So, ladies and gentlemen, it is time and if Colleen, my co-chair, might make her way to the front, please, then we can get this session started. As you know, the talks are very limited in time, 12 minutes, followed by discussion, and we are not as tight in the schedule as we have been in previous years, but we should try to stick more or less to the allocated times. So my name is Peter Kohl, I have the pleasure of co-chairing with Colleen Clancy, who is over there. Can you wave? No? And I would like to hand over to Indhuja to give the first presentation. Good afternoon, everyone. First, I would like to introduce myself. I am Indhuja Perumal-Vanaja, PhD student at the University of Padua, Italy. Today, I will be sharing with you data about sympathetic innervation in arrhythmogenic cardiomyopathy, a disease well-known to the majority of the people present here. First, I would like to bring greetings from my professor, Tanya Zalia, who cannot be present here today, and she hopes you'll appreciate our data. Okay, let's start. Arrhythmogenic cardiomyopathy, it's a disease, familial cardiac disease, accounting for most cases of stress-related arrhythmic sudden cardiac sudden death, predominantly in the young and the athletes. In about 50% of the genetically diagnosed cases, ACM is caused due to the mutations in the desmosomal proteins, which in cardiomyocytes build desmosomes to connect adjacent cells. Exercise is a well-appreciated risk factor, triggering arrhythmias and also accelerating the heart remodeling, which, depending on the genotype, may affect the right ventricle or the left ventricle, or even both. The typical hallmarks of an ACM heart remodeling is characterized by cardiomyocyte cell death, tissue inflammation, and replacement of the muscle with fibrofatty lesions, which compromises the contractile function and favoring arrhythmias in ACM hearts. However, the picture is not quite simple, because the clinical representation of ACM is highly variable, making it very hard to diagnose. For example, lethal arrhythmias may be the first disease manifestation in apparently very normal hearts with no remodeling, and on the other hand, there are patients affected by ACM who never experience any arrhythmic episodes. As a consequence, this makes the ACM pathogenesis still obscure, and current treatments only focus on reducing the arrhythmia incidence for these patients. So we have to look for the unconventional. Up until now, ACM research focused on ACM as a disease of cardiomyocytes, which carry the desmosomes, but recent work, including ours, demonstrate that cardiac and extracardiac cells express desmosomal proteins, even though they are not organized into the desmosomes. So we decided to approach ACM as a multicellular disorder. Now, approaching ACM as a multicellular disorder apparently decreases the chances of therapy. However, if we want to remove an invading evil, we need to shoot at its roots. So in our lab, we identified them in the sympathetic neurons that mediate inter-organ connections and control all the cell types that are involved in ACM, which express drugable sympathetic neurotransmitter receptors. But why the sympathetic neurons? Sympathetic neurons are suspects for a number of reasons. First, they highly innervate the mammalian heart, as you can see here from the 3D reconstructions of murine and human neuronal networks, and they also connect all cardiac cell types and impact on their electrophysiology. Second, alteration in the distribution of the sympathetic neurons in the heart is associated to increased arrhythmogenesis. Third, as described by me earlier, that exercise is a risk factor in ACM and is accompanied by increased sympathetic neuron activity. And more recently, we described that the sympathetic neurotransmitter neuropeptide Y is elevated in the ACM patient and enhances arrhythmogenesis in human cardiac stromal cells, which we observed in vitro. Does this body of evidence say that sympathetic neurons may be additional cells that are affected in ACM? So we asked ourselves, are sympathetic neurons affected in ACM and involved in the disease mechanisms? And here, now, I will be focusing on the first question, which is, are sympathetic neurons affected in ACM? In a recently published manuscript in the Journal of Physiology, we demonstrated that the sympathetic neurons isolated from the normal mice, superior cervical ganglia of the normal mice, as you can see here from the biochemical assays, express desmosomal proteins. Among these various desmosomal proteins, we focused on DSG2, one of the genes which is responsible for a biventricle form of ACM remodeling. And further confocal immunofluorescence confirmed the presence of DSG2 in murine stellate ganglion sections, and notably, more interestingly, also in human stellate ganglion sections. In vitro assays showed the DSG2 immunoreactivity in the sympathetic neuron axonal process, and also in the sen nucleus. Given that this DSG2 is expressed endogenously in the sympathetic neurons, the neuronal cells, this suggests that in the DSG2 mutant carriers, these sympathetic neurons will carry this mutant genes, and thus maybe are additional cells that are affected in ACM. And this is true, because we assess sympathetic neurons from DSG2 mutant mice, from the superior cervical ganglia of DSG2 mutant mice, which is a model for biventricular form of ACM. And as you can see here, in the second lane, there were profound morphological abnormalities in the sympathetic neurons. In particular, they were thicker, shorter, less ramified compared to the controls, and also they had irregularly distributed vericocities. A similar phenotype was also observed in normal cells where DSG2 mutation was introduced through viral infection. So at this point, we decided to see the innervation pattern in these DSG2 mutant hearts. For this, we analyzed DSG2 mutant mice of one month, corresponding to early disease stages where the heart has no remodeling, and DSG2 mutant mice at six months, corresponding to an advanced disease stage where there is extensive heart remodeling. And confocal immunofluorescence in the heart showed that the sympathetic neurons were thicker, which were already evident at one month. And at six months, you can see very clearly that the sympathetic neurons have a different morphology with thicker abnormalities. And quantitative analysis showed that there is a hyperinnervation, there is an increased sympathetic neuron density in the interventricular septum and right ventricle at one month. And the innervation is higher in all the heart regions at six months. This innervation, this hyperinnervation was also accompanied with a different neuronal topology, which you can see here from a colorimetric map, which represents the average neuronal density in the myocardium, represented by color code blue for low density and red or purple for higher densities. While in the normal control hearts, you can see that the sympathetic neurons are homogeneously distributed with the increased innervation density in the epicardium, when you look at the DSG2 mutant mice, there is hyperinnervated area interspread with less innervated areas. So we decided to characterize the neuronal network topology in 3D. So for this, we performed some whole-mount confocal immunofluorescence in tissue-clarified heart blocks of control and DSG2 mutant mice at advanced disease stages in both the subepicardial and the subendocardial regions. Here is an example of the 3D reconstruction. As you can see in the control hearts, the subepicardial region has a higher innervation compared to the subendocardial region. Now if you look at the DSG2 mutant hearts, the subepicardium is hyperinnervated. And I would like to remind you that this is the place where the remodeling start in these ACM hearts. And also, the quantitative analysis confirmed the hyperinnovation, as you can see here. And we also analyzed damaged areas in the myocardium, where you can see disarranged neurons, which is heterogeneously distributed in and around the myocardial remodeling. A similar phenotype was also observed in the samples of human AC patients, where the sympathetic neurons tend to concentrate around the remodeled areas. So now I would like to conclude by saying that sympathetic neurons are affected by both ACM mutations and context-dependent factors, and that abnormal cardiac sympathetic innervation is a new structural hallmark, at least in DSG2-linked ACM. Now the question is, is there a link between neurons and heart remodeling? To address this point, we are currently assessing what effects DSG2 mutant have on this sympathetic neuron electrophysiology. And the preliminary data show an altered calcium dynamics upon nicotine treatment, suggesting an abandoned release of neurotransmitter. And as said earlier, we found that neuropeptide Y is elevated also in the plasma of ACM patients. And now we are reappraising the role of neuropeptide Y in ACM. And I hope to present more interesting data in the next conference. Now I would like to thank my professors, Tania Zalia and Marco Mungillo, and all my lab members and all the collaborators. And thank you all for your attention. So this talk is open for questions. Please come to the microphone and introduce yourself. I'm Jeff Safitz from Beth Israel Deaconess in Boston. So this is an interesting story. I hadn't really thought much about sympathetic neurons. But you're looking at these mice rather late in the game. By six months, they have a very extensive phenotype. The phenotype really begins in this mouse model at about eight weeks of age. And by 14 weeks of age, it's really very well developed. So my question is, are all the changes you're seeing really driving early stages of disease? Or is this the result of late stage remodeling, do you think? I think it's something temporal, because you can see the sympathetic neuron morphology already altered at one month. And I think this sympathetic neuron drives this remodeling, ACM, heart remodeling, with time or something like that, if I understand. But do you see these changes in sympathetic neurons early on, say in six weeks? One month mice also. Oh, okay. Thank you. Okay. Thank you. Carolyn Remme, Amsterdam. Very interesting data, sorry. What is the normal function of desmoglein-2 in these neurons? And how does this, for instance, change the way these neurons are actually making contacts with cardiomyocytes? Do you know anything about that? Well, there are some literature that they participate in axonal process sprouting into the younger age, and also maybe with the release of neurotransmitters. But we are currently looking also into this aspect, and maybe next time we meet I'll give more insights on this. And are there other desmogleins present in these neurons? Yeah. Like I showed you, there are also other desmosomal proteins. No desmoglein type. Is it only desmoglein-2, or is there other? We looked at only desmoglein-2, so I'm not sure. Thank you. So, one final question. Omer Bernfeld from Michigan. Very interesting. Thank you. I have a question regarding the arrhythmogenicity of your observation. Will you agree that it's reasonable to speculate that you have increased ectopies, electrical ectopies in these hearts, and do you have evidence for that? Yeah. I don't have it now, but these mice have increased PBCs, ectopic beats, sustained arrhythmias. We have some telemetry data, but I currently have... To contrast it with impaired conduction? Yes. Okay. Yeah, yeah. Thank you. Thank you. So, thank you very much for this presentation. Thank you. Our next speaker is Joris de Groot, who will introduce us to fibroblast adipocyte interaction. So, let me help to get this going. This is the monitor. Thank you. Thank you very much, and thank you for inviting me. Let me see what happens here. Yes. I will talk to you about the interaction, the crosstalk between fibroblast adipocytes and myocytes in the setting of cardiac arrhythmogenesis with the focus of atrial fibrillation. These are my disclosures, and the interest in the topic came from the work of the group of Presch-Sanders a while ago, where they showed in sheep, in this case, that were obese, that were fed with a high-energy diet, that there were differences in the electrophysiology of the right and the left atrium, and more particular, that there was a significant reduction in conduction velocity along all different segments of the atrium, and that went hand-in-hand with an increase in interstitial fibrosis in the atria and ventricles of these hearts, and this was also significant. Subsequently, they looked at patients undergoing ablation for AF, and they screened patients, and they find obesity as a BMI of more than 27, so this was a kind of lenient definition, and these patients underwent mapping as part of their standard procedure, standard ablation procedure for AF, and what they found is that in obese patients, there were indeed more areas of low voltage and of slow conduction, and these were significant as well, specifically for the conduction velocity and for the fractionation of the electrograms. So why does this occur, and why does obesity or extra fat tissue change the electrophysiological and structural substrate of atrial fibrillation? David Dong, one of our PhD students, recently wrote this review where he clearly indicated the different types of crosstalk that are possible in the heart, and to make this story a little easier, there is indirect crosstalk where you can have otocrine, endocrine, or paracrine interaction between cells or of circulating components that interact with cells, but there's also direct crosstalk between different cell types via gap junctions, mechanical junctions, or even nanotubes, and I like this figure very much because I'm thinking about atrial fibrillation, and that's just a detail. If you look at all these interactions that take place, and that take place before actually anything happens to the myocyte or the atrium. So adipocytes affect coagulation. They affect fibrosis by influencing myofibroblast, but also inflammation, and all these factors together are supposed to add to the substrate of atrial fibrillation. So one of the questions that we ask is what is actually in the fat tissue of patients with atrial fibrillation? So we did a proteomic analysis of fat tissue, and that's here, but also of the secretome of the epicardial fat tissue of tissue of patients with and without atrial fibrillation, and the volcano plot that you see here shows many, many, many different proteins, but importantly, there is a differential expression of some of these proteins in AF compared to non-AF, and that has consequences for the genesis of fibrosis. Eva Moerlendijk subjected isolated fibroblasts to the secretome of patients with AF and without AF, and this is just an example of collagen 1 and fibronectin, and she showed that there was an increased expression of the production of these profibrotic factors in the fibroblasts when subjected to secretome of patients with AF versus non-AF, and this was significant. And was there a trend towards a more severe phenotype? The yellow dots here represent the paroxysmal AF, and if you want to see it, and I do, then it looks like the effect is larger in persistent than in paroxysmal AF, but that was not significant. When subjecting these findings to a biological process analysis, the most important interactions between the differentially expressed proteins related to immune response and immune system. So what happens if you subject cardiac tissue or cardiac cells to the secretome of patients with atrial fibrillation, and that is what Orianne Hernot did. She used isolated neonatal retrotracular myocytes that she plated on a multi-electrode device and she subjected these cultures to the secretome of patients with atrial fibrillation, and she compared control medium with subcutaneous fat and epicardial fat and could show that indeed epicardial adipose tissue slows conduction, and that was significant, but also increases conduction velocity heterogeneity. She also shows that when a microelectrode was impaled in the cell culture, that was significant depolarization of the resting membrane potential and the shorter action potential duration. In patch collapse experience, the latter could not be confirmed, but indeed there was some depolarization of the resting membrane, there was a shortening of the action potential, and there was a decrease in the upstroke of the action potential. When specifically looking at the currents involved, she found that IK1 was significantly reduced with no change in IK or the calcium current, the L-type calcium current. Next what Orianne did is looking at the expression of the different genes that are associated with the composition of the action potential with PCR, and again she showed that the KCNJ2 including the IK1 current was decreased and the SCN3B, a sodium current, was increased. Also interestingly, these changes were not limited to the ion currents, also the gene for the gap junctions and the connexin 43 were actually reduced in the cultures that were subjected to epicardial adipose tissue from AF patients. So the next question is what is actually in that secretome? There are many, many different proteins expressed, but one that stands out is MPO, myeloperoxidase, which is a product of neutrophils and associated with inflammation of the atrium. And Eva again looked at the expression, or this was actually immunohistochemistry, looked at the presence of the MPO in the epicardial adipose tissue of patients with persistent AF, which is the more severe form, paroxysmal AF, patients without AF, or patients without AF who would develop AF in the future from a specific study that we did that we had the tissue from. She could show that in the EAT more MPO is present in persistent atrial fibrillation, that it's localized sub-epicardially, and that there were no differences in the myocardial MPO. And interestingly, she also showed that looking at the MPO here in brown in persistent AF, which is clearly visible, there was a large difference between persistent AF and no AF. But also in the patients who did not have AF yet, but who were bound to develop AF in the years following the procedure, there was already the position of MPO present, suggesting that this may play a role in the pathogenesis of AF here, and in the interaction that epicardial adipose tissue or adipocytes may have with myofibroblasts and with myocytes. And indeed, the next step, of course, was to subject these neonatal retroventricular myocyte cells to MPO specifically, and that's what Rashad Alshama did. This is a control activation map where activation runs from red to blue, and this is a activation map of tissue or cell culture subjected to MPO, where you can appreciate that conduction velocity is, well, at least more heterogeneous here. And she also found patterns of lines of block where there was continuous conduction and reentry within these small cell cultures. Interestingly, there was also an overexpression of fibroblasts here in green in the MPO-treated cells compared to the control cells. And looking at the local electrograms, there was more fractionation in MPO. There was depolarization of the resting membrane and shortening of the action potential duration, as was also seen in the cells that were subjected to the full saccharotome of epicardial tissue. So, this brings me to my conclusions. Epicardial adipose tissue is associated with electrophysiological changes in obese sheep and obese patients with atrial fibrillation. That's the observation. There is crosstalk between adipocytes, fibroblasts, and myocytes through direct and indirect interactions. And the EAT saccharotome of patients with AF induced expression of extracellular matrix proteins in fibroblasts. Moreover, EAT saccharotome of AF patients decreases IK1 and connexin 43 in cardiomyocytes, resulting in arrhythmogenic conduction slowing and conduction block. And mass spectrometry of EAT of patients with and without AF reveals overexpression of immune-related proteins in AF, where MPO stands out in particular. MPO in itself induced fibroblast proliferation, conduction slowing, and reentry, similar to the full saccharotome of epicardial adipose tissue. Thank you for your attention. Thank you very much, Joris. The talk is open for questions, and please introduce yourself. Thank you for the presentation, Omer Bernfeld, Michigan. A comment and maybe a question regarding the relationship between conduction velocity and coupling, primarily connexins. First of all, the conduction velocity dependence on the electrical coupling is bimodality. So it's not monotonous. So that may induce the heterogeneity that you see. So you have to be careful when you quantify conduction velocity and relate it to amount of connexins. The other thing is that in the atria, you probably have to look at connexin 40 or so. You looked at 43. So do you have data on that? Well, to start with your first question, I did not imply that the conduction velocity slowing was caused by a reduction in connexin 43. But we observed a slower upstroke of the action potential, which is probably more important here, and a reduction in CX43. So I cannot dissect the two. This was done in neonatal retroventricular myocytes. So you're completely right that the model used here was a ventricular cell model. And we did not look at connexin 43. Or actually, it was not different, but we did not demonstrate the difference. But obviously, the question would be, what would you see if you do this in atrial tissue or in atrial cells? I agree. Thank you. Joanna Montgomery, University of Auckland. Thank you. I really enjoyed that. I was wondering, do you know the fate of the patients that will have AEF in the future, whether they become persistent or paroxysmal? Because I noticed it was interesting that you saw quite big differences between the persistent and the paroxysmal in your quantification. But the ones that were yet to be diagnosed were often in the middle. So yes, I was wondering on your thoughts on that. Do you have that information? We don't have this information. This was a study of 150 patients undergoing standard cardiothoracic surgery, of whom we took the left atrial appendage for research. And we followed them up for the development of atrial fibrillation. And only 18 out of 150 developed atrial fibrillation. We did not specify that in persistent or paroxysmal. But importantly, this occurrence of the arrhythmia was months or years later than the moment that the left atrial appendage was taken. It's just interesting that that was so different from the paroxysmal ones. Thank you. So perhaps a final question from me. You highlighted that you had an overexpression of sodium channels in the context of depolarization. Is that a compensatory response because you have fewer channels available? And how does it relate to conduction velocity? I find it a very interesting question, but I would not know the answer to that. It's very well possible, of course. But we took the conduction velocity measured from the multicellular electrode as a basis and then looked at observationally at which changes occurred that could align with the things that we found. To answer the question that you asked, you should specifically look at overexpression or underexpression or blockade of this current, which we didn't. Thanks. OK. Well, then I think we thank our speaker and hand over to you. And so I am delighted to introduce my esteemed co-chair, Dr. Peter Cole, who will be presenting on the topic of targeting heterocellular coupling as antiarrhythmic therapy. Take it away, Peter. Yeah. Thank you very much. And again, it was a title that was given to me. I do not have any conflict of interest that's worth reporting. What I would like to highlight is just a very brief scientific historic excuse before I show some unpublished data. We know since the 1960s that two myocytes interconnected by non-myocytes, presumably fibroblasts, can beat synchronously. And if you destroy the fibroblast bridge, they stop beating synchronously. The reason for that is that myocytes and fibroblasts can electrically couple. You take a fibroblast and a myocyte on a patch pipette, bring them together, wait a few minutes, and then you see what is shown here. The myocyte action potential clamps the fibroblast. And that, of course, is then a driver that can allow you to conduct excitation in a passive manner, completely electrotonically, as has been shown in beautiful experiments by Gaudasius here who showed that up to 300 micrometers can be bridged passively between myocytes via conduction through non-excitable cells. So the connections here matter, and Martin Rook had shown that very nicely. We are used to looking at connections in intercalated disks, but it's important to realize they are also between fibroblasts shown here in the middle panel or at the point of contact of myocytes and fibroblasts. Here 40, connection 43 in the ventricles, but it's true for the atrium. It's also true for connection 40. And if you look at the percentages of co-localized connections in the heart, connection 40 might actually be the one that is more prominently coupling non-myocytes and myocytes. Now in disease, there is substantial remodeling here. These are animals that have received a ventricular infarction to cause left ventricular remodeling, but also the intraventricular node is heavily remodeled, as can be shown here. And so it is a highly dynamic self-regulatory system. To sum up this little historic excurse, connections of multiple types are absolutely common at points of contact of different cells. The relevance, dynamics, and regulation of this are currently ill-explored. And what is important to realize is that the presence of connection is not yet proof of electrical coupling. But equally, the lack of connections doesn't rule out that there is electrical interaction. This can either be by a few channels that may simply be too small to be recognized, but also by a faptic or capacitative coupling, which we've heard about today in the second session and we will hear about more in the final talk of this session. Now what does it look like when non-myocytes contribute to conduction? I think this is a beautiful example. It comes from Godfrey Smith's team in Glasgow. It is a rabbit heart with a transmural infarct. And if you stimulate electrically just outside the infarct border zone, you can see propagation of electrical excitation into the infarct, into what clinicians like to call dead tissue. So you can see action potential waveforms here at the points lettered in the alphabet. And this is what the substrate looks like. In the post ischemic infarct, you have myocyte islands. These are islands rather than strands in most cases. Nonetheless, there is electrical activity. One question by the reviewers was, is it from the subendocardial layer that survives? But if you ablate that layer, you still see pretty much the same thing. And so the reason or the question then is, do myocytes in the border zone and the connections that they form with non-myocytes allow passive conduction to surviving cardiomyocytes in the center of the infarct? For example, here is a cell that is still maintaining its cross-striation weeks after the infarction. So it could be possible that the non-myocytes are little telegraph lines. The surviving myocytes might be repeater stations that recondition the signal. And that may be why the excitation managed to cross the transmural scar in the experiments by Godfrey Smith. What does repeater station mean? Well, if you have a strand of myocytes and then a bridge of non-myocytes, at some point they will not excite downstream cardiac cells here in this model after seven inserted non-myocytes. But if in position five you put a cardiac myocyte, that will be excited above threshold, will regenerate the signal. And so now you can continue to conduct for any lengths that you prefer. So this is a possibility that might explain the data from Godfrey Smith's lab. And it is quite tricky to investigate that in vivo or in situ using classic electrode techniques, because if you were recording something like this second green action potential shape here, you wouldn't know whether you are in a myocyte or in a non-myocyte. So how to approach the topic then? One way of studying it is to inject cells that are excitable, skeletal myoblasts that do not normally express connexin 43. And if you do that in a mouse, in fact, you see it in green where they are, you run a fast-pacing protocol and see ventricular tachycardia or fibrillation episodes in these mice. And you see it in all of the mice that the bond team of Fleischmann tested. But if you make these myoblasts express connexin 43, they integrate electrically, and you reduce the likelihood of causing a ventricular fibrillation to just one-third of control. So this is possibly the case that here these cells become the repeater stations needed to make the scar electrically invisible. And that actually perhaps should guide one way of thinking about how to use heterocellular interactions for potential therapeutic benefit. If you think of the opposite of ablations, say, for atrial fibrillation, you would want to exclude any such mechanism. And so in ablations, you would perhaps want to down-regulate connexins and make sure that no myocytes survives the procedure. All of this, of course, is indirect evidence. So how can one track action potential dynamics then in non-myocytes? Well, you can use optogenetics, and we will again hear in the final talk how that can be done beautifully in both directions, observing and steering. We used a voltage-sensitive protein to observe in a cell-specific manner the membrane potential in non-myocytes, and this VSFP has a big advantage over many other of our tools. The transgene is not fluxed, but flexed, and I didn't know what that means until I learned about the VSFP. Essentially, the genetic sequence is inserted in the inverted reading frame. There is no accidental readout. And that then makes for a very highly targeted expression. Here you see a superficial cryoablation in a mouse heart, and you can see the histology here in red, the myocytes, and blue, the nuclei, in green, the cells expressing VSFP. And you can see that only 0.07% of myocytes co-expressed the voltage sensor, the voltage reporter. So that is highly cell-specific, selective, and if you now do optical mapping, what you can see is action potentials. So you see action potentials that are reported by a protein that is not in cardiac myocytes, and so this shows that in vivo there can be this action potential clamping of non-myocytes, and it shows that the direction of cardiomyocytes affecting non-myocyte electrophysiology is real and happens in vivo. How about the other direction, the non-myocyte to cardiomyocyte coupling? Now, the icing on the cake will come in the final talk, but I will show just a little bit of what we have done. We tried the same about 10 years ago, expressing channelrhodopsin under the control of WT1, which is a non-myocyte promoter. We were able to trigger excitation in the hearts, we were really quite pleased, and then we looked at the histology, and what we found is that while in most of the heart we had fairly cell-type-specific expression, there were patches where cardiomyocytes also expressed the channelrhodopsin, and therefore no conclusions could be drawn on the basis of these kinds of experiments. So what you need is inducible systems that are allowing one to express the channelrhodopsins not from the germline onwards but later in life, and again we will see a beautiful example later on on how that can be used. We used it in particular to explore the three-dimensional architecture of cells in the heart after clearing. On the left-hand side you see the fibroblasts in ventricular myocardium, on the right-hand side macrophages, and if you look at them in a little bit more detail then you see that these are these beautiful intricate cells. Here you see a fibroblast, and here you see a macrophage, and what I want to highlight by going between fibroblast and macrophage is that by simply looking at the single cell morphology, we wouldn't be able to tell who is who. And that's true also in a quantitative manner. If you look at cell volume, there is no substantial, no significant difference. Cell surface area is also generally similar. What differs is the extent of cells present, so you have in most parts of the heart many more fibroblasts than macrophages, and the interesting thing is that fibroblasts form interconnected networks while macrophages are solitary cells. So in a healthy heart, no two macrophages seem to touch each other, at least we haven't seen that. So that is then one of the conclusions, that the fibroblasts form these beautiful interconnected networks with very tiny, fine processes, both they and the macrophages that are solitary cells express surface connections, and yeah, difficult to distinguish unless you have cell-specific markers. How about lesions then? We did cryo-lesions, and our data here is 28 days post-cryo-lesion, so a little bit later than what we will see in the next presentation. You see that fibroblast density goes up, also in remote tissue after a lesion in the heart. That is not the case for macrophages. We only see the increase directly in the scar zone, and we now tried to use light to change the electrophysiology of these constructs. We remained unable to trigger excitation in these hearts, but if you look at the restitution properties, you can see that in remote tissue shown in green, in the scar border zone in yellow, and in the center of the scar, illuminating fibroblasts to depolarize them using channelrhodopsins changes restitution in a significant manner in the heart, whereas for macrophages that is not the case. So what I would like to conclude then here on this data is that the effects of non-myocytes on cardiac electrophysiology are really exciting. They may be quite subtle. It may take approaches like these to explore them, and it is very likely the case that they differ across the different cell types, cardiac regions, lesion models, so ischemic or cryoablation, different repeater station presence, but also duration of remodeling after the lesion may be associated with differential responses. So to conclude then, non-myocytes form the majority of cardiac cells, and that is why we are talking about them. Non-myocytes and fibroblasts and macrophages express connections, and they can definitely be coupled to cardiac myocytes in C2 and in vitro. Fibroblasts are interconnected networks. Macrophages are unitary cells in healthy tissue. And in the absence of remodeling, the effects on cardiac electrophysiology can be quite subtle. In disease, the effects on electrophysiology change, and it is well beyond what we classically read in the textbooks, namely that non-myocytes form obstacles to conduction. They do much more, but what exactly they do is something that further research will have to address. And the mechanisms, the homeostatic and pathological changes, and certainly the therapeutic potential remains to be explored. I would like to send best regards from the whole team in Freiburg, and in particular from Callum and Franzi. Both are independent group leaders, and I showed predominantly their data. And we would like to invite you to come to Freiburg. It is a beautiful town, and we organized a meeting in September that some of you may find interesting. Thank you very much. Thank you, Peter. Such a great talk. As we sort of wait for people to come up to the microphone, I'll take the chair's prerogative. So this is something I've been thinking about for a long time, about the context of the non-cardiac myocytes in the heart, and in which situations they act as sources, as sinks, or as shunts, and how that may change depending on their orientation, or their organization, or their degree of maturity, or their response to a variety of factors that they experience in the environment. Can you tell us more about that? Well, it is one of these questions, like the last question I asked the previous speaker, where essentially you have to guess a little bit, right? So what we know is that non-myocytes and myocytes can electrically couple, and they can interact via other ways. Tunneling nanotubes have been shown. Capacitative interaction has been proposed. Aphaptic coupling, we have seen the evidence between myocytes. We will see the evidence between non-myocytes and myocytes in the next talk. So in terms of electrophysiological effects, yes, non-myocytes may divide and may lead to, and will lead to zones of conduction block, but we need to be careful of distinguishing fibrosis and presence of non-myocytes. We often use the term interchangeably, and fibrosis refers to the non-cellular part that certainly insulates. And then non-myocytes can either be in parallel or in series with myocytes, and the effects will differ. If they are in parallel, they will be the load, they may change excitability, refractoriness. If they are in series, like in the cell culture experiments, they may conduct, and if that then is in an area where you also have myocytes present, then all kinds of things can happen, and I think that has yet to be explored, both using computational models and experiments. It's ripe for a model study. Absolutely. We need a little more data from you, Peter. Peter, great talk as usual. I had a question regarding your insights or opinion about Connexin 43 or other gap junctions versus nanotubes, which you have published on. I mean, Connexin 43 we know, but when do you think nanotubes may play a role in this context? Yeah, that is another very difficult question that I find very hard to answer. So the nanotubes, we are currently exploring whether there is direct cytosolic interaction or not. So far, we have very little evidence to suggest that cardiac fibroblasts and cardiac myocytes actually form a continuous connection in vivo. In vitro, yes. In vitro, we see fragments of mitochondria traveling from non-myocytes to myocytes through these equivalents, I guess, of tunneling nanotubes. In principle, I think what the tunneling nanotubes are doing is they scan the surface of a cardiac myocyte, identify where that myocyte needs to be integrated into the exocellular matrix, and then they start to deposit collagen. And we can show that, and we have seen that that seems to be happening. Of course, you need the resolution of EM, three-dimensional EM, which means dynamic observations are very, very difficult, and it's a very slow process. The tip of tunneling nanotubes, we have now a technique where we think we can harvest them, and that is what the next step hopefully will give us an insight into the proteome of tunneling nanotubes. If they express connections, then there may be another domain of electrical interaction that may perhaps be very small or small enough to be hard to detect, perhaps with new imaging techniques that may be possible. But you see, I'm hand-waving. I really do not know a precise answer to your very insightful question. Great. Let's thank the speaker. And our last presenter of the session, from UCLA, Arjun Deb, who will be enlightening us on steering cardiac electrophysiology with light-activated ion channels in non-myocytes. I'd like to start off by thanking Peter and the organizers for inviting me here. Personally, I've learned a lot from Peter's work over the last few decades, and so it's really exciting to present this work. So I'll tell you a story of how fibroblasts can directly excite myocytes and contribute to arrhythmogenesis in a pathological heart. Now, I don't have to inform this audience that fibrosis is one of the leading causes of death, and you can see here, if you have fibrosis after, in a diseased heart, you're much more prone to develop sudden cardiac death than if you don't have fibrosis. And this, by the way, is independent of your cardiac function. So, even if two people have the same EF, the person with greater fibrosis has a much higher probability of having sudden cardiac death. Okay, so how fibrosis causes arrhythmogenesis. And I'm not gonna go into all mechanisms, but what I'm gonna try to focus on is what we think could be an interesting mechanism which could be playing out in vivo. And I want to point out that the structure of the heart radically changes. As you can see here, there are very few myocytes in scar tissue, and most of it is fibroblasts. And as Peter and others have alluded in the last few sessions, that fibroblasts are known to communicate with cardiac muscle through gap junctions. And so, this is the model which we are trying to propose, which means that, let's say, a fibroblast which is colored in brown is communicating with the red myocyte. And now, let's imagine that the fibroblast undergoes some kind of stretch, could be ROS, could be membrane stretch, and it depolarizes. So if it depolarizes, there is a flow of current from the fibroblast to the myocyte, and so the myocyte membrane potential now rises, and it's more susceptible to arrhythmogenesis. So that's the fundamental concept. And as Peter and others have shown, the other way that it's possible, but in vivo, this model has never really been proved, and so that's the question we're gonna answer in the next 10 minutes. So that's the question, can the non-excitable fibroblasts in scar tissue alter cardiac muscle excitability and contribute to arrhythmogenesis? And how do we answer this question? So, we adopted an optogenetics approach. As many of you in this audience very well know, the optogenetics essentially involves expression of a nonspecific light-sensitive cationic channel. So when you shine a specific wavelength of light, the channel opens, and cations flow through, and there is a current moving into the cell, so the cell gets depolarized, and the membrane potential rises. So that's just the basic concept of optogenetics, and it's mostly been used in neurosciences, and it's been reported in the New York Times, so it's fairly common among scientific use. So, what we did was, we generated a mouse where the myocyte, which is the excitable cell, does not express any optogenetic channel. In contrast, the fibroblast, defined by a very specific marker called TCF21, is induced to express the channel. And we call this mouse the CFCHR2 mouse, the hydrocardic fibroblast-channeled rhodopsin mouse. Now, it's very important to understand this hypothesis. So, for example, if you have a mixture of myocytes and fibroblasts, and imagine these fibroblasts are connected to the myocytes by some form of electrical communication, let's say gap junctions here, and you got the optogenetic channel sitting on the surface of the cardiac fibroblasts. There is no optogenetic channel on the myocyte, so now, when you flash light on this system, the only channel which gets activated is the optogenetic channel on the fibroblast. Now, this is absolutely key to appreciate this, that there's no other excitable light which can activate the myocyte in this system. And the whole experiment hinges on this. So, if you do that, then what you can do, if current will move into the fibroblast, now, if the fibroblast electrically communicates with the myocyte, then the myocyte should get depolarized or should fire, and you should be able to record it. And if that's the case, that's bona fide proof that the fibroblast must communicate with the myocyte. So, that's the experiment. So, this is key. As the myocyte is not being stimulated, any evidence of alteration of myocyte excitability when the fibroblast is excited would provide proof of this. So, I'll quickly go through this because this is just the mouse. So, this is a TCF21 Creef. The fibroblasts express the channel rhodopsin, which is green. The myocytes are in red. You can see panel, top panel is uninjured myocytes. So, you have myocytes and fibroblasts. And then you can see in this panel, which is injured and in higher magnification, there's a lot less red cells. Myocytes have got killed or dropped out, and it's mainly fibroblasts proliferating. So, that's our system. There's a scar. You optically stimulate the heart, and then you record at the base of the heart away from the scar tissue. And that's what it exactly looks like. So, you shine light there, and you record from the base of the heart. So, here's the first set of data. So, if you can, the blue bars are pulses of optical stimulation. So, you stim the heart at seven hertz, which is 420 beats a minute. And you can see that the pre-stim rate was 360. As soon as you shine blue light on the fibroblasts, the heart now picks up at 420. You stop, it exactly goes back to where it was. And you can do this even at faster stimulation rates of nine hertz. And this video illustrates this. Now, you can see the blue light will come on in a flash, and you can see the heart rate picking up. And then as soon as the blue light stops, the heart goes back to where it was beating. Okay, so this is a benefited proof that the fibroblasts in the Langendorff prep is able to excite the myocyte. But this is something also important, which Peter was alluding to in his talk. If you'd carry out the same experiment in an uninjured portion of the heart, so there is no scar tissue, then you're unable to excite the myocyte. So this is very, very important. You need a critical mass of fibroblasts in scar tissue for this phenomena to take place. Okay, so this is just ECG, but you can see that the normal activation is A to V, right? A is black and blue, V is ventricle, so A to V. But now you're exciting from the apex, so it's V and A. So it changes, so this is another proof that the excitation of the heart is actually following your pattern of optogenetic stimulation. So now how about the live animal? So this was a Langendorff I showed you. So we do this in a live animal. So we infect the animal, seven days later we come back, the animal is anesthetized, so the heart rate is low, and it's just a simple surface electrocardiogram. So just like what you would have in a doctor's office. So it's a summation of all cardiac electrical activity. And so now you can see that you're stimming at seven hertz per minute. The heart rate before stimulation was 390, now the heart is beating exactly at 420 beats per minute. And what is important is the A to V has also changed here, the ratio, so V comes first and A goes after what you would expect if you're stimulating from the apex. So this shows that it can happen even in the live anesthetized animal. And again, if you stimulate an uninjured portion of the heart, you cannot excite the heart. So that's an important concept. So next, okay, so you're stimming the fibroblast and it's generating cardiac myocytes. The question is whether it's generating mechanical contraction and there's blood being actually pumped out. So to do this, we put a pressure sensitive catheter in the carotid artery. So the carotid is connected to the heart. So every time the heart pumps, there's a pressure pulse in the carotid. And so you can see here that if you stim at this rate, this is the pressure pressure on the carotid, you can see that the carotid pressure wave coincides with the optical stim. And it also leads to self-sustaining arrhythmias. So for example, if you stim, stim, stim, and then you stop here, as shown here, you can see that the heart goes into bigeminy. Okay, they're twin paired beatings. It doesn't really go back into the normal before as it was in a sinus rhythm. And this we've seen particularly with hearts, which are like 24, 28 weeks after infarction, the scar has matured. And here, for example, is the other way around. You see heart block, there is no ventricular pulsation after a period of stimming. And then this was interesting. So we had various forms of arrhythmias take place. And this was you stim the heart very, very fast at eight hertz a minute. So the heart is beating at 480 and you stop. And as you stop, there is a ventricular standstill. Okay, so these just show that if you have an ability to stimulate the fibroblast, and these fibroblasts can then communicate to the myocyte, it can profoundly affect myocyte excitability. And in some cases, given appropriate boundary conditions, it can lead to sustained arrhythmogenesis. So I'm behind time, so I'm gonna end off with the concept of efaptic and gap junctions. So the question is, what is the form of communication? And whether it's gap junctional or not, Peter has shown and others have shown that CX43 is an important gap junctional communication. So we knocked out CX43 by generating triple transgenic mouse where the fibroblast is knocked out for CX43, but does express the connection, sorry, the channelrhodopsin. And here you can see even these animals which have CX43 deleted in the fibroblast, you stim at seven hertz per minute, and the heart still responds. So at least for this phenomena, connection 43 is not needed. So we deleted connection 43, and as Peter said, there may be other connections. So we did single cell RNA-seq and identified four more connections, 40, 45, et cetera, which have been reported in the literature as well, and we knocked them out sequentially with CRISPR. And we do this in an in vitro experiment. The fibroblasts are still able to excite the myocytes after stimming. I'll just end off with efaptic here. I got one minute left. So people have talked about efaptic a lot this morning, and I want to make it really very, very simple because I'm not an electrophysiologist. And it's a phenomena described in the 1940s, and the experiment was this. There are two neurons, squid axons, lying side by side, okay, imagine two axons lying side by side. And you stimulate axon A, let's say the top axon. But you're recording in axon B, which is the axon lying side by side. There is no connection between axon A and axon B. But you're recording in axon B, and you see an action potential propagate through axon B. So there is no electrical synapse between these two neurons. And so this word was coined much later, which is called an efaps, a synapse which behaves electrically, and therefore efaps. And that's how this term efaptive connection comes on. All of it really depends on the extracellular space in between the two cells. That's all that matters. And so I'm going to skip a lot of the slides, and I'm going to show at the end, mathematical modeling shows both are possible in our system. And I want to end with this. So I think Steve, earlier this morning, Paul Singh pointed out that our mathematical modeling suggests that when efaptic is sufficient, and your gap junction on the y-axis is very, very low, the heart can still conduct. When gap junction is very high, doesn't matter what the efaptic is, the heart can conduct. But in the middle, when there is a gray area, that's where the two really play a synergistic role. And how many cardiac fibroblasts do you need to efaps? The mathematical model suggests you need five. So five fibroblasts must connect to a myocyte to make this happen. Okay, I won't go into validation of models, but I'll go on to the last part where this is unpublished data, and we've shown that maybe the mathematical model suggests that there were sodium channels at the efaptic cleft, which are important, and we used a very low dose of tetrodotoxin, and can show that this can be, efaptic connection can be targeted. So to conclude, what I have shown today is that there is the fibroblast, most important take-home message is that fibroblasts can communicate with myocytes, electrically, and these phenomena are only present in a diseased heart, or at least physiologically important in a diseased heart, and that efaptic and gabjunctal coupling are mutually redundant and synergistic. Now, this is the most important part of my talk. My post-doc, E.J. Wang, she joined my lab five, six years ago. She did not know what an action potential was, and in five years she did this. She was an extremely bright woman, she had a paper in science, but she's on the job market. So if you want to recruit a very promising recruit, a candidate, send me an email and I'll forward it to her. Thank you very much. Thank you. Thank you so much for such an informative talk. I wonder, in the mathematical models, what's the membrane potential in the fibroblast? I think it was about minus 40. And how much does that fluctuate? We actually patch clamped, so we know exactly. I think it goes up from minus 40 to 10. Plus 10. Yeah. Caroline Remmer, Amsterdam. This is my incomplete knowledge of myofibroblast and fibroblast, but if you are using this model and you have, in an MI setting, you have myofibroblast formation, right? Do they still contain the channelrhodopsin? And how much would that sort of contribute to this, what you're seeing? Or is it just the fibroblast that you're targeting? Yeah, I can answer the question. So the TCF21 marker actually generates, has been shown by other people to be a precursor of the myofibroblast. So the periostin positive myofibroblast originate from TCF21 trace cells. So in our system, and we did that because the reviewers asked us to, most of the cells actually, many of them, I don't remember what percentage, but it was quite a high percentage of cells. More than 50% of them expressed myofibroblast markers. Okay. And do you think there's also a potential role for factors that are released by these cells instead of just the pure electrical coupling that could also contribute to what you're seeing? Yeah, it's a great question. We struggled for this for six months. We tried various experimental, and I think there is. We have done some experiments which suggest that the secretome or the fibroblast dramatically affects myocyte excitability in the system. I just haven't pointed it out. Thank you. Joost de Groot from Amsterdam as well. I liked your talk very much. Thank you. Thank you. But you showed that this interaction is only relevant in the setting of a diseased heart, in the setting of a scar, where the proportion of fibroblasts compared to the proportion of myocytes is changed. And whether or not this is a direct coupling through gap junctions or effective coupling, it's a safety issue, I think, that you need a bulk of fibroblasts or fewer myocytes, as Peter Cole alluded on. So as a clinical electrophysiologist, what should I do? Should I get rid of that last myocyte in the scar, or should I try to treat or affect or cure the fibroblasts? Yeah. So it's a good question, and I can tell you what we think. In fact, we wrote it in the paper as well. So the basic concept you have as an ablationist is to reduce the number of excitable cells in the substrate. That's why you ablate. So you kill as many myocytes as you can within the scar tissue, of course guided by mapping. So you reduce the excitability of the scar. So the question really is, if you take out all myocytes, every single myocyte, then the scar is not excitable. So for sure. But if you are not 100% successful, and if you leave behind a few myocytes in a critical region which you don't know when you finish ablation, then this is possible. Because now, with the other myocytes killed off, you have reignited the fibrotic process. So now this can happen. And so we said that in the paper, we argued that these things may have clinical significance, even though nobody's going in with the channelrhodopsin and exciting fibroblasts. But if you have a critical mass of fibroblasts resulting from your ablation around a few excitable myocytes, then it's possible that you can excite the myocytes and throw in an extra few PVCs. Can I ask one more question? So, the second thing that I noticed was that you had cardiac standstill after simulating the ventricle. Do you have any clue why that was the case? And I ask this because in one of your tracings, you showed the dissociation of the P waves in the QRS complex, and there was a VA conduction block. Yeah. But why does the AV node, why doesn't it work anymore after you stimulate the ventricles? Yeah, so I think the reason behind AV block is probably concealed conduction. So you remember you're stimulating the scar from the apex, right? You're optically stimulating, so you are creating a wave of depolarization, which is originating from the apex. So we think that it's probably concealed conduction, which is probably because you're repetitively firing. Phase four block. Yes, so it's causing the forward AV block. Now, as for your question about why there's sudden cardiac standstill, I think it's because the myocardium is heavily diseased because we've let these animals go for 24 weeks. So for example, if you in a pacing lab would rapidly stim, stim, stim, stim, stim with a pacing catheter and suddenly stop, if the myocardium is diseased, you won't get, you know, an extra systole coming in and the heartbeat originating. So I think that's what we are seeing. So it's not something very specific to our optical stimulation. If you took those hearts and did the same experiment with a pacing catheter, you would get the same phenotype. Thank you. Thank you for your talk, very convincing. So you showed that you need an infarct in order to be able to pace the myocardium through the fibroblasts. And I was wondering, in a control heart, have you measured conduction velocity while stimulating the fibroblasts? Because it could be that the fibroblast brings the cardiomyocytes closer to the threshold so that you can still have enhanced conduction. Did you look at that or? No, we did not, but we're doing it right now. And I think Peter in the last talk showed some data as well, where he showed that there was an effect on the action potential duration. So yeah, so I think that's, what you're saying is correct, that there may be subclinically observable effects on the myocyte, which if you don't rigorously record with, you know, an optical mapping or some kind of a conduction velocity measurement, you would miss it. But in that sense, it would be antiarrhythmic. Yes. So that would have, depending on the, yeah. Thank you. Feels to me like a classic source-sink relationship. Yes, yes. So you probably don't need an infarct, you just need a sufficient number of fibroblasts in order to be able to continue to conduct. Absolutely. It's the number of fibroblasts to myocyte ratio, which is critical. Well, thanks all the speakers for such a wonderful session. And we'll adjourn.

cardiac health

arrhythmias

fibrosis

sympathetic neurons

arrhythmogenic cardiomyopathy

electrophysiology

fibroblasts

optogenetics

cellular interactions

antiarrhythmic therapy