Good morning, everyone. Thank you for joining us for today's session on the basic sciences and translational medicine. Today we're going to have a first part of talks, which is welcome to the neighborhood diversity of heterocellular interactions of the heart. We have an exciting platform of speakers today that are going to talk about the different cell populations in the heart that play a role in both its function and its electrophysiology. We'll have a talk on fibroblasts, on endothelial cells, on glial and neuronal cells, and also on immune cells that contribute to this system. I'm Michelle Tallquist, I'll be one of the moderators, and my partner here is... My name is Igor Efimov. I would like also to remind you that you can ask questions as we go, so please open your mobile app for HRS, and you can use question and answer button there. So I will see your questions right here, and we can read it to the speaker after each talk. So I guess we will start, and actually the first speaker is my co-moderator. Michelle Tallquist from University of Hawaii, and she will present a talk entitled Molecular Identity and the Abundance of Non-Myocytes in the Heart. Please. Well, good morning everyone, and I'd like to thank the organizers for inviting me to today's session. I have a disclosure that I will not talk about electrophysiology at all today, but I will tell you about some of our studies that we've been working on that relate to the role of fibroblasts during disease processes. For a while now, it's been known that previous counts of cells within the heart were overestimating the fibroblasts and underestimating the other cell populations. Here's just a diagram of the cell populations that exist in both the murine and the human hearts. These studies have been done, and what these studies realized was that the endothelial cell really is the most abundant cell of all of the cells in the heart, and then of the other cells, fibroblasts are really only 10%. They're not as heterogeneous as previously thought. They really do come from a common origin, and they're not as abundant. And of course, then we have the other cell populations, the myeloid populations, and the studies here didn't actually even look at the glial or neuronal cells, and we'll hear a little bit about those later on today. Another point that I'd like to make is when we're thinking about the heart and its heterocellular nature, each region is very different. So while fibroblasts come from common origins, the dependence on their gene expression and their activities is really related to its location and their neighbors, their cellular neighbors. So for example, a fibroblast within the ventricular wall is going to be very different from a fibroblast in a valve or in the atrium. And so it's really important to not lump the cells together as one group. Each behaves in a different manner. I think this is most illustrated by this in vivo model within the mouse system where we introduced three different types of injury, and you can see an expansion of fibroblasts in very different areas of the heart, indicating that each of those areas is going to have an extremely different fibroblast activity and response based on cellular numbers, as well as where the immune cells are present. There's been a lot of interest in looking at the role of fibroblasts during therapeutic processes and trying to target identifying signaling pathways or specific functions of fibroblasts that we might be able to contain or to reinforce to produce benefits in the heart. And along with these studies, you can see that there are many different options to go about targeting fibroblasts and looking at their activities. What we decided to do experimentally is take a more brute force approach and just discover what is the role of a fibroblast in an adult heart. And in this system, we're using diphtheria toxin to deplete out fibroblasts within the mouse heart. And we're doing that using a CreLOX P system, and we have a lineage tracing that allows us to tag these fibroblasts and look at how many cells remain in the heart after these depletions have been done. And so in the heart, you can see here we're using PDGF receptor alpha. That's the Cre system that we're using to deplete out the fibroblasts. And if we look by Western blot for the presence of those cells expressing PDGF receptor alpha, you see that it's near zero in these hearts. We still have a few remaining hearts, which is about 10 to 15% of the fibroblasts remain in the hearts. I include the lung picture to remind me to tell you that when we're doing this, we're actually removing fibroblasts from a vast majority of the organs. Any of the cells that are expressing PDGF receptor alpha will be expressing diphtheria toxin and, in theory, die. So when I'm talking about the heart, it's not a specific fibroblast ablation in the heart, but basically a depletion across the body. And one of the surprising aspects of this is that when we look at the other cell populations in the heart by flow cytometry, we see that there isn't a major disturbance in either endothelial cells or the immune cell component, although we can bring the fibroblast populations down very dramatically. This suggests to us that the heart is resilient with the sensitivity of loss of fibroblasts. And also, if we look in the—and I apologize for using ablated here. It should be depleted, but we've often used ablated for the loss of fibroblasts. And when we look at this system, we see that we have a loss of collagen-producing cells. And these hold true for up to a year in time. So the summary so far is that when you remove fibroblasts, other cells do not seem to change in their numbers, new cells do not start to express collagen, and we have no sensing system in the heart for the remaining fibroblasts to replace the loss of fibroblasts. If we look simply one month after a depletion of the fibroblast, we see that the collagen matrix and the basement membrane surrounding the cardiomyocytes is relatively normal. We see laminin, collagen-1, collagen-4, and collagen-6 are relatively indistinguishable at one month post-ablation. If we start to look seven months post-ablation, we start to see irregularities in the patterning of the collagen matrix, but overall, the structure of the heart looks very normal. And if we look by echo at the functioning of the heart up to a year in age, there's no difference. If we look at cardiomyocyte hypertrophy, there's no difference. So the take-home message is removal of fibroblasts from these hearts. The hearts really seem to be, they do not sense the loss of those fibroblasts. Because we had this availability of a reduced heart, we decided to induce myocardial infarction by LAD ligation in these hearts, and one surprising result is that we were able to maintain cardiac function in these hearts. The remaining fibroblasts, the 10 to 15% of the fibroblasts that remained in the heart were able to go to the site of the injury and populate and deposit matrix such that we only got one or two ruptures in the multitude of animals that we've done. If we look at the function after this removal of fibroblasts and LAD ligation, we see that compared to controls, our mutant hearts really retain more functionality by echo than the mutant hearts, I mean than the control hearts. And if we look at diastolic function, we see that the depleted hearts maintain diastolic function, while as anticipated, the control hearts lose diastolic function. This of course was with 10 or 15% of the fibroblasts remaining, we decided to see if we could push the system a little bit, and instead of using one copy of diphtheria toxin, we are now expressing two copies of diphtheria toxin, and now we can bring the fibroblast levels down to about 5% of the normal fibroblast levels in the heart. If we stay in for PDGF receptor alpha expression, you see that there are no PDGF receptor alpha expressing cells in these hearts, and again, by quantification of collagen expressing cells, this is 8 to 50 weeks post deletion, we see that there's still no compensatory mechanism for loss of those fibroblasts. In these hearts, contrary to the single copy of diphtheria toxin, we do have changes in both the amount of collagen that's present, and the amount of collagen type 6 that's present, as well as the overall organization and structure of the basement membrane surrounding the cardiomyocytes. If we bring the fibroblast levels down to near zero, we really start to see overt phenotypes in the basal level of these hearts. But of course, when you have injury, you're not going to have the facility to be able to remove fibroblasts ahead of time. The next thing that we did is using the same system, we decided to use different time points of ablation to look at what happens with the functionality of these hearts when we remove fibroblasts one day after injury, one week after injury, and two weeks after injury. We thought, as previous people have shown, that one day after injury, it's very likely that we're going to get a lot of rupture. At day seven, we thought that this is after the majority of the fibroblast proliferation, but potentially we would be inhibiting some of the collagen in matrix remodeling that might be detrimental in the heart, and then we might actually rescue part of the functionality. And then by day 14, we were uncertain what we would see, and we thought potentially this would be a long-term beneficial outcome, similar to what we had when we ablated fibroblasts prior. As expected, if we remove fibroblasts one day after injury, we see that there's a lot of rupture of the hearts and animals dying. If we remove the fibroblasts at day seven or day 14, we have much better survival that's on the range of those of the control hearts, and so we anticipate that we're going to see differences in the overall matrix structure of the hearts. And this is just an example of the amount of deletion that we get when we're removing fibroblasts at these various time points. So at day one, we're very efficient, day seven, very efficient. By day 14, we're not able to remove as the highest percentage of fibroblasts as we were at the earlier time points. And if we look at the collagen here, we're using a special technique called collagen hybridizing peptide to look at denatured collagen. Here we show it in green, and the intensity of that green is identified here in this fire plot. If we look at the amount of denatured collagen that's present at 28 days with each of these time points of ablation, we see when we remove fibroblasts immediately after ablation and allow the animals to go, we see that basically the collagen is not forming fibrillar collagen. It is denatured and still disorganized. This is also the case at day seven, but if we do day 14 removal of fibroblasts, we see that the collagen matrix is relatively similar to what we see with the control. If we look at ejection fraction, not surprisingly, those animals that survived with a day one ablation, we have worse ejection fraction. Possibly surprisingly, if we remove fibroblasts even day seven after injury, we still have a fairly poor outcome with regards to heart functionality. And then if we remove fibroblasts at day 14, or compared to the control, we see the same functionality, suggesting that at least at 28 days after injury, two weeks of fibroblast removal doesn't afford any benefit at this time point. And then lastly, I just want to mention, thinking of the heterocellular nature of this system, if we do remove fibroblasts, we have an expansion of the CD206 positive macrophage population. If we look at the gene expression profiling by single cell nucleus sequencing, we can see that we have an entirely different gene expression profile of the cells remaining in the heart, suggesting that indeed fibroblasts are communicating with these immune cells, and that if loss of fibroblasts occurs, we have a significant change in the remaining cell expression by those immune cells. So what I've told you about today is that hearts can sustain fibroblast loss, collagen scaffold is relatively stable, there's no compensatory mechanism, either from other cells or new fibroblast populations, or even the existing fibroblasts, reduction of fibroblast removal, the effects of it are varied depending on the timing and the magnitude of the removal of fibroblasts. And then finally, mice with fibroblast reduction up to 90% exhibit progressive alterations in the ECM composition, and we're currently looking at those animals. I'd like to stop there and answer any questions. Thank you. Thank you. So please introduce yourself before you ask your questions. Again, this is a model system. I think that's very strong, and I think there might be a small population of cardiomyocytes in the P1, P3 heart that express collagen, but I think if there are any specific populations and it's in a unique scenario, I don't think in the majority of all the single cell studies that have been done that the cardiomyocytes are producing a major amount of collagen. They certainly cannot organize it the way fibroblasts do. They don't have all of the other components, so even if they do, they're not making collagen that's very useful. Yeah, so we actually did those experiments and looked at the alterations in the absence of MI with regards to how the immune cells. We do see a huge increase, not huge, sorry, we see a 1% increase in the number of neutrophils that come in. That's very obvious any day after we remove the fibroblasts. But surprisingly, we don't see any shift in macrophage populations, and we don't even see macrophages clearing out the fibroblasts. So we do see a slight neutrophil shift at any time point when we remove the fibroblasts. That's gone by two days, and if you look at seven days, you don't see any changes in cell number. I'm sorry, we're a little bit behind the schedule, so can you help us? Oh, okay. Okay. I'll talk to you later. Two short questions, two short answers. Peter Cole, Freiburg. Thanks a lot. I have two short questions. One, you highlighted... No, no, no. What do you want for the next one? Well, he got two. You highlighted that in the depletion model to 15%, you mentioned that fibroblasts moved to the injury zone. Are you convinced that they move there, or do they multiply in the injury zone? Yeah. When we counted and we looked at the proliferation by EDU staining, it really looks like they're moving from the other areas. Again, the mechanism of this, but if you look at the proliferation rate at the site of injury it does not go up for the remaining fibroblasts, so it really does look like they're being recruited from the surrounding tissues. Applying the depletion after the fact, are you reducing fibroblast numbers, or are you preventing an increase? Sorry, this was- Yeah, we are reducing fibroblast numbers. They're going down dramatically, especially the P1, there's like zero fibroblasts in those tissues. P14? P14, we're reducing them, but probably not as significantly. Please, short question. Very brief question. Do you have any information on what's going on in the atria in your model, which is fascinating? We don't. I wish we did. I keep telling people to study the atria, but we don't. Please do. Thank you, everyone. Thank you. Thank you so much to the organizers for inviting me. I hope the talk shows up. Yeah, great. So I want to talk today about immune cells and their role in arrhythmogenesis and rhythm disorders. And I think many of you have probably heard about this new concept of electroimmunology, which follows the idea that immune cells are causally involved in arrhythmogenesis and are not only like a bystander in cardiac diseases. So most of the work in this new or arising field has been done on macrophages and comes from the Narendorf Lab, which has shown here that macrophages really can causally contribute to electrical conduction in the healthy heart. And in cardiac diseases, we have many phenotypic changes of macrophages. We have new macrophages getting recruited to the site of injury. And we also have deaths of the resident macrophage population. And all of this has been also shown by the Narendorf Lab here that this contributes to atrial fibrillation. But today, I want to focus also on other immune cells, which are also participating in cardiac diseases. And I think the disease we are understanding best in the immunology field right now is the myocardial infarction. So if you think about the pathology of our underlying myocardial infarction, you can envision here that neutrophils are the first cell population to come in, to get recruited to the site of injury within minutes and hours after myocardial infarction. And then we have myeloid cell populations like monocytes and macrophages, which only get recruited much later to the heart, which is more weeks and months after the event. And they contribute to proliferation and repair and also help with scar maturation and stabilization. So if you now think about the sudden cardiac deaths, which is described in patients suffering from myocardial infarction, we have some of them which experience arrhythmia and related sudden cardiac deaths early after myocardial infarction. But we also have some which have that much later. But if you now think about the dynamics of the inflammation, you can tell that these types of sudden cardiac deaths are probably not really related to only one type of immune cells, but are more presented by different or individual cell populations. So in this study here, we looked into the spontaneous ventricular tachycardia happening after myocardial infarction in a mouse model. So we refer to that mouse model here as the so-called storm model because it qualifies as an electrical storm. The model is based on two pillars. So we feed the mice with a potassium deficient diet so that they really have established hypokalemia. And after three weeks, they get a large myocardial infarction. And the combination of these two triggers here, so the potassium deficient diet, hypokalemia, and myocardial infarction will lead to these spontaneous ventricular tachycardia episodes. And this is here just a two-minute time episode from one of these mice here. So all these dense episodes here from the ECGs are ventricular tachycardia. And you can tell that most of these ventricular tachycardia episodes really happen hours after myocardial infarction. So if you now put that together with the inflammation or what I just have shown you about the dynamics of inflammation underlying myocardial infarction, the earliest cell population, which is probably contributing to these early type of arrhythmias here in that model, would be the neutrophil population. And that's why we now came up with a new experiment where we depleted neutrophils from the blood pool in the moment when we induced the myocardial infarction. So that is usually done by using antibodies against neutrophil surface proteins. And that really leads to the temporal depletion of the neutrophil blood pool. And if the neutrophils are not anymore in the blood, they can also not be recruited to the site of injury anymore. And if we now combine this tool here, so the neutrophil depletion tool was our storm model, which has the ventricular tachycardia episodes, this basically leads to the fact that we can reduce ventricular tachycardia burden by almost 80% if we take the neutrophils out of the equation, which kind of establishes causality here. When we look into patients, and I think that is even more interesting here, we studied the association between neutrophils and myocardial infarction patient here in an actual STEMI patient cohort from Oxford, almost 200 very well phenotyped myocardial infarction patients. So all of this here is corrected for troponin, MRI, all the metrics you can think of. And we plotted the neutrophil counts from these patients here by their arrhythmia severity. And there's really a trend. So the more neutrophils these patients had, so the higher the levels were, the more severe their arrhythmia phenotypes were. And you can also take that one step further here. And this is an independent cohort, which we had analyzed together with the Master and Earl Briggams, where we had almost 800 patients with myocardial infarction. And we just split the patient cohort here in higher and lower neutrophil counts and plotted their survival readout. So this is a combined survival readout for sudden cardiac deaths and also deaths from any course here over the first year after myocardial infarction. And you can tell that patients with higher neutrophil counts had a fourfold increased risk to die after myocardial infarction. But obviously, this is not specific to the infarct itself, but it's specific to the neutrophil count. So what we are doing right now, we are in the process of trying to understand that much better. So we are running a prospective clinical trial to understand the role of neutrophils in myocardial infarction patients. And the recruiting of that study here is ongoing. So this is very preliminary data. But I think it is very helpful to understand the neutrophil dynamics and how they contribute to arrhythmia in the MI situation. So we are taking blood from patients which get hospitalized with myocardial infarction, STEMI and NSTEMI patients. And we are also monitoring their ECG prospectively. So we don't know yet who of the patients will have which type of arrhythmias. But the study is a prospective study, as I said. And the ECG recording is ongoing with an event recorder for the next three years. So we have all the neutrophil biology. We are assessing that ad hoc with the blood samples. And only retrospectively we can analyze how patients with a certain type of arrhythmias are prone in terms of their neutrophil dynamics. So this is first results here from the study here. So you can tell that NSTEMI and STEMI patients have elevated CD45 positive leukocyte cells here. And this is particularly true at hospitalization for the neutrophils. So you can tell that there is this early neutrophil peak at hospitalization. The main difference, by the way, between NSTEMI and STEMI patients is the time they need to get to the hospital actually because NSTEMI patients usually take much longer to get to the hospital because the symptoms are not that strong. So that is our explanation why they probably have a lower neutrophil peak compared to the STEMI patients which usually come in one or two hours after MI. And this is data from the flow cytometry here. So we are analyzing neutrophil activation markers and wanna understand the dynamics of those because neutrophil activation markers are kind of responsible for the function of these neutrophils. And which came to a surprise to us was this one here. So it's the TRM1 receptor or in humans it's called CD354. So it seems to be reduced in patients with myocardial infarction and even more in STEMI patients. And this is a receptor neutrophils need to get recruited to the site of injury. What is also interesting here that this receptor never gets back to normal even if you follow up these patients for 180 days. This level will always be under the level of the healthy donors. And we are obviously also trying to understand right now what are specific like other proteins not only expressed on a surface but also other protein expression and which is different between NSTEMI and STEMI patients. So if you now think about this TRM1 receptor, yeah, so it probably qualifies as a good target to prevent neutrophils to get recruited to the site of injury. And this is here in in vitro experiment was a transfer migration assay. Probably some of you have seen that before. So you basically put a chemo attractant protein here. So the chemokine, yeah, which is needed by neutrophils to get attracted to any site of injury and it's also expressed by cardiomyocytes to get them recruited to the heart. And you can do that in vitro and then you can just count the neutrophils which migrate towards the chemo attractant gradient. And if we do that in presence of a TRM1 inhibitor, yeah, so this is now a small molecule inhibiting this receptor we found interesting. You can tell that the TRM1 inhibitor here prevents neutrophils from getting recruited to the other side. So they cannot react to the chemo attractant anymore. And this is probably something which is now taking it one step further to the clinical scenario. So we combined this TRM1 inhibitor with our storm model and just here the sanity check. So the troponin and infarct sizes are pretty much the same between both groups. So the TRM1 inhibitor doesn't help with the infarct size, but it prevents the neutrophils from getting into the heart. And if we do that here, so it's a pharmacological approach, we can prevent again the ventricular tachycardia burden by 80% in our storm model. So this is something which we are following up on right now and hope to get that closer to the clinical scenario in future. So what I was trying to tell you today is that neutrophils and especially, but also immune cells in general really matter and contribute to arrhythmogenesis after myocardial infarction. Thank you very much. Questions, please. Very nice. So I'm still me, Sam Dudley from Minnesota. Maybe not, I don't know. Anyway, so macrophages cause arrhythmia by secreting IO1 mostly, which then eventually causes oxidation of the right receptor and triggered activity. So neutrophils also secrete IO1. Have you looked at whether IO1 is mediating the neutrophil effect? And why are you measuring EKGs for three years? Neutrophils are gonna leave the heart within a week, maybe two, if you're lucky. Yeah, thank you. That's a very good question, actually. So we think that the neutrophil, that is also already described in mice, that the neutrophil dynamics are not getting back to normal even years after the event. So that's why we are following up on ECGs. Mechanisms, so pro-inflammation and chemokines, that is definitely a mechanism. Neutrophils would also have, if you look into macrophages and neutrophils, that's probably the same, but neutrophils are much more aggressive. So in an acute phase, they also do lots of oxidative stress, and we think that this could be one of the main mechanisms underlying neutrophils' action in arrhythmogenesis. So that's probably a difference between neutrophils and macrophages. More questions? If not, let's thank the speaker. Thank you. Now I'd like to introduce the next talk. The next talk is by Joanna Montgomery from the Department of Physiology in Auckland. The title of her talk is Glial and Neuronal Cell Influences on Cardiac Electrophysiology. Good morning and thank you so much Michelle and Igor for putting together this session and inviting me to be part of it. So today I'm going to talk to you about mostly neurons and a little bit of glia and looking at how they may contribute to atrial fibrillation. So my background is in neuroscience and my lab has a major interest in how neurons communicate and then how that goes wrong in different diseases. And of course neurons are quite different from many other cells in the body. They're highly polarized cells and they have these elaborate extensions which means that they can extend throughout the brain and also out to the periphery. And one of the major properties of neurons is that they exhibit plasticity. So what this means is that they can change their level of communication by changing synaptic strength. So this is illustrated in cartoon form here where we see an increase in synaptic strength known as long-term potentiation which results in either post-synaptic mechanisms of increased receptors or pre-synaptic mechanisms of increased transmitter release. And then with time we see an increase in synapse density. Now of course neural networks don't just exist in the brain. We know they exist all over the body and particularly in the heart. So here we see the innovation of the heart from the brain through the parasympathetic and sympathetic nervous systems. And if we zoom into the heart and look on the surface, particularly on the atria, we see these dense neuronal networks and these are called ganglionated plexi or GP. And if we do some acetylcholine esterase staining, we can see these beautiful neural networks covering the muscle tissue. So this is cholinergic nerve fibers. So given my background in plasticity and these neural networks and knowing the importance of these neurons in AF, this raises a question of is there plasticity in these neural networks and could this plasticity actually be a route for either disease or a target for modulation. That remains an ongoing large question in our lab. We started this work in the spontaneously hypertensive rat, which also shows spontaneous AF episodes either with carbicol stimulation or with electrical stimulation. And we developed techniques to do wholesale patch clamp electrophysiology on these neurons that you can see here. And we could measure spontaneous activity happening in them. And what we found is that there's an increase in the spontaneous activity in this spontaneously hypertensive rat. And then doing some imaging studies, we can see that there's an increase in the density of synapses. And also in the postsynaptic receptors, like I showed you before, which happens with LTP, we see an increase in alpha-7 acetylcholine receptors located at the synapses on these neurons. So it seems that these neurons have shifted to a potentiated state. We also developed some techniques to do calcium imaging in these neurons. And here's some examples here of some evoked activity happening in the GP neurons. And what we found in the SHRs, there was an increase in the amount of fluorescence indicative of an increase in calcium signaling and an increase in the duration of these calcium signals. Interestingly, in the glia, although this is spontaneous activity, we didn't see that difference at all. So there was no difference in the amplitude or in the frequency of events. We then advanced this work to look in human tissue. So we have been gaining right atrial GP samples. And you can see here, this is what these GP look like from the human tissue. So these are the neurons. And these are embedded in the epicardial fat tissue. And then on the right here, you can see in green, is the cholinergic neurons. And in the red are glial cells that wrap really, really tightly around these neurons. And for most of the data I'm going to show you today, this was the patient population. So we compared people undergoing cardiac surgery that were either non-AF or AF, and then tried as much as possible to balance age, gender, and comorbidities. And what we found in these neurons, first of all, importantly, is that it didn't seem that they were sick or different in AF patients versus non-AF patients. Their fundamental physiological properties were very similar in terms of resting membrane potentials and kinetics in response to single action potentials. We did see some significant differences when we compared rat versus human in these neurons. One of the major ones being capacitance. And when you fill the neurons while you're doing the recordings, you can see why that is. So the rat GP neurons are very simple, one or two neurites, very short neurites. Whereas the neurons from human tissue are much more elaborate and complex with large dendritic trees. The other thing we observe is that the dendritic trees wrap, and the axon fibers wrap around the cell body where we know the glia are, and a very dense amount of synapses happening here too. What we found, though, if we push these human neurons more, we started seeing some differences in their physiology in AF versus non-AF patients. So this is seen here, and we see different populations of these neurons based on their action potential firing frequencies. And we see with AF that we see a higher frequency of these higher firing neurons. We also can see that these neurons are more excitable by measuring their Rio base, which is defined as the amount of current required to fire an action potential. And we see this is significantly lower in AF patients, meaning their threshold for firing is much lower as well. Also we can look at accommodation. So neurons either will accommodate or not. And so accommodating means that they will slow their action potential firing with time. And then some neurons will just basically keep blasting the whole time. And what we see in AF patients is that there's a higher proportion of these non-accommodating high-firing frequency neurons. We also, together with the Bioengineering Institute, have been doing some extended volume confocal imaging. So on the left here, this is an example of peripheral staining of some GPs from rat tissue. And if we zoom in here, we can see some beautiful clusters of GP neurons amongst these dense nerve fibers. And on the right here is the human tissue. So this is what we call a scout image, which is a lower resolution but high area. And so you can see within the epicardial fat tissue, these clusters of neurons are scattered throughout the tissue. And if we zoom in on these GP, we can see clearly. So in green here is the cholinergic parasympathetic neurons. And red is the TH-positive sympathetic neurons. And they're resulting in coming nerve fibers. So doing some quantification on these neurons, we found, first of all, that there's differences in the types of neurons in AF versus non-AF. So we see that there are fewer of the cholinergic and more of the sympathetic neurons, and then more of, sorry, less of the dual-phenotype neurons in patients with AF. Looking at changes in synaptic density, we see that there's an increase, again, in AF versus non-AF, whilst no change in the actual density of the neurons, and an increase in the size of the synapses as well. So we're seeing this very similar structural plasticity that's been defined in the brain. Now we also know in the brain that NMDA receptors are really important for synaptic plasticity. Of course, these are glutamatergic synapses, whereas we're looking at a largely cholinergic system. So we were just intrigued to see if these receptors are here. And indeed, they are. So this is, in red, the gluN1 subunit of the NMDA receptor. So we can see punctate staining over the surface of the neurons. And in some neurons, we also see punctate staining within the nucleus as well. So we're just trying to figure out whether we see differences in AF and non-AF with that. But we can see a significant difference in the expression of these receptors in AF patients. We've also been looking at activation of neurons and glia in human tissue. So on the top here, looking at these beautiful glial cells marked with gap junctions here, wrapping around the neuron. And then what we can see with CFOS staining, so this shows activation of cell types. We can see this can happen very clearly in the glia, and then also sitting within the nuclei of the GP neurons as well. In some initial quantifications, data sets about 70% to 80% complete at the moment, we see a much higher frequency of CFOS-positive nuclei in AF tissue, showing that these really are activated, both neurons and glia. So what we found is this shift to a potentiated state that we see in AF, and this is both in human tissue and in rat tissue. So these neurons have a higher excitability, their synaptic function and density has increased. And so now we really want to know, well, what role does plasticity pathways play in trying to reverse these? If we can shift these synapses, if AF shifts these synapses into a potentiated state, can we actually shift them back to a depotentiated state? We know in the brain that glia are really important and play a major role in modulating neuronal excitability and plasticity, but we don't really know a lot about what they're doing in the heart in these GP neurons. Clearly they're really important, just their structural relationship is so tight with the neurons, and so we really are digging into that a lot more. And so what we know so far is that both glial and neuronal activation has increased with AF. And we really need to just not focus on one cell type here, and we need a multifactorial understanding of their communication. And also not forgetting the other part of this partnership, which is the fat cells as well, which is something else that we are doing some work on in the lab currently as well. So I think that triad of cells are really important in regulating heart rhythm. And I'll just finish by thanking the people that contributed to this work and the funders and our clinical team as well. Thank you. Please state where you're located. Yeah, David Patterson, University of Oxford. Nice data, Joanna. Really interested in the human data that you showed. So with the increased excitability, which is really convincing, do you think is there any evidence for M current downregulation? Have you looked at that at all? We haven't yet, but that's exactly what we want to dig into is now. It's like, what actual channels are underpinning this? Because that's going to really show us where we can more specifically target. Yeah, that would be an obvious target, because the pharmacology is there for treating patients with pain. So it would be really interesting to see whether you can affect excitability with those known drugs. Thanks, David. You may have mentioned this, but in the AF heart, there would be areas of non-myocytes, fibrosis or whatever other. Are the nerve terminals and the glia populating areas of non-myocytes? And what would be the effect of neuronal activity on the behavior of those cells that now we know affect so much the electrophysiology of the cells? Yeah, absolutely. So the neurons are actually buried in the epicardial fat, the cell bodies are, and then their processes then extend out. So in terms of where the actual excitability of that's all buried in the epicardial fat, in terms of where their processes go on the non-myocyte areas, that we don't know. So we just get the epicardial fat samples. We don't get the larger samples, although we are talking with Igor about some larger samples that we may be able to get. So atrial fibrillation, there's a lot of data now that atrial fibrillation is a metainflammatory state with macrophage activation and incorporation in the atria. So is it possible that the changes that you're seeing are secondary to the inflammation, or alternatively, the neuronal effect is actually on the white cell behavior rather than on the myocytes themselves? I agree. That would be really interesting to look at. So in my other life, I'm actually a neuroscientist in the brain also. And we are actually working on some neuroinflammation work there. So we have all the markers. So I agree, it'd actually be really cool to actually look at those markers in these GP neurons as well. Anna Finnegar from Northwestern Chicago. Thanks for this beautiful talk. I'm always excited to see GPs doing that well. I was wondering, in your rat model, did you notice any regional differences in your findings? Meaning a GP on the right atrium, does it behave exactly the same as our GP on the left atrial posterior wall? Yeah, so for the rat, we could actually measure from both whole different sites, which is obviously we can't do from the human. And we didn't see any differences. OK, thank you. If there are no more questions, thank you so much, Joanna. Thank you. And the final speaker for today is Fernando Santana from UC Davis Health. And the title of his talk is Microvascular Cardiomyocyte Cross-Treatment. Okay, well, good morning to everyone, and just wonderful talks preceding mine, and always a tough act to follow when you have such a distinguished group of presenters. So I'll do my best. My job here is to really tell you a little bit about some of the work that my lab at UC Davis has been doing for the last, I would say, five years in terms of the vascular anatomy of mostly the SA node, but really has impact just in multiple, in other regions of the heart. And I will tell you that this project started more or less at the beginning of the pandemic when a group of friends were connecting via Zoom every Friday just to keep our social network going. And most of the people in that group were people working on neurovascular coupling. And it's just that the relationship between blood flow and neuronal activity. And one of the questions that came up during that many meetings that, as I said, was a social lifeline was how expensive an action potential was or not. And to my surprise, there was a great deal of polarization in the group. There were half of it was like, it doesn't cost anything at all, to others that felt that it was really, really expensive. And that's why we have evolved very tight mechanisms for functional hyperemia. So that really triggered a chain of events that really culminated with some of the work that I'll tell you in a second. And you're all familiar with the cartoons like this one here. The cardiac cycle starts under normal conditions in the SA node via membrane clock and calcium clock mechanisms worked out by Ed Lakata's group over many years. And those action potentials escape. And that cycle in human is about 100,000 times per day. In a mouse, it's about a million times. And the amount of energy that goes into each one of those beats in the form of ATP is substantial. ATP is consumed, obviously, for cross-breach cycling, for signaling pathways like cyclic AMP generation, obviously, for the maintenance of ionic gradients like calcium, sodium, potassium that really are critical for the action potential. And just classic work, beautiful work back in the 90s summarized in this review by Liz Murphy and Steinberger in 2008 using NMR suggested that in the work in myocardium, ATP levels and here I just want you to focus on the purple line are relatively high under steady state conditions about 8 to 9 millimolar ATP and don't change that much in a beat-to-beat fashion. But you know, if you induce ischemia, those levels of ATP go down and then they can recover somewhat. As a graduate student, that's what I grew up with and I just really went with and just really elegant work that was in my mind at least set up everything that I, a lot of the things that I thought about related to cardiac energetics. So what I'll do today is just really come up with a different hypothesis, with something I think is new and it's that vascular patterning modulates beat-to-beat energetics and excitability. And again, this is just really being inspired by my neurovascular coupling people. And one of the things that I'll show you is that we observe that contrary to past work that there are actually beat-to-beat fluctuations in intracellular ATP at both at the base maker cells and also working myocytes and really these fluctuations seem to be patterned and correlated with vascular density, mitochondrial volume, and cell firing behavior. In other words, what I think it's happening is or what I will be trying to propose with this is that what you have is a new way of thinking about what sets excitability. It's not the only factor, but it could be, we think, an important factor. Cardiac rhythm and contraction are not just myocyte autonomous, but they're vascularly orchestrated. So when you have a system with a high vascular density, high mitochondrial support, you can support a high work capacity. We all learn, like with neurons, that cardiac cells, neurons don't have large amounts of energetic reserves to keep going and hence this really tight relationship between blood flow and neuronal firing. And the opposite, obviously it's true, is that when you have low vascular density, you have low mitochondrial support, the work capacity of those surrounding cells is likely to be lower. So let me just tell you that we started looking at the microvascular architecture of the SA node maybe about five years ago. I wish I could tell you that this was just motivated by my deep, you know, or a deeply rooted interest in how it is organized in the node, really, but it was just a postdoc looking for a project and the only thing I could come up with was this. And what he revealed to me during these experiments was actually quite surprising, and in the context of the conversations that I told you about every Friday, really triggered a lot of testable hypotheses. So what he found was that there was a great degree of, at least in the mouse SA node, there was a great degree of heterogeneity with regards to the microvessels. Capillaries were much, much denser in the superior region of the SA node than in the inferior region of the node, and it turns out that not only are seemingly more capillaries in the superior section of the node, there seems also more HCN4 positive cells, which I am going to use to define SA node myocytes, than in the inferior region of the node in the mouse. And I will not talk about pericytes, but pericytes are these really interesting cells with lots of attention from the neurovascular group of people on how they can control blood flow through capillary networks. They have the capacity to contract, and in my view, relatively understudied. Lots of pericytes in the superior region of the node, very fewer in the inferior section of the node. The density of pericytes is much higher than in most regions of the brain, so that's going to be an interesting cell type to study. So as we saw that, I said, you know, a couple of outcomes can be true here. One hypothesis is that cells throughout the node, for example, fire or have the intrinsic capacity to fire action potentials more or less at the same frequency, and that once they are in the node, in the intact part, that really the vascular supply to these regions is what sets the firing capacity. That's not what we found. Actually, what we found was that there was a degree of heterogeneity, as many people in this room, Igor, also Ed Licata's group, have shown of firing behaviors in the SA node. Cells isolated from the superior section of the node seem to have the capacity to fire action potentials, and some of them in a very regular manner. They're not perfectly regular, but they're regular enough. With some cells that fire irregularly with a lot of subthreshold voltage fluctuations, and then some cells that are quiet with some subthreshold voltage fluctuations, and then, boom, they go through a train of action potentials of a high frequency and periodicity. The inferior section of the node was mostly populated by, well, it had a significant number of cells, 42, but a lower level of cells that would fire action potentials. They tended to fire action potentials at a lower frequency, and then tons of cells, 36%, or about 36%, only fire subthreshold voltage fluctuations, and about 21% were silent. Here you have the superior section of the node, tons of capillaries up there, populated by myocytes that can fire action potentials at a high rate, and the inferior region of the node, much fewer capillaries, but also cells that seem to be much more quiet, and they're not firing at a high frequency. With that, we decided to just say, okay, let's just take a look at really what's going on with ATP in these cells now that there are a couple of new tools that have come out by groups both in Cornell and also UCLA, and these indicators are pretty good in the sense that they're single wavelength, they don't depend on FRET, so they're relatively easier to use. In particular, IATP developed by the UCLA group is relatively insensitive to ADP and also AMP, and relatively low levels of pH sensitivity, but mostly in the cytosol, but they also, the group in Cornell, developed a variant of the IATP that goes straight to the mitochondria. So the KD of these indicators, the apparent dissociation constant, it's about 1.2 millimolar ATP. So if ATP levels were 8 to 10 millimolar, these indicators should have been saturated, but that's not the case, that's not what we see. So first of all, what we see is that there are bit-to-bit fluctuations of ATP in the assay node. So this is a 2x vivo preparation, we're imaging this system, and contraction has been arrested with blevistatin using a two-photon microscope, and here you're seeing these oscillations happening in this particular preparation at around 4.2 hertz, and if you go to the inferior region of the node of the same assay node, what we saw was that the frequency of these fluctuations was lower, and also much smaller in amplitude. And then when you really average out these values, what you get is that, on average, at least, ATP levels are higher in the superior section of the region of the node than in the inferior region of the node. Remember, that's the region of the node that has the cells that are firing action potentials at a much higher rate. It's not true, it's just not true, it's also true for ventricular myocytes. The situation there is a little bit more complicated, but what I will say is that, associated with action potentials, many cells exhibited what we call mode 1 type of ATP transient, an increase in ATP cytosolic levels with each one of the action potentials. And there was a second type of cell that produced what we call mode 2 type of ATP transients, and there, instead of an increase in ATP, there was a decrease in ATP associated with each one of these action potentials. So here, what you're seeing is a general picture that ATP levels are not as high as we thought they were, based on prior studies, and also oscillate in a bit-to-bit fashion during each one of the cardiac cycles throughout the heart. I'm about to finish here, okay, all right. So I will just end with this slide, because I'm running out of time, but what I will tell you is that we have also been able to measure oscillations of ATP in the mitochondria. Here what I'm showing you is one type of fluctuations, the mode 1, which is very similar to what we see with the cytosol with mitochondrial ATP levels going up with each calcium transient, and then the opposite happening in mode 2. So with that, what I want to really convey, I'll wrap up by conveying the message that there is a tight and seemingly strong relationship between vascular supply, blood flow, and electrical activity of the underlying cells, and that I will posit that under multiple pathological conditions when you see vascular varifaction, the ensuing hypoxia that emerges from that trigger changes in the myocyte that then can alter its own electrical excitability, and hence some of the things that we see during heart failure with reduced ejection fraction, but also with preserved ejection fraction. And with that, I will stop, and thank you, guys, for having me. Many thanks, Peter Kohlfreiburg, I enjoyed that. The difference in automaticity may be related to the lack of mechanical influences. At least that seems to be the consensus in pacemaking, that if you have these activities that are either intermittent or irregular, you apply a little bit of stretch as you would normally have in the sinus node, and the cells start to fire beautifully and rhythmically. Have you considered that option, have you tried that? I haven't tried an experiment to just test that, not only in the context of electrical activity but calcium and also ATP. I would say that at least an alternative hypothesis or model needs to be considered, is that on top or maybe instead of the entrainment model that we have embraced and for which there is a lot of data, I will proceed that those cells that don't necessarily fire action potentials are there to really serve as noise generators that help cells, the surroundings, connecting cells to reach threshold. Neuroscientists call that stochastic resonance, so when you have a periodic oscillator, some of those oscillations will not reach a threshold, but then when you have noise that really pushes over the edge, it hits the system. In that context, you will see that heterogeneity not as a problem but rather as a benefit. What I would argue is that the regions of the node that have fewer microvessels are less like, I would argue, that have less energy to just fire at the higher frequencies of the super vascularized regions of the node. Just a thought. Thanks. Hi there. I'm Nikki Posnack from Children's National, so this is really amazing work. I noticed that your traces weren't overlaid, so I had a question about how, like what the timing is. Is it, is the ATP being released in response to that calcium upstroke or like how does that timing work? It's a really, thank you for the comment, and I'll say that's a really good question. One of the, so we have done that and there is a delay of a few, I would say about ten milliseconds between one signal and before ATP starts going up or it starts going down. I would caution, because the reviewers of the paper really push us on this one, that the kinetics of the GFP-based ATP are much lower than the calcium indicator. So even with that in mind, I think it is, it's real, but from a purely biophysical point of view, I have to exert some caution on that one, but it makes sense that there is some delay. And did you mention using blebistatin? Yes. That was my other question. Okay, got it. Yes. Thank you. Especially with the node. Not for the ventricular myocytes isolated, but for the node, yes. Hi there. Raghavan, Ohio State. Fernando. Really, really beautiful work. Two quick questions, or one, I guess. Are there, between the mode one and mode two cells, do you notice any differences in sodium potassium ATPase and our SERCA expressions? That's a good question. I personally haven't done the experiment. We think that the major driver for mode one or two is the expression of mitofusings. So cells that express a much lower level of mitofusing have weaker mitochondria SR coupling, and those are the ones that likely, those are the ones that display mode two type of kinetic, transients, ATP transients. I guess the second question is, when you see changes in rhythm in the myocytes, do you see any changes in the endothelium? Good question. No, we haven't looked at it. We should. Thanks. Yeah. Thank you. Okay, last brief question, please. Fernando, very cool talk. It's very beautiful. If I understand it's mice or mouse model, is it specific for mouse is coronary, cyanide or not coronary, artery, anatomies, because it's going from superior to inferior part, from head to tail, sinus node, because in the human, it's majority cases is entering through receptor, more inferior part, and where we have actually on the whole of the head, center tail, we have more supply, actually preferential to the tail of cyanide or not. Or you have different anatomy in mice. I'm saying it's, is it spacious, specific observation? Good questions. In general, first of all, obviously, we haven't looked with this level of resolution. But is it consistent presently, like consistent artery in all your mouse, you have that, right? It is very consistent among, between mice and different strains. And what I would say is that in terms of, even though the exact location of the SA node artery within the node shifts quite a bit, at least some of the classic work that has been done, and it's quite, some of it is quite old, it shows that at least in humans, there are some elements that overlap in terms of the microvascular anatomy, with regards to just the large artery going, you know, from the top to bottom. Whether the microvasculature too will exert that, we don't know for sure just yet, but I will tell you that at least expression of markers for endothelial, you know, capillary endothelial cells in the pig seem to be higher, most are expressed to a higher extent in the head of the SA node than in the tail of the SA node, which it would be. Thank you so much. Yeah. Thank you. Okay. That concludes this session. I'd like to thank all the speakers for excellent talks and staying on time, and the audience for such great participation. Thank you so much. Thank you.

heart function

electrophysiology

fibroblasts

endothelial cells

immune cells

myocardial infarction

arrhythmias

neural networks

ganglionated plexi

microvascular structures