 in running for a while now, aims to give early career researchers and online platform in this crazy time to continue to share their research as an alternative to in-person gatherings. I am Deidler Weigel. I'm Director at the Max Planck Institute in Tübingen and I'm also Deputy Editor at E-Life. I've been with E-Life from the very beginning since 2011. I'm original Developmental Biologist by training, working on Drosophila and later switched to Plants and now work on Plant Evolution and Immunity. So today we'll have three talks again. We'll hear from Enan Morales Navarrete, who is a junior researcher at Pontificia Universidad Catolica de Ecuador in Quito. He'll talk to us about multi-scale analysis of liver tissue organization. He is a recent awardee of one of our travel grants, which he can use at the next in-person opportunity. And meanwhile, he is gonna present the talk to us that he was planning to give at the BioSB20. Next up will be Carolina Makovsky, who recently became a postdoc at UC San Diego in the US. She'll be presenting work from her PhD, which she carried out at McGill in Montreal, Canada. She'll be speaking to us about probing myelin in first episode of Psychosis with MRI, a framework to understand negative symptoms and verbal memory. And this talk again was initially scheduled as part of a symposium at the Society for Biological Psychiatry as OVP Conference in New York. Where Carolina was also to be awarded in Early Career Investigator Award. Hey, hello. And then finally, we'll have Philip Lange, Assistant Professor in Canada Research Chair at the University of British Columbia in Vancouver, Canada. And he'll talk to us about the proteome landscape of pediatric leukemia in patients and xenograft models. So each of the talks is gonna be 10 minutes. And after the 10 minute talks, we'll have five minutes of questions for each speaker. And then we'll move on to the next talk. So to ask a question, any of you participants can type it into the chat on Zoom or directly into the Google document, which is linked in the Zoom chat window. And we are joined today in the background by Miranda, Anya and Naomi from E-Life, who are going to support us. And they'll help me to line up your questions. If you're able to, I'll invite you by name to ask your questions and the team will unmute you so you can do that. Otherwise, I can just read your questions out loud and include your name where possible. And then we'll also have this open notes document which is another place for you to contribute shared public notes. We welcome you to do so and to list yourself as contributor in the list above the speakers for today's webinar. So thanks a lot for doing this. Finally, I'd just like to let you know that we are recording the webinar and also live streaming it on YouTube. And during this live webinar, we ask you that you please be respectful, honest, inclusive, accommodating, appreciative and open to learning from everyone else. Do not attack the mean disrupt harass or threaten others or encourage such behavior. And if you feel uncomfortable or unwelcome on any of these webinars, please contact E-Life by email to events at elivesciences.org. And the inbox is being watched by Anya Starris from E-Life. And of course we reserve the right to ask anyone to leave and or to deny access to a subsequent webinar on Zoom. If you need any help at any point, please send Miranya, Mirando or Anya a chat message directly using Zoom. So in Zoom there's this functionality where you can send private messages. All right. So first we'll hear from Ernan about multi-scale analysis of liver tissue organization over to you, Ernan. Okay, just give me one second. Can you see my screen now? Yeah, here it is. Can you hear me? It is on, perfect. Thank you. Okay, first I want to thank the organizer for giving me the opportunity to present my work today as part of the E-Life online research talks. So now we'll tell you about the multi-scale analysis of liver tissue organization. It was, this was a project I was working on during my stay at the Google Professor Marino Serial at the Max Plan Institute of Molecular Biology and Genetics. So let's start. One of the major challenges in cell and developmental biology is to understand how cells can form complex dimensional tissues. Probably nowadays one of the most powerful to address this question is microscopy. Here I'm showing you an example of a confocal image in three dimensions of liver tissue. As you can see here, we can observe the different cells forming this complex tissue in 3D. But most of the time, just visualizing these components is not enough to understand the structure of the tissue. So we have to go one step forward and we actually have to perform a digital reconstruction of the tissues so that we can perform also a quantitative analysis. So here I'm showing you an example of the reconstruction of the different cells forming the liver tissue in 3D. So once we have the different components, the different cells forming the tissue, so we want to understand what is the actual tissue organization. So what we think is that this problem is like to try to solve a puzzle in 3D. So where we actually have all the pieces, like meaning the cells and the different components of the tissue, we know the shape of the different components. We know the final configuration of the tissue and what we want to do is try to find out where are the rules that we have to follow with all these pieces together to perform a functionally dimensional complex tissue. So during this talk, I will tell you about how we try to apply these concepts into the analysis of a specific tissue that is liver tissue. So for this, let me first give you a very introduction into liver. So liver is a vital organ that is in charge of several functions like bicyclism or metabolism. The minimum structural unit of the liver is called the liver lobbyol. So we can understand the liver lobbyol as a two-pipe system. So on one side, we have the sinusoidal network. So the blood is entering to the liver, to the portal vein, going through the sinusoidal network and going out towards the central vein. So we have another network, the canaliquin network that is transporting bile in the opposite direction towards the bile duct and out of the liver. So between these two networks, we have the parasites. The parasites are epithelial cells that have a very peculiar and unique apical basal polarity. Every single parasite is facing several sinusoids which allows for the proper exchange of nutrients with blood and the apical member of this parasite between the parasite is forming the bicanaliquids which together give rise to the bicanaliquin network. We want to understand how this organization at several levels can give rise to a global tissue organization. For this, as I mentioned before, the first step is try to get the actual shape of the different components. So a few years ago, we developed a pipeline for the imaging and reconstruction of these different components of the tissue and multiple scales. Briefly, so in the first step, we image the tissue, like a big portion of the tissue, meaning around one cubic millimeter of tissue with lower solution. This allow us to reconstruct a big structures like the veins, the central vein here in Cian and the portal vein in orange. So in the second part, we take a smaller piece of the tissue here, these white boxes that we can see here. So if we do the imaging higher solution, we are able to reconstruct a cellular level components like bicanaliquin here in green and the insular network in magenta and also individual cells. And we can do it with such a resolution that we can even identify subcellular components. This is an example of a nepotocyte where we can identify the apical membrane here in green, the basal membrane in magenta and the lateral membrane in gray. So we did this implementation in a software that is called motion tracking, that is free to use. So now let me show you a full reconstruction of the small part of delivery of your, so the central portal vein access. From here to the end of the presentation, you will always have the central vein in Cian on the left side and the portal vein orange on the right side. So this reconstruction of bicanaliquin network, as you can see here, is a complex three-dimensional structure, very similar to the sinusoidal network. These two networks are never touching each other, but they are globally organized between central and portal vein. This is the nuclei of the different cells that are from the tissue and this is the reconstruction of the cell surface of the different parasites. As I mentioned before, every single parasite is in contact with these two networks. As you can see here with the bicanaliquin and the sinusoidal network. And this also defines the apical and basal patches of every single cell. As you can see here in this example, the parasite have what is called a multipolar organization of the apical and basal compartments. So in this project what we want to understand is how this multipolar organization that looks a bit complicated to explain at the cellular level can give rise to two networks that are functional at a global level, meaning a tissue level between central and portal vein. To do so, the first thing that we have to do is to describe mathematically and the polarity of the parasites. So for this we had a collaboration with Benjamin Friedrich and Frank Hewley here from the Max Plan Institute of Physics of Complex Systems. So with them we use tools from the physics to mathematically describe the polarity in the case of cells with the multipolar organization. So just let me briefly explain you. In the case of simple polarity, imagine that you have one cell where you have only one patch of any market could be the apical market, for example. And then mathematically it can be easily described with a vector. So it's going from the center of the cell towards the place where you have the maximum accumulation of the market. In the case of cells where you have not only one patch of the market but several ones like in a parasite, we cannot describe this anymore with a vector. We have to use something more complicated. We use in this case what is called a tensor. So a tensor can be described graphically with two axes. So one axis is going to describe the orientation where we have the maximum accumulation of the market and the second axis where we have the minimum accumulation. So let me show you two extreme cases. So imagine that you have a cell where you have only two patches. So the first axis that we are going to call it bipolar axis or A1 is showing you the direction where you have the maximum accumulation of the market. Another example, so imagine that you have not two patches but a ring-like structure over the cell. So A1 cannot be defined because you have all over the surface. But you can define a second axis that is called A2 that we call ring-like axis, which give you the direction where you don't have accumulation of the market or the minimum accumulation. In the real case, like in a parasite, these two axes are uniquely defined. So it means that we can define the bipolar axis and the ring-like axis in a unique way for every single cell. In this case, for the organization of the apical membrane, we can do the same for the basal membrane as well. So for the cell that we saw before, so the bipolar axis is showing the maximum accumulation of the apical surface. And the ring-like axis is showing where we basically don't have apical surface. So once we have the mathematical description of the polarity that we call biaxial-cell polarity because we describe the polarity with by two axes, so we wanted to see how this is organized over the liberal of yield. So here, we just plot the biaxial in one of these axes, the bipolar axis, so you have the same membrane, the portal membrane, and every single axis represents one aparocyte. So from this picture, it's still difficult to see if there is some pattern of organization. But actually, if you only do some local averaging, meaning that you remove the local noise, we could already observe the underlying pattern of organization. It's actually not that there is no organization at all, but there is a pattern behind the organization of the aparocyte polarity. And if you can see, if you see to this organization here, through the mines to a flux line that are going from one side to the other. So we can quantify this. So for this, we simulate the flux, our flux process from center to portal vein, and then we can measure how well these axes are organized, aligned with this. So this is shown here. So this is the quantification. This is another parameter. Negative values mean they're perpendicular. Positive values are parallel. So to sum up, we found the polarity is aligned with the flux of this diffusive process. And this is very similar. So we can do something similar to the Neville, for example, a sinusoidal Neville, and we found a similar result. So it means that both the circularity were parasites on the sinusoidal Neville, a couple, and probably locally organized. So sum up. So today, I show you how we are starting from microscopy images, in particular confocal microscopy images. We can reconstruct the different components of the tissue. And we can use this information to derive some principles behind the organization of the tissue. In this case, we talk about multi-scale organization because we see at cell level, a cellular level, bi-axial polarization of these cells. So at the multi-cell level, we have a coordination between the polarity, the circularity, and the organization of the sinusoidal Neville to different cell types, which at tissue level will serve like a long range organization. In terms of physics, it is called a liquid crystal organization. It's not a perfect organization like your crystal, but it's not totally dissolved like in a liquid. We have something in the middle. We have local noise, but we have some long range organization. So with this, I just want to say my former lab and Marino with Marino Serial and my collaborators at BKS and Pinnacle University of Tennessee. And now I'm happy to take any questions from you. Great. Thank you very much, Anand. Awesome. All right. Let me ask the first question. So this is really amazing what you can do with modern microscopy and reconstructing these networks and the cell surfaces and whatnot. So you did this with hepatocytes. Could you easily transfer this to other tissues as well? Yes. So actually, so in the beginning, we applied this only for the reconstruction of liver tissue, but it also worked for the nicest of kidney tissue or lung. That's something that we are very proud of. So the methods are very general, so it can be easily translated to any other tissue. And so I guess the answer to the other question is chicken and egg type of problem. So you have this sinusoidal network, and then you have the cells. So what comes first is that this sinusoidal network that forms the scaffold structure that then guides hepatocyte polarity or is it the other way around that the hepatocytes control the organization of this network? So that is actually a very good question. So what is describing the literature is that the sinusoidal are making like this scaffold in the beginning. So because they are transporting blood and doing development there, they want to form first. And then they guide the polarization of the parasite. However, we found that it's actually not only one direction, but it's a two-direction organization. So to analyze a little bit more, what we did is we added one protein that is Integrin Beta 1. So Integrin Beta 1 is basically in the cellular matrix. So we did the unknown down the nanoparticles in the parasite. And surprisingly, what we found is that the polarity is still establishing a parasite. However, this liquid crystal organization is gone. So here we see an example of the knockdown of the Integrin Beta 1. And we also observed that the organization of the sinusoid is gone after this knockdown. So since the knockdown was specifically for the parasite, so it means that also the polarization of the parasite is altering the organization of the sinusoidal network. So we think that it's a two-way interaction. We have a certain communication between the two types of cells. Is there precedence in any other tissue where you have a similar relationship between the individual cells and then some supracellular network? So I find that no nothing. But I don't know. OK, all right, cool. All right, great. Wonderful. Let me check. I think that those were all the questions. All right, awesome. Thank you very much, Annan. Thank you, thank you very much. All right, so we have Carolinas slides up. So Carolina already introduced her from UC San Diego. San Diego has a special place in my heart. I was a professor there for 10 years. So I really love San Diego. Carolina, please take it away. Excellent. Let me just share my screen here. Can everyone see this? And can everyone hear me? Excellent. So thank you so much, Dietle, for the introduction and Eli for allowing me this opportunity to share my research in these crazy times. But it's really great to be able to connect virtually. So yes, San Diego is amazing. Montreal is as well. This is where this work was actually conducted since I just actually moved to San Diego. So we're going to shift still looking at imaging techniques. But now in humans, in a psychiatric application. So I'm working here with first episode of psychosis patients. And we want to look at neuroanatomy, particularly myelin, to better understand predictors of poor functional outcome in psychosis. Let's make sure I can advance perfect. So I want you to go back way back into time and think back to when you were a late adolescent. So when you were a teenager and to early adulthood, you can appreciate that this is a time of immense growth. Your brain actually is quite plastic and actively developing during this time. There's a lot going on at this time that can really set the stage for the rest of your life as well. And this unfortunately is actually when psychosis tends to have its onset. And when people think of psychosis, they often think of positive symptoms. So these are symptoms that are kind of a blurring of what's real and what's not, things like hallucinations and delusions. And these are treated quite effectively with anti-psychotic medication. But there are other factors that are, I don't want to say ignored, but we don't have great treatments for. And these include things like negative symptoms, so like a lack of motivation, social withdrawal, as well as cognitive deficits. And in particular, and what's sort of an enigma in psychiatry is that verbal memory deficits tend to have even more pronounced deficits in patients with psychosis compared to other cognitive domains. So these are the factors that I really want to focus on in this talk. Now, when it comes to the anatomy, every single region has basically been implicated in psychosis. It is a disorder of disconnectivity. But the hippocampus stands out as potentially a tangible marker that we could use for therapeutic purposes. This is actually a figure from Carol Tominga in 2010, which showed that the hippocampus might be central to psychotic symptoms, so positive symptoms. And she basically showed that there's specific lesions within the circuit that can lead to those positive symptoms. And when we look at imaging studies, again, the hippocampus kind of comes out at the forefront where there's the largest effect sizes with volumetric differences. So lower hippocampal volumes in both patients with schizophrenia and bipolar disorder. So there's something to be said about the hippocampus. So for this talk, there's two primary aims. One is the hippocampus connected to other cortical abnormalities. So we wanted to do this by probing intercortical myelin, which I'll show in a moment. And if the hippocampus indeed is central to some of these cortical abnormalities, is it also related to verbal memory and negative symptoms? These predictors of poor functional outcome. So there's lots in the cortex that we can look at. And I'm really gonna be focusing on compartment A here. So these are intercortical myelin fibers. And the reason being is this is a highly plastic compartment of the brain. And also when we look at the proportion of these fibers in the brain compared to other long range white matter fibers, they're actually very, very highly expressed in the brain. So there's more intercortical myelin fibers than say long range fibers. So as I mentioned, this was a study done in Montreal. So we are taking these first episode of psychosis patients from a clinic called PEP Montreal. This is a longitudinal data set, which is really, really valuable and very hard to get in first episode of psychosis. So we're tracking these patients. I have 27 patients and 29 controls with really high resolution imaging. And we track them for about 18 month time period after their entry to the clinic. So the bottom here, you have clinical time points after their first episode of psychosis. And then we have imaging time points currently at the top. And then we also get our clinical and cognitive assessments. So you can see there, we have pretty closely spaced time points to really be able to characterize fine changes in neuroanatomy as well as cognition and symptoms. And these are patients that are between the ages of 18 to 35. So that early adulthood time period I was saying. Okay, so how are we gonna get at myelin deficits in vivo in patients noninvasively? So for this, we turned to a MRI or magnetic resonance imaging technique called quantitative T1 mapping. Where essentially each of these pixels or they're actually in 3D. So each boxel in the slice actually holds a relaxation time in milliseconds. I won't get into the physics here, but all you need to know is that the shorter the relaxation time, it's more likely that there's more fat or white matter in the signals. This is why it comes up really dark on this image. And the longer relaxation times tend to have more gray matter. And it's actually been shown that a large contributor to this T-run relaxation time is myelin content. So whenever you see QT1 on my slides, I'm just gonna interpret it from just for clarity as myelin content. Okay, so we, as I mentioned, we really wanna understand hippocampal centrality in relation to intercortical myelin. So to do this, we have a really high resolution T2-weighted image here that really lets us gauge hippocampal subfields, but we also had an atlas that allowed us to resolve the white matter that is output from the hippocampus. So this includes things like the alveus, fimbria, fornix and mammillary body. And at the level of the cortex, we can probe intercortical myelin layers across many, many portions of the brain, but just for simplicity, we then reduce the data to 62 cortical regions. So we can get these myelin content for cortical and hippocampal regions. So we have longitudinal data. So in order to handle this, what we did is we really first wanted to see is the hippocampal central to cortical abnormalities. So we adapted a method from Ellen Evans Group at McGill University, where we can assess between two regions of interest, this similarity or dissimilarity in their trajectory of myelin across time. So you can imagine ROI is subfield CA1, and this is prefrontal cortex. How similar or dissimilar are the myelin changes across time? And in this way, we can actually get for each person or each participant a subject-specific matrix when we look at all pairwise comparisons. And once we got these matrices, we calculated something called the participation coefficient. And in graph theory terms, this is a measure of centrality. So just to give you kind of a visual of this, this is obviously airport activity pre-coronavirus times, but for those that are familiar, Chicago, here is a very, very popular, well-connected airport. So you can argue that has very high centrality. It'll get you anywhere you want. Whereas an airport such as Missoula International, it has very low centrality. So we can kind of adopt this approach to the brain and we can say, okay, this is our hippocampus with all of its regions, and how is it highly connected or not as connected to these other cortical networks in the brain? So we actually did indeed find lower centrality in first episodes of psychosis patients in this hippocampal unit compared to other cortical networks. And when we looked post-talk, it was really driven by these output structures of the hippocampus, including CA1, alveosporinics, and the millery body, bilaterally, as well as a little bit more left hemisphere representation. Okay, so that's all fine. In Dandy, we have the centrality measure, but I really wanted to bring it back to clinical outcome. So at the behavioral level, we knew from our data and also previously that negative symptoms and verbal memory deficits are related. So having a worsening verbal memory trajectory over time is associated with worsening negative symptoms. And what's really interesting is that lower hippocampal centrality is also actually significantly associated with worsening negative symptoms and worsening verbal memory over time. And once again, when we look post-talk, these results are really driven by centrality of the left CA1, alveosporinics, millery body, and right fibria. Once again, these output structures, mild and rich regions of the hippocampus. And just to bring this all together in a model, we actually found that verbal memory or changes in verbal memory over time mediated this association between hippocampal centrality and changes in negative symptoms. So to sum up, we found that there's disturbed covariance between the hippocampus and the cortex and this might actually stem from mild and rich output regions of the hippocampus in these first episodes, psychosis patients. And we are showing here for the first time, not only is the hippocampus central to positive symptoms as has been believed before, but it's also central to negative symptoms and verbal memory might help to explain this relationship. And the hippocampus, as we all know, is highly plastic. So it positions it well as a potential therapeutic target. And I don't want to get into every single future direction here, but what made me really excited is that we know that hippocampus, for instance, can actually change with exercise. So we know that exercise can actually increase neurogenesis in the hippocampus. So perhaps this could be a really beneficial treatment avenue for these patients with negative symptoms since antipsychotic medications aren't helping them. And finally, there is transdiagnostic potential of this work. You may have noticed I really didn't use any diagnostic categories here. We can actually see verbal memory deficits and negative symptoms in many psychiatric and neurological disorders like epilepsy, Alzheimer's disorder, obsessive-compulsive disorder, just to name a few. So with that, I will wrap up. I'd like to thank you for your attention. This was work supervised by Allen Evans and Matan Lepage in Montreal. You can see all my co-authors there in this paper. The revisions were just submitted last week to hippocampus. So hopefully we'll see a good outcome and I'm happy to take questions. Thanks so much. Great. Thank you. Thank you, Carolina. Thank you for being very much on time as well. So these patients, are they medicated? Yeah. So what's really unique about the clinic that we worked with is in order to meet criteria for a clinic, they had to be relatively anti-psychotic medication naive. So over the course of their life, they could only have a cumulative dose of one month of anti-psychotic medication. So this means that when we did take patients to the clinic and started scanning them, we could really rule out anti-psychotic confounds. Now, of course, once they had their first psychotic episode, they were put on anti-psychotic medication. It's a naturalistic study design. So it's not like they're just randomly, or they're not assigned to a particular medication. But because we worked so closely with clinicians, we could really track down their dosage, their medication adherence. So we actually did do a supplementary analysis controlling for their anti-psychotic medication dosage, which we had very well recorded. For the purposes of the hippocampal centrality analysis, it did not impact our centrality measure at all. But there are some studies that have suggested that a psychotic medication can have an impact on intracortical myelin. But it didn't seem to impact this study. So, you know, being far away, so you say psychosis, is that just one disease? Or isn't it many diseases, psychosis? Yeah, so psychotic disorders definitely have, they're quite heterogeneous. There's a wide spectrum. So most patients will go on to have a diagnosis of schizophrenia. Others will have bipolar disorder. There's also, some will have a diagnosis of major depression with psychotic features. And then others actually have a first episode and they're okay for the large part. So it was just this transient first. So, you know, sometimes it's just the first six months and then they're put on a psychotic medication and maybe there's also some other factors in their life that just required more clinical attention. And, you know, that's, I wish it was more of a common case, but about, I'd say about 15 to 20% of patients actually have a first episode that resolves itself with, at least with early intervention. And I think that's also a really big plus of this clinic is that we are catching cases early. But yeah, absolutely. There's many psychiatric diagnostic categories here, but I think the field is trying to shift away from looking at these diagnostic categories and instead looking at severity of symptoms and how it's related in a more dimensional manner to the brain and outcomes. So from your work, does that speak to, you know, how important is how early this is caught basically? So how early you got diagnosed? Yeah, I think that principle really stands for many disorders, you know, early detection is really key for better outcomes and for more successful interventions. So absolutely the same stance here and it has been actually shown that having these first episodes, psychosis clinics in place, there's quite a few actually in Canada does indeed actually reduce the severity of psychotic symptoms and relapses in the future. So this is a really important part. And now there's studies that are being designed such as the one I'm working with right now actually called the ABCD studies. So this is adolescent brain cognitive development in the United States, tracking 11,000 kids and their families across the country for 10 years. And the idea is we're tracking them from nine to 21. And there's gonna be unfortunately all sorts of, there's gonna be some kids that do develop psychiatric symptoms, including psychosis during that time, but then we might be able to really find early markers, both at the level of the brain, also in terms of behavior, school performance, et cetera, that we can help catch these disorders earlier. We have one last question obviously from somebody who understands this in more detail than I do. And question is, could there be some influence of verbal memory on social function that could contribute to motivational deficits and other negative symptoms? Absolutely. And I think that's why we are seeing that verbal memory negative symptoms are linked. So, and it's not clear if one comes before the other, but you can imagine that if someone has verbal memory problems, they might not be able to express their thoughts as well. And they might be a motivated to actually engage in social interactions because of that, or they might be embarrassed so they don't talk as much. And then that leads to some negative symptoms as well. Sort of looking at the literature, you do find that before the onset of psychosis, so in clinical high risk, you see verbal memory deficits already. So it's possible that verbal memory might be the one that comes before. You also see verbal memory slightly to a lower extent, but lower verbal memory deficits even in individuals or family members or first degree relatives of patients with schizophrenia or psychosis. So it could be that, yeah, exactly, verbal memory might influence negative symptoms, but absolutely it is very much intertwined with social interaction and communication. We need to kind of disentangle those pieces a bit further. Okay, great. Well, thank you again. Thank you again, Carolina, wonderful. Thank you. All right, our final speaker today is Philip. And he'll talk to us about proteome landscape of pediatric cancer in patients and xenograft models. Philip, over to you. Thank you, Detlef and the fabulous team for having me here and Carolina and Hanan for the fantastic, interesting talks for me. And so we'll switch gear a little bit, but we stay with the scheme of trying to identify problems earlier. And so we'll now look a bit at cancer, specifically cancer in kids and proteomics. And so you probably all heard about the new idea of precision medicine or by now I'm not so new anymore where you find a genetic mutation. It's usually genetic mutation and match a specific drug to it. Previously, what we usually did is we found a common phenotype. It's a cancer or it's a cancer in the lung or it's a cancer in bone. And that's the best treatment we can do for it. Now we can do a bit better and find a mutation and find a drug that might hit this susceptibility. That's the idea of it. It's a nice idea. The question is how well does it work? Well, the positive news here is that in the adults and even in the kids, we're doing very well at finding such mutations that we could potentially target with a drug. That's what we call actionable. But then when it comes to how many of these we can actually treat with these drugs, it drops dramatically. So only 23% of the adult ones that have a drug identified get treated and only 8% of the kids get treated. And if we look at the outcome in this MITA study, we don't have data for the kids yet, but for the adults, really only 7% show a positive response. And now this has a number of reasons but one of the reason is really that we look at this late. Usually it's in the relapse, it's in the high risk patients. And at the time when we find this mutation, this susceptibility, they are already in a very dire state and oftentimes don't make it to the time when we can actually give them the drug regulatory slowdowns there and the disease is progressing rapidly. So we're late. And in this cycle that I've painted here, where we go from a diagnose, we can treat fairly well initially in a monitor and then we relapse eventually. We also have a psychological problem where you have this constant fear of, it all starts all over again, the cycle starts all over again. And at BC Jones Hospital in Vancouver, Canada, we started initiative to think about breaking this and really switch this to preparing for a relapse. So understanding the kids cancer before it comes back, coming up with an idea about the relapse before. And what we do for this is, well, we look, we enroll all kids at diagnosis and look at diagnosis biopsies and remission biopsies, try to find on genomic level and now also on proteomic level, specific differences in these tumors that make them susceptible. And then also expand these cells. We're only doing this for leukemias so far, we expand these cells in a xenograft model in a mouse model where we get more of these cells because one of the problems really with kids is that you don't get much material that you can study. And then we can study the biology of each kid and we can study susceptibility and really treat them with drugs in the lab and we will be eventually ready for a relapse. Now, this is a setting that sounds extremely costly and implausible, but luckily, even though childhood cancer is a huge problem for the ones who get it, it's still a rare disease. So number-wise, it's actually something that you can do. And you have to remember that the cost occurred if a kid has to live with the late effects of a cancer treatment for decades on is actually not a small one. So the question for this talk today really is that we have to answer how well do these mice that carry the kids cancer cells reflect this cell biology that it has in the body? It's sitting in a very different host in a very different environment. So how much has changed? We know that the mutations are retained. We know that the general drug susceptibility is retained, but if we want to study more of it, we need to know do they actually have the same makeup? And the critical component in its biology or what we use as a proxy here is the protein makeup. So what we did here is we studied 10 BALL patients and 10 T cell leukemia, four T cell leukemia patients expanded them in several mice. And then we looked at the protein. We also looked at post-sensational modification, namely phosphorylation, most will be familiar with that. And we looked at protein proteolysis cleavage by proteases. And I want to segue into that for a second here. So proteins can be cleaved by proteases. This can have huge functional impact. And to monitor this, because most won't be familiar with that, what you can actually do is you can look at the newly created start of a protein. So if you cleave a protein, you will create a new immunoterminous. And now we developed a new method that is very sensitive at enriching these N-terminal parts. We call them N-termini of proteins. And so we can now get thousands of those out of very small amounts of material. This is 10,000 T cells or respectively a few leukemic cells from a patient biopsy for example. And we can do this with a nice small coefficient of variation, which is of course always helpful. So if we now look at this from a high level perspective and just look at the protein, quantified protein of a cell, a kid's cell out of the patient and out of the mouse, do they align or are more the cells coming out of the mouse group together versus the cells coming out of the patient group together? What we see here in high dimensional reduction is that all the diagnostic BLL samples fall together or most of them, most of the BLL relapse samples fall together, the T cell leukemias fall together. And this is regardless if they come out of a human or if they come out of a mouse. So really on the global perspective, we retain the features of these cells very well. We see that cell lines are actually very poor model for these disease, they are far out from everything. And we also see if we look at normal blood monocytes and these leukemia examples here that had a very small blast count, so cancer cell count, they all fall together with normals. That is a nice control. So now the next question we wanted to ask was, do they retain features that determine or differentiate between a primary or a sample and a relapse sample? And so we looked at what's significantly different between a primary and a kid and a relapse and a kid and then looked if the ones in the mouse on the protein level cluster together in the same way. And you see here these purple boxes are always a match patient and xenograft model sample and they all fall nicely together, which is very encouraging and these are primaries and relapses. Now there are some differences that we found though between patients and xenograft expanded ones and this is primarily in response to the immune system, which is not surprising or is actually expected because I didn't tell you that these mice are immunocompromised so that they don't reject these cells. So you would absolutely expect a cell to respond differently to anything that's coming from an immune stimulus essentially. But interestingly, we also found that we have increased cell cycle and probably more expansion, more growth of these cells in the mouse versus the patient. And we confirmed this in a pairwise manner using a number of the classical proliferative markers. Now, looking at prosensitization modification, we say essentially the same thing. They also retain specific features. We lose a bit of the patient specificity here and I can talk about this more. Now, what is interesting, what we see is that many of the proteins that don't change. So here we have protein changes, phosphorylation changes, proteolytic changes. Many of the, oh, I have to hurry up here a bit. Many of the proteins that don't change actually change on the post-sensational level. And this is probably because a change on a protein can lead in a cascade down to changes on the modification. So if you knock out an enzyme, this has a massive effect on the proteins it modifies or even a ripple down effect if we hit an inhibitor. And now we can of course also use this to study the drug effect, not only by seeing if a cell dies or doesn't die, but also by looking at which pathways like phosphorylation signaling here is now activated or deactivated in a specific patient. So to better understand possible compensation mechanisms that might occur in the relapse and prepare for that. And with that, I would like to summarize that essentially we have shown that mouse models of pediatric leukemic BL and TLL are accurate models with a few differences but by large, they are very nice models to really study a patient specific model off a pediatric leukemia. And with that, I would like to thank everybody who did that work and would also like to let you know that Vancouver is a beautiful place and there are positions available. Thank you, happy to take questions. Awesome, thank you, thank you, Phillip. So it's comforting that these Xenografts and the patient samples overall are quite similar but you do see these differences. What about using humanized mice? Would they be useful at all? So do you think it's an interaction between the mouse immune system and the Xenografts that make them diverge from the patient samples? In some of the ways it's absolutely the lack of the immune system or remaining mouse immune system that makes this difference. And using humanized mice would probably change this. It is possible to do this for pediatric leukemias. It's not really necessary and adds an extra level essentially. Our collaborators on this, Dr. Gregory is doing that for other cancers now so that we can expand other cancers because most of the others don't actually even want to grow in mice properly. So for some of these having the humanized system in place will not only be beneficial but also give you the more accurate picture, absolutely. So are there Xenografts that fail? I assume they don't always work. And if that's the case, do you think there's some interesting biology basically? There are definitely some that fail and there seems to be some indication that maybe the rate also of the growth is somewhat correlated to the aggressiveness of a particular clone, which would then in turn mean that the ones that don't grow don't grow are problematic. But it's also always a question that I would think at what time you actually get this biopsy and so you have, to imagine that you are getting this out of the bone and it's a very crowded situation usually by the time you diagnose these kids and it's very packed with blasts. And that's probably the reason why they stopped proliferating at this point when you actually biopsy them while they're still growing in the mouse where you're hitting them at a different time point. And so some are probably just not in a state to really make it through that process of biopsy and keeping them moving them into a mouse quickly enough. That's cool. So figuring that out what is actually a clonal effect and what is actually the effect of timing is probably difficult. So we have a question from Hossain. I'm trying to see like an unmute him. I asked, let me see, do I see him? Attendees, okay, okay, okay, okay. I think I'm almost there, okay. So let's see, does this work? No, I think I'm, because I'm not the organizer, I think I cannot unmute him. So I'm sorry Hossain, so I'll ask your question. So when you do this phosphor proteome analysis, so the proteome changes and then phosphorylation changes. So do you normalize phosphorylation changes by protein change or how do you do this? We don't strictly normalize it by protein change but we control for protein change. So we essentially primarily focus on the ones that are not changed on the protein level and then change on the phosphorylation level. And that's of course an important thing to look at and not get fooled by. At the end of the day, it's not always a easy answer which one is the more meaningful one because even if a phosphorylation is up and the protein is up, this might have the same effect as if only the phosphorylation were up. So if it comes down to the fact how much of that phosphorylated site is there, the way how you get there is maybe not important. So just normalizing it is, I think not always the best answer but you have to look at it. Yeah, got it, got it. Yeah, yeah. So it gets it, yeah, it's also a function of whether the two forms compete with each other in one form. So, okay. I was saying I hope I asked that correctly so I unfortunately cannot unmute you if you can maybe unmute yourself. I hope I did, he says thanks. Okay, so that was good. All right, so I'm looking at the questions. I think that's about it. Okay, so thank you again, all three of you and then Carolina, Phillip. Thank you, awesome. Thank you for sharing your work with us today and thank you to all the attendees for listening and asking questions. So we have our next slide come up. Yep, next slide is coming up. So this brings our scheduled online research talk series to a close. I actually shared the first one I think and the last one now. It's clear that lots of colleagues really do appreciate the chance both to present and listen to talks in line instead of in person. That's really I think something that will stay with us. Although I do have to say I've given a couple of online talks myself. I've been just as nervous as I am in person. So that hasn't changed. So we definitely plan to continue these. And if any of you listeners and in person talk canceled or postponed or do not have an opportunity to share it online elsewhere, we welcome for you to register your interest in being a speaker using the link on our screen right now eliveside.org wait list. And it's also posted into the chat, I think. And to learn about the next online talks, please follow elivecommunity on Twitter and sign up for our early career researcher ECR newsletter. And for updates, you can also follow elivecommunity at elivecommunity on Twitter. Final thank you. We can thank each other. And again, once more, thank you very, very much. So we're keeping this open this Zoom for 20 more minutes to give you a chance to chat directly with the speakers. We'll see who is staying with us. If you'd like to stay, please do so. Everyone else, thank you for joining us. You're welcome to sign off and we'll be ending the YouTube live stream now. All right, great. Hey, thank you very much, guys. Were you guys nervous? I...