Just a nice example of using mouse modeling to help understand the molecular pathophysiology of a human disease, even if it is a relatively rare human disease. So I think this audience obviously knows about the impact of cardiovascular disease and sudden death. There's more than half a million deaths each year from cardiovascular disease, and various estimates put about half of those are sudden cardiac in nature, usually ventricular fibrillation, ventricular tachycardia, often in the setting of coronary artery disease. There are lots of mechanisms that lead to sudden cardiac death. I've listed some of the relatively uncommon, more Mendelian types here on the left side of the slide, primary channelopathies like long QT syndrome, hypertrophic cardiomyopathies, usually these are due to mutations in sarcomeric proteins, dilated cardiomyopathies, these are kind of a mix of causes, of genetic causes, arrhythmogenic cardiomyopathy, usually due to junctional proteins. Obviously much more common are the complex causes that lead to ischemic heart disease. I often joke when I made this slide, a 60-year-old guy was supposed to be an old guy having become a senior citizen not that long ago. He doesn't seem that old anymore and maybe I actually should replace that picture with somebody even older. But obviously that's what you usually see in the clinic. The rare ones are by definition much more uncommon. So when I was going through residency and a little bit into fellowship, and this is too small to read, we were in the middle of the era of anti-arrhythmic drug therapy for patients who were at risk of sudden cardiac death, usually from ischemic heart disease. And this was a very common outcome from some of those trials. This is from the CAST trial. And if you look closely, you'll see that the placebo group survived better than the treatment group. If you're a clinical investigator, this is not exactly the result that you're looking for. But this was really typical of drug treatment trials in that era and really was the death knell for a lot of anti-arrhythmic therapies until amiodarone and some other drugs came into the fore. And somewhat by default, the implantable defibrillator has become the mainstay of therapy for patients at risk of sudden cardiac death. But I think as you all know, there are a lot of problems with defibrillators, battery changes, lead failures, infection of devices and the leads. They may shock inappropriately. The cost is quite high for the implantation. And perhaps most concerning is most patients at risk don't benefit. The number to treat is relatively high. And with shocks going off in patients that it's not saving, the cost-benefit analysis is quite complicated for this therapy. And if you think about it even more, patients are getting treated after they're developing VT, rather than having a pharmacotherapy or some other basic therapy that prevents the VF or VT from developing in the first place. Now on the flip side, conduction slowing as opposed to VTVF is also problematic. There are 200,000 or more pacemakers implanted in the United States each year. Many of these, the majority for sinus node dysfunction or complete AV block. And as again, as I think most of you know, conduction slowing in the setting of depressed ejection fraction is a risk factor in of itself for increased mortality. This is from the Vest trial. And if you just look at the QRS duration, which is an indirect and imperfect measure of conduction velocity and conduction slowing, you can see that those that have the broadest QRS duration have the worst outcomes. And this is in patients who had an ejection fraction less than 35. But it's an independent predictor for increased mortality. So we, for these many reasons, because of all the morbidity and mortality associated with conduction system disease, we've been studying this for a number of years. On the left side is just a cartoon outlining the important elements of the conduction disease. And again, this should be really familiar for everyone in the audience. The sinus node, which is setting the pace. The AV node, which allows for a bit of a delay so that blood can pump from the atria into the ventricular chambers. Then the hisprikinji system, his bundle, the bundle branches in the prikinji fibers, basically act like the electrical superhighway to bring an electrical impulse to the bulk of the ventricular myocardium in a fairly synchronous fashion. And that allows for optimum cardiac output. You want most of the muscle cells to be beating, to be contracting fairly synchronously. And that's illustrated in the movie on the right side there as the impulse comes down the bundle branches into the terminal prikinji fibers and we get excitation and contraction. Well, how does this all happen? Obviously, we think of cardiomyocytes perhaps homogeneously, but actually they're quite heterogeneous. Pacemaker cells, atrial cells, we have hisprikinji cells, we have ventricular myocytes, and then we have the supporting cells in the heart, the fibroblasts, the endocardial cells, the endothelial smooth muscle cells. But of the excitable cells, which I'm showing on the right side here, each of these have a specialized function because of the genes that they express. So obviously, pacemaker cells have to make pacemaker currents. They have action potentials that allow for automaticity. They have a base for depolarization and this is underlined by basically a molecular signature so that they're expressing high levels of the HCN pacemaker channel. They have gap junction channels that are unique and if we compare that, for instance, to atrial or ventricular myocytes, they have to rapidly activate and contract and so they have high levels of the sodium channel because that's very important for the myocyte to excite. The prikinji cells are perhaps the most interesting because they sort of have two properties. They conduct electricity extremely rapidly, the fastest in all of the conduction system. So they have the highest levels of NAV1.5, which is the cardiac sodium channel. It's the gene that's mutated in some patients with Brugata syndrome or LongQT type III. They also have a specific gap junction channel, KinexM40, which is kind of a large caliper channel so that electricity can go from one cell to the next cell and sort of give it that superhighway property so that it conduct electricity quite quickly. But they also make some pacemaker currents. They also express the HCN4 channel and it gives them a little bit of automaticity and that can lead to triggered arrhythmias when it's a bit overactive. So again, there's a lot of diversity in the molecular underpinnings, the molecular signature of each of the excitable cells in the heart, but that's what's responsible for their unique action potentials, their unique function. Everything is well suited for this whole system to act as a network and excite the heart and allow it to contract in optimal fashion. So we've been interested in the Purkinje Fiber Network for a number of years now. This was originally described by an anatomist, Johannes Purkinje, who actually thought the Purkinje Network was a cartilaginous part of the heart, really did not understand that it had anything to do with cardiac conduction. It wasn't until Tuara, a number of years later, made the connection between the anatomy of the heart and what the physiology might be. And this is a picture that he drew more than 100 years ago now of a human heart that was opened up. This is really done under the microscope and you can see it's a beautiful rendering of the very proximal bundle branches then sort of ramifying out into the terminal Purkinje Fiber Network. So why have we become interested in this? Well, because I think it's becoming more and more recognized, particularly in folks who are spending a lot of time in the EP lab, that Purkinje cells in the Purkinje Network can trigger arrhythmias. In the rare forms, it's been implicated in long QT syndrome, particularly long QT type III, with sodium channel mutations. There's data suggesting it plays an important role in catecholaminergic polymorphic VT, familial forms of intricular fibrillation, such as those with mutations in the IRX3 transcription factor, but also in common forms of heart disease like heart failure, drug induced long QT syndrome, or even an ischemic heart disease. And I think probably Hussigar's group has made the most impact on identifying Purkinje cells triggers and showing the importance of trying to ablate those and getting rid of the VT and VF that are seen in some pictures. And this is just from a recent review of his. You can see the Purkinje spikes that are coming before the QRS and so you start getting a short run of VT that degenerates into more of a VF. And on the right side, again, maybe the mouse will work, you can see the little Purkinje trigger here before the QRS and then it gets ablated. You get rid of the Purkinje trigger and the VT often goes away. And they and other labs have had fairly good success if you can identify the Purkinje cell, Purkinje network as the trigger of long-term curative outcomes in patients with these Purkinje cell triggered arrhythmias. So over the years, we've tried to develop a number of genetic tools that could help us study the developmental biology, the physiology and the pathophysiology of the conduction system. So on the left is the picture I just showed you. On the right is a mouse heart that expresses a fluorescent protein in the specialized cardiac conduction system. And I think the most remarkable thing is I often feel like I'm reinventing the wheel. I mean, Tuara did a pretty good job over a hundred years ago describing the Purkinje fiber network. The resemblance to what we're seeing in a mouse heart is really quite remarkable. I guess also the fact that a mouse and human seem that similar evolutionarily is quite remarkable as well. But these genetic tools gave us a new mechanism to visualize, to isolate Purkinje cells and to study them that we weren't able to have before. And just as an example, again, going from human to mouse and mouse back to human, if we isolate Purkinje cells from the mouse heart and we can do that from mice that have mutations that are similar or identical to human diseases, we can see that we get some of the same cellular abnormalities that we see in human patients. So for instance, in the middle, we've taken Purkinje cells from a mouse that has a mutation in the ryanidine receptor that mimics CPVT. And we get all of these delayed after depolarizations, which triggers for ventricular arrhythmias. And if we take a Purkinje cell from a mouse with long QT type III with a mutation in the sodium channel, we can see these early after depolarizations. And so we can use the mouse to model very nicely what we see as the pathophysiology in humans, and obviously it's a lot easier to study murine cells and breed up colonies of mice than it is to be able to do these sorts of cellular studies in patients. So we've operated for a long time now under a general paradigm that developmental processes can inform disease mechanisms. And I think probably Eric Olson's group showed this the best in their studies of cardiac development and cardiac hypertrophy. Many of the pathways, many of the genes and proteins that are used to develop the heart are kind of reused in abnormal ways when we have pathophysiology. So some of the growth genes that help heart form are overexpressed or abnormally expressed when the heart is subjected to different stressors and it leads to abnormal hypertrophy, and occasionally hyperplasia, but more often than not hypertrophy. And we thought that same paradigm could apply to our studies of the conduction system if we could understand how these cells developed, how the network formed. Maybe that would give us some insight into diseases of the conduction system that are occurring in adult humans. So this is, I should have started this a little earlier, but this is a really nice movie just showing how the heart goes from a heart tube into the four chambered heart. And again, that process and understanding that's been really informative. And on the bottom, we wanted to understand how we could go from an embryonic heart to an adult heart with Purkinje Fiber Networks and how that all came to pass. So this work started with an MD-PhD student in my lab, quite a number of years ago, Stacey Rentschler, who's a faculty member at WashU right now. And it started serendipitously with a mouse that was not made to study the conduction system, but a reporter gene, which is a gene that can turn cells of different color, which was being used really to study the brain. When that transgenic mouse was made, the gene landed in a funny spot, but it seemed to land in a funny spot that had allowed it to be expressed in the heart and specifically in the heart's conduction system. So when this mouse grows up, it looks a little like this in terms of its heart, and we could see here the left bundle going out to the Purkinje Fibers over here, the right side, right bundle, into the right ventricular chamber, and the right Purkinje Fiber Network. And again, this happened totally by accident, but I suppose chance favors the prepared mind, and we were able to run with this and do a lot of different experiments. So the first question we wanted to ask is, what are the signals in the developing heart that are important for taking sort of regular old myocytes and urging them and encouraging them to become Purkinje-like cells? So if we start in the middle here, we know that the Purkinje cells form immediately underneath the endocardium. So this very thin pink layer here are the endocardial cells, and these blue-stained cells are the Purkinje cells. So it's suggested that there might be some interactions, and this is all taking place here in the kind of inner zone, the trabecular zone of the heart. This is the area on the interior here, and the endocardium is kind of lining all of this trabecular myocardium. But it's suggested that the endocardium might be secreting a factor that landed on these trabecular cells, turned on a gene program that pushed them toward a Purkinje cell lineage. And so we wanted to basically test that hypothesis, and so what we did is we took embryos, mouse embryos, they're inside the sac here, here. We took out the embryo and then we took the heart out of the embryo and put it in a tissue culture dish, and they're pretty happy. This is one of those embryonic hearts. They just sit in the dish for a couple of days, beating away, and we thought we could treat them with different growth factors that we thought endocardial cells might be secreting. And since these hearts had that blue-color marker gene in there, the assay was basically to see if we could make the color, the blue color stronger and spread out. And if it did, that may be some indication that that was an important growth factor for forming Purkinje cells. And so this is a control heart after a day or two in culture. These are hearts that are treated with a number of secreted factors that come out of endothelial or endocardial cells. Endothelin I, angiotensin II, insulin-like growth factor I. Endothelin I made the hearts a little bigger, but really, they didn't change the pattern of the blue color, the betagalactosidus expression at all. Then we tested noregulin, which is another very potent growth factor that's secreted by the endocardium. This is the same pathway that gets dysregulated in breast cancer when there are mutations in the herb B2 and herb B4 receptors. That's what noregulin lands on. So noregulin is secreted by endocardium and the underlying trabecular myocytes make the receptors for this. And when we bathed these hearts in noregulin, it basically turned most of the heart blue, suggesting it was pushing them toward a prokinge-like cell fate, or at least turning on a gene expression program that prokinge cells express. So that's great. Blue colors are nice, but what about real genes? Is it really making them express prokinge cells? So I told you before that prokinge cells make a lot of sodium channels and they make a lot of a special gap junction protein called Konexin-40. And so this is antibody staining for those two proteins. So the pinkish color here is the Konexin-40 gap junction channel protein. And the pink color on this one is all of the sodium channel protein. And it's very enriched in the trabecular myocardium as is Konexin-40. If we treat these hearts with noregulin and then stain for them, you can see that virtually all of the cells start making high levels of Konexin-40 and Nav-1.5, the sodium channel protein. So it's very consistent with the beta-galactosidase blue dye staining. And again, it suggests we're turning on this gene expression program. So I'm not going to go through this in detail. I have in the past and people glaze over, but noregulin has a very complicated downstream signaling pathway within the cell. We wanted to figure that out. You don't need to figure that out, so I'm just going to kind of cut to the chase of the punch line here. But we used a variety of techniques. We used chemical inhibitors of some of these kinases in the pathway. We did gene expression screening, so we separated the prokinge cells from the rest of the working myocytes and looked at what genes are highly enriched in the prokinge cells. And we sort of compared these orthogonal approaches to identify a transcription factor that might be highly enriched in these developing prokinge cells. On that kind of screen, we identified a protein called ETV1. It's a transcription factor. It's activated by getting phosphorylated. It's best known in sort of the neuroscience literature. And if you knock out mice with this, there's abnormalities in motor neurons. This was really interesting because for many years, people thought prokinge cells were nerves. And if you think about that sort of electrical superhigh where I was describing, it was like this nervous network that's conducting electricity very quickly. So the fact that knockout mice were known to have a neuronal phenotype was kind of a nice little tidbit suggesting that maybe we were on to something. But nobody had looked in the heart before. But ETV1 was one of the really highest expressed transcription factors in prokinge cells altogether. So the next thing we looked at is a mouse where everywhere it's supposed to make this ETV1 protein, it makes this beta-galactosidase blue color. And it gives us a readout of when and where it's expressed. And this is an embryonic part. So mouse gestation is 21 days. So this is about halfway through. This is the day of birth, postnatal day one, three weeks, postnatal day 21, and then 10 weeks after life. And so everywhere there's blue is kind of a nascent prokinge cell progenitor and with time a more mature prokinge cell. So this is, you can see they're forming nicely in the trabecular region of the heart. As the heart septates more you can see that it's starting to turn into what you'd expect where you can see the left bundle and the prokinge fiber network. And then it just gets more and more refined and sculpted until it really looks like a mature prokinge network. You'll also see that it's highly expressed in the atria. And the atria trabecular myocardium is kind of the second fastest cell type in the heart. So you've got the hisprokinge cells, then you've got the pectinated trabecular atria myocardium and then you've got the atrial and ventricular working myocardium. So ETV1 seems to mark the fast conducting cells of the heart. And I'll get back to that in a little bit. But if this is really important then there should be a problem when you're not making ETV1. And I told you there was a knockout. The neuroscientists, they look at problems. That's what their phenotype is. But we got hold of some of these knockout mice and wanted to look in cardiac physiology. And so we can perform EKGs in mice. And this is the P wave. This is the QRS. Now the T wave comes quite early in the mouse electrocardiogram. Repolarization starts very soon after depolarization begins and so there's kind of a blurring between the QRS and the T wave. But this is what it looks like in here. Here's the next P wave coming up. If we look at the EKG and the knockout mice we see a few interesting things. First, the P wave is a little bit broader. The PR interval is prolonged. The QRS looks a little bit like a right bundle branch block with an RSR prime. And why would that? And this is all sort of quantified here. So this gives numbers to what I just described to you. But if you think about where ETV1 is expressed in the atria and the hisprokinji system, so if you had slow conduction in the atria your P wave would broaden out. If you have slow conduction in the hisprokinji system you're going to get this PR prolongation and QRS prolongation. So the physiology very nicely coincided with the anatomy and the localization of the ETV1 program. So why is this all happening? What's the basis for it? So there are two main reasons why loss of function of ETV1 is causing this phenotype. First, if we actually look at the conduction system, here's the wild type on the left. Again, we've splayed open the left ventricle of a mouse heart with this GFP fluorescent reporter gene. You can see sort of this exuberant protingy fiber network. But in the mutant heart it looks like we're just missing a lot of protingy cells. The GFP, the green signal, is markedly reduced. So either the cells are there and they're not making the reporter gene or the cells are sort of missing altogether. But it's what we call very hypoplastic. And you could imagine if excitation of the heart is dependent upon electricity kind of rushing down all of these cells the heart on the right would have a problem and this could easily account for this kind of QRS prolongation that we're seeing here. There's even more to that. So if you take out the few remaining green cells and patch clamp them, which means you record the electrical currents in them this is the experiment that we did on this slide. So in a normal mouse heart if you look at the protingy cells and currents are plotted a little funny but the more negative it is here the more inward current the bigger the more current there is. So protingy cells have the most sodium current. The atrial cells as I told you were the second fastest conducting cells. So they have an intermediate level and then the ventricular myocytes still have a lot but they have less than the other two cell types. And if you look at knockout cells protingy cells, atrial cells, ventricular cells all of the sort of added oomph that we see in protingy cells and atrial cells is gone. They all make just about the same level of sodium current. So we've kind of lost the special properties of these fast conducting cells. The protingy cells are acting like regular working ventricular myocytes. The atrial cells have down regulated the sodium channel and so it makes sense that ETV1 is kind of a transcription factor that lands on genes that are important for very rapid conduction. Like the sodium channel, like the Konexin-40 gap junction protein, turns them on so that they can conduct electricity very quickly. So, you know these genetic diseases, these mouse models are fine as far as they go but is there any real relationship to the more common sorts of diseases that we see whether those are in mice or in humans. So this is an experiment actually it's published now that David Park in our group did where he took a mouse and then imposed a pressure overload by doing aortic banding. So the equivalent of aortic stenosis or severe hypertension, just put a ligature around part of the aorta so that the aftelot is increased a lot and this is a sham surgery done here and you could see at least on the outside of the heart the ETV1 gene is highly expressed in the atria as I showed you these are just the isolated atria and if you open up the heart you can see all the blue prokinzy cells here so what happens if you perform the aortic banding and then take a look well you can see and it's really interesting kind of cool that the left atria markedly turns off the ETV1 gene expression right so the atria is not blue anymore and again since aortic stenosis or the tach banding is initially really just a load on the left side of the heart this nicely demonstrates that the right side which is at least early in developing a hypertensive phenotype there's no heart failure on the right side and so the expression is preserved again on the inside of the left ventricle which is seeing all of this aftelot we get a marked down regulation there very the blue color is very reduced compared to the control so this suggests that even in common forms of disease there may be a link between getting exposed to that stressor turning off this transcription factor dialing down the sodium channel and as I told you in the beginning slow conduction is really a marker for having increased mortality morbidity and mortality in the setting of the failing heart and so I think it just brings out the importance of this entire pathway and David's gone on in a recent paper in circulation to look at the atrial tissue from a number of patients with a range of ejection fractions and you can see again with human studies things are always messier than they are with mouse studies but there's a significant relationship between ejection fraction and the expression of ETV1 so as the heart function declines there's less ETV1 expressed but even more an even stronger relationship is the relationship between the amount of ETV1 that's expressed and SCN5A which is the sodium channel right so if you make a lot of ETV1 you make a lot of sodium channel so it's exactly consistent with what we found in our murine studies but here we're actually using human cells and again suggests that stressors that lead to decreased heart function dial down ETV1 and dial down some of these really important genes that can lead to these syndromes well what about sort of additional data on top of that so I think we are familiar with GWAS studies GWAS basically says are there sequence variants that are associated with a trait or a disease I'm interested in alright so I want to study diabetes I'm going to look at 100,000 people I'm going to scan their genome and do the ones with diabetes have a different genetic signature or a sequence variant or a mutation that's not found as frequently in the general population so we did something with Dan Rodin in his group at Vanderbilt called a FIWAS study and this basically flips that paradigm around so they've over the years taken all their patients who come into the Vanderbilt system and coded everybody from the electronic health record and done genomic analysis on all of them and looked at all of the SNPs that they have across the genome and what FIWAS does is asks what traits are associated with a SNP of interest so one looks at the gene and one looks at the disease the FIWAS sort of flips that around the other way so on a normal GWAS plot at the bottom is the chromosomal location and on the y-axis is the probability and the higher up the more probability is the tighter the association with a FIWAS study down on the bottom is basically the disease and so this goes from system to system infectious diseases, neoplastic disease, cardiovascular disease and if there's an association then the p-value is a very small number this is presented as negative log of the p-value and so there were really only one or two diseases that were associated with a sequence variant in the ETV1 gene and they looked at African Americans and European Americans separately and if you look at what the disease is bundle branch block and heart block so if you think about going to all the possible diseases a patient could have that are coded in the EHR the only ones that associate with a sequence variant in ETV1 are heart block so it's really nice additional evidence that took us from kind of a murine model to murine pathophysiology to human clinical relevance with both kind of genetic proof as well as proof in acquired heart diseases so can we go the other way can we make things better ETV1, you get rid of it you have less of it, conduction is slow not so good for you so we thought can we do a gain of function experiment so here we take neonatal rat myocytes that are very easy to culture and we can infect them with viruses that either over express just a green protein as a control or a virus that over expresses ETV1 and on the right it's just one of these cultures they beat quite happily for a couple days in culture and if we infect them you can see here this is connected 40 is the gap junction protein so at control there's a little bit but when we over express ETV1 we make tons of this connection 40 gap junction protein and same thing here with the sodium channel at baseline you know we make sodium channels not surprising in myocytes but it really blasts it out when we start infecting it with ETV1 all of this red color is additional and if we measure sodium currents again the baseline level almost doubles when we over express ETV1 and that's quantified here and then if we kind of look at all genes that are involved in kind of excitability all the ion channel subunits and other genes these are actually the expression levels of real ventricular myocytes and these are real prikinji cells and if something is red it means it's making more of it for the prikinji cell than in the myocyte and the other way around if we look at the neonatal rat myocytes that are not infected or are infected with ETV1 we can see that the colors of these over expressing ones really correspond very nicely to a bonafide prikinji cell so this one single transcription factor is sufficient to kind of turn on a whole gene expression program that at least makes it act like a prikinji cell whether it's really changed its fate or not or more work but at least it's behaving physiologically like one which is sort of an important finding we wanted to move from a rodent model to a human model so these are human pluripotent stem cells that are differentiated into cardiomyocytes basically the exact same experiment I just showed you we over express ETV1 and in fact it turns on connects in 40 it turns on the sodium channel kind of the same result in a human myocyte as we did in a rodent and we get a lot more sodium current as well so I think one could think ahead about whether this strategy might be useful translationally to improve conduction in a heart that is diseased and has slow conduction QRS prolongation so just to summarize this part this new regular ETV cascade is critical for the formation function of the specialized conduction system it activates a genetic program that's responsible for the unique physiology of the network high level expression I focused on the gap junction protein in the sodium channel but there are many other ion channel subunits that also are regulated as they should be it's down regulated with pathologic stressors pressure overload suggesting a link to acquired heart disease and a human sequence variant a SNP in this gene is associated with bundle branch blocks suggesting additional relevance to human disease and I told you this tells you it might be a novel therapeutic target if we can turn on this whole signaling cascade and over express ETV1 in the right cell types at the right point in time it might improve conduction well as we thought more and more about neregulin we know it comes from the endocardium and it lands on these trabecular myocytes and it encourages them to become a prikinji like cell but it made us think are there additional crosstalk between these two cell types again the endocardium is kind of plastered right up against the trabecular myocardium and growth factors and other cell adhesion molecules there are all sorts of ways they can be talking to one another and might some of these additional pathways be important in the signaling there are a number of known factors bone morphogenetic protein 10 angiopoietin VEGF, different FGFs WINS, notch signaling there's all sorts of candidates in here we decided to focus on BMP10 partly because it's very highly expressed early in development by the trabecular cells and it lands on the endocardium and Gordon Keller whose group is up in Canada now had a nice paper in cell stem cell showing that BMP signaling is very important if you take human IPS cells to push them toward behaving like endocardial cells so it's already known BMP10 is really critical for normal heart development and heart growth but this gave us even more encouragement that that might be important molecule in terms of the kind of reciprocal talking between these two cell types so what might BMP be doing in the endocardium is it turning on a transcription factor or pathway and what might else besides the regular might the endocardium be secreting that could land on the trabecular cell and again encourage it to become a purkinje like cell so we wanted first to screen for all of the transcription factors and endocardial cells that responded to BMP and this is a little complicated but let me take you through it we took an embryonic heart and again we cultured it in the little petri dish and we treated it with a drug that inhibits all of BMP signaling right so either the control dish or this inhibitor and then we did gene profiling we looked at every gene that went up or down and if it went down with this inhibitor it suggested that BMP normally makes it go up right and then we did the gene profiling and then we looked at every transcription factor that was expressed in the right direction that the LDN inhibitor made it go down and we got this list here and the second one on the list was a gene that I had no idea had anything to do with the heart sounds like ETV1 but this one's ZBF1 it's a little bit of alphabet soup but let me just tell you a little bit about that it's another transcription factor it's key it's called early B cell factor one because it's critical for B cells to develop right and when you knock it out you don't get B cells again just like the neuroscientists look at neurological things and don't think about the heart you know the immunologists look at B cells and T cells and don't think about the heart but in more recent years this gene has also been implicated in the developmental biology of other cell types neurons osteoblast, adipocytes in the kidney and glomerular cells there's some interesting work in flies and in mice suggesting that it has important developmental roles so we got some antibodies to EVF1 to see where it was expressed now this is an embryonic heart day 13 and a half just to orient you this is the developing compact zone of the heart so that's the outer layer and all of these are the trabeculae the inner layers of the heart and these pinkish purple cells those are the nuclei of the endocardial cells right that are lining the trabeculae here and there's an occasional cell elsewhere but the most prominent expression of this EVF1 protein is in the nuclei of the endocardial cells in the heart so that's pretty good because that's what the screen was looking for it was looking for transcription factors that are expressed in the endocardium so what do these mice do so it was known before if you knock out EVF1 the mouse is runted basically the whole mouse is smaller every organ is a little bit smaller if you take out the heart the heart is also somewhat proportionally smaller than it is in the wild type it turns out actually the heart is a little thicker for its size than the wild type if you do the wall thickness compared to the entire weight of the heart or the dimensions of the heart it's a little bit either hypertrophy or hyperplastic a little bit bigger than normal but this is what the conduction system looks like on the inside of these EVF1 knockouts so again this is a similar picture from before with this exuberant Purkinje fiber network the EVF1s are very hypoplastic dysmorphic looking it looks like it has continuity in the right bundle branch kind of a right bundle branch anatomical block here but there are areas where there's just completely devoid of the GFP positive Purkinje cells so what do we do when we do an EKG here again here's the control and here's the EVF1 knockout we get the same RSR prime right bundle branch block type pattern we don't see any changes in the P wave we don't see PR prolongation to the same extent and that's because it's not expressed in the atria to any extent this is much more circumscribed to the ventricular conduction system and if we quantify that we can see that the QRS on average is significantly longer than in the wild type mice this is a technique called optical mapping this actually is a mouse heart here in Langendorf mode maybe some of you have seen this but we're basically hooking the heart up to plumbing to keep it happy while we do this experiment and then we infuse the heart with a die that responds to voltage so when the heart depolarizes the die gives off a signal and we can capture that with a camera and so this image right here our camera is looking at the outside surface of the heart and this red area here and red area there are kind of where the left bundle branch and the Purkinje fibers over the left ventricle excite the left side of the heart and this is on the right side so we get these two breakthroughs where the terminal Purkinje cells are activating the heart and then these kind of coalesce and flow from apex to base to depolarize the whole ventricular myocardium and that's why we get apex to base squeeze out to the great vessels that's the direction you want to maximally optimize cardiac output if we look at a couple of knockouts they're very abnormal looking we get activation on the left here but we don't see any right bundle branch breakthrough here as if there is a bundle branch block this other one we don't even know what's going on maybe this one had an even more abnormal conduction system but it's definitely not breaking through where it's supposed to so there's a very severe physiological defect in the conduction system here we also did echocardiograms on these mice and we actually did strain imaging so this is a wild type and these green lines that are coming in and out for those of you who aren't familiar with it that's basically the reactor of the strain and you can tell it's pretty synchronous right so they're all pointing out and then they're all pointing in because the heart is contracting in quite a synchronous fashion if we look at one of these knockout hearts you can see it's pretty willy-nilly right and to my understanding of the literature this is really one of the only interesting models of dyssynchrony right where it's a purely electrical phenotype that's leading to this abnormal activation of the ventricle and we're getting some of the walls coming in and contracting and exciting and contracting at different times in the cardiac cycle we haven't studied this much further but you might imagine that this could go on to develop more of a heart failure phenotype from the dyssynchrony and would be an interesting model to see about some sort of resynchronization therapy perhaps well what's going wrong with these hearts from sort of a developmental biology point of view this is a section again through a wild type and a knockout the red stain is for troponin so that tells you where all the myocytes are HABP is a marker of a binding protein that the endocardium secretes so that kind of gives you an idea of the outline of the endocardial cells and DAPI is just a blue stain for the nuclei so you can see a few things here in the knockout there seem to be more cells the compact zone is thicker the whole trabecular region just seems more robust if you look at just the green staining it seems like maybe there's more endocardial cells if we do a more H&E staining the trabecula and again most of you have never seen these before you're going to have to take my word for it but the trabecula are kind of immature looking there's a little bubble here between the endocardium and the muscle there's space that's kind of filled with cardiac jelly that should be getting resorbed and it's not if you look here at the higher power so there's a problem there cells are seem to be dividing too much there's increased proliferation of endocardium and myocardium and there's an immaturity and that's a common theme in developmental biology cells divide, divide, divide and then they get signals to differentiate and come out of the cell cycle and then they do their thing whether it's a Purkinje cell or a neuron or a myocyte here it looks like that balance is thrown off the cells are dividing too much but they're not going on to become totally differentiated so what might the endocardial cells be making you know so ebf1 is a transcription factor in it what genes is it turning on and off that's signaling back to the trabecular cells to be doing this correctly or incorrectly when it's missing so what we did here is take the endocardial cells we can sort them from a normal mouse or one that doesn't make ebf1 we can again look at the genes that are turned on and turned off and what we did is focus on parachrom factors on secreted factors you know what might it be making that the endocardial cell can be pumping out that lands on the myocyte and what's different when the ebf1 is missing and two factors kind of stuck out to us one of them is something called DKK1 and it's a very potent inhibitor of wind signaling again I'm not going to give you a lesson on wind signaling but it's a very potent growth and differentiation factor it's important a lot of lineages but DKK1 is markedly decreased in the knockout suggesting there might be more wind signaling and then another protein called HGF parasite growth factor it has a lot of activities but also can act as a wind activator and we're making more of this so we have less of an inhibitor and more of an activator it's kind of a perfect storm to give us more wind signaling and so that is certainly a candidate a signaling pathway that might be dysregulated in these developing hearts so we have a model here where the trabecular cells make BMP10 it turns on ebf1 in the endocardium it regulates the levels of these two secreted proteins and when everything's great we get normal growth and differentiation but when ebf1 is missing we turn down DKK1 we turn on HGF1 these lead to this aberrant hyperplasia, aberrant differentiation and instead of trabecular cells kind of going on to becoming prokinzy like cells they just divide a little bit too much more it's more complicated than this we have some data that the mick protein is involved in this pathway as well and we're trying to tie that all together so in terms of ebf1 it's a BMP inducible transcription factor it's also in the cartoblast and we're looking at that separately it directly regulates endocardial and directly we mean cellatonisly it means it's doing its thing in that cell but it indirectly regulates the way the myocytes behave by causing the endocardial cells to secret proteins that land on the myocytes and in its absence the cells proliferate abnormally there's aberrant differentiation it's most prominent in the prokinzy cells maybe because the endocardium is very close to it and so any changes in the expression levels are going to impact the cells that are really its closest neighbors but we get this very profound hypoplasia of the conduction system and a loss of some of these prokinzy cells so just to wrap it up the conduction system it's really an important trigger of life threatening arrhythmias it's a common sight of conduction slowing and heart block we've been using another a number of approaches to try to unravel the molecular project the genetic architecture that governs conduction system formation physiology pathophysiology and you know I always say this holds promise everything holds promise but you know we're hoping that this will lead to some additional translational discoveries so that we're not as reliant on devices such as defibrillators or ablation that actually we can come up with therapeutics for some of these important diseases of arrhythmogenesis most important slide for people who've been working on this David Park is an MD-PhD electrophysiologist clinical electrophysiologist but also a basic scientist Akshay Shakar was involved in the ETV1 story and Eugene Kim another cardiologist has mostly been involved in the EBF1 story and then a bunch of additional people and collaborators from here and there so thank you