 Yes. Hello and welcome, everyone, for the SOS-Extrusion talk. Once again, I would like to remind you that these talks are part of the Worldwide Neuroinitiative. It is a sharing platform that was created at the beginning of the COVID situation where neuroscientists from all fields and around the globe exchange about their work. So I really encourage you to check the WWN website. You will find plenty of ongoing and past talks that will probably interest you and your colleagues. So you will find all the necessary links in the description. Today, I'm glad to receive Evelyn Sanagor from Newcastle University. Evelyn obtained her PhD from the Hebrew University in Jerusalem where she studied regeneration in cat spin-on motor neurons. She then moved to the U.S. for a positive position as a NIH where she worked on development on locomotor networks in Checambrio spinal cord. She then moved to California where she worked on light response in developing turtles retina as a smith, Kettlewell I research institute in San Francisco. Now she is based in the UK as a professor in retinal neuroscience where her lab specializes in retinal plasticity in development indices. Hello, Evelyn. How are you doing today? Hello, Maxime. Well, thank you very much. Thank to you. So do I share my screen now? Okay. Here we are. Is that okay? Oh, no, not yet. There we go. Okay. I go. Yeah. Good. So thank you very much for inviting me to present my work in this forum. I think it's a fantastic initiative and allows us somehow to keep in contact in these difficult times. So today, I would like to tell you about some very novel results from my lab that show some quite compelling evidence that there might be a very strong link between early neonatal spontaneous activity and angiogenesis. So as probably most of you know, everywhere in the immatural central nervous system, there is very intense spontaneous bursting activity around the time of birth. And wherever it has been looked for, it has been found. I'm talking about the retina, spinal cord, cortex, hippocampus, etc. And of course, it takes different forms in different parts of the CNS. And there are many studies that show that this early activity is involved in neurite outgrowth, synaptogenesis, circuit formation. And very interestingly, so this activity is present only for relatively short and very well periods called critical periods that occur long before sensory experience is even possible. And but the refinement of neural circuitry continues long after it disappears. So the question is whether that early activity before experience sensory experience could have additional roles during the maturation of the CNS at early stages. Now, wherever it happens, that spontaneous activity really takes the foremost very intense spontaneous bursting intense firing, which is metabolically very demanding and neurons in general are the most metabolically demanding cells in the organism. So they are very reliant on oxygen supply from blood vessels. And it turns out that both the neural and vascular networks develop early in development in tight just a position during the short critical periods and they form what is called the neurovascular unit. And here you can see in this diagram, you can see what the neurovascular unit consists of. So it consists of neurons, blood vessels where you have endothelial cells, pericytes, basal membranes, and astrocytes, and microglia. So all these components make form the neurovascular unit that allows communication between the neurons and the blood vessels. So the question is whether these early very strong hotspots of activity could attract growing blood vessels. And here you can see spontaneous activity in the immature retina. So you have this very strong burst of activity. And the rationale for thinking that is that intense impulse activity will result in local hypoxia. That is due to high oxygen consumption that is necessary to activate ATP dependent transporters that will restore ionic concentrations to resting levels. And hypoxia triggers angiogenesis via a regulation of transcription factors called hypoxia inducible transcription factors or HIF. And in the retina, for instance, HIF-1 upregulates the production of vascular endothelial growth factor or VEGF which is released from astrocytes and that promotes angiogenesis. So as I think all of you know, in the immature retina, ganglion cells generate spontaneous bursts of activity that are correlated between neighbors. And that results in waves that propagate across the ganglion cell layer. And in mouse, which is the model I'm going to talk to you today about, so gestation lasts about 21 days, during which the retina will develop from the optic vesicles and the optic cup. And then, sorry, the spontaneous waves are going to start very shortly before birth. Initially, they are mediated by gap junctions. So these are the stage 1 waves. And then immediately after birth, they become driven by acetylcholine, by cholinergic connections from the starburst and macular cells. These are the stage 2 waves. And then, once the glutamatergic synapses from bipolar cells mature, the wave control switches from acetylcholine to glutamate. And that will occur at postnatal day 10. So these are the stage 3 glutamatergic waves. And then the spontaneous waves will disappear at eye opening around P12, so that there won't be any interactions between the spontaneous activity and visual experience. Now, during that entire period of spontaneous activity, we know that this entire period is a critical period for activity-dependent wiring of the visual system. But what I would like to talk about today is I want to investigate whether there is also activity-dependent angiogenesis during that period. So what about angiogenesis in the mouse retina? So you can see here the developmental progression of, well, first of all, of cell birth, proliferation and angiogenesis here. So birth is here. This is embryonic. This is postnatal. And before birth, the retina is completely avascular, no blood vessels in the retina. And then it starts exactly at birth. You have a blood vessel that will enter the retina through the optic nerve head, and they will propagate in the ganglion cell layer. So this is the superficial plexus. So here you can see that. So we have labeled, you can label blood vessels with isolectin B4. And here you can see that superficial vasculature as prognatal D3, 4, 5 and 6. And you see that it grows with development. And you can see the outline of the retina here. So the blood vessels, I mean, that superficial vasculature will reach the periphery by P8. And then after that, you are going to have, when the muller cells mature, they will guide blood vessels to plunge into, to start forming the deep plexus here. So that will start from P9 to about P12. And after that, finally, there will be the formation of intermediate vessels here. And of course, at the same time, you also have maturation of the retina blood barrier here. So the plexus grows with development. And then if we look at the changes in blood vessel branching, which can be done using Scholl analysis, and that's on the image J. So basically the algorithm draws concentric circles all around the vasculature, starting from the optic nerve head to the end. And then it counts the number of intersections from blood vessels on every one of these concentric circles. And this is what is plotted here for four retinas, a P3, P4, P5 and P6. So well, you can see that the radius of the plexus increases, but you see that there is more and more, there are more and more branches also. And reaching a peak value at about around two-third of the length of the plexus. And then there is a sharp drop. So angiogenesis coincides with the period of retinal waves. So that brings us to think that maybe the waves could drive angiogenesis. Going back to that hypothesis I presented to you earlier, that angiogenesis may be driven by hypoxia due to a strong neural activity. And so the superficial plexus, which starts at P1 and reaches the periphery at P8, coincides exactly with the period of the stage two cholinergic waves. And then the deep plexus begins at the end of the cholinergic waves, but also overlaps with the glutamatergic stage three waves. So there is only one paper that I know about that has looked at that. And it's a paper that was published last year in Nature Communications. This is a group from Salk or UC San Diego, I can't remember. Anyway, they have manipulated the cholinergic activity in neonatal mice, and using various approaches, genetic and pharmacological approaches, and found that it impairs the deep plexus development, and also the blood retina barrier, but not the superficial plexus that develops immediately after birth. So the question is whether there is some non cholinergic specific activity, a neural activity that may be involved in angiogenesis in the superficial plexus. And so here I need to introduce you to this very serendipitous discovery that was done by Jean de Montigny, who was a PhD in my lab. So Jean was looking at the development of starburst amacrysine mosaics. So he used a choliner acetyl transferase antibody to label the starburst amacrysine cells, and then looked at retinal hall mounds, so he could calculate the density of the cells at different stages of development. And then to his amazement, to our amazement, he found these cellular clusters, very large cells that seem to label for choliner acetyl transferase, which is shown in green here. And so these cells form an annulus around the optic nerve head, and that annulus expands with development. So you can see a P2, P3, P5, and they reach the periphery by a P7 or P8. And here you can see the progression of these clusters from center to periphery. So just by calculating the ratio of the distance from the cluster to the optic nerve head to the distance between the optic nerve head and the periphery, and you see that as development progresses, the clusters are moving further and further towards the periphery. And the biggest, the most pronounced changes occur between P3 and P4 and then P6, and then it stabilizes. And from P8, the cluster cells begin to disintegrate and they completely disappear by P10. You can't see them anymore. And so they coincide with the end of the cholinergic waves. So of course, our first question when we initially discovered these cells, what are these cells? So based on immunolabelling, we initially thought that there were cholinergic ganglion cells because they showed a signal with a cholina-stheil transferase and also with our BPMS, which is the antibody used to label ganglion cells. But then much more recently, and by more recently, I mean, because Jean initially discovered these cells in January 2019, and then this past March, about two, three weeks before we went into lockdown, we realized that the cells are actually autofluorescent. And this was discovered by Vidya Krishnamurti in the lab. So that means that all the immunolabelling that we had done really is kind of meaningless. And the identity remains elusive. So what we know is that they are in the ganglion cell layer. So here, well, here we use the chat antibody, although they de-fluorescent many different wavelengths. Even if you don't use any antibody, you will see them. And I will show you that on another picture later. But so you see the cluster cells here in the ganglion cell layer together with the starburst amachine cells in green and then the ganglion cells in red. But then if you go further down to the inner nuclear layer, you see only starburst amachine cells. So they are in the ganglion cell layer. And you can see that here, that's the same area in this little animation. And since initially we thought that they were cholinergic, we wanted to see whether they express the vesicular acetylcholine transporter, vasht, which you can see here. You can see very nicely the plexus formed by the starburst amachine cells using vasht. And I mean, so you can see that the cluster cells seem to make very intimate contacts with the starburst amachine cells. But so the only certainty we have from this is that they are in the ganglion cell layer. And they seem to make contact with other cells. But we don't know what they are. So are they a subtype of transient ganglion cells? We don't know that yet. So the cluster cells and the superficial vascular plexus expand simultaneously during the period of stage 2 cholinergic wave. So is there a link between them? And yes, there is a very strong link. It turns out that the clusters localize, co-localize precisely with the edge of the vasculature. So here you can see the clusters. And then here you can see the blood vessels in the same retina. And here's the overlap. So they are just underneath the edge of the vasculature. Although we never see them ahead of the vasculature. And that's actually could be an important point for discussion later. And not only that, but they expand in synchrony. So as both of them move to the periphery at the same rate so that the clusters are always seen exactly under the outer edge of the growing vasculature. And so what about the stage 2 waves? I mean all the studies that have been done on waves by many people suggest that the the cholinergic waves are initiated in random locations. And they've been also some nice modeling about that. But all the experimental data that shows that has been using only limited areas of the retina to record the waves either with calcium imaging or with multi-electrode arrays that do not cover the entire retina. But so maybe the origins are not random. Maybe they also follow some centrifugal developmental pattern. And so we record waves from the ganglion cell layer using a large-scale high-density multi-electrode array. And some of you who are listening to this talk have that same system. So this is a system from Cribrin. So that array has 4096 electrodes. And it really provides a pan-retinal perspective of network activity at fantastic spatial temporal resolution. And here you can see a corner of that array. So here you can see individual electrodes, where here the electrode pitch is 42 micron. And on top of it, there is a retina with 51 expressing ganglion cells, expressing some green fluorescent protein. So you can visualize the cell. So you can see that these electrodes really give near cellular resolution. And so we use this system to record retinal waves. But then the thing is that with that array, with a 42 micrometer pitch, you cannot cover the whole retina. But Cribrin has other arrays with an electrode pitch of 81 micrometer. And so the spatial resolution is lower. But the advantage is that the entire retina sits on the active electrode. And you can see that here this is a P6 retina. And you can see here the outline of the retina lying on the array, ganglion cell layer facing down on the electrode. And you can see the waves that were initiated and how they propagate across the array. So that allows us to see all the waves that are generated in the retina. And so using that approach, well, we're using an analysis that we published in 2014 in a paper where we looked at the ontogeny of retinal waves, but published in the Journal of Physiology six years ago. And so we can quantify a lot of wave parameters, and we can also localize their origins. So this is what we used for this project here, using the big array, the array that encompasses the whole retina. And so we can measure, we can localize the origin of every wave. So we, we call for half an hour or an hour, and then we process the data. And when you can see here is a retina just post-recording photographed on the MEA. And then after that, of course, we do the analysis. And then once we have the wave origin, which are the little green dots here, we can overlay the two images. So we can see exactly where the waves originate from. And then to quantify where the wave origins are, in terms of center versus periphery, we draw an ellipse around the outer edge of the retina, and then a concentric one, which is half size. And then we count how many wave origin, which are these little dots here in the central part, with that the gray mask, as opposed to the peripheral part. And we take only the areas that are covered by the retina, because of course we have tails in the retina to flatten it. So there is, so we use only the part that are covered by the retina. So, and then we can calculate the ratio of wave origins between periphery and center. And that's what you can see here. And we have done that from P2 to P13, to eye-opening. And in blue, you see the stage two cholinergic waves. Green is the glutamatergic waves. And what you see here is the periphery versus center ratio. And as you can see, there is a very strong developmental change. And again, with a very big change occurring between P3 and P4. And then around that time, the ratio reaches values that are above one, meaning that there are more waves originating from the periphery than from the center. And then later on, that ratio really disappears. And it's certainly not there anymore during the glutamatergic waves, which are a completely different beast. I mean, there are small hotspots of activity that tile the retina. So that suggests that there is a very strong trend from center to periphery. And just to verify that, more recently, what Jean did was a Monte Carlo randomization of the wave origins for every retina, repeating that 10 times. And then we calculate the ratio. And by doing that, we completely lose that trend. On average, we always find a ratio that is around one, meaning that the origins are equally distributed between center and periphery. So in real life, we find that the wave origins are not random at all. They really follow a center to periphery pattern, just like the cell clusters and like the blood vessels. And so just to summarize that, so we see synchronized centrifugal progression of the superficial vasculature, the clusters, and the waves. And then I just, when I was preparing this talk, I just went back to data that we published in our 2014 paper, where we looked at the size of the waves, measure either at the center of activity trajectory length or the area of the waves. We find that, so these are postnatal days here. And we find that, again, they reach peak size around P6, P7, and then they decrease in size. And amazingly, you can see that in all cases, you see that there is a strong change occurring between P3 and P4. And then after that, there is a decrease. And of course, after P7, and of course, in the case of the clusters, it's not that you have a decrease, but the clusters simply disappear. They disintegrate. So could the cluster cells trigger the wave onset? And so this is when we started a very nice collaboration with Tim Gholish's lab. So this analysis was done by Fernando Rosenblit in Tim's lab. So what Fernando did is electrical imaging during the waves. So basically, it did spike trigger averaging of the electrical signals in all the neighboring channels. And taking into consideration only spikes that are convincingly part of waves, because there is, of course, also some random activity. So here you can see one channel where, I mean, this is the spike that is used for STA. And this is the spike here. And then in quite a few channels, like this one here, you see that on some adjacent channels, there is some dipole activity. What does it mean? It means that there is a positive deflection that has a z-score above 5 that occurs before a smaller negative deflection. And so all the cells, I mean, the electrodes where this was detected in this particular retina are indicated with this green mask here. So could they reflect activity in the cluster cells? I mean, they are smaller. The signals are smaller. So they could represent some graded potentials. I mean, not a spike. We don't know exactly what happened. And here you can see a time lapse movie of the same. So you see that this dipole activity occurs at the same time as the spikes. And you can see it very clearly. So this is the time for the STA here, time zero. But many of the channels do not show that. They show only negative deflections of smaller amplitude around the channel used for STA. So these are presumably just reflecting propagating spikes during a wave. So how do these regions with positive dipoles relate to wave origins? Like I've shown you a couple of slides back. So this is shown here for four retinas, two P4 and two P5 retinas. For each case, we recorded one retina with a large MEA, which the advantage is that we can record from the entire retina. But the spatial resolution is not as good as the one with the smaller pitch that you have here. And so in green, you see the wave origins that were, the wave that occurred during that one hour recording. And then what you see overlapped here, overlaid on the wave origins is the pixels, the XY locations where there was a significant positive deflection preceding the time of a negative deflection. So in other words, these significant dipole signals. And as you can see, there is a pretty good correlation between the location of the wave origins and these dipoles here. And it's probably easier to see in the case in the retina that were recorded with the higher spatial resolution like this one and this one here. On the other hand, if now they plot all the projections over all electrodes, regardless of whether there is a positive dipole, you see that the creation is much more blurred here. There are many more electrodes that show a signal here. So it may be that this dipole does reflect activity in the cluster cells. Now, if the cluster cells trigger spontaneous activity and that induces hypoxia. So we would expect the hypoxic conditions to be strong in the area of the clusters near the growing tips of the blood vessels. So what we have done here is, and I hope you can see it because I'm not sure how good the video is. I can certainly see it very well on my screen here at home. We have used an antibody against the hypoxia-inducing factor 1-alpha and it's the oldest green specks here. And so here you see cluster cells and the blood vessels in magenta and then all the green specks here is the HIF 1-alpha. And you see that there is a very strong expression in the vicinity of the cluster cells. And the other thing that we can see here is that there is a very pronounced decrease in gradients from these areas towards the periphery to areas of the retina where there is no vasculature yet. So that does suggest that there is hypoxia around the cluster cells and at the growing tip of the blood vessels. So the question is, so if activity promotes angiogenesis and the clusters are basically where the activity originates from, so you would expect angiogenesis to be more pronounced in the area of the clusters. And this is a very nice work that was done by Cote Nitton and M.R.E. student who was in the lab this year. So what Cote Nitton did, she did some shawl analysis in retinal segments going from the optic nerve head to the periphery. And she looked at the number of intersections on the concentric circles but only in the outer half of this segment because that's where the clusters are. The clusters are not near the optic nerve head. And she has done that in areas with clusters and areas without clusters at the same eccentricity. And what she found is that at every age she has done that. She found that the areas where there were clusters or C plus always had a higher vascular density meaning more intersections of blood vessels on the shawl circles than in areas at the same eccentricity but without clusters. So this is true at all ages where it was done. And here you can see the pool data. So there is more angiogenesis under the clusters. So they disappear. These cells completely vanish by p10. So how do they disappear? Well it turns out that they undergo microglial phagocytosis. So here you can see cluster cells in, I mean here they are actually used without any antibody. This is just like autofluorescence and overlays with labelling for microglia using the I by 1 antibody. So microglia are the resident macrophages of the CNS. And so this is very different from ganglion cells because ganglion cells in the nonadalretina the majority of ganglion cells die of program cell death or apoptosis. So 70 to 80 percent of ganglion cells disappear. But we haven't seen any sign of apoptosis. We should probably do more of that. We haven't looked at that very much in detail so far. So microglia are extremely dynamic cells and they change shape in different conditions. So this is a resting microglia here that you can see. So they have this elongated shape with a few processors but once they're activated and the engulf cells they become much rounder and with shorter processors. And you can see that here. Now interestingly and you see that most of the cluster cells are surrounded by the microglia. And interestingly we find that the intensity of the fluorescence in the cluster cells appears anti-correlated to the level of phagocytosis. So here for instance you see cells that are very bright but you don't see any sign of microglia around them. While some other cells here have a much fainter signal and are completely surrounded by microglia cells and you can see that everywhere. And where it looks like the cluster cells actually attract microglia. And here in these two micrographs here you can see that very nicely. I mean here so the microglia are labeled in orange here blood vessels in magenta and then in green whiteish we have the cluster cells and you can already see that there are more microglia near the cluster cells. And again Courtney has quantified that and it turns out that the density of microglia is significantly higher in areas where there are clusters than in areas of the same eccentricity but without clusters. And not only that but she found this beautiful positive correlation between the density of the cluster cell because not all clusters have the same density. Here you see for instance a cluster with a high density and here one with a lower density. So she found that there is a nice positive correlation between the density of the cluster cells and the density of microglia. And she also found that the only place where the microglial density significantly higher is in the cluster area. So the way she did that was by growing axes going from the optic nerve head to the periphery and dividing those into four areas. One containing a cluster and one ahead of the cluster further in the periphery, one just behind the cluster and one near the optic nerve head and she has done that in areas with clusters or areas without clusters. And this is what you can see here. And she found that only at the clusters there are significantly more microglia in areas with clusters than in areas without clusters at the same eccentricity. Everywhere else there is no difference. So again it looks like the cluster cells are really attracting microglia in their vicinity. So finally I wanted to know I said well maybe these cluster cells are some specialized ganglion cells. So what would chronic changes in neonatal activity in the ganglion cell layer if we can change that chronically how would it affect angiogenesis and the clusters? So luckily we have in the lab a mouse model that is used for another project that was just perfect for that. And this is a model where the majority of ganglion cells, so these are BRM3B expressing ganglion cell which is about 70 to 80 percent of the entire ganglion cell population express excitatory designer receptor exclusively activated by designer drugs or dreads. So these excitatory dreads are dread GQ. So this is pharmacogenetics. And so these dread GQs are activated when a designer drug binds to them and the one we use close up in an oxide or CNO which has no endogenous receptors in the retina. And so when dread expressing ganglion cells sends CNO they will depolarize and fire like crazy. And the nice thing is that the effect is prolonged and CNO can be applied systemically for chronic studies. And here you can see, so this is done in vitro, here you can see waves in a P6 retina. So here you see the raster plots over all active electrodes using the big MEA. So here each one of those represents a wave and here you can see the average firing rate during each wave. And then if we add CNO to the same retina we find that the wave frequency increases and there is a lot more background activity also. So there is a lot more activity in the ganglion cell layer. So what we have done is inject CNO chronically in pubs from birth to P6 and then sacrifice them at P6 and look at the blood vessels and the clusters. And again these are very preliminary results. Here we are. So here we can see, so these are both P6 retinas from the same litter. The top is one where there were dreads in the BRNPB expressing ganglion cell and in this one down there weren't. So this is a controlled litter made, no dreads. And as you can see, and you can see the outline of the retina here, as you can see the plexus is, the growth of the plexus is somehow aborted. It doesn't grow as far as it should when there is more activity. And you see that the clusters are still remaining under the edge of the vasculature, although they seem to be much more spread out. And so there is definitely an effect on the vasculature and on the clusters when you increase activity in the ganglion cell layer, which will of course increase the amount of hypoxia. So not only the plexus is shorter, but there is also more proliferation of blood vessels, as you can see here. So here you have, so these are four retinas from the same, from four, sorry, from four pubs from the same litter. So two that were dread positive and two dread negative. So you can see that the plexus is shorter and you, but you see that the peak branching is higher in the dread positive. So shorter plexus and more branches. And you can see that here as well, just putting all the maximum number of intersections is higher in the dread positive retinas. So this was done by Courtney and by Dia, another M.H.S. student in the lab. So to summarize, so we have discovered the transient population of cellular clusters in the neonatal retina. And so far, so the current evidence suggests that, first of all, that these cells may be electrically active and trigger the spontaneous waves, that they may be a specialized transient population of ganglion cells, that they may guide angiogenesis in the superficial vascular plexus by generating hypoxic conditions. And that they may send an eat-me signal to microglia once growing blood vessels reach them. Because if they really generate very intense activity, I mean, this is metabolically very expensive. So if the role is to attract blood vessels, once the blood vessels have reached them, you don't need them anymore. So you want to get rid of them. And then chronic increasing activity levels in the ganglion cell layer during the first postnatal week hinders the vasculature development, but at the same time promotes angiogenesis by increasing the number of branches in these shorter plexuses. And actually, there is a published evidence that shows that the retinal superficial vascular does not develop at all in the absence of ganglion cells. So this has been shown in the MAT5 knockout retina, where there are, I think, less than 5% of ganglion cells. And there is no vasculature. These retinas remain avascular. Also in the BRN-3B knockout, when you knock out BRN-3B, you have only 30% or 30%, 20% or 30% of ganglion cells. And then in a disorder in humans, it's called anencephaly, which is a developmental defect of the cerebral cortex that results from the failure of the closure of the anterior neural tube. Well, in neonates, it turns out that, well, they have no ganglion cells and no vasculature also. So these cluster cells may provide the first evidence that there are specialized transient neuronal populations in the CNS that guide angiogenesis through neural activity. And of course, I think what I've presented you probably opens more questions than answers questions. So first of all, we still don't know what is the cellular identity of these cells. And we're going to have to do a single cell sequencing to find that out. We don't know why the autofluoress. So suggestions are that it could be lipofusine, which is expressed by cells and it can be a sign of stress. And it may be that these cluster cells are initially not fluorescent and that they become when they are, let's say, most active and that they become fluorescent only when they are contacted by the blood vessels and the microglia. We don't know that. And the only way we will be able to tell that is by combining imaging together with activity recordings. And we, of course, we need also to find more direct evidence for involvement in retinal waves. And again, that will have to be done using imaging. And then we need to understand the cell signaling mechanism that link the neural activity, angiogenesis, and the disappearance of these cells. And then finally, and very importantly, these cells might be linked to a disorder called retinopathy of prematurity, which is a devastating disorder that often can lead to blindness. And it's a disorder that is due to abnormally high exposure to oxygen in premature infants. So it would be very interesting to see whether there is a link between these cells and what they do to angiogenesis and the retinopathy of prematurity. And if we understand better what's happening, this could eventually lead to a cure, which would be fantastic, of course. So I just want to thank the great people who have been involved in this project. So Jean de Montigny discovered the clusters and he did all the initial labelling, which turned out not to be very useful, unfortunately, but we didn't know. And he was also involved in the recordings. And then I was very lucky to have three master students in the lab this year, mostly working on image analysis from home during the lockdown. So Courtney, Dia, and Dimitris Boussoulas-Certedakis. And then Vidya Krishnamurti, postdoc in the lab, was also involved in the project. And then I would like to thank Robert Jackson and Gerrit Hilgen in the lab for some technical help. And then very interesting discussions with a local colleague, Gavin Florian Nindalako. And then with the ophthalmologist, mostly Roxane Illier, who is a local ophthalmologist who specializes in retinopathy of prematurity. And she's very excited about this project. And then the bioimaging unit, thanks to all the beautiful kits that we were able to take all these nice pictures. And then in Göttingen, Fernando Rosenbrit and Tim Gholish for all the fantastic work they did with the electrical imaging. And then I would like to give a very special thanks to Jeremy Kay from Duke, because I presented this data exactly a year ago at SFN in Chicago. But I knew nothing about blood vessels at that time. I just presented the clusters and the correlations with the waves. And Jeremy came to see my poster. And we had a very interesting discussion. And he's the one who told me, well, you know, it's very interesting because the angiogenesis occurs at exactly the same time. It seems to follow exactly the same time course. And I had no idea because I never had any interest in blood vessels, I have to admit. And as soon as I got back home, I started reading and feeling the immense gaps in my knowledge. And I think I've learned a lot in the past year about that. And it's really thanks to Jeremy that this project has progressed the way it has. So thank you, Jeremy. And then I would like to thank the funders. New Castle University was funded all the students and the BBSRC. So thank you very much for your attention. And that's it. So I'm ready to take questions. Well, thank you. Thank you, Evelyn. Thank you very much. That was a beautiful talk with lots of beautiful pictures and data. I'm so very excited to be leading the questions now. Maxim has become in this post. Hence, you see me. So there's quite a few questions. And I think why don't we go through them? I actually have one myself, but I'm going to skip that till the end. So the first question is from Marvin as I fight here in Brighton, is the activity in the regions into which the wave progress reduced directly after a wave has happened or passed through that area? So that again that where the... So the ideas you've got a wave and he's asking is if the Garmien cells where the wave was passing through, if they are suppressed after the wave has gone through? You mean if they're going to a refractory period? Yeah, something like that. No, no, not really, but well for a long quite a time. I mean, I think the reason he's asking that is because in our work we see that sometimes. Yeah, it's not very strong. And then in the pharmacogenetics, when we use the excitatory dread, I mean, we see activity all the time. And so, well, I know that many people believe that there is a very strong refractory period in the waves. I have to admit, we have not been able to see that in our recordings. And I know I'm not very popular when I say that, but you know, this is what we see. And yeah, okay, oh, yes. Okay, so I think that certainly answers that. So we've got second question by Serena Richet. Okay, I'm not going to try to pronounce that name. Sorry Serena. Did you try to see if they don't disappear if you have a depletion of microglia cells? By they, I think she means the blood vessels. So if there is depletion of blood vessels of microglia? No, if you have a microglia depletion, if the blood vessels then don't disappear. I don't know. I don't know yet. You know, I mean, this is, there are many things I don't know yet because this is all, it's all very preliminary. And obviously, now I'm in the process of trying to get funding to take this project much further. So there are many, many unanswered questions. Okay, great. So then we've got one from Marla Fela. Jeremy Kaye published that astrocytes follow RGC and RGC's and guide on genogenesis. Do you know what astrocytes look like around the cluster? Well, yes, we have labelled retinas for astrocytes and we don't really see any, I mean, the astrocytes are there before everything else. You know, I mean, they form like, you know, like runways. They penetrate, they invade the retina long before the clusters and the blood vessels. And we haven't been able to see anything compelling, you know, like anything special about the cluster areas in the astrocytes. Okay. But you know, certainly to be followed, but you know, at first, no, we don't see anything. Oh, great. So actually, the next one is a comment by Steve Masse on my question. So I'm going to ask my question and then read out his comment as well. So my question was, if you've compared your results from the mouse with other species, well, A in primate, of course, which has obvious applications, but then the other one is in species where there are no blood vessels in the eye, like in birds. And then Steve points out that they also don't have them in the rabbit, which of course is a very good point. Well, I don't know. I mean, you know, so I think, well, you have cheek embryos, right? So maybe you could, all you have to do is, you know, just look at retinal hormones and you don't even need to use any antibodies or anything. I mean, just see if there are autofluorescent cells. I don't know. But, you know, well, I'm not saying that these cluster cells might be the only way to drive the spontaneous activity. And, you know, obviously, the mechanisms are different in different species. And, you know, like, I mean, I've worked on the turtle retina for a long time and, you know, and so these are vascular as well. And, well, there, I mean, you know, there is complete overlap between the cholinergic and glutamatergic activity. And, you know, it's, I don't know. So this may be something that is really unique to mammals. But I cannot tell you, but it will be very interesting to tell that. But I think that, you know, what I find really amazing is that this spontaneous activity is really present everywhere in the developing CNS, taking different forms. And it's present during such, you know, it has very specific patterns in different areas. And it's present only during very limited periods. And we know that this is the period during which the blood vessels grow as well. So it may well be, the thing is that it's more difficult to study angiogenesis in other parts of the CNS because vasculature grows in three dimensions. And so it's very, it's difficult to visualize. And until I became interested in retinal vasculature, you know, I didn't realize that is actually a classic model to study angiogenesis because, you know, everything is plain now, and it's easy to study. So many people use the retinal as a model. So very lucky. But I think that, you know, there may be, I mean, so for instance, in the cortex, there are the sub plates, neurons, which are transients, and they are there. And we know that they participate in spontaneous activity that are extremely active and they really guide CNS formation, but then they disappear. So, you know, they may also be there to guide angiogenesis. I don't know. So it would be very nice to try to expand that to other systems also. So so many questions. It's only the tip of the iceberg, I think. I like the fact that everything can be somehow linked, you know, like everything happens in constant. It makes so much sense, you know, and you try to make use of what you have for different purposes. Yeah, absolutely. So I've got one more question by Hugo Caligaro. And after that, I think we will close this official part and we will invite everyone into the Zoom who wants to hang out afterwards. Marla is already there. And yes, so let's just ask Hugo's question. So what he's asking is recent papers from Dr. Richard Lang suggest a role of metanopsin and neuropsin both expressed in the RGCs in angiogenesis and hyeloid regression. Could the process you describe be light dependent? Yes, there is a very nice paper in nature. I think that I suppose that the people they are referring to. And yes, and so it looks like, but that happens mostly before birth. It looks like, well, there is, you know, light that is coming to the womb from outside. And that light is sufficient to activate the IPRGCs. And, you know, so they will generate activity. So, and oh, yes, because of course, there are some, what I haven't told you is that there are some, you know, like embryonic blood vessels, the hyeloids, blood vessels, and these retract from the retina at birth and give space for the what is with Matthew into the blood, the material vasculature. And so it looks like the IPRGCs are very much linked to that, to the, you know, to the early blood vessels. Now, whether they have an impact on what I've told you about, maybe that would be very interesting. It would be interesting to maybe to use a melanopsin knockout mouse and see whether, you know, there are clusters and what happens to, you know, and how all these things are linked. I don't know. But that could be easily done, I guess. Yeah. Okay, well, thank you very much again, Evelyn. So, as I said, we will now close this official part. Stay, stay. And so we will, so anyone who wants to can just jump into this zoom now, and hopefully we'll close the actual video recording so that what we say afterwards will not be recorded. That's the hope. Okay, so let's do that. So everyone, please jump into the zoom if you feel so inclined. I see something from Michael O'Donovan. Oh, a super presentation. Evelyn, very interesting and provocative ideas. Do you see that? Do you know whether the metabolites of CNO are active on the vasculature? I have no idea. The hind limb afference