Scientist Stories: Barclay Morrision, Neuronal Dysfunction in Brain Slice Cultures after Concussion

Good morning, everyone. Thank you for coming to the IBB seminar series this morning. Today I am delighted to introduce Barkley Morrison, who is an expert in traumatic brain injury, and I've gotten to know him over the last year or two via a bunch of Zoom calls with BM Seed, which is a company that we both are collaborating with and have collaborated with in the past, which is an in vitro approach for studying traumatic brain injury. But Barkley is a professor of biomedical engineering and the director of the Neurotrauma and Repair Laboratory at Columbia. He's also the vice dean of undergraduate programs, which sounds like a very cool title. His research focuses on the biomechanics of traumatic brain injury at both the tissue levels, at the biomechanics of tissue injury, as well as some of my favorite topics, the cellular pathways that govern the injury itself. He currently serves as a council member and president of the International Research Council on Biomechanics of Injury. He's a board member of Football Research Incorporated and is on the editorial board of Journal of, uh, Journal of Neurotrauma and Brain Multiphysics. He received his bachelor's in biomedical engineering from Johns Hopkins and his PhD, uh, from the University of Pennsylvania. Did his postdoc at Southampton University in the UK and started at Columbia in 2003. He is the recipient of many awards, including the Rickard Scalack Best Paper Award from the ASME and was recently given the award for the Great Teacher Award from the Society of Columbia Graduates. So that's very cool. Without further ado, I'm going to give it over to Barclay, and we look forward to your talk.

Great. Thank you, Levi. And I want to start by saying thanks to Levi and Michelle, both for inviting me here and giving me this opportunity to share some of our work with all of you. Happy to keep this more on the informal side. Raise your hands, interrupt, ask questions throughout the talk. Happy to take the questions. Gives me a chance to collect my thoughts. So as Levi said, I study traumatic brain injury. My lab studies TBI. And just like to start with a statistics slide to put TBI into some context. You've got some great experts here, so maybe you are familiar with some of this already. But essentially, TBI, latest stats maybe put this at over 2 million TBIs happen every year. Some of those result in emergency room visits, hospitalizations, but really the big numbers here are the deaths and disabilities. So every year about 85,000 people sustain a permanent long-term or permanent disability due to a head injury, and it also results in about 65,000 deaths every year. All of these permanent disabilities add up, and so over 5 million Americans currently have a long-term disability because of TBI. And the other really devastating thing about traumatic brain injury is that it, it impacts people really in the prime of their lives. Um, you know, over 60-70% of these injuries occur in the age groups from 15 to 45-year-olds. So these lifelong disabilities affect somebody for, you know, 30, 40, 50 years depending upon when they got the injury. And so CDC has estimated annual costs of TBI upwards of $75, $76 million. And then this is also— TBI is a major factor in, in the U.S. military, not Perhaps for the reasons you might expect, most of these injuries are mild. These are some statistics from the DOD. 80% TBI in the military is, is mild, and about half of those are due to exposure to blast. So, so not shrapnel. Helmets do a pretty good job of stopping shrapnel, but quite a few are due to blast where that helmet isn't designed to stop that shockwave. So what we try to do is, in our lab, is try to reduce those socioeconomic costs associated with TBI. And we've taken sort of a two-pronged approach, one trying to understand the biomechanics of the injury so that we can help others design better safety systems to prevent those injuries from occurring in the first place. We look at things like tolerance criteria How much deformation or how much pressure can the brain withstand before something goes wrong? We also look at mechanical properties for computational modeling. You need to know those appropriate mechanical properties for those models to be able to accurately predict the tissue response. But then we also try to understand the biological cascades that are set in motion by those mechanical stimuli that lead to the injury. Because a lot of the pathology develops over time. The initiating event maybe lasts 40 milliseconds, 100 milliseconds at the longest. And yet that sets in motion extended biological cascades that may take hours, minutes, hours, days, or even longer to unfold. And if we can understand what that pathology is, we can potentially come in with new therapies, drugs, identify the targets, and stop that progression, all with trying to reduce these socioeconomic costs. And then, as Levi mentioned, we also develop new tools. And some of this technology has found its way into this company, BM Seed, which makes these stretchable microelectrode arrays. But I won't be talking about those today. So what happens mechanically in a fall, a car accident, an impact? And we can go back to the '70s, this sort of well-cited paper by Nahum et al., where they took a postmortem human subject head, instrumented with pressure transducers, accelerometers, impacted it with an impactor, and measured what happened. And two things happened. First, there's a pressure transient that propagates through the brain tissue. And depending— the pressure transient is location dependent relative to the impact site. This happened to be a frontal impact. So the pressures are highest up on the frontal side and actually become negative on the posterior side, which is the genesis of a coup-contracoup lesion. Um, pet peeve of mine is engineers, biomechanics. Um, your brain doesn't bounce around inside your skull. And a head injury, you don't get a countercoup injury because your brain, you hit the front, your brain bounces off the front side of your skull and hits the backside of your skull. And you get an injury on the backside. You ask a neurosurgeon, he'll tell you that it's the head bouncing around. Sanjay Gupta on CNN's got videos of this. Um, it's— that's not what happens. What happens on the backside is right here, you get this negative pressure on the backside, and that leads to potentially cavitation, but at least deformation on the backside. That is where you get your contrecoup injury from. So next time you're talking to your neurosurgeon friend, Tell him, ask him if he thinks the brain bounces around inside the skull. Unless Michelle's talked to him, he'll undoubtedly say it does. Anyway, so this pressure transient, it's proportional to the linear head acceleration, extremely linear, you can see here. And it's a very fast transient, maybe lasts 5 milliseconds. And it's in fact passes before the head even starts to move. And then what else happens after that pressure transient goes through and the skull begins to move? We also get brain deformation shear. Brain is basically incompressible. It's encased in a rigid container, your skull. Really the only way for it to deform is through shear. And that shear is induced through rotational acceleration. And back in the '70s, '80s, Omaya, Generale, Thibault really pushed this idea that it was the rotational acceleration that led to— in their studies, they were looking for trying to produce something that they called concussion. At that time, concussion was defined as traumatic unconsciousness, which is no longer how we define concussion today. Concussion doesn't require unconsciousness, no loss of consciousness, and is really just a change in mental status, mental function. So for their definition, they were able to induce concussion, but it was really diffuse axonal injury, which we would recognize as a much more severe injury today. And so this started this controversy, this split in the field, where the linear acceleration and the pressure transient that was generated by that linear acceleration was really disregarded. And everybody since then has really been focusing on this rotational acceleration and, and shear of the brain tissue. And so we've sort of come back to revisit this a bit since we have been studying blast traumatic brain injury pressure waves as they transit through the brain and We've found that these two injury mechanisms, these biomechanical mechanisms lead to different pathologies where, um, we find that the linear accelerations, which would give rise to these pressure transients, lead to something more mild, like a concussion. And the tissue deformation arising from the rotational accelerations lead to something more severe where we get cell death. And so we've been looking at these two injury mechanisms, but what I'm going to talk about today is mostly about the pressure transient and using that as a model for concussion and trying to understand how those biomechanics translate into pathology. And we've— in our studies, there's lots of ways to try to study this phenomenon. You can use sports, you know, the the field of play as a living lab and use animal models. We've focused on tissue culture or in vitro models of TBI, and they give us some advantages, but you also need to understand their limitations. So in terms of advantages, in vitro models allow us to control the biomechanics of the injury very precisely. We can apply either a deformation to the tissue or a pressure transient and separate those two injury causations, look at how they— what kind of pathologies that they introduce separately. If you design the model correctly, it can be very reproducible, and it can also give you a way to independently verify what you're doing to the tissue. You can also test CNS components in isolation. You could injure neurons, you could injure microglia, you could put them together, injure them together, and see how the the tissue types interact. Obviously, in vitro approach has caveats. The biggest probably is whatever that tissue preparation is that you're studying, that you're injuring, you need to understand those limitations. Does it really represent the in vivo tissue? And if it doesn't, how are those limitations affecting the biological response? Obviously, there's no systemic contribution to the injury response. You could add back in circulating cells or systemic cytokines, and that could be another advantage where you control that addition. But without adding that, you have to keep this in mind that it's an isolated system. And then whatever you do do mechanically to the tissue really needs to represent the in vivo biomechanics. Potentially what you're studying has no real correlate to the in vivo situation. So our tissue culture system of interest is an organotypic brain slice culture. Here we take the region of interest out of the brain of a rat pup, in this case, either hippocampus over here or cortex. We cut that into thin sections about 400 microns thick, and then we culture that whole slice of tissue. So all of the different cell types are represented. We don't disrupt the anatomy and the interconnections between all of those cells. So we think it represents a pretty good ex vivo representation of the brain. Here's an acute hippocampal slice, and then these are all taken 2 weeks after the culturing. So they maintain the structure of the hippocampus quite well. It's very recognizable. Pyramidal cell layer is all tight. Granule cells are there. They're all wired up. This is MAP2 staining dendrites, interconnections between the cells. And then it also remains a couple hundred microns thick in vitro. So it's really this three-dimensional living construct of the brain. And then as I said, we can also culture other parts of the brain cortex. We've got a little bit of experience with the cerebellum as well. Yeah, please, please.

Sure, sure.

Yeah, yeah. So the question is, with concussion being mild and patients getting mild concussions, how do we know whether diffuse axonal injury happens in those patients because we can't get their brains to examine them? So there are, there have been studies where people do get a head injury, it's a mild head injury, and then for whatever reason they do die of something else afterwards. So there actually have been some human studies. And in the more mild, the diffuse axonal injury really isn't present. There's some high-resolution MRI that can be done to begin to pick up very small lesions. And you don't see those lesions in the more mild cases. We also then have animal models, so large animal models like pig model here, pig model up at Penn, where you can also reproduce those biomechanics in a larger brain that's more similar to human, and then sacrifice the animal and take the brain out. So there's both clinical as well as experimental evidence there. Any other questions? Okay, great. So one of the advantages of, or I guess one of the techniques that we've used more recently in studying the outcome after these more mild concussion-like injuries is studying the electrophysiological activity of the neurons in our slice. So here's a picture up here of one of our slice cultures, and we've got it sitting on a 60 channel microelectrode array. It's an 8x8 grid, but it doesn't have electrodes in the corners. And here we can interrogate and measure the function of the entire hippocampal circuit. The reason why we focus on hippocampus is because it's very vulnerable in TBI. It's essential for learning and memory, and those learning and memory deficits after head injury are likely due to damage to the hippocampal circuits. So we can record throughout this whole structure. We can stimulate different pairs of electrodes and we can measure, you know, network activity. We can also induce plasticity in this slice, something called long-term potentiation, which is thought to be the cellular molecular correlate of learning and memory. We do that by stimulating the Schaeffer collateral that connects— that's the output from CA3 that connects to the CA1. We stimulate that with a high-frequency theta burst and then wait some time, and the response in the CA1 is enhanced by that LTP. So over here in blue is a baseline evoked recording, and then red is after inducing LTP. And then this is just a time course of that whole process. We do a baseline theta stimulation, this high-frequency stimulation. And then as the circuit recovers from that, the response is now increased. And then we can quantify that as a percent change above baseline as this ability to change, as ability to learn really in a dish. So we have two injury models. We have a tissue, tissue strain model where we grow the slices on a deformable substrate, pull that substrate down over a hollow cylindrical indenter, and that causes the substrate to stretch and the tissue on the substrate stretches as well. We pull the well down over the indenter so that the membrane stays in the same plane of focus and we take high-speed video. We only do one injury. The video is just looping over. And then we can quantify what the actual tissue deformation was. We can go up to quite severe strains, 50% biaxial strain, and at high rates, 50 reciprocal seconds. This is all under feedback control, so we can do those, dial in exactly the injury that we want. And then for applying this pressure pulse, the model we have uses a shock tube, so we, we High-pressure gas fills up the driver section. It's separated from the driven section by a frangible membrane, Mylar membrane, and that membrane bursts at a certain pressure, and all of that compressed gas comes out all at once and generates a shockwave down this tube. We don't just blast our cultures in air. We actually put them in this fluid-filled receiver to simulate the surrounding, brain tissue and skull complex. It also prevents reflections off the floor. It has some baffles in it. And so with this, we're able to either change the driver gas, the size of the driver, or the thickness of the membrane, and we get a wide range of pressure pulses. And we can quantify those pressure pulses by the peak pressure— where's my cursor— peak pressure, the duration of that pulse, and then the area under that curve gives us the impulse. And then doing some quick dirty back calculations to the pressures, intracranial pressures in Nahum's study, we can relate these to linear accelerations. And we can go quite high pressures, which probably at some point this doesn't linearly— this is more interpolation at this point at these higher accelerations, but So what happens in response to these two different modes of injury, these two different fundamental biomechanical injuries? So with the tissue strain, we've characterized this and worked with this model for many, many years. This is the only data I'll show from that. But what the point I wanted to get across here is that even at pretty low strains, we see substantial cell death in this model. So even at 10% Lagrangian strain of the substrate, the tissue rather, we're seeing substantial amounts of cell death, which really isn't a concussion. This is a more severe injury. Getting back to sort of Generale Omayas work where that rotational acceleration was causing diffuse axonal injury. And other lesions that they could see. Yeah. Oh yeah.

Hi, thanks. Just a quick question just for reference.

Yeah.

So in normal daily activity, what level of strain from turning head, getting up, things like that, where would that be on that graph?

Yeah, probably around, you know, 5%. Yeah. So So this, you know, with this being an in vitro model, we've got some background level of cell death. So I'd really look at kind of the delta between these two. So, you know, 5% is not going to really do anything to you, but 10, we see 10, you're going to, it's going to take a kind of a pretty big hit to produce 10% tissue strain. So, looking at this pressure model that might mimic more of the linear acceleration, this transient pressure pulse through the tissue, these different levels— this level 9 is an extremely large pressure pulse— we see very little cell death. We have to get up to this level 9, the highest, basically, level that we can drive our device in order to see any appreciable cell death. But what is interesting is that the function of the neurons in the absence of any cell death, even at these low exposures, can't undergo LTP. This neuronal— this molecular correlate of learning and memory is impaired a little bit, even at this level 1 exposure. But, you know, at level 2, we're starting to see significant deficits. And by level 4, it's completely wiped out. What's interesting is that these deficits take a little bit of time to develop and they can spontaneously recover. So similar to a concussion, depending upon the severity of the impact, you know, you're likely to recover in a, in a shorter amount of time than from a more severe impact. So here we've got shams in blue, dark blue. We've got this mild level 4 exposure blast wave pressure impulse in the purple, and then level 9, this much more severe, where we just start to see some cell death. In the turquoise. And what we see if we measure LTP immediately after the exposure, we can still induce LTP within that first hour. But if we wait longer than that, we see this significant deficit. And then here, what's interesting, if we wait long enough, the exposure, the deficit from the less severe exposure spontaneously recovers, whereas the more severe didn't recover within those 10 days. We didn't go out any further. It may spontaneously recover. It may not because of the cell death, though. Yeah, Candace.

I don't think this is serum-free media. And are there modulators in the media that you have investigated that affect both the recovery level and the time course?

Yeah, that's a great question of what's in the media. The injury itself is done in serum-free media, but then it goes back into serum media after that. It's quite a lot of serum. 25% horse serum, and we have no idea what's in it. I can tell you that progesterone is neuroprotective. So is estrogen. And we got to go up to 1 nanomolar progesterone. I actually don't know what circulating— typical circulating hormone levels are. Relative to 1 nanomolar. Certainly it's not that high, because if we put 1 nanomolar in, we see protection. But it's one of the fallacies of using serum. You never know what's in it for sure. So we asked ourselves what molecularly might be going on, what's, what's getting broken by this pressure transient. And LTP is an extremely complex process, tons of different protein players in it. This is just some of them I've got listed here. These were some of the ones that we were able to look at. And what we found, just to keep the story short, is after looking at all of those different constituents, was there was only one of the things that we looked at that was disrupted by the injury before trying to induce LTP. So over here is before trying to induce LTP. This is a chemical form of LTP, adding glycine, activate NMDA, get calcium to come in and induce LTP. And this is Westerns for this PSD95 protein. And it was significantly reduced. The expression was reduced after the injury. And when we tried to induce LTP, it wouldn't come back up. And so PSD95 is an interesting molecule in that PSD95 stands for postsynaptic density, 95 for its molecular weight. But it's a scaffolding protein that connects to the cytoskeleton on one side, and then it allows AMPA receptors, which are essential to the LTP process, to anchor into the synapse to be functional, to make the response to the same amount of glutamate coming in to make it a larger postsynaptic response. And so what we thought was, you know, after injury, this anchoring, the AMPA receptors were no longer able to anchor into the postsynaptic density. So even though stimuli were coming in, calcium may have been coming in, AMPA receptors being upregulated, but they couldn't get translated or they couldn't get moved and locked into the postsynaptic density because of the lack of PSD95. So we looked around for drugs that might be able to upregulate PSD95, and we found a drug called riflumilast. It's a phosphodiesterase inhibitor. Which means that it increases intracellular concentrations of cyclic AMP by preventing its degradation. And in a different study, increasing cyclic AMP levels was shown to increase expression of PSD95. And so we tested that in our injury model treated with refluromelast, and we were able to upregulate PSD95 even when we injured the tissue. And that upregulation of PSD95 allowed us— allowed for the induction of LTP. So able to upregulate the protein, and then functionally, we rescued LTP. So just a dose-response curve here. Roflumilast is a particular PDE4 isoform inhibitor. We also tried a couple of other pan-PD phosphodiesterase inhibitors, pecamlast, ibrutilast. They were also protective in terms of protecting LTP. But then we tried an off-target inhibitor to a different phosphodiesterase isoform, and it was not neuroprotective or did not protect function. So, really implicating phosphodiesterase 4 upregulation, downregulation, and decrease in cyclic AMP as a crucial step in this pathology to a pressure transient. The other nice thing I mentioned we can do with the microelectrode arrays is to look at the network function. Yeah, go ahead, Michelle.

Before you move on, I just had a question about these inhibitors. Have any of them been tried in vivo models of TBI?

Not— yes, they have. Miami group has since— not these, but they have new molecules that are phosphodiesterase inhibitors, and they work. They work in vivo. Nice thing about— Riflumilast is— it's actually FDA approved for COPD. But I think it actually doesn't get into the BBB, across the BBB, because these have emesis. They make people throw up when they get in centrally. So riflumilast doesn't get in centrally. And so it doesn't have that negative side effect. So I don't know about the Miami compound if it had any emetic effects in the— they injected directly in the brain. Yeah. So maybe not translatable in its current form. We can also look at the whole network activity with our microelectrode arrays. And here we can build a correlation matrix on an electrode-to-electrode basis to really see which electrodes are firing together, what group of neurons are firing together, and get a, I don't know, an overview of the organization within this circuit. There's lots of ways to come up with that correlation matrix. We use something called the spike-time Tiling coefficient. But basically what it gives you is a correlation matrix of, on an electrode to electrode basis, it's a symmetric matrix the way that we're calculating it. And then we can order these electrodes based on their anatomy. So we've got all the dentate gyrus electrodes here on the diagonal, CA3 here, CA1. So these are intra, Yeah, intra-regional connections. And then on the off-diagonals are the inter-regional connections, seeing how the circuit is wired up at a more macroscopic scale. These are just some raster plots. So we take these raster plots and calculate this correlation matrix. And you can see that there's a structure to it. We can also challenge the— the network with bicuculline in this case, and to see how we can alter that firing. Bicuculline is known to cause bursting in neural circuits, and you can see that in the correlogram as well, that these electrodes start firing together. What we find after injury is that this correlation is reduced without challenging with the bicucullin. And then when we challenge with bicucullin, it enhances that difference between the sham circuit. And you can see visually here that you don't get that enhancement of correlation between the electrodes. So we use this as an LTP on one side, learning and memory plasticity, and then looking at this as a as an overview of the network and its functionality. Then we can calculate things like connectivity and clustering coefficient. How much of the network is firing together gives you connectivity. And then clustering is, for a given electrode, how connected to the surrounding electrodes is it? And just here is just saying that we can quantify that correlation matrix give it a number so that we can do some statistics on it. And here we can see on the baseline without adding the bicuculline, we've got a significant decrease in this connectivity that we saw visually. And then with the bicuculline challenge, we can enhance that difference between the injured circuit and the uninjured circuit. So we use this, and especially since it was GABA that was Using bicuculline, it's an antagonist of the GABA channel, the chloride channel, which is an inhibitory neurotransmitter, inhibitory channel. So we, because of this big difference that we saw using bicuculline, we wanted to see if there was an effect of the injury on GABA channels. GABA channels are large pentameric channels, two alpha, two beta subunits, and a third subunit, delta, gamma, rho, sigma, maybe I'm missing one. And so we wanted to look at those subunits and see if injury was doing something to these subunits and affecting the kind of GABA channels that were being expressed. And we saw a decrease in expression of two of these alpha subunits that we were quantifying. Didn't see any changes in the beta. And then we looked in the literature, found a drug that was known to upregulate GABA subunits. This is SAHA, is the short name. It's a histone deacetylase inhibitor, which is known to upregulate GABA subunits. We threw some of that in, and we see a significant upregulation in our alpha-1 subunit, not in the others. And then we looked at this physiologically. And again, what we saw that a decrease after the injury in this connectivity clustering coefficient, but Saha rescued that. Didn't bring it all the way back up to sham levels, but it rescued it. And then also in the bicucullin challenge, it also increased. The response to bicuculline. So this involvement of GABA channels could be a— especially in concussion, in these more mild injuries, could be one of the molecular mechanisms leading to, leading to the physiological effects of the injury. So now that we had this mild injury model, we also wanted to look at what happens in repetitive injuries. So repetitive concussions. And here we've just got our single injuries in this category, level 1, level 2, or level 4. You see this dose response of the injury severity. And then if we combine them, so this is 2 injuries with 1 day in between the injuries, we see this significant accumulation of these 2 more mild injuries. Level 2 completely wipes out LTP, whereas before it didn't significantly decrease it. So we have this now, a model where we can study repetitive effects. We want to know what was that interval, that critical interval between those two injuries. And if we go out to 3 days between the two injuries, so we're looking at a level 2 now, Level 2 exposure. If we wait 3 days between those 2 injuries, we still completely wipe out LTP. But if we wait long enough, in this case 6 days, probably after the tissue is recovered, um, back up to an injury response that's more similar to a single exposure rather than this cumulative effect of multiple exposures. We can look at this in many different ways. Um, looking at multiple injuries, looking at different, um, time points after the injury. So again, here, two level 1 or two level 2 injuries applied 3 days apart. But if we wait long enough in the recovery— I'm sorry, they're applied 1 day apart. But if we wait long enough in the recovery, we see spontaneous recovery of these deficits. So again, you know, potentially this mild injury, we see cumulative effects. Here again, we're, we're multiple injuries, uh, with one-day intervals. But if we wait long enough, we get spontaneous recoveries. And then go to even more rapid injuries. Here we're applying 3 injuries in the span of only 10 minutes., but we've dropped the severity down to this more mild injury exposure. And here again, we see this dose response, but if we wait long enough after the last injury, we get spontaneous recovery again. So concussions, one concussion isn't going to give you CTE. Multiple concussions before you recover might be the reason for CTE. And if you wait long enough in between those concussions, you can probably recover. And if you have a new concussion in the future after recovering, it's probably not going to be additive to the earlier one. So we asked ourselves, what could be— what could be the memory? What could be the sensor, the transducer in the tissue that's remembering these multiple or accumulating these multiple injuries? And we turned our attention to glia. We looked at both astrocytes and microglia, and in our hands at least, it was really the microglia that seem to get activated more with each exposure. So this is just looking at immuno for IBA1, marker for microglial activation, semi-quantitative plus, plus, plus quantification of the immunos. And we see a significant increase in microglia. And then we want to ask what's happening with the microglia. So, they're becoming activated. Can we quantify that activation in some more detail? And so, we looked at two markers. One, a pro-inflammatory microglial marker, CD68, and an anti-inflammatory microglial marker, CD206. And we wanted to ask, you know, are these— do these change in a dose response with the number of injuries? Are they correlated with the number of injuries. And then, you know, maybe there's something there that's causative. So here, just Westerns for CD68, we see this increase in CD68, this pro-inflammatory phenotype, which does increase with the number of exposures. And then our CD206, this anti-inflammatory marker, decreases with the number of exposures. So microglia are becoming activated. They're undergoing this phenotypic switch from a quiescent to a more inflammatory state. And so we asked, well, what if we got rid of the microglia? If they are having—

turning—

making this change in phenotype, is that a good thing or a bad thing? And so we eliminated microglia or depleted them, didn't eliminate them completely, depleted them in our slice cultures with two different strategies. One, Chlodrosome, it's encapsulated chlodronate. The microglia engulf that, encapsulate it. It gets released, and then it kills off the microglia. Doesn't kill off all the cells because only the microglia are engulfing it. And then pexidartinib is an antibody to a microglial growth factor receptor. And so it also eliminates the microglia. And this is just Westerns to show that quantitatively. Yeah, there's a question in the back. I think one of the earlier slides when you were saying that microglia and astrocytes, they're all day 1 after the tumor was damaged. Is there only immunostaining after day 1 after the injury? Is there one day after that? Yeah. No, this was— This was 1-day inter-injury interval, so the 2 injuries were given 1 day apart. But this immuno, I want to say, was 4 days after the second injury. So we eliminated the microglia and want to see what effect of eliminating or reducing, depleting the microglia has on LTP after injury. And what we saw is with the clodosome, we reduced that LTP deficit.

Reducing—

depleting microglia is not a great thing. In our shams, we saw a decrease in LTP as well. But this was just kind of proof of concept. But LTP was enhanced or rescued or wasn't as decreased after injury. Also, pexidartinib had using a different drug through a different mechanism, a similar effect. So not that this is a translatable therapy, but it gives us an idea that microglia are a real player in, in the pathology after the injury. So we've done a few studies of trying to modulate the microglial function and response to the injury. One of the first things we tried were just NSAIDs. COX inhibitors. So, ibuprofen rescued LTP, but it had a very tight dose response. A little too much, and it didn't work. Too little, it didn't work. And then we tried inhibitors to different isoforms, COX-1, COX-2, found similar results. They both work, but again, had this very tight dose response. And if you look to the literature, concussion, mild TBI, NSAIDs. There's a lot of controversy there, and it might be because of this tight therapeutic window, which we can reproduce in our cultures. So, given that tight therapeutic window, we look for some other therapeutic targets downstream of COX activation generates prostaglandins. And so, we looked at trying to inhibit a prostaglandin receptor. EP3 is the most highly expressed in the hippocampus, and luckily there's a nice inhibitor for it. Throw that in and we get a pretty robust response from inhibiting the EP3 receptor, preserves LTP after the injury. And then lastly, we tried a new molecule, MW189. This is from Linda Van Eldick at University of Kentucky.

It's—

they believe it works by attenuating pro-inflammatory cytokines. So we thought this might be a great molecule to test given the role of microglia, we think. And this MW189 in this repetitive injury model was also effective at rescuing LTP. And that was true. If we treated for the whole time starting after the first injury, or if we just gave for the 24 hours between the two injuries. So fresh serum after the second injury, that was enough to potentially blunt the cumulative effect on the microglia. So they sort of didn't get over that activation hump to cause a pathology. But that's just speculation. All right. So I'll wrap it up there, leave some time for questions. So to summarize, I think there was this old controversy or separation between linear acceleration pressure, rotational acceleration tissue deformation. And I think we ignored pressure for a long time. And I think the pressure pulse actually does do something. I think it does something more mild. Than the tissue deformation. Pressure pulse doesn't really lead to cell death where the tissue deformation can and does. Didn't show the data here, but in our hands, when we stretch the tissue, astrocytes seem to be more activated after that injury, whereas the microglia seem to be more sensitive to that pressure pulse. And those initiating biomechanics would then potentially suggest different therapeutic strategies depending upon the injury severity. And then we can reproduce these neurophysiological deficits with that pressure pulse. Looking towards repetitive injuries, the dysfunction, at least on LTP, is dependent upon not only the severity of the exposure but also on the interval between multiple exposures.. And in most cases, that deficit will spontaneously recover. So it gives us a nice model, I think, of concussion. And then presented a number of different drugs, different therapeutic targets that seem to improve outcome. So I just want to finish by acknowledging collaborators at Penn and Duke who were instrumental with the with the pressure pulse work, funding sources, ARL, Alan Frontier's group, and then graduate students whose work I presented here. So, thank you very much. And got some time for questions still.

So, thanks. I'll start off. So, is it relevant or do you think it's relevant to do combination? Are you able to do combination injuries where you're Um, because sometimes, you know, in especially in the military, you won't have just a pure pressure pulse.

Yeah, that's a great question. Um, right, we, we designed the, the pressure model to try to eliminate as much tissue deformation as possible in order just to answer a question: does the primary blast injury do something, or is it just people fall over and hit their heads when they get blown up? We've talked about combining them, but we haven't combined it yet.

What do you think would happen?

What do I think happen? I think the stretch would probably be more severe and you would just get a stretch injury response at that point. I know your microglia probably get activated because of the cell death. Anyway in a more delayed fashion. So I don't know if you'd really see something different when you combine the two. We've thought about doing it, but we haven't done it yet.

Yeah.

And there was a question in the front over here.

Can you wait for the mic? Because there's people online. We want to make sure they can hear.

How time dependent are all the treatments?

Like if spontaneous recovery happens within a week, administering them late?

That's a great question. What's the therapeutic window for having a beneficial effect? So we have looked at it with a few of these. With the refluromelast, we could go out to 6 hours, but only 6 hours. Anything longer than that and it didn't work. That's actually a fairly long therapeutic window in traumatic brain injury, but that's the longest that we've seen. These others, We don't ever add anything, treat shorter than an hour. So all of these beneficial effects were waiting an hour after the injury before treating the cultures. But it's fairly quick. What we haven't looked at is whether we can speed the recovery process, right? So do the injury, let the deficit set in, and then if we start treating at 24 hours, do we get faster recovery than what we see? So we haven't tried that yet. That would be interesting. Any other questions?

I have one.

I'm curious.

What you think the mechanism of the microglial activation is.

Yeah, it's a great question. What are the microglial doing? So we know that if we just put IL-1 beta into the tissue culture media, we inhibit LTP. So I think it has a lot to do with the cytokines that they're releasing. What the cytokines are doing to the neurons, I don't know, at least not in the terms of LTP.

Yeah.

I guess in the other half of that is the microglia, are they becoming activated in response to the mechanical insult itself or are the astrocytes and microglia responding or releasing things?

I think the microglia are responding to the mechanical stimulus. And probably responding first, other people have looked at tolerance of different cell types. And the neurons are kind of resistant to cell death from stretching them. I mean, Doug and those guys stretched their cultures pretty far. You've stretched— I mean, you've shown, I think, that the astrocytes die before the neurons do in co-culture. So I do think it's a system thing happening, right? You've got cells interacting. Microglia are supposed to be responsive because they're sentinels and they're cleaning up stuff where you don't want your neurons to be sensitive. You want them just to be the circuitry and the wiring. So I kind of do think that microglia might be, you know, sort of the initiator. And that's why we depleted them to see if they would do something, just wiping them out. But I think downstream of, say, cytokine expression, I think there are studies that have looked at that when you have activated microglia, you get pruning of synapses. And so I don't know what those mechanisms are, if anybody's worked that out, if that's again just in response to cytokine release or if they actually go in there and do something physical. Great.

All speculation.

Excellent. Thank you. Excellent talk. It was very interesting. So I was kind of curious because your results on showing like multiple, like level 0, level 1 injury and they're recovering later, but have you ever done like an injury, waited a few days and then done another injury? Because there's that concept. I don't like to use it, but the preconditioning and how the cells respond later to an injury.

Yeah, so we have gone out to 6 days between 2 injuries, 2 exposures. And with 6 days, the response, at least for LTP, is more like the response to a single injury. So that— oops.

Yeah, no, I guess— sorry, I meant like with a— sorry, I meant going from like a level 0, 1 In my head, to like a level 2 or level 3.

We haven't mixed them up. We haven't mixed them up just because then your experimental matrix explodes. Yeah, I was just curious. Yeah, we haven't. That would be another interesting, interesting study. Yeah. These, these level, like that zero exposure level is more on the occupational level. So by firing a howitzer or firing a mortar is what that's reproducing. So not being on the receiving end of an explosion, but a shockwave from a muzzle. And so there is interest for sort of training purposes, how many mortars should these guys be firing and women be firing during training?

Yeah, that's what's interesting. I was kind of interested in that sense too, because there is a limit. I don't remember what the limit is. It is, right? They're only supposed to do so many before it leads to what they consider a permanent injury.

Yeah, totally right. And that's part of what ARL is funding us for, is getting to, you know, going from a dish to people. But, you know, what is cumulative? How long do you have to wait? How many can you sustain before—

Yeah, it's interesting. It'd be fast. I'm sure the military is very interested in those findings. Yeah.

Well, we really appreciate it. That was excellent. It was interesting to see all the pharmacological studies, which I hadn't been up to date on. So thank you for sharing that. And let's thank Dr. Marcin.

Thanks, everybody.