Our first speaker is Igor Efimov from George Washington University, and he's going to discuss electrical propagation in the cardiomyopathic heart. Well, thank you so much, the organizers and the chairs. I really appreciate the opportunity to present our data. So I would like to share with you today some data we've obtained studying live human hearts, which we have a robust program in our laboratory already for over 10 years, where we receive explanted human hearts, which we can do both by informatics or mixed studies, but also functional studies. And the question I would like to discuss today is what is the fundamental mechanism by which myopathic heart can sustain ventricular fibrillation, ventricular tachycardia. So fundamentally, we know of two major mechanisms shown here. One was formulated by Thomas Lewis back in 1925, which is basically a single or a finite number of reentrant rotors, which can be... Can I control it? Is it moving my mouse? Do you see my mouse there? Well, then it's not going to help. I changed... Okay, here we go. You can see that, right? Yes. Okay. Sorry about that. It's only in Silicon Valley. So the two major mechanisms, as I already said, the top one is what Thomas Lewis formulated in 1925, and then Gordon Moore in 1950s proposed multiple wavelets. So over these years, with many, many groups, ourselves included, we studied essentially what molecules are responsible, and for probably the past 50 years, most laboratories have been focused on a single molecule, blaming everything on one particular protein or one phosphorylation site. And when early work involving qPCR was conducted, like this paper we did in 2013 with Gene Narbonne at WashU, we, of course, confirmed what was shown previously in a number of animal models, such as SIRCA2A is downregulated, both in ischemic and non-ischemic cardiomyopathy. KCHIP2 is tremendously downregulated, probably number one suspect. And then, of course, also Connexin-43. It became evidence it's too complex to blame one molecule. And so subsequent studies, like, for example, this study, again, from WashU, led by Doug Mann, showed that when we compare three different kinds of cardiomyopathy, so black color shows non-failure. These are donor hearts, which are as healthy as you can get for basic studies. Ischemic cardiomyopathy, dilated cardiomyopathy, and septic cardiomyopathy. So you can see that on this principal component analysis, ischemic and dilated cardiomyopathy hearts, they converge. They are non-distinguishable from one another. Black dots corresponding to donor hearts are different, and septic cardiomyopathy is also very different. But still, the picture is very complex. You're talking about hundreds or even thousands of genes which are up or downregulated, and finding one which will be your favorite therapeutic target was clearly very difficult. However, using bioinformatics, you can look at it and find what expected functional consequences of this tremendously complex remodeling process would be. For example, if you look at contractile apparatus, you see decline, essentially, in contractility. Cell-cell junctions. I'm sure Mario Del Mar will teach us about it in the next presentation. But you can see, as we already showed before, connexin 43 is downregulated. You have plakophilin, desmoplakin, also downregulated. You have calcium handling, which has changed. So it's already something which you can test and perhaps even try to pursue as a therapeutic strategy. So this unpublished paper, led by Palak Shah from Inova Hospital in Virginia, and also a very large cohort of hearts from Utah, just again reinforced the same message that you get hundreds and hundreds of genes which are up or downregulated. But what was now clearly shown in this study, that regardless of which cardiomyopathy you start with, it could be non-ischemic or ischemic cardiomyopathy, at the end stage disease, they completely converge. Again, if you look at all the hearts in this cohort, on the principal component analysis plot, you can see this big spot is basically all different types of etiologies of cardiomyopathy, and they are different from donor hearts. So it was very difficult to think about strategy. How do we define what exactly remodeling process is responsible for arithmogenic outcome of cardiomyopathic remodeling process? So most recently, we started looking at more high-level processes in a regulatory network. I'm not sure in one study we just published, we found that these thousands of genes which converge only to about 50 transcription factors, which actually control them, and this is already something you can maybe focus on with some meaning. Another technology is so-called CAGE, or cap analysis of gene expression. In this case, you are looking at specifically at activity of promoters and enhancers. We pursued this line of inquiry, and we found that in the human heart, in the donor human hearts, we built first library for such hearts, we found 13,000 promoters, and 1,000 of them was new, and also we found 4,400 new enhancers, and 3,000 of those enhancers are new. What was really fascinating looking at this dataset, for the first time we could compare publicly available SNPs from multiple GWAS studies related to heart disease against our database of promoters and enhancers, and we found that at least 20 percent of SNPs related to heart disease are actually in the promoter region. So these are regulatory networks. These are not protein coding areas which would result in malfunctioning protein, but rather it's really about expression or no expression of an important gene. So what are the functional consequences of that? So in our previous studies, we showed that regardless of cardiomyopathy, these are end-stage failing hearts, human hearts. You can see that in non-failing hearts, it takes about 50 milliseconds to propagate from endocardium to epicardium, while in the failing hearts, this time doubles. So conduction velocity effectively essentially slows down by 50 percent. And this, in part, is due to Connexin-43 downregulation, but also, as I showed you, there are multiple targets related to mechanical junction, sodium channels, et cetera. What happens to repolarization? So repolarization, again, in myopathic hearts, in humans, slows down. You can see here a comparison, black action potential in the donor heart and red action potential in a typical failing heart. And you can see also that this prolongation of action potential duration is fairly small, cannot be really compared, probably, with the long QT syndrome. But what does it have to do with arithmogenesis and maintenance of arrhythmia? Going back again to classics, when George Mines introduced concept of wavelength, which is a product of conduction velocity and refractory period or the surrogate action potential duration. So when we look at, essentially, maintenance of arrhythmia, wavelength must be less than path length. Otherwise, reentry circuit will not fit in the available myocardium. So this one-dimensional concept is simplistic, of course. How do we translate it to actual human myocardium and to actual human cardiomyopathy, especially considering that you have thousands and thousands of genes, which are up or down, regulated? It's a very complex picture. So we conducted the following study, which involved 30 human hearts, 18 hearts from male and 12 hearts from female. You can see here clinical characteristics, BMI on average was 29, and these are none of them had cardiac causes of death. These are all caused by brain death. Preparation which we used is called VEG preparation. So when you excise a segment of left ventricular myocardium with a branch of coronary artery, which is cannulated, so we will close down all the small branches to prevent, essentially, drop in blood pressure, it's perfused. This sample is about 7 centimeters in length, 3 centimeters in width, and 2 centimeters thickness of normal human heart. Preparation is placed in a tissue chamber, and we're imaging this from four sides using voltage-sensitive dye imaging technique. So this technique allows us to measure full profile of action potential propagation, repolarization, and we chose a pharmacological approach because, obviously, we couldn't impose changes in gene expression in an acute experiment to modulate wavelength, and we chose it by changing action potential duration. We used, essentially, pinacidyl variant concentrations, which is an IKATP opener, and in the end globenclamide to reverse changes in the action potential duration. And to characterize arrhythmia, we also used a phase approach where you basically, using Hilbert transform, you change action potential recordings into phase. So first, we verified that in this study that because pinacidyl is a very specific opener of IKATP channels, there is no change in conduction velocity in this case. So we see, first of all, we found that the longitudinal conduction velocity along the fiber orientation in this direction is approximately 70 to 80 centimeters in the human heart at slow pacing, slow frequencies. But transverse and transmural conduction velocities are actually different, slightly different, which is meaningful. But there is no difference with pinacidyl or globenclamide. They do not change conduction velocities. However, if you look at action potential durations, there is a dramatic change. So normal action potential shown here in red, green you can see almost like a mouse morphology of action potential when 25 micromolar concentration of pinacidyl and then globenclamide partially reversed because it partially blocks IKATP channels, and these are histograms for all recordings in the field of view. So if you look at action potential duration as a function of pacing cycle length, of course you have restitution properties, meaning with acceleration of frequency, the action potential duration shortens. You can see in control in pinacidyl and globenclamide. So of course there is a concentration dependence of APD. We looked at gender differences, and initially it was a trend that female action potential is somewhat longer, but on our data sample it was not statistically significant. And of course this difference also was most pronounced in slow cycle length, or slow frequencies at like two seconds and more. So now going back to what George Mainz introduced as a wavelength, he was talking about product of conduction velocity and refractory period, but of course in the three-dimensional myocardium you have three different conduction velocities. You have along the fiber, across the fiber, and transmurally. Therefore you have to talk about three-dimensional wavelength, not one-dimensional. So we computed longitudinal wavelength, transverse wavelength, and transmural wavelength. So if you multiply these three, you introduce a volume wavelength because myocardium is three-dimensional. And basically our hypothesis now is that in order for ventricular fibrillation, ventricular tachycardia, to be sustained, you must have a volume of myocardium larger than this three-dimensional volume of wavelength at shortest pacing coupling interval, when you can basically tachy-pace heart into arrhythmia. So wavelength, as you see here, in control, longitudinal transverse and transmural, with pinacidyl it drops down dramatically, and that's why short membrane action potential would be pro-arrhythmic, because you shorten the wavelength, and glabenclamide would reverse that. So let's see how our theory actually works. So we propose this vulnerability plot where you take ratio between available volume of myocardium to measure how many cubic centimeters of your myocardium you have, its tissue volume, and you divide it by wavelength volume, which of course depends on cycle length at which you pace the heart. With increasing frequency or decreasing coupling interval, curve will take this shape. So if you are below one, meaning that your wavelength volume is exceeding essentially your tissue volume, it will be safe zone, you cannot induce arrhythmia. If you're above one, this is a vulnerable zone, you can induce arrhythmia. So here is the data. So if we take control, curve remains well under this safe zone, so you cannot induce arrhythmia. We did not see arrhythmia except one particular case, which was arrhythmia just immediately after we dissected the preparation. It was probably ischemic in the beginning, but then it essentially went away. With 15 micromolar pinacidil, curve still remained under this safe zone. We could not induce arrhythmia in any of the 30 hearts we studied. With 25 micromolar pinacidil, when action potential shortened dramatically, you can see curve now has a very large branch above this identity line, and you are in vulnerable zone for short coupling intervals, and indeed, we induced arrhythmias in many, many hearts. So then for 50, 100 micromolar pinacidil, also it was inducible, and then we applied glabenclamide, again, curve dropped down, and we could not induce arrhythmias anymore. Therefore, our conclusion here is that, yes, indeed, the only way you can induce arrhythmia in a three-dimensional myocardium, if your tissue volume exceeds wavelength volume in a reasonable range of your pacing frequencies. So let's see now how this arrhythmia looks like. This is a very messy plot. I'll show just the movie. These are actual raw recordings of fluorescence. We have four pictures because we have left transmural, epicardial, right transmural, and endocardial for all four sides. It's very hard to follow. Therefore, we used phase transformation with Hilbert, and then we removed information about phase and only show wave fronts and wave breaks, or you can call them phase singularities if you like to do so. So this is how it looks like. Again, this shows fluorescence. This one, phase and wave fronts, and here, only wave fronts and phase singularities, and you can see, for example, in this area, there's a fairly stable reentry. Let me try it again. So this is a source, but you can also see that there is not a single well-identified source, but rather multiple wavelets, so it seems like Gordon Moore was right about ventricular fibrillation in this particular model of human cardiomyopathy. So these are characteristics, so we had essentially approximately two to three, two to four wave fronts and phase singularities in these several square centimeters of our field of view, and we never saw one dominant rotor. So let me conclude that regardless of what type of cardiomyopathy leads to arrhythmogenesis, ultimately, the general rule, arrhythmia will be inducible only if three-dimensional myocardium ventricular arrhythmias can be only sustained if ventricular volume is larger than certain coefficient kappa multiplied by wavelength volume. Kappa depends on how you define action potential duration or refractoriness. We had defined that 80 percent of repolarization, in this case, kappa was four, but if you define it in some other way, it will be a different coefficient. And also we see that in a human, myopathic heart and models of it with VTVF is sustained by rotors and multiple wavelets. Thank you very much. I'd like to give credit to my students who really work hard with this human program. Thank you. Questions, please. Thanks very much for the beautiful work. Can I just ask you, the things that you've worked out, do you think they could change the management of the difficult patient who's stuck on ECMO in intractable VF and you're trying to get them out of the things that you've learned might guide a pharmacotherapy depending on the circumstances of what caused it? You know, my current reading that the process we pursued for probably the past 50 years trying to target one particular molecule, design a drug against sodium channel, calcium channel, potassium channel, is probably not a very smart one. So we do have essentially more high-level control circuits, including transcription factors, which was, for example, recently very nicely highlighted by discovery that PTX2, which initially we didn't even know what it was, but it turns out to be number one SNP responsible for atrial fibrillation. So I think that learning about transcriptome would be critically important, and then learning essentially about promoters and enhancers which control that would be also critically important. So I think we have something there. But other than that, we work hard on developing novel, actually, electropherapies, so low-energy electropherapies for such arrhythmias. We've developed low-energy defibrillation for both atrial and ventricular arrhythmias. So do they cause less electroporation? Well, basically, you know, we have just concluded the atrial fibrillation human trial, where we were able to show that in patients with paroxysmal atrial fibrillation, we can convert at about half a joule energy compared to what was experienced previously in control, which was about six to seven joules, so it's about one order of magnitude reduction. And we have now also several cases of ventricular defibrillation. We just started our clinical case, and also it's very encouraging. Interesting. Time. Good work, Igor. I appreciate your, especially the human studies is tough here. In human ventricles, you're typically going to have a lot more ventricular mass. The wavelength issue isn't really a problem. You can induce ventricular fibrillation in those hearts, unless you're doing a wedge prep or a rabbit heart or a rat heart or something. So what's the value? What do you learn from this analysis with the wavelength analysis in three dimensions in a human heart? Right. Well, I think what is important is, I'd like to make, again, a distinction. This study was not about how to induce, but rather how to sustain. So the mechanism of induction could be very, very different. And actually, now we do have tools to study that as well. I just saw beautiful data from Bordeaux, from Michel Heisager's group. They managed to study nine patients in which they were able to catch spontaneous onset of ventricle fibrillation in patients wearing ECGI vest. So they could actually map the process of onset. And it always starts with a well-formed, very synchronous VT-like pattern. This early stage, when you have essentially a trigger in your substrate, the survival of first circuits will be determined basically by this mechanism. Ultimately, you're right. A heart is big enough that actually wavelength will fit in the whole heart. But initially, it has to have what Art Winfrey called elbow room to survive this initial few rotations. And that's what these studies with ECGI show. So it does have meaning. And of course, it's important that due to cardiomyopathy, you will change wavelength condition in this local environment for first re-entrant circuit to survive, and then essentially to spawn new waves throughout the heart. Thank you. Igor, great, as always, and just a quick comment in arrhythmogenic right ventricular cardiomyopathy. I think the general consensus is that the most dangerous period for ventricular fibrillation is not when the cardiomyopathy has progressed and the fibrofat infiltration has invaded a lot of tissue, but quite the contrary, when there is more ventricular mass that can then degenerate towards VF. So it would be very interesting to think of a possible way of assessing these ratios of volume as a potential evaluation for risk factor towards VF of patients with ARVC. Exactly. It's a very nice observation and goes very much along with what I said. And indeed, now we can, using approaches like what Natalia Trajanova developed using MRI assessment of myocardial mass in different cardiomyopathies and measurements of functional conduction velocity and repolarization could really give a good handle on essentially survival of first waves produced by ectopic activity, for example. Thank you. Thank you very much.

electrical propagation

cardiomyopathic heart

ventricular fibrillation

ventricular tachycardia

arrhythmias

myopathic heart