on Mechanisms of Ventricular Arrhythmias and Acquired Heart Disease. I am Kevin Donohue, and my co-pilot is... Robin Shaw, on time. Thank you for coming. For those of you who've looked at the schedule, thank you for resisting the planned happy hour that is four to 5.30. Maybe I shouldn't have mentioned that. But it's going on somewhere not here. But this is even more fun. So I'm supposed to tell you it's my pleasure to welcome you to San Diego and Heart Rhythm 2025, the 46th Annual Meeting of the Heart Rhythm Society. If you've not already done so, please download the mobile app from your preferred app store. This is how you can participate in live Q&A during sessions. Please scan the QR code on the screen, not there, and to access the session's Q&A, or just come to a microphone. One thing they don't tell you that I am telling you, please introduce yourself because we're all friends, but some of us have not met yet. And that way we'll know who you are and where you're from when you're asking questions. And please join the discussion. So our first speaker is Bjorn Nollman from Vanderbilt who is going to tell us about calcium handling abnormalities. Well, I thank the organizer very much for inviting me to present here. And I've been tasked to talk about calcium handling abnormalities as individual or part of a combined mechanism in acquired ventricular disease. Quite a mouthful. Well, I'll try. So I would like to start out making this complex title much simpler. So there's two types of calcium handling in the heart. I would like to term physiological versus pathological. Physiological calcium release is when the calcium channel triggers our calcium release that's required for excitation-contraction coupling. That's the normal physiological calcium release that we all have to have. Now there's a certain type of calcium release which has been classified when you have gain-of-function mutations in RYR2 that then cause unregulated intracellular calcium release which I have termed, and others, pathological calcium release. And this pathological calcium release is directly linked to ventricular and atrial arrhythmias. A couple years ago we discovered that an old drug, fleconide, actually inhibits this pathological calcium release and is very effective in preventing ventricular arrhythmias in patients with cardiopulmonary polymorphic ventricular tachycardia. These patients have gain-of-function mutations in RYR2. And then over the years, a number of studies, we actually really determined that this dual sodium and RYR blocker is actually the RYR block that's critically important for its antiarrhythmic effect in CPVT in this genetic disorder. Why am I telling you about this genetic disorder is because in structural heart disease, in acquired heart disease, in coronary heart disease, heart failure, atrial fibrillation, a large number of studies have identified gain-of-function and unregulated calcium release due to postventrational modifications. And a number of them listed that together also cause calcium leak, diastolic calcium leak, which then can cause DAD, so the late-after polarization through an electrogenic mechanism by a sodium-calcium exchanger. These DADs have, for decades, been associated with triggered ectopic beats, so it's a way for having a ventricular trigger. However, they can also contribute to the substrate in that they cause areas of slow conduction, which together, a trigger and a substrate can cause ventricular arrhythmias, atrial fibrillation. So the big question that remains is whether this association with RYR hyperactivity in structural heart disease has any mechanistic influence on atrial ventricular arrhythmias. And so we hypothesize that it does, and to test this hypothesize, we want to employ an RYR inhibitor that should be antiarrhythmic in settings of structural heart disease. Now, fortunately, there is a clinical-approved RYR inhibitor that blocks both the skeletal muscle RYR as well as the cardiac RYR dantrolene. Now, it's clinically used primarily to treat malignant hyperthermia, which is due to gain-of-function mutations in the skeletal muscle RYR. So using this as a tool in the last couple of years in a series of studies, we've shown that RYR inhibition with dantrolene prevents VT in animal models and in the ex-vemo human heart, and this was in the guinea pig non-ischemic heart failure model, in a ischemic heart failure model where we showed that, and then in collaboration with Igor, who's also in the audience, in an ex-vemo human heart model, that RYR inhibition with dantrolene prevents ventricular arrhythmias in these models of structural heart disease that are all associated with RYR hyperactivity. Now, but the big question was, does it do the same thing in humans? So about four years ago, in collaboration with Bill Stevenson and Ben Shoemaker, who's also in the audience here, Maj and a number of other fellows, we set out to do a randomized clinical trial of RYR2 inhibition with dantrolene to treat VT in structural heart disease. We enrolled patients that had either coronary artery disease with prior infarct, immunogram use, or reduced left ventricular ejection fraction that were referred to VT ablation. This summarizes the results, and this will be also presented this hour as an oral abstract at 4.48. So the study ran from 2020 to March 24. We screened 500 patients, enrolled 69, and 51 patients complete the protocol. 29 were randomized to dantrolene, 22 were randomized to placebo. So our study had three aims. We wanted to assess the effect of dantrolene on conduction and refactoriness, and also on cardiac output and blood pressure. After all, they're blocking a calcium release channel. And finally, an arrhythmia outcome measure, whether there's an effect on inducible ventricular arrhythmias. And this will be presented elsewhere, but briefly, there were really no significant effect of dantrolene on conduction velocity or refactoriness, and importantly, there was no effect on cardiac output and blood pressure in this patient population with essentially heart failure, structural heart disease. And the aim three was looking at the VT inducibility before and after either study drug or placebo, and as you can see here, in the dantrolene group, there was a drastic reduction from 40% to 14% in VT inducibility, whereas in the placebo group, there was no effect on the VT inducibility in these studies. So from that, we concluded that RYR2 block with dantrolene is safe in patients with structural heart disease, and that dantrolene protects against VT induction by programmed stimulation in heart failure patients. So the problem with dantrolene is that it is also an RYR1 block, and so it causes skeletal muscle weakness, it causes fatal liver toxicity if you lose it long-term, so it's not an ideal drug. So a couple years ago, in collaboration with our chemist Jeff Johnson, we set out to find new small molecules that can modulate RYR2 activity. And so we came across this natural product, it's a depsipeptide, and it turns out the unnatural enantiomer of this natural product is a potent and selective RYR2 inhibitor, which is shown here, and importantly, it actually acts as a partial inhibitor in a random binding assay, and inhibits pathologic calcium release also as a partial suppressor. Importantly, this cyclic depsipeptide has antiarrhythmic activity in vivo, this was in our CPVT model. We've since then gone on and studied this molecule a little bit further, and we're surprised to find that for a peptide, a cyclic peptide, it actually has, it's quite druggable, it's stable in plasma, it has half-life around seven hours in rodents, which is probably easily once a day dosing in humans, it's orally bioavailable, very much like cyclosporine, it's metabolized by 3A4, it does not have phase two metabolism. And it shows in vivo, antiarrhythmic potency in nanograms per mil, and very importantly, it does not affect skeletal muscle, unlike a dantrolene, so it shows no evidence for skeletal muscle weakness. We've begun a study using this molecule to probe the role of RYR hyperactivity in a number of disease models, I'm not gonna go into that in great detail here, but just to show you that in this same model of a mouse chronic ischemic heart disease, so basically where we induce a coronary infarct, let the mouse survive for four weeks, have an established scar, and then we do programmed stim to test for VT induction, and just like dantrolene, infertisolide works just as well to prevent monomorphic VT in this model, suggesting that RYR hyperactivity not only contributes to the trigger, but also to the substrate for VT in a structural heart disease model, as also suggested, obviously, by our human data. So with that, I think it's time to update this very old classification that was proposed in 1970, the Von Williams Classification of Antirhythmic Drugs, and add a class five, the OR2 channel blockade, as we actually proposed in our recent edition of the textbook of pharmacology. So with that, I thank the people in my lab and my group, and the people at Vanderbilt and all the collaborators for their help with all these studies, and the funding sources, and this gives us plenty of time for questions. Thank you. These are the online questions. These are the online questions. Thank you, Bjorn. Questions from the audience. I can start off. Is that okay? You know, I had two questions. You answered the first question. When is that gonna be added to Von Williams? And you actually have an authoritative take on when that. Yes. And then the second is really the meaning behind that, is are we talking for, and the data are beautiful, blocking right in receptors? Is this what triggered arrhythmias? Can we go after reentrant arrhythmias? You know, in terms of mechanism, all of the above, how do, where is their potency considered relative to class one or class threes? So these are all questions, good questions. So we did this study to ask the fundamental question, does it prevent reentrant arrhythmias? Does it modulate the substrate? And we now have shown this in like three different animal models and most recently in the human trial because it prevented monomorphic inducible VT. So that is a, it clearly affects the substrate. Now, why and how? Again, in modeling data, if you have a local area of spontaneous calcium release which causes a subthreshold membrane depolarization, DAD, so we know these spontaneous calcium releases occur in thousands of millions of cells. So even if it doesn't trigger VT, but it will change the membrane potential and that can either accelerate conduction or it can slow down conduction depending on where you are in that membrane potential or conduction velocity relationship. Furthermore, it can also activate it in especially in diseased heart, calcium activated potassium channels which will shorten the refractory period. And that's what we actually found in our animal model that the refractoriness was shorter and eventually normalized the ERP refractory period. So this, from multiple sources of evidence, suggests that IR modulation is not just the classical trigger, but it also seems to stabilize the substrate. It's beautiful. Hi, Pat Nafke with Catholic Physician, also grad student at Medical College of Wisconsin doing computational modeling. So did you look at, like, were you able to separate out the effect of the, say, the acupotential effects versus the substrate effects? To, you know, like, I mean, I'm just, you kind of get both with the calcium overload, right? But, you know, did you try to see if one of those are, you need both of those at the same time or if it's just strictly a substrate effect? So it clearly prevents, in the animal models, the triggers. Clearly, the number of PVCs went down, which is easy to show. And in terms of the substrate, we measured conduction velocity in the XP movement how we really didn't see an effect. We measured it in our clinical trial. We didn't see much of an effect. But again, this could be very local to wherever the entry circuit is. So the answer is we really don't know. In a mouse model, and this work was published, the post ischemic model, the acupotentials were shorter, and after dantrolene, they became normal. In the guinea pig model, the guinea pigs actually have a longer acupotential than heart failure, and treatment with dantrolene shortened them. So it basically normalized them in both species. Now, guinea pigs and mice have very different acupotential regulations, you know, high plateau, low plateau. Calcium release affects it differently. So that's all I can say to do, and we can follow up after this in more details. Thank you. is Anna Finninger from Northwestern University, who will talk about the contradiction of conduction abnormalities to scar-related VT. Thank you, I was hoping I could, you better keep going, you're on your right. It's your turn. Well thank you all for being here and not at the happy hour. These are my disclosures. And I'm excited to talk about VT because I spend most of my time thinking about atrial fibrillation and sometimes it's nice to think about an arrhythmia that's a bit more organized than atrial fibrillation. So today I'm going to share some of our work, actually share with Kevin Donahue here on the role of contribution of conduction abnormalities to specifically scar-related ventricular tachycardia. So let's talk with a clinical case. This is a 67-year-old male who has ischemic cardiomyopathy and who had recurrent VT despite amiodarone treatment. He came to the EP lab, was easily inducible, but as often happens, had a hemodynamic instability with VT. He had two predominant VT morphologies, which you can see here, either right bundle superior axis, which is VT1, or a right bundle inferior axis, which is VT2. First looking at his voltage map, he had a large, a lot of scar, but with a predominant lateral scar with an irregular border here. And because of a hemodynamic instability, we first started looking at isochronal late activation. You can see here the propagation of this ILM and then this overall gives us a ILM map such as this one where you can see a deceleration zone at the anterior edge of the scar here. And we can see here the difference even within the same catheter placement between the green splines, which are earlier, and the red splines, which are really close but much later. And so knowing this, we could place our catheter right at this spot here, the green dot, which is where the deceleration zone was. And in VT, these here at this spline up here indicated by P, which led to termination of the VT with a local capture but no global capture consistent with presence in the isthmus. And even more, the next capture beat here resembled VT1, suggesting that the same isthmus was responsible for both VTs. And so knowing this, we ablated at the site of entrainment, which slowed and terminated VT2, and VT1 was no longer inducible. So this is a best-case scenario, it doesn't happen always, but a scenario where the ablation strategy guided by conduction information, or ILAM, it was shown in clinical trials that the deceleration zone in sinus rhythm can predict critical sites for reentry, and that ablation of these sites can lead to functional changes in the propagation. And in a trial here, looking at patients up to 12 months, showed freedom from VT in about 70% of patients. Of course, this is not randomized, so we don't know if this is the best strategy, but it seems promising. And so clinically, targeting conduction abnormalities seems to be a fairly effective strategy for VT ablation. However, we still don't really understand what the mechanisms are that can lead to abnormal conduction. And you'll say, of course we know. Obviously, you know there is scar, we have, there can be physical disruptions, fibrosis, fat, we can have sourcing mismatch, or even the cells themselves could have changes in connection expression, or the function of gap junctions can be different. So if you talk about this to people in the field, it seems like we already know what's going on, but do we? If we look at the data, what is out there, first looking at gap junctions, fairly old work's been really trying to focus on this and understand this, and it seemed that actually, the size of gap junctions is actually not different between regions that are infarcted or not infarcted in the myocardium. There is, seems to be a pattern with lateralization of connexin-43 immunosignal in the infarct border zone. However, the question is whether this is truly related to the VT isthmus, or maybe might be more ubiquitous in all forms of myocardial remodeling. Here is a study on the right showing that there is a reduction in connexin-43 density in all forms of cardiomyopathy, whether it's dilated cardiomyopathy, ischemic cardiomyopathy, or even post-myocarditis. So there might be an association between connexin-43 localization and VT, but the association between whether there's a functional role still remains unclear. How about fibrosis? We know that there is fibrosis and scar, and that there is some, that it might be involved. However, if we look, there's an old study looking at patients who had a scar but had no VT, or patients who had scar and had either subacutic chronic VT, it seemed that the amount of fibrosis in the myocardium could not distinguish patients who had VT from the ones who did not. There's exactly the same amount of fibrosis or necrosis in the patchy infarct of these patients. So we're back at this. What are the mechanisms for abnormal conduction in ventricular tachycardia, and what differentiates VT substrate from other regions with scar, but that are not participating in the VT substrate? And I think a main limitation of the prior studies is that it's actually exceedingly hard to appropriately measure conduction velocity, just collection of velocity in the myocardium, and to really relate what's happening in the tissue versus what's happening functionally with VT, especially in humans. And so here, the goal of the study was to try to answer these questions by using a large animal model where the infarct is very defined, we can map extensively, and then we can obtain tissue. So here we used the porcine model of post-infarct VT. Animals, we form an MI here by occluding the mid-LED with a balloon. We get an acute ST elevation MI. The animals recover, and then after five weeks, we can perform an invasive EP study, verify that they do have inducible monomorphic VT, and then spend a lot of time determining where is that VT, where is their circuit? And then, knowing this, we can separate the tissue, the hard tissue, into wedges that contain the VT circuit, wedges that have scarred our border zone but do not contain the VT circuit, and then remote myocardium. And these wedges can be exposed ex vivo to optical mapping to confirm presence of a VT circuit or not, and then also obviously used for tissue analysis. So first, looking at ex vivo optical mapping, we can also measure conduction velocity there in greater detail than in vivo. We found that while there was only a marginal reduction in the conduction velocity in the longitudinal axis in wedges that contain the VT circuit versus wedges that did not contain the VT circuit, there was no effect on the transverse conduction velocity, and the result is that the anisotropy is actually different. And so there is an attenuation or reduction in a normal anisotropy in the wedges that contain the VT circuit compared to the ones that didn't. So our next question were what can explain this change in conduction properties in the wedge at the tissue level? First, we looked at fibrosis with mesentrichrome, and while we obviously found scar in blue in the border zone wedges, the amount of scar and even the presence of kind of cells and bundles of myocytes within that scar could not differentiate the wedges that contain the VT circuit here in red to the wedges that did not contain the VT circuit in orange. Next, we looked at the cells themselves, right? The cells act as cables, and if we look at the cell size, the diameter of the, kind of in the short axis of cells, while there's a decent heterogeneity in all regions with high difference with large cells on the bottom and small cells on the top, there was virtually no difference between regions examined, and even between scar area and non-scar areas. The cells themselves did not appear to be of slightly different sizes. So the next question is what connects cells? The number one kind of culprit for conduction between cells is connexin 43, and it was previously shown that the GJA1, which is the gene encoding connexin 43, is globally, the expression is globally reduced in the scar area, but that did not distinguish regions that lead to VT versus regions that do not lead to VT. But that's the gene, so what happens at the protein level? And for this, we can look at the protein connexin 43 localization. And while the majority of connexin 43 should be localized at the intercalated disk, in pathological settings, it can be lateralized. And so here, in this example, I'm showing you connexin 43 immunosignal in green. We have another marker in cadherin here in red. And if we look at the fiber orientation here, indicated by the arrow on the side, we know that what is perpendicular to the fiber orientation is consistent with intercalated disk localization, whereas what is parallel to the fiber orientation is suggestive of lateralization of connexin 43 to the lateral membrane, which is seen in pathological settings. So then we looked at our three different groups. So uninfarcted tissue, wedges that have SCAR border zone but no VT, or wedges that included the VT circuit. And I don't know how well it projects here, but you can see more green that's parallel with the fibers here in the VT circuit area compared to, can I see this? Okay, well, I'll guide you through this. So on the right, more lateralized, on the left, more lateralized green signal to the lateral membrane, that's parallel with the arrows, whereas in the non-VT scar, we see less lateralization. And on the right, I'm showing you the summary statistics where we see about 40% of the total membrane signal is lateralized in the VT circuit area compared to non-VT scar and uninfracted tissue. Now, if you look at this, you'll say, well, yes, the green is lateralized, but so is the red. So what happens to other junction proteins? And so we perform the same analysis looking at encadherin, which is an adherence junction protein, and we see actually the same pattern. Encadherin is lateralized in the wedges containing the VT circuit compared to non-VT scar. So we have connexin gap junctions, we have adherence junctions. What else, are other proteins, do they travel together? And it seems that that was shown in other contexts, that connexin 43 and desmosomal proteins seem to travel together. And this was the case, so we looked at PKP2, which is one of the components of the desmosome. And here again, we could find the PKP2 that's perpendicular to the cell fiber, consistent with the intercalated disk, and this projects even less well, but also PKP2 that's parallel to the membrane, to the fibers, and so that's consistent with lateral membrane signal. And here again, in the VT circuit, it seems that there is a lateralization of desmosomal protein PKP2 compared to non-VT scar. So, so far of this kind of work in progress, I shared with you that connexin abnormalities in the VT substrate are difficult to assess. We do see structural changes in the scar border zone, but that does not distinguish from the VT substrate, from the non-VT substrate, which would suggest that fibrosis and other structural changes are required, but perhaps not sufficient to lead to a VT substrate. However, if we look at the subcellular structure, we see that the VT substrate is characterized by remodeling of cell junction proteins, which includes at least gap junctions, adhering junctions, and desmosomes. And so our hypothesis is that these changes participate in the altered anisotropy and reentry. I'd like to thank my growing lab members at Northwestern, as well as Kevin Donahue and his group, who were essential to this work, and thank you all for being here. over here is to put in a question to give time for people to come to the microphone. So I'm happy to play that role. It was a wonderful presentation. Canonically, the redistribution of connection to lateral membrane is considered ultra-trafficking. Before that, it was sort of movement in the membrane, but nonetheless, these hemichandles were lateralized. And, but when you throw in catherine and you throw in desmosomal proteins, just to ask the sort of obvious question, which is the 800-pound gorilla in the room, I guess, with his work, and it's very interesting, is the inter-clated disc moving to lateral membrane? I mean, and do you have some EM studies or you must have surely thought about that, wouldn't you? Yeah, so that's a, right, it's an excellent question. So we think lateralized connection 43 is acting as hemichandles. Here, it suggests, based on the data that we have, which is paraffin-embedded tissue, that there is cell-cell contact, and so that these lateralized connections may act as full gap junctions and not as hemichandles, right? And so the question is, why is this happening, and is there just depolarization of the whole cell, right? Are the, instead of having the inter-clated disc just as the end, do they start to move to the sides? And I don't have EM data, I wish I did, to really answer that question, but it sure does look like that, that the cells become less polarized in terms of lateral versus kind of end junctions, yeah. Very different model, interesting. applications of this work. So now we target those areas, right, of deacceleration modulation. Do you see any room for more targeted modulation of VT substrate based on all of the work that you've been doing? Yeah, that's a very good question. So the, kind of, I think the end question is, can we cure or treat VT without causing more scar, right? And so if the substrate is more physiological, you know, a cell expression problem, we might be able to fix that by changing gene expression, right? And I think I'm gonna have to cite Kevin Donahue's work where they showed that in animal models, if you overexpress connexins in the ventricles with gene therapy, you can, you know, improve that, you can reduce VT in an animal model. And so my hope is that one day, we'll figure out how to do gene therapy appropriately, durably, you know, with all the questions involved, and we may be able to not induce scar, and so, you know, not oblique more, yeah. Thank you. And so the next speaker is Chris Rippinger, and neuronal myocyte interactions in diseased ventricles, arrhythmia triggers, the staining mechanism, or both. All right, thank you so much, and thank you to the organizers for inviting me to speak today. So yes, I was given this title of neuron myocyte interactions, arrhythmia triggers, sustaining mechanisms, or both. If you just wanna like check out for the next 15 minutes, I can just tell you right now, it's probably both, so. But if you wanna stick around, I'll tell you a little bit about why. So I'm going to discuss mostly structural and functional remodeling of the sympathetic nerves following myocardial infarction today. And remodeling occurs in many different forms, but most relevant to what I'm gonna talk about today is the fact that after myocardial infarction, the infarct regions can be hypo-innervated, or have loss of nerves in the infarct region. And that's because of the presence of something called chondroitin sulfate proteoglycans, or CSPGs, and my collaborator, Beth Haubecker, has shown this in multiple different models, primarily in rodents. We just showed this just last year in the rabbit model as well. And so that's what's shown, can you see my, no. All right, now you can see it. And yeah, these don't project super well, but in the red is the CSPGs in the infarct area. And what happens is the sympathetic neurons have a receptor for the CSPGs, they try to grow back in, the CSPGs brine to that receptor, and they stop, leaving the infarct area denervated. But despite that, the border zone regions can actually be hyper-innervated, because there's lots of nerve growth factor in and around the infarct and border zone. So this is just some data from the UCLA group, showing all of these very small nerve fibers in the border zone, so you can get quite heterogeneous sympathetic innervation. And another thing that happens is something called cholinergic transdifferentiation. So this is a process by which the sympathetic nerves can actually, for a period of time, start to make the parasympathetic transmitter acetylcholine. And so that's shown here. There's norepinephrine content on the top, and you can see in the scar, two to three weeks after MI, the norepinephrine content goes down, and that would be expected, because there's not a lot of sympathetic nerves there. But if we look at the acetylcholine content in the scar, it remains about the same, but in the viable tissue away from the scar, you actually get this increase in acetylcholine release, particularly around one to two weeks, which then subsides by about three weeks. And that's because the sympathetic nerves are actually making and releasing acetylcholine for a little while there after MI. So when we think about sustaining mechanisms, and so things that might promote the maintenance of arrhythmias, we know that MI increases action potential duration dispersion and dispersion of repolarization, just because the presence of the infarct and all of that remodeled tissue. And we found that nerve loss in the infarct, or denervation, can exacerbate this heterogeneity during sympathetic activity. So here is a cartoon of just two MIs. This would be the normal situation where the infarct remains denervated. And then this is a mouse model in which the receptor for CSPGs in the neurons was knocked out. So the neurons essentially don't see the CSPGs, and they can grow back in to the infarct area. And when we add a beta agonist such as isoproteranol, we can see that these denervated infarcts have a really dramatic response, and the action potential shortens a lot, whereas the re-enervated infarcts just have a more of a moderate physiological response. And we quantified that here. And so that led us to sort of this idea that the denervation must lead to some sort of super sensitivity to catecholamines. So normally the heart is richly enervated, that loss of nerves in the infarct region likely leads to something happening within the cardiomyocytes themselves, that then they become super sensitive to catecholamines. And this can then exacerbate APD dispersion and creating the substrate for arrhythmias. However, we also found a few years ago that that process of transdifferentiation, or the sympathetic nerve-released acetylcholine, may counteract that heterogeneity just a little bit. And so in a different mouse model, we had mice in which choline acetyltransferase, or CHAT, was knocked out just from the sympathetic neurons. So the parasympathetic neurons are still making acetylcholine as normal, but the sympathetic neurons were not. We created MI in those animals. And we found when we knocked that out, the action potential duration dispersion, or heterogeneity, was actually increased quite a bit, and that can be shown in these maps of action potential duration here. And so we think what happens is that from about one to two weeks post-MI, that sympathetic nerve-released acetylcholine may counteract some of this heterogeneity and could sort of be transiently protective, reducing some of the APD dispersion. If we move on to think about arrhythmia triggers, again, we see that these denervated infarcts exhibit calcium mishandling and triggered activity. So back to this model of the denervated and reinnervated infarcts, the denervated infarcts, when they're exposed to isoproteranol, have diastolic calcium elevation, which can then lead to triggered activity in PVCs that's quantified here. Those PVCs could be blocked with the beta-blocker propranolol, showing that they are mediated through beta-adrenergic receptors. When we took a look at the maps, we saw that the PVCs were arising from this infarct area. So due to this supersensitive response to catecholamines or beta-agonists, there's calcium mishandling and triggered activity. So then more recently, we've really moved on to try to better understand what might be the mechanisms of this supersensitivity. Why are these sort of amplified responses in the denervated infarcts? And so if we think about that, this is just a little cartoon of a sympathetic neuron. It releases norepinephrine. That binds to beta-1-adrenergic receptors, which stimulates adenylacyclase to make cyclic AMP. Cyclic AMP then goes on to activate PKA. PKA phosphorylates all the downstream ion channels and calcium handling targets, which may be leading to some of these functional consequences. So we thought if cyclic AMP is a key second messenger in this signaling cascade, maybe we could take a look and see if perhaps amplification of this second messenger is contributing to this supersensitive response. And I also just want to point out that these things called phosphodiesterases, or PDEs, they're responsible for rapidly breaking down cyclic AMP and turning that signal off, and that's gonna be important in a minute. So to take a look at this and look at mechanisms of supersensitivity to see if this involves cyclic AMP, we used a FRET sensor, which changes its fluorescence when binding cyclic AMP. We generated a mouse model with cardiac-specific expression of this cyclic AMP FRET sensor, and then we modified one of our optical mapping systems to take a look at whole-heart FRET imaging in Langendorff perfused hearts, and that's just shown here. And when we modified that system, we could add beta-adrenergic agonists, or catecholamines, and then we could take a look at whole-heart cyclic AMP responses. So we could now look at this second messenger, and then we modified that system at the same time to also be imaging voltage and calcium to look at functional outputs. So we wanted to ask the kind of relatively simple question, does elevated cyclic AMP activity underlie this post-MI beta-adrenergic supersensitivity that we've been seeing? Is that sort of an amplification point in leading to supersensitivity? So we performed myocardial infarction in these mice that expressed the cyclic AMP FRET sensor, and in the top, we have a sham animal in which we've done a bolus of norepinephrine, so we just wash norepinephrine into the perfusate, it goes in and out, and the time course of that signal is shown in black here, so you can see it rapidly increases and then rapidly decreases. And in the MI, when we do that, contrary to what I thought would happen, I thought we'd be making more cyclic AMP, that's not what happened, we actually make a little bit less. But what was really striking were these sort of decay kinetics were really, really slow. So the cyclic AMP is not being rapidly broken down as it should be, it's sort of lingering around a bit too long, and those decay kinetics are just quantified here. We also found that it was this period after the MI, or after the norepinephrine application that we started to see a lot of triggered activity during these later phases when cyclic AMP should be broken down, and it's not, and that's quantified here. And again, we saw these arrhythmia triggers arising from the infarct region. Now, we did a slightly different experiment, but we encouraged the nerves to grow back. This time, instead of using a genetic model, we applied a drug after MI to the animals that blocks the receptor on neurons for CSPGs, so the neurons, again, don't sort of see the CSPGs, they can grow back into the infarct. We repeated that same experiment, and although it didn't rescue the peak cyclic AMP levels, it really sort of rescued these decay kinetics. So again, we had really rapid, fast breakdown of cyclic AMP, which was then associated with far fewer PVCs, so it sort of prevented those arrhythmia triggers when we can rapidly degrade this second messenger. So this led us to wonder why would innervation impact the breakdown of cyclic AMP in the cardiac myocytes. Doesn't really make a lot of sense why that should happen, but we think it lies in the creation of these signaling complexes. So this phosphodiesterase, the enzyme that breaks down cyclic AMP, is held right at the beta-1 receptor through an anchoring protein called SAP-97, and some work from several years ago from Brian Kobilka's group in co-culture, when they took cardiac myocytes and sympathetic neurons, grew them together in co-culture, and they took a look at these signaling complexes, and in particular, this SAP-97 molecule within the cardiac myocyte, they found that it was expressed in the cardiac myocyte only where the neuron was laying right over the cardiac myocyte. So something about the physical presence or signaling presence of that neuron might help the arrangement of these adrenergic signaling complexes in the cardiac myocytes. So we think probably what's happening is when the nerve is lost, that signaling complex gets disrupted, PDE can dissociate, and then we've got too much cyclic AMP hanging around for too long. It's possibly also diffusing to other locations within the cell and signaling in ways that are causing arrhythmias. And then in just my last minute here, I just wanted to address one other mechanism. Does this nerve loss impact norepinephrine reuptake? So when the sympathetic nerves release norepinephrine, about 90% of it is taken right back up by the sympathetic neurons and recycled, so it's efficiently cleared away from the receptor. So we also thought, well, if those nerves are gone, maybe the catecholamines are just signaling too long at the cardiac myocytes, and it's not being appropriately taken back up by those nerves, and that could be another reason why the cyclic AMP signaling is lasting too long. Maybe there's just signaling at the receptor for too long. And so to address that, we created another mouse model of a sensor called the Grab Norepinephrine Sensor. So it's a non-functional alpha-adrenergic receptor expressed in cardiac myocytes linked to a fluorophore. When it binds a catecholamine, it lights up, and it can tell us about catecholamine signaling within this neurocardiac junction. And so we made these mice, we took their hearts, we added norepinephrine, and they light up. And then we did myocardial infarction in these hearts, and for these experiments, we isolated the mouse heart with the sympathetic nerves and the spinal cord intact, so that's what's shown here, when we can put a catheter to the first thoracic vertebrae, stimulate the sympathetic nerves. When we did that after MI in these hearts, these norepinephrine-sensing hearts, we can see during nerve stimulation, the infarct remains dark. So that confirms that they're denervated, there's not a lot of catecholamines being released in the infarct. However, if we washed on catecholamines, if we gave a bolus of norepinephrine, we can see the infarct lights up quite a bit, so we know that there's lots of surviving cardiomyocytes in there. But in this case, the norepinephrine lingered at the infarct for much longer, this stayed bright for a longer period of time, it wasn't efficiently washed out and cleared away by those neurons, and we think this could be another mechanism by which the signaling is kind of lingering for too long, and so this is something, a new idea that we're pursuing with these new sensor mice. So to finish up here, we think changes to post-semi nerve structure and function likely contribute to both arrhythmia triggers and sustaining mechanisms. Unlike what we originally hypothesized, we think this denervation supersensitivity is not necessarily associated with amplified cyclic AMP signaling of the second messenger, but potentially with impaired breakdown. We think that restoring sympathetic innervation may increase the breakdown of cyclic AMP by restoring these signaling complexes and phosphodiesterase activity, and there also might be another contributing factor of impaired norepinephrine reuptake when those nerves are gone, and norepinephrine staying at the receptor for too long. And so I'll end there and acknowledge all of my talented group and collaborators and funding. Thank you. So from a translational standpoint, if you had to put your nickel down on one thing that you would reverse or prevent, what do you think it would be? So I think, translationally, I think that there could be some potential in growing these nerves back. And when we first did it 10 years ago, together with my collaborator, Beth Haubecker, she said, I wanna grow the nerves back and see what happens. I was like, that seems like a terrible idea. Because we think of sympathetic activity as being proarrhythmic. But in this case, I think restoring that normal levels of sympathetic signaling, maybe making the nerves more homogenous so we don't have heterogeneous nerve density could be helpful. In your last mouse model, when you showed nicely the spinal stimulation increases in norepinephrine release and uptake, I couldn't help but notice that the right atrium was really bright. Is that a unique finding in the post-MI model? Or did you see that in normal mice as well? And what do you think that means? So I think we do see this in normal mice as well. And I don't know if the right is really that much brighter than the left. It could have been just exactly how the heart is positioned there. But we think it's because the atria are so rich with sympathetic innervation. And so we see both the norepinephrine and lots of cyclic AMP being made in the atria versus the ventricles when we do this type of stimulation. So we actually have another project where we're just focusing on the atria because I think there's a lot of work to be done there. That was fantastic, Crystal. Just on your figures, you showed PDE4D. And I'm just wondering if you've looked in detail at is it PDE4D that's the key or have you looked at other subtypes as well? So we think, yes, we have looked. I didn't present that data today just in the interest of time. We've looked at some of the other cardiac subtypes, PDE2, PDE3, PDE4. And actually, after MI, none of them are down. They're all robustly expressed after MI, which is why we think maybe it's where they're expressed and in that complex is important. We're focusing on 4D because we've seen some sex differences in 4D. And we also think 4D is anchored there to the beta-1 receptor, so it might be important at that complex. But we are looking at the major cardiac isoforms as well, yeah, thank you. Thank you. So our final speaker of the session is Saman Nazarian from the University of Pennsylvania, and he will be speaking about the contributions of non-myocyte cellular factors to VT mechanism. Right, well, let me start by thanking Kevin and Robin for the invitation to be here, and to all of you for staying till the end of the day. My talk is a bit more clinical, and those were my disclosures. Let's see, I'll make a laser pointer here. Okay, so, let's start with reentry requirements. We're all familiar with this. What we need is an anatomical obstacle, and a circuit of adequate size, which is dictated by not just the physical size of the circuit, but also the conduction velocity, the refractory period, or duration of the action potential, and then we need unidirectional block for the circuit to start, and as electrophysiologists, we're all obviously very interested in understanding the properties, not just of the obstacle, but also the surviving myocardium and how it's affected by the obstacle, and that may be, as was mentioned in earlier talks, composed of things other than just fibrosis. Now, we can look at the setup for reentry fairly well with 3D imaging. I've been in the space of looking at high-resolution MRI for BT circuits for quite some time now, and there are software packages available that can take a 3D image and develop from that image. These are not electroanatomic maps, but rather reconstructed MRI images that show you the potential channels or corridors. I should clarify for this audience that thinks of channels differently, corridors through this fibrotic or infracted area, and the problem with this is it can give you a place to go look, but often, if you actually do a lot of retabulations and take images, you'll notice that there's a large infracted region and hundreds of channels to look at, and that doesn't really save you time. It actually gives you more headaches for trying to find these. And one thing that we noticed about, I guess, 10 years ago now is that there is a lot of fat inside infracted tissue. We didn't notice this. This was known to pathologists. This was known to radiologists, but really, in the AP field, I think to me and to most of my colleagues, this was relatively new. And so we thought, well, if we look at some of these patients that are sent to us and we get an MRI and we assume that the regions of infarct are all fibrosis, how much of that is really fibrosis and how much of it is fat, like pathologists have known for a long time? And when you get a CT scan and you compare very similar cuts as close as you can get and take into account that anything that has basically an intensity less than 30, which would look black on the CT image and is inside the myocardium is fat, you realize that significant proportions of infarcted tissue that we were just considering was fibrosis is actually fat. How much of this occurs quickly, how much of it late? Really, at a macro level, you have to wait four years for that amount of fat to be visible. But if you take infarcted animals in a swine model, for example, within four months, you'll start to see some fat. So you do have to wait to see this sort of thing. And what's interesting is four to 10 years, the timeframe that it takes to develop this kind of lipomatous metapoiesia inside the scar is actually the timeframe that it takes to develop monomorphic VT clinically. Patients after an infarct don't develop VT in the first year. Most of them take four to 10 years to develop monomorphic ventricular pectoral. So the next task we took was to take this to the patient level and really study this. And what we thought we would do is, for patients, take an MRI and using that MRI, reconstruct potential corridors that are traversing the infarct tissue. And then get a CAT scan and look at the corridors that actually traverse not only infarct tissue, but the fat and see if there's anything distinguishing about those corridors and the way they behave or in the way they participate in ventricular tachycardia. So in total, we had, for this first phase study, 30 patients post-MI. You can see the median number of years post-MI was actually 17 years. So it takes a while for these infarcts to mature. And I think this is really important. Out of 381 corridors that we segmented out of these complex infarcts, 84 of them in 30 patients participated in VT circuits. And 99% of these 84 corridors were surrounded by lipomatous metaplasia. In comparison, only 4% of non-VT corridors were surrounded by lipomatous metaplasia or fat. Now, we tried to figure out why this might be. And one hypothesis was that, well, the fat might be a better insulating agent for that extracellular current. I'm a little bit worried talking to a very basic audience about this. Because as clinicians, we can't actually measure current, obviously, and we're really estimating this using very inexact methodology by using Ohm's law, because we know the unipolar voltage and we know the unipolar impedance. But for what it's worth, estimating this, we find that it's a lot more stable when you measure, when you estimate current along the path on multiple points on these corridors that traverse fat, as opposed to corridors that are only going through fibrosis. So it seems like that could be a potential explanation. Also, when we look at these corridors and measure conduction velocity across the corridor, these are conduction velocities measured during baseline rhythm, not during VT. You find that, in general, the conduction velocity is far lower. And also, the standard deviation of the conduction velocity is lower, compared to the conduction velocity as impulses traverse corridors that are not going through lipomatous metaplasia. So there's an association of slow conduction velocity with corridors that are bounded by fat. And when we try to estimate the repolarization period using the ARI, which you can obtain from the unipolar electrogram, what you find is that corridors that traverse lipomatous metaplasia have longer activation recovery intervals as a reasonable surrogate of the repolarization interval. And how could that be relevant, of course, in terms of developing unidirectional block as an impulse traverses that tissue? It could be very, very important. So let me show you an example. This is a patient of mine that has an incredible story, actually, a really, really incredible person. He had his first MI at a young age by having an accidental nail gun trauma to his chest. He developed a pneumothorax and also lacerated his LAD, had an anterior MI, and I followed him for years before he actually started developing VT. But he had VT many years after his MI, and one of my partners first actually took him for his first VT ablation and really did a pretty thorough job of chasing down several VTs and doing largely a substrate ablation. The next day, at our institution, we're in the habit of performing NIPS. After the procedure, we performed the NIPS, and he was still inducible for VT. Nevertheless, we started him on sotal law and wanted to see how he would do with that one ablation. He did have a recurrence, and it was a similar VT morphology. And so I brought him back to the lab and very readily could still induce this VT. So let's look at his substrate. Here's a CT scan. You can see this anterior infarct with a very thinned out apical segment. And here's the area that I'm hoping you'll appreciate, has very low Hounsfield units, and is either air or fat, and there's no air inside the myocardial wall, so this is fat. And you can just see it with the visible eye, but you can also use software to really bring in the area that has fat deposition. So here, at the very beginning of the case, we noticed some pretty interesting, very delayed fractionated electrograms there. And when I pace, first at 50 milliamps, I have this morphology, and many years ago I would have moved on, but I've learned that you really have to change your voltage, change your output, and look at what you capture locally as opposed to large field. And when you drop to 25 milliamps, you start to get a little closer to the VT morphology, and when you drop your output to less than 10 milliamps, look at how close to the VT morphology that gets. Really, I think, demonstrating nicely how this tissue with fat deposition in it is compartmentalized, and as you drop your output, you can capture specific quarters within that tissue. And here's during VT. I'm not gonna pretend that I can say where the post-pace interval is here. It's very fractionated, and we could make it look good, but suffice it to say that there's certainly concealed fusion here, and probably a good post-pacing interval based on that. And here's how we avoided his VT. This is, I wanna underscore again, look at all the area of scar. And we really just honed in on this area where the fat deposition is. That's where you can go immediately to start fishing. And it was hemodynamically unstable, but because we had good pace morphology information, we induced VT, and then immediately came on, and it terminated very quickly. And just to show you the time it took, here at the beginning, you can see induction. We were very easily able to induce it with doubles. As soon as we induced it, we came on ablation, and it's a clean break with prolongation of the cycle length, which made us feel pretty good. And sure enough, he was not inducible after that. So certainly very, very useful in the setting of ischemic cardiomyopathy. Now what about non-ischemic cardiomyopathy? We've done some work with this, and it's dependent on the ideology of non-ischemic cardiomyopathy. In sarcoidosis, there's a fair bit of fat deposition. In other types of non-ischemic cardiomyopathy also, you can see fat deposition. It's in smaller amounts, and it's more heterogeneous. It's very much, as you can see from this AHA plot, deposited in the more basal sections, sectors of the heart. And that's, of course, from our experience, you know it's very important. The basal-most segments of the heart are very important in non-ischemic VTs. And we've shown that these corridors that we segment from the scar images, if they traverse the lipomatous metaplegia, they're much more likely to participate in VT. Here's a violin plot showing the volume of fat surrounding corridors that participate in VT as opposed to corridors that don't participate in VT. And in fact, you can use, we've come up with thresholds that you can use to better understand the likelihood of a corridor participating in VT based on how much fat is deposited there. So let me show you a quick case of a non-ischemic cardiomyopathy patient now. This is a 77-year-old, also non-ischemic cardiomyopathy, very large dilated LV, also had a prior VT ablation, which was very thorough. You can see really pretty dense substrate modification of this very lateral basal area. And he had VT recurrence despite being on amiodarone. And so I brought him back to the lab. And here's his MRI. You can see this basal lateral area of scar fairly well on the MRI. There's some artifact from the defibrillator, but pretty easy to identify that. And here also on the four chamber image, this area of late gadolinium enhancement. Large area, it's not a small area. And if we reconstruct that scar, you can see a fairly large area. And as I mentioned before, you get all these different corridors. Well, which one do I want to focus on? Which part of this, it was already completely ablated before. And it's nice to be able to focus on one area. And so here we obtained a CT. And I'm just gonna show this so that you can appreciate the fat deposition in this non-ischemic patient. So this is scrolling through this lateral basal portion of the left ventricle. And as it goes through, you can see these islands of fat coming in. It's much smaller, granted, than an ischemic cardiomyopathy where you can just see patches of fat inside that myocardium. But these areas of fat are real. There's no question that it's there. It can't be anything but fat if it's inside the myocardium here and it's that dark. And so we can segment that. And it focuses on us on a much, much smaller region of the scar. And here I have the patient in the lab. There's an agilis in the coronary sinus and two multipolar catheters in two lateral branches. This lower one is where the fat deposition was. And I'll show you on the map image as well. And here in this VT, he actually had two different morphologies of VT. And this is the bottom set of electrograms are from this lower multipolar catheter. And you can just very readily appreciate how much of the cycle length is represented on these poles that are going through that region with the fat. And here's the entrainment from one of those catheters. Very much a good response. And I actually initially in this patient thought, well, that branch is going right through the fat region and it has great signals. So let's use some alcohol, put in alcohol. You can see that it's bright after alcohol injection on intracardiac echo, but it's loaded. It didn't get rid of it for whatever reason. It probably didn't infuse into the entire tissue. And so then I obtained epicardial access. And just to show you here, look at the very large area of scar, the dark black spots, these are the delayed potentials. And they generally correspond to these corridors that we're seeing that we had segmented from the MRI. And so it is a very large area. Now compare that to the fat map. And what I really wanna show you is look at the ILM map and the fat map. I was really pleased with this. It really shows that the slowing and the conduction issues are very much concentrated in that region with the fat deposition. And very delayed potentials here to just corroborate the finding from the ILM map. And again, excellent pace morphology for the VT here and the pacing there. And ablation at this site, epicardially rendered both of those VTs non-inducible. So I'll stop here and conclude that infarct tissue is really very heterogeneous, consisting of not just fibrosis, but also fat. And it matters what's bounding the corridor. The corridor really behaves differently dependent on the boundaries. What's the mechanism of that? I think I'll have to depend on the smarter people in this room to figure that out. But there's clearly differences when the corridors are bounded by fibrosis versus fat tissue. And non-ischemic scar myocardium can also contain fat. It's often in smaller amounts and more diffuse increments, but really does seem to have potentially important electrophysiologic effects. And it depends very much on the type of non-ischemic substrate. And there's a lot of functional information that we can gain from these images to help with our mapping and ablation procedures. So I thank you for your interest and attention and look forward to further discussion. Thank you. Thank you. afterwards, but that brings the session to an end. So thank you for hanging out with us through a very long but hopefully very productive day.

ventricular arrhythmias

Heart Rhythm 2025

calcium handling

flecainide

dantrolene

scar-related ventricular tachycardia

sympathetic nerve remodeling

non-myocyte cellular factors

ischemic cardiomyopathies

ablation strategies