Scientist Stories: Sonia Lombroso, Engineering replacements for Microglia

So hello everyone. Thank you so much for joining us today. My name is Dominika Domi, and I'm a first-time moderator, hopefully not the last one. I'm super pumped for, for today's meeting, and today's presenter is going to be Sonia Lombroso, who is a—

oh, sorry—

who is a pharmacology PhD candidate working under the mentorship of Dr. Chris Bennett at the University of Pennsylvania. And Sonia's thesis work focuses on the development of cell-based therapies for the treatment of neurological disorders. And most recently, Sonia, along with co-authors, engineered a human CSF1 receptor with inhibitor resistance that can be used to replace microglia without the requirement of chemotherapy or irradiation, a finding that we will hear more about today, which makes me very excited. And Sonia, beside that, is also particularly interested in understanding factors involved in microglial identity. She focuses on answering questions like, does cell origin matter to microglia identity? As she believed that this type of questioning will be necessary to create the next generation of cell therapies. So I will give it to Sonia right now. And I would like one quick announcement. If you have any questions, please type them in either chat box or raise a hand after the end of the talk and don't be shy to ask the question yourself. So, Sonia, take it away.

Thanks, Dominika. And thanks so much for having me. I gave this caveat before. My dog is being a little chaotic. So, I apologize if you hear any background noises. He's well cared for, I promise, just needy. So, as was mentioned, I'm a graduate student at UPenn in the Bennett Lab where we are a neuroimmunology lab. Primarily studying microglia. And today I'm really excited to talk to you a little bit about this project in which we engineered a human CSF1 receptor to be inhibitor resistant with the goal of replacing microglia. And so I want to start first with an acknowledgment slide specifically for this project because it was so collaborative. It was kind of a happy turn of scientific events in which two labs, the Bennett lab and the Blerton-Jones lab, were working on the same question and really actually had a lot of the same answers. And so we ended up merging forces. And I think because of that, because of this collaboration, it's a good example of collaborative science in which our findings are much stronger because of that collaboration. So in terms of people to highlight, I wanted to highlight Jean-Paul, who's the co-first author on this paper, as well as Graham Peat and Dave Marzan, who are researchers in the Bennett Lab who worked on this project before I even got there. Great. Wonderful. So as mentioned, we'll be discussing kind of some work from this paper recently published in Journal of Experimental Medicine titled Engineering an Inhibitor-Resistant Human CSF1R Variant for Microglia Replacement. And for a brief overview for how I think today is going to go, so I'm going to give you a little background on microglia and what we think of as microglia replacement therapies. I'm going to talk briefly about some models and tools you might not be familiar with. Then I'll get into basically our murine studies of these experiments and our human iPSC-derived microglia studies. And then finally, I'll briefly summarize, talk about some future directions. And I'd love to hear questions and your thoughts on the work as well. So in the Bennett lab, we like to start kind of every project and presentation with kind of a large overarching question. So here, really, the driving question is how can we improve cell therapies for brain diseases? We think that one way, one cell in particular that might be well poised to answer this is microglia. What are microglia? Well, they're early brain occupants that stem from yolk sac progenitor cells. They're the resident macrophage and primary immune cell of the central nervous system occupying about 10% of the brain. Microglia are dynamic and process-bearing. What I mean by that is that they have this really ramified morphology in which they form complex branches and offshoots that allow them to occupy a lot of space in the brain. Here I'm showing you an image of microglia stained for IBO1, which is a pan-macrophage marker shown in red, along with DAPI, which is a nuclear marker. And so all of these red cells are brain macrophages. And what you can appreciate is that all of this black space, it's not actually just empty space. It's filled with other brain cells, such as neurons and glial cells and all sorts of things. So these microglia are really in the brain milieu and occupying a lot of space there. And while microglia are often thought about in terms of their immune function— functioning like phagocytosis or clearance of cellular debris, they're also extremely specialized for other functional requirements of the brain in order to maintain CNS homeostasis. And so some of these other functions are the ability to eliminate and remodel synapses, or even promoting neural survival, or even eliciting programmed cell death. So healthy microglia kind of help maintain homeostasis. But upon inflammation and injury, microglia can adopt these kind of amoeboid morphologies. And while it's often not really well understood if this morphology is causational or correlational, these diseased microglia or microglia in diseases have been associated with many different types of diseases. And so we're really interested in understanding if we can target certain factors in these cells, if we can ameliorate some of these diseases. And so you can imagine that one way to do this is like a targeted approach. So just editing or targeting one gene that's known to be dysregulated in these microglia in disease states. But another approach is actually replacing this dysregulated microglia in its entirety. And we kind of think about these things in terms of, or call them microglia replacement therapies. So if we can replace these disease microglia with healthy surrogates, perhaps we can ameliorate some of the symptoms or some of these diseases. And one particular class of diseases that we think is really well poised to be addressed by these types of therapies is a class of diseases called leukodystrophies, of which Krabbe disease is one. And so Krabbe disease is a severe neurodegenerative condition that's caused by deficiency in the GALC enzyme, which is an enzyme critical for the breakdown of toxic fats like psychosine and galactoserums. Serumamide. And in humans, Krabbe disease has an incidence rate about 1 in 100,000, and its symptoms include things like seizures and cognitive impairment and ataxia. And there's a range of severities, but in its most severe early onset form, when left untreated, Krabbe disease will cause death by 2 years of age. And there's no cure for Krabbe disease. There's no known cure for Krabbe disease. And there's only a treatment that is effective presymptomatically. And I'll talk a little bit more about this treatment in the subsequent slide. But in addition to this, like the cortical atrophy you can see in these patients of— sorry, MRIs of patients affected by Krabbe disease. Krabbe disease is also defined by reactive brain macrophages that have this altered morphology and increased lipid storage, which increases with disease progression. So, here we're showing, we're looking at hematoxylin and eosin stain of a macrophage. A CAC with Krabbe disease. And I hope you can appreciate that the morphology of these brain macrophages is really different than the ones that I showed on the slide before. So, these are very round. They have very few processes. They're full of lipids. And often they're actually multinucleated. And it's not— the role of these cells in the disease is not well understood. It is not well understood. And so, this definitely needs more study to understand this. And so I mentioned before, so luckily for us, there's a well-validated mouse model of Krabbe disease called the twitcher mouse, which arose from a spontaneous mutation that causes a truncated mRNA. And if left untreated, these mice die around P35. And in the twitcher mice, you can see that they have kind of this failure to thrive. They weigh a lot less than their wild-type counterparts. Have basically this motor deficit in which they have this twitching behavior. And with progression of disease, you can also see the hallmarks of brain pathogenesis as well. So there's an increase in demyelination, there's an increase of cytokine production. These are relative, just very unhealthy mice that recapitulate disease. And as I mentioned before, there's only a presymptomatic treatment for Krabbe disease and that treatment is hematopoietic stem cell transplantation. And while HSCT is beneficial if given presymptomatically, it has a couple caveats. The first being that presymptomatic for some patients means within the first several months of life. And the second is that in order to be effective, HSCTs depend on toxic preconditioning such as chemotherapy and irradiation. And that's required to deplete the endogenous macrophage niches in order to get donor cells to engraft there. And so this is toxic not only during treatment, but also has many health complications later down the line for these patients. Furthermore, the mechanism of action for HSCTs in Krabbe disease is not super well understood. I mean, there's still a lot of debate on this. One putative mechanism is the idea of cross-correction. So it's the idea that some of these healthy cells make their way to the brain and they provide a functional copy of GALC, the enzyme that's dysregulated in this disease, and that ameliorates the lack of this enzyme in other cells. However, with this hypothesis, one caveat is that as it currently stands, even after irradiation and chemotherapy, donor cell engraftment is not particularly high in mice or in humans. And so here, researchers varied either pretreatment method, either using irradiation or busulfan, as well as donor cells delivery, either intravenously or intracranially. And what you can see is that even at the longest time point after stem cell transplant with the most kind of aggressive pretreatment and intracranial transplantation, the level of chimerism is still relatively low in this mouse model. And this has been corroborated by postmortem human tissue of patients receiving bone marrow transplants. And so these researchers took advantage of sex mismatched donors in which female patients received male bone marrow transplantations. And similarly, it was seen that while the relative number of host cells was depleted, the actual number of donor cells was still relatively low. And I don't mean to imply that cross-correction is the only hypothesis for how HSCTs work in Krabbe disease. The Feltri lab and others have provided a lot of data that suggests that the mechanism of action of macrophages in this disease is actually through other functions such as like phagocytosis. But either way, whichever hypothesis ends up being true, either this idea of cross-correction or other macrophage functioning, the idea that if we could increase chimerism chimerism that would increase disease symptoms still remains. So basically, the question is, how can we increase brain chimerism? How can we increase microglia into the brain? And so we have a couple models in the field that we can do this. One is a genetic model. So you can genetically deplete microglia. And this is kind of picking your lab's— your favorite flavor of genetic depletion. In our lab, we use a CX3:CR1 Cre crossed with a CSF1R:Flux:Flux. So in the brain, CX3:CR1 Cre should be primarily microglia expressed, and CSF1R is primarily restricted to myeloid cells. So this allows for a pretty specific knockdown of microglia with tamoxifen treatment. In our lab, we tamoxifen treat on P1 and P2. And at this point, so you're depleting the host microglia shown in purple. And at this point, these brains can basically— these mice can basically be used as in vivo incubators for whichever cell— donor cell of interest. And so this could be monocytes or it could be microglia, it could be other macrophages of really any origin. And then you can— these donor cells can engraft and fill the brain, and you can analyze percent chimerism at later time points. And this model has been really well validated in our lab by a postdoc, Carly O'Brien, where she shows that she's able to get over 85% of donor cell engraftment in these mice. In addition to genetic models, we also have chemical and pharmacological models for microglia depletion. So many researchers use whole bone marrow radiation paired with transplant. And in these models, similarly, you get widespread engraftment of donor cells. So you might ask, okay, so we have these two different flavors of microglia depletion. Why do we need anything else? Well, genetic models obviously are not very translatable to human therapies. And chemical models for microglia depletion can often be toxic and are not tissue-specific. And so in order to kind of add an additional tool to the arsenal of how to deplete microglia, we started looking at ways to create CSF1R inhibitor resistance. And so first, just kind of what is the CSF1R receptor? It's a tyrosine kinase receptor whose signaling is required for macrophage survival. Endogenously, there are two known ligands, CSF1 and IL-34, which can bind to the extracellular CSF1 receptor, a part of the CSF1 receptor, which allows for kinase cascades for cell survival. And researchers have taken advantage of this, this need for CSF1R signaling for survival by creating small molecule inhibitors that bind to the kinase domain of CSF1R and block this signaling cascade. And these inhibitors, one example, they're called Plexicon. You'll hear me talk about this a lot in this talk. These small molecule inhibitors have been shown to lead to massive depletion of microglia. And many other people have shown this. And we have recapitulated this in our lab with a neonatal depletion model in which we show a dose-dependent depletion of microglia. And we do this through subcutaneous injections from day of birth to whatever, really whatever time point you want. And this depletion is also effective across many different brain regions. Okay, so we have this model, this pharmacological inhibitor of CSF1 depletes microglia. Well, why can't we just use that to replace them? Well, two main regions. The first of which is that microglia have this really amazing ability to repopulate the niche after depletion. And so these researchers basically depleted microglia with PLEX for 14 days, at which point they looked at repopulation by endogenous microglia at several different time points. And so what they're showing here is that the microglia are shown by these white dots represented of CX3CR1 expression, GFP expression. And so at day 0, you can appreciate that the relative number of microglia has been massively depleted. But then as soon as you get to 3 days post-inhibitor cessation, you can already see an appreciable amount of rebounding or repopulation of these endogenous microglia. And by day 7, nearly the entire microglia niches is repopulated. And so while these inhibitors are really effective to deplete microglia, these microglia have this amazing capacity to repopulate. And I mentioned there's a second reason why CSF1R inhibitors limit ability to use them for microglia replacement. And that's because, if you can imagine, this inhibitor is blocking CSF1R signaling on also not only on endogenous microglia, but also on the donor cells, right? So it will also kill the donor cells if the animals are continued to be treated. So from this, we hypothesized that if we could just create an inhibitor-resistant CSF1 receptor, perhaps we could boost donor cell engraftment in the CNS. And to do this, we used a couple different tools, which I'll just briefly describe here just because you might not be familiar with them. The first of which is a murine pro-B cell line called BAF3s, which is dependent on RTK signaling for survival. And this line has been classically used to study RTK signaling previously. And what's really— Cool about the cell line is that you can switch its receptor tyrosine kinase dependency from IL-3 to other ligands and other RTK signaling pathways. So we switched this dependency from IL-3 to CSF1 to study the effects of different CSF1 receptors on survival. And the second mouse model that we use are these ER-HOXB8 macrophages. I'll refer to them as either HOXB8s or CIMs, which stands for conditionally immortalized macrophages. And these are macrophage progenitor cells that are kept in their progenitor state through viral-mediated expression of OXPH. And these are really useful cells because they can be made from really any mouse line. So we make these cells from mouse models that lack CSF1R receptor called the CSF1R KO. And this mouse basically has no endogenous CSF1R receptor. So we're able to transduce these cells so that they only express the human CSF1R variant. So that's very useful. We've also shown that these CIMs are able to readily engraft in an open brain niche. So here we're showing CIM engraftment in a CSF1R knockout host. So these hosts have a brain devoid of microglia. And these cells are able to robustly engraft here. And they adopt microglial morphology as well as expression of IBA1. We also use iPSC-derived microglia, which I'll talk a little bit about how we made these lines later on in the talk. And then finally, one really interesting thing about the CSF1 receptor is that the murine receptor can be signaled through either the murine CSF1 or human CSF1. But the human CSF1 receptor depends on human ligand signaling. So in order to study the human receptor, we need a mouse with a human ligand knocked in. So we have that mouse by itself. We also have it on our mouse model of Krabbe disease, as well as corresponding immunocompromised lines, which are necessary for studying iPSC-derived microglia. Here is just an image of these HOXP8s, which can also easily be used in culture. They differentiate into bone marrow-derived macrophage-like cells. So here's kind of our working model of our hypothesis. So if we have an inhibitor on board, endogenous microglia cannot signal through the CSF1 receptor. This leads to host cell depletion. Then if we can transplant donor cells that have an inhibitor resistance to this PLEX, the small molecule inhibitor, these donor cells can preferentially engraft and not only engraft, but also expand and replace the host cells. And so in order to try to figure out how we could make these inhibitor-resistant receptors, we profiled the crystal structure of PLEX bound to CSF1R receptor with a structural chemist and used this to identify potential residues for mutation. So here we're looking at the binding pocket with a few potential residues called out. This is Plex shown in magenta with the CSF1R receptor shown in gray. And on the right, we're looking at the binding pocket of Plex in the receptor with one specific mutation, the 795 position, and how steric hindrance at this mutation, either for alanine, cytosine, or valine, might disrupt the binding pocket of the kinase. So with the structural chemist, we identified about 20 proposed mutations. And we then generated target mutants by transducing these mutants into BAF3s. We then performed IC50 curves with Plex inhibitor, and we found that one particular variant, G795A, conferred resistance relative to other mutants as well as wild type. We then moved this mutation into our murine macrophage line and performed IC50 curves where we also concluded that G795A G795A also conferred inhibitor resistance in these macrophages as well. We wanted to test if G795A also conferred resistance in primary macrophages. So we transduced bone marrow-derived macrophages with the IRRs, the inhibitor-resistant receptor. And we saw also that this inhibitor resistance was able to occur in these primary macrophages as well. And importantly, using endpoint data, we saw that these G795A-expressing macrophages had similar growth profiles in just the ligand only, as well as they were unable to grow without the presence of ligand. So there was no evidence for a loss of function here, or ligand-independent growth. So next, we wanted to test if G795A had an effect on kinase signaling, because basically, it's shown that this receptor allowed for inhibitor resistance. But since G795A is an RTK, we wanted to see, does this affect some downstream kinase signaling? So we— to do this, we performed a pulse chase reaction where we differentiated G795A and wild type expressing macrophages. And after differentiating, we serum starved these cells for 4 hours in the presence of varying amounts of Plex. And then added— after that 4-hour period, we then washed away that inhibitor and provided ligand. And I'll first take you through the wild type expressing cells on the left. And so without the presence of inhibitor, these cells are not able to activate the receptor. The receptor is not phosphorylated. And the downstream target ERK is also not phosphorylated. Now, when you get presence of the ligand, you get activation and phosphorylation of that receptor and phosphorylation of downstream target ERK. Now, with the wild-type cells, with any appreciable amount of PLEX added into the solution, you see an aberration or a depletion of this activation of this receptor. As well as phosphorylation of downstream target ERK. And with increasing concentrations of inhibitor, you see that this is maintained in these wild type cells. Similarly, G795A expressing cells are not— the receptor is not phosphorylated in the presence of no ligand. There's no phosphorylation of downstream target ERK without ligand. And as soon as you provide ligand, this receptor is phosphorylated and there's phosphorylation of downstream target ERK. Now, in contrast, G795A expressing cells continue to show activation of the receptor and phosphorylation of downstream targets with continued increasing levels of PLEX. Now, at really high concentrations of PLEX, this activation is inhibited. So, from this, we concluded that there's a normal functioning of human CSF1R receptor expressing cells, as well as G795A seemed to not have any effect on this one downstream target of this receptor. There's no obvious extrakines activity without human CSF1R, which is a big point as well.

Okay.

So to briefly summarize the first part of this talk, we showed that G795A confers resistance against CSF1R inhibitors when expressed in vitro in murine macrophages. Importantly, we saw that this had no discernible gain or loss of kinase functioning in these macrophages. So our next question was, does G795A, a human CSF1R receptor, allow for preferential engraftment of murine macrophages in the CNS? And so just one more piece of background here is that many in the field have shown that if you just do a transplant of macrophages, either bone marrow-derived or true microglia, into a non-depleted brain, you don't get replacement of microglia. And so we're just recapitulating that with the added piece of evidence that this is also true for these conditional immortalized macrophages. So you don't get engraftment in a non-depleted brain. And this is with the presence of plexus cells. Pre-treatment or without that. Sorry, one second. My dog is being crazy. Let me just try to stop him from doing what he's doing.

Stop.

Sorry about that.

Okay.

So where was I? Yes. So we— So that these conditionally immortalized macrophages can only engraft in the open microglial niche. So next, we wanted to see if we could get these macrophages to engraft in the neonatal brain. And so we used this neonatal depletion model in which we gave subcutaneous injections of inhibitor as well as cell transplants at P0. And what we saw was really robust engraftment of G795A-expressing macrophages, which was really exciting. We saw that this engraftment was dependent on G795A. So the G795A-expressing cells engrafted, wild-type cells did not. And we also saw that this was dependent on PLEX. So G795A macrophages couldn't engraft in the non-depleted brain. Really importantly, we saw that these cells also persisted after we stopped inhibitor treatment. So when we took these animals off PLEX, these cells were able to persist. And so here we took the animals off PLEX for 7 days and then harvested them. And you can see that these cells, donor cells are shown in green because they're GFP expressing. So these donor cells are able to persist. And what these callouts here are showing is that these G795A expressing cells adopt a very microglial-like morphology. Again, they're expressing AIBO1. And in the regions that are not engrafted, we get repopulation of endogenous cells. But the regions that are engrafted with our donor cells, we get relatively low repopulation of endogenous cells. And we also showed that this engraftment wasn't dependent on immunocompromised mice. So we were able to recapitulate this in immunocompetent hosts. Similarly, it was dependent on G795A expression and also PLEX treatment. So G795A-expressing cells do not engraft in the untreated brain. So since the neonatal brain is relatively more permissible than the adult brain, we also wanted to test if we could get these cells to engraft in the adult brain. And what we found was similar engraftment in the adult brain with G795A and PLEX dependency. So in this model, it's a little bit different because they're adults. So we pretreat these animals with PLEX diet for about 14 days prior to cell injection. We do these cell injections through stereotaxic transplants and then allow these cells to either engraft for 14 days, which is our on-PLEX time point, or we, at this point, we either harvest the animals for our on-PLEX time point, or we allow them to engraft for another 7 days on our off-PLEX time point. And we were really excited to see that on PLEX, these animals had significant engraftment of G795A-expressing cells. And again, these cells adopted a microglia-like morphology expressing IDO1. And importantly, wild-type expressing cells did not engraft in these adults either. And then we saw after 7 days after inhibitor treatment was stopped that these cells were really able to persist and fill the brain niche. Quite well. So this is callouts of G79.5a expressing HOXB8s after 7 days of inhibitor cessation. And we also saw that these cells express both resident and ontogenic markers. So these cells are able to express TMEM119, which is a microglia-specific or a brain macrophage-specific marker. And they also express MS4A7, which is a marker of peripherally derived macrophages. So just to briefly recap the murine macrophage experiments in vivo. So we were able to see that G795A in murine macrophages can engraft and expand in both the neonatal and adult brain. This engraftment was done in a plex-dependent manner, and these cells were able to persist after the removal of inhibitor. So our next question is, okay, does G795A confer inhibitor resistance to human microglia? So the next steps for us to test this, we and the Blerton-Jones lab basically introduced mutations into the human microglia, into human microglia using a CRISPR-Cas9 approach where we edited the endogenous CSF1R loci. And we created homozygous clones. And so as I mentioned, this was done at UCI Irvine and at Penn. So we now have two independently generated iPSC lines harboring this mutation. These iPSCs then underwent hematopoietic differentiation using standard protocols. And then this is where we can take two approaches, the first of which we can differentiate these cells into i-microglia in vivo. And this is something that the Blairen Jones lab is really, really well versed in. And the other way we can do this is by transplanting these cells as common myeloid progenitors into the mouse brain and kind of let the brain do what it does best and kind of have it induce microglia state in these cells. And as I mentioned before, because these are human cells, this is dependent on immunocompromised mouse. So we use a RADDL2 knockout mouse model here with the human ligand knocked in. So this is basically work that John Paul in the Blerton-Jones lab did, where he saw that G795A expressing i-microglia are resistant to CSF1 inhibition. And he did this through a time-lapse imaging of caspase activation. So basically, wild-type i-microglia, when exposed to different types of CSF1 inhibitors, show caspase 3/7 activation, which are both activated during apoptosis. So there's a cell death occurring in the wild-type expressing cells in presence of inhibitors. And this activation was not seen in G795A expressing cells. And this is true for other CSF1 inhibitors as well. So there's a clear resistance to PLEX of these G795A-expressing cells. But our next question was, okay, they're resistant, but what's their gene expression look like? Is this really altering their gene expression? And so John Paul did bulk RNA sequencing on wild type and G795A-expressing microglia in culture. And he found basically very few differences either at the whole transcriptome level or at the microglia signature gene level at baseline. So these are iMicroglia cultures. And at baseline, there's very few differences. And so the next question was, okay, well, what about in response to PLEX? And as expected, wild type expressing cells have massive gene expression changes in response to PLEX, whereas G795A expressing cells have actually no gene expression changes in response to PLEX. So they're very inhibitor resistant in that the inhibitor presence of the inhibitor does not change their transcriptional profile. So this was all done in vitro. So the next question is, okay, well, is this also true in vivo? And so Jean-Paul transplanted these cells into host animals where he then waited 60 days and then re-harvested these cells and did whole transcriptome analysis using bulk RNA-seq and found that similarly G795A and wild type expressing cells had a very similar transcriptional profile, both on the whole transcriptome level and microglia signature gene. And then in vivo, you know, there were a few genes that were different. There's about 7 genes total that were different between wild type and G795A expression cells in vivo.

Okay.

So next we wanted to see on the whole transcriptome level, it seems like there's not really an effect of expressing G795A on these cells. But what about in kind of— we know that microglia are really important in immune response. So what about in immune signaling? Is there an effect of expressing these cells on that? So to do this, we— or Jean-Paul did transplants of these cells and then allowed them to engraft for 60 days, at which point he exposed them to LPS, these animals to LPS injections. And basically what he saw was that there was no difference between G795A-expressing microglia and wild-type microglia. So I'll take you through. Here, I'm just showing G795A-expressing cells. But what we see is that upon LPS treatment, there's a downregulation of IBA1 and P2OR8/12. That's known in the field. There's also an upregulation of CD45, as well as MX1, which is an interferon response gene. So these G795A-expressing microglia are responding very similarly to LPS as wild-type microglia. And so you might ask, okay, well, what do wild-type human cells look like when they're transplanted into the brain? And what happens to them when you transplant them? So Jean-Paul transplanted wild-type expressing microglia into these animals and showed that with no treat— So I'll just orient you a little bit. So here we're having brain slices in which we're looking at IBO1, which is a pan-macrophage marker shown in green, and KU80, which is a human nuclear marker. So KU80 will be marking any human cells, whereas IBO will be marking any macrophages in general. And so with no treatment, what we see is that the human cells kind of are basically localized to the site of injection. The wild-type cells are localized to the site of injection. But as soon as you start increasing— as soon as you start providing PLEX, as early as 10 days, you see a real depletion of not only these murine macrophages, but also these wild-type expressing human iPSCs-derived microglia. And by 30 days, you see basically a complete reduction of any brain macrophages, murine or wild-type. And that's true for 60 days as well. Okay. But what about G795A? Well, we saw that G795A iPSC-derived microglia were able to fully replace the endogenous microglia. So similarly, upon no treatment, the cells were localized to the site of injection. After 10 days of treatment, however, you can see that these cells are starting to expand from the site of injection, whereas the murine macrophages are being killed. By 30 days, you can see that there's a persistence of these human cells, and they're not only persisting, but they're proliferating and expanding. You can see that they're filling most of the brain niche by 30 days, and by 60 days, they fill the entire brain niche. Importantly, we also showed that these cells persist after the removal of inhibitor. So these animals were treated for 30 days with the inhibitor and then taken off the inhibitor for 30 days, at which point you can see that these cells also engraft throughout the whole brain. And then also we're just showing— this is done in the Bennett lab where we showed that the human microglia— we transplanted human microglia at P0, allowed them to engraft for 60 days, and then PLEX-treated these animals and showed that they fully replaced the murine macrophages. Here we're looking at human CD45 by murine CD45. And so human cells will fall within this gate and the murine cells will fall here. We also recapitulated the G795A microglia brain engraftment with our iPSC-derived line, showing that wild-type cells do not engraft. By 30 and 60 days, there's a lot of brain engraftment, and then these cells do persist similarly off-plex for us as well. So just in summary of the G795A engineered human macrophage studies, we saw that G795A microglia have normal gene transcription compared to wild-type microglia at baseline. The transcriptome of these G795A-expressing microglia is not altered in the presence of CSF1R inhibitors. G795A and wild-type microglia have similar protein response into LPS, one immune signaling paradigm. And engineered human microglia engraft robustly when expressing the G795A mutation, but not wild-type expression. And that these G795A-expressing human microglia persist even after the removal of inhibitor. So just broadly concluding the talk so far. So we found that G795A substitution in the CSF1R receptor confers resistance to CSF1R pharmacological inhibitors without a discernible gain or loss of gene expression or kinase functioning. These G795A-murine macrophages and human microglia engraft in an inhibitor-treated dependent manner, inhibitor-dependent manner. And that these both G795A murine-expressing macrophages and human macrophages persist even after the removal of inhibitor. So we're really excited about this work, and there's a lot of future discussion about directions. And I would love to hear what the audience thinks. Some things that we think could be directly targeted is returning to that question of Krabbe disease, using this tool to basically provide a functional copy of GALC to these to these animals and see how that affects Krabbe disease. That's something we're actively working on. But more broadly, you could use these cells to really payload delivery for any range of neurodegenerative diseases. One that comes to mind is obviously Alzheimer's disease, but I'm really interested to hear how the field thinks about this tool in this regard. And then finally, one thing that I think is really interesting about this tool is regardless if you think it will work to deliver a payload or not in disease setting, It'll be really critical to study the role of human microglia in a wide array of neurological disorders via humanized mouse models. So you can really study now without irradiation, without chemotherapy, how human cells respond to immune signaling, but also just diseases in general.

Perfect.

So again, just one more time on the adoption slide. Thanks again to my co-author, Jean-Paul. And really the entirety of the Bennett lab and the Blair and Jones lab. Mariko Bennett was integral to this project as well, as well as Robert Spitali as well. My funding sources and just stop on this slide here.

Thank you so much, Sonia. It was really great presentation. Very intimidating amount of data, I have to say, but great to great to get to know this model more. So I encourage people to put their question in the chat or like raise their hands. I personally— okay, so we have one person that raised a hand. Delancey, do you want to ask your question? Or— yes.

Hello, Sonia, thank you so much. This is incredible work, and I'm like really enjoying reading every piece of it, and the listening is That's also amazing.

Oh, thanks so much. I hope my dog wasn't too distracting.

No, not at all.

Not at all.

I was just wondering, like, after you implanted the human microglia into the mouse brain, I was just wondering if you guys, like, check any changes in the microenvironment in the brain, like, such as, like, any changes in the cytokine, chemokine levels, or like any immune interactions with other cells like T cells in case, or— like the monocytes?

No, it's a really great question. And aside from the LPS studies, we haven't done any profiling. So that definitely remains to be seen what's happening there. I mean, yeah, I think it's a great question. And definitely future research needs to be done there.

Thank you.

Thanks so much.

While waiting for other questions, I have actually two questions myself. One is a very broad one. And one is very specific. You wrote in your paper that, however, future studies should finally map the consequence of the mutation on the cells, on the CF1R signaling, Donaldson longevity, and microglial function. So that's my first question. Are you doing any more research experiments towards that? And do you find any differences in how the cells behave with that mutation that you introduced?

Yeah.

So we haven't found any other differences aside from those few genes that we reported in the paper. I think, as in the paper, we mentioned more broadly understanding how these cells have altered— if they have altered responses in immune signaling is really important, I think. So we showed IHC data from after LPS. But right now, we're in the pipeline is RNA sequencing after LPS to have a broader understanding of what's happening on the transcriptome level of these cells. Cells in immune response. So LPS is one thing that comes to my mind, but other models of demyelinating diseases and kind of more complex neurological disorders and diseases, I think, will be really pertinent here. So we could use the cupazone model for demyelination and kind of to kind of investigate if the functions of these cells are similar to wild-type cells. I think that would be really critical.

And yeah, Yeah, amazing. I will have one short question and then I will give it to Anne. You said that also the other lab created the CRISPR Cas9 with the introduced mutation. So I want to ask if the clones are isogenic, so with the same genetic background, or if not, whether you see the differences between those clones that both labs generated, uh, based on the genetic background of the, uh, of the cells?

Yeah, so in the paper we only included the cell line from the Blurt and Jones lab.

Okay.

Um, so I think future— we'll have to do future studies to compare these two cell lines. I think that will involve basically us shipping our cells to each other, um, to really do some further studies on, on that. I think, um, yeah, perfect.

Thank you. And, um, I'm giving the The question is—

Hi, Sonia. Thanks for the talk. It was really great. And I had a question about, in terms of human treatment, would you still need to do a depletion protocol before treatment? And like, what kind of side effects might that have on a human? Yeah, it's a great question. Short answer is yes, they'll still need to be a depletion in order to deplete the microglia niche. And PLEX is actually used in the clinical setting already to deplete microglia and macrophages in patients. But it's a really great question is like, what is the effect of prolonged PLEX exposure on humans and on mouse bodies? From mouse models, we know that it does— prolonged PLEX exposure does have some effects. I think it affects OPC cell number. Although it didn't affect oligodendrocyte cell number, it does seem to have an effect on OPCs. That's a really important question, I think, understanding what What is the minimum amount of depletion that we can use in order to use this inhibitor-resistant receptor would be a very important question in my mind to answer. Then furthermore talking about how this works in humans in general is one thing that maybe someone will ask this question, but something that comes to my mind is, is this even a useful technology for treating things like Alzheimer's disease? We know that there's so much pathology in that brain and actually mouse models has have shown that in many AD mouse models, as well as tauopathy models, microglia are actually more resistant to pharmacological inhibition or depletion. And so kind of using this model to study that a little bit more, I think will be really pertinent here. And then finally, I'll just also add one other thought is that if we're transplanting these healthy surrogate cells into a brain that's completely messed up, what's going to be the response of these cells to that brain environment? And I think that's going to be something that needs to be studied on a disease-by-disease basis. So lots of work to be done still to look at this in different disease models.

Awesome. Thank you.

Thanks, Sam.

We have one question from Tejbeer. Thanks for a great seminar. Do you know by what mechanisms do the engrafted human microglia expand in the mouse brain? If yes, do they undergo proliferation or?

We know that they are actively proliferating at the basically, if you remember in those images I showed you, there's a very obvious border of the human iPSC-derived cells. We did some staining there to show that those cells that are on that border, you can imagine it being a wall of expansion. Those cells are Ki-67 positive, so they are actively proliferating. Beyond that, I'm not— we're not really quite sure of the mechanisms. I think it's a really interesting question. It's also a question that can be answered with other mouse models like of microglia depletion, like genetic mouse models, which is something we're also actively working on in the lab.

Sure. Next question from Du. It is fascinating work. Microglia play complex role in both detrimental and protective directions in AD brains. Can therapeutic implantation of iMGL or monocytes in AD patients play a confounding role in disease progression? Any speculation?

Yeah, I think it's a great question and something I think I alluded to previously is that, you know, there are some diseases in which I think you see a clear benefit of this tool. So when you're, if it's a monogenic disease in which you can just provide the missing enzyme or any other payload that like can compensate in the brain environment in terms of more complicated diseases like AD, I think it does get more complex, right? Because you have a brain environment that that might be messed up in many ways that are kind of causing the dysregulation of the microglia cells themselves. And so if it does, replacing these dysregulated microglia with healthy circuits, will that just lead to further dysregulation of those cells, of those donor cells? I'm not sure. Definitely something to think about.

Question from Betsy. Yeah, sorry.

So I'll also add another way I think that this tool could be used in terms of AD is people are making CAR macrophages already. To target AD. So this tool could also be used as a way to kind of really engraft those CAR macrophages. And our lab's definitely working on that as well.

Perfect. Question from Benty. Thank you for a great talk. What about the perivascular macrophages? Do you also deplete them? And what about the repopulation?

That's a great question. So we in this paper did not look at repopulation of perivascular macrophages. In genetic models in our lab, explicitly the CX3 CR1-CSF1R-FLOX animal, we do see repopulation of meningeal macrophages. I think further validation and further studying will definitely have to be done in this model to know if FLOX depletion with inhibitor-resistant expressing cells replaces those cells. I will say, in a very like N-of-1 experiment recently, we have seen that these cells are able to migrate pretty far far down the spinal column. But that doesn't really speak to replacement of perivascular macrophages.

Perfect. We have many questions rolling. So, I'm going to give it to Jerry. And then I think we can take one more or two more questions.

Thank you. That was a great talk. I have a naive question about microglia depletion. So, plaques induce apoptosis of microglia. I'm just wondering what cell types are responsible for cleanup the debris if all microglia are gone? So, one possibility is probably astrocytes. And that related to my second question.

What do—

how do astrocytes respond to microglial depletion and repopulation in your model? Thank you.

That's a really phenomenal question. So, basically, the question is, is how are other brain types, really astrocytes, responding to depletion via PLEX and repopulation with these cells? And short answer is we don't know. We definitely need to look into that. That's all I can say is that we really don't know. Really great question though.

This is a fairly new model, so it's a lot of stuff that needs to be done, but it's also very exciting because also it opens a way to multiple avenues of research, which actually that's what makes it so exciting about this study. A question from girl. Thanks for a great study and talk. Is this study could be improved transplanting yolk yolk sac macrophage precursors instead of hematopoietic precursor cells?

Yeah, that's a really good question. So kind of to rephrase the question, should we go earlier in kind of cell identity? Should we be replacing with yolk sac cells potentially? I mean, we could also just try transplanting the— see what happens if we transplant the iPSC line that's expressing the inhibitor resistance receptor. So just some background on this question. Microglia, I mentioned really briefly that they have unique origins to other macrophages. So microglia stem from yolk sac progenitor cells. In the mouse model, they're seeded around embryonic day 7.5, whereas other tissue-resident macrophages in peripheral organs are seeded by later stages of hematopoiesis. And interestingly, those other populations are often continuously replaced by bone marrow-derived macrophages. And Chris has shown— Chris Bennett has shown that you can take those peripherally derived macrophages and transplant them into the CNS, and they become very microglial-like. They upregulate about 95% of microglial identity genes. But they still don't get to that 100% of microglial identity. And so the question remains, kind of, does that matter? Is there a functional difference between those between those peripherally derived macrophages and bona fide microglia. And I think jury's still out. But it's a good point, is should we be going towards, like, yolk sac progenitor cell for transplants? And I think for some of our murine work, murine macrophage work, I think that's definitely the direction that we're going towards. We are— Kelsey, a graduate student in the lab, is trying to make HOXP8s from yolk sac progenitor cells. So that's kind of— tackling that question a little bit.

Great, great talk, Sonia. Thank you so much for, for talking to us today, um, and thank you so much for everyone joining the seminar today. If we have no— oh, two participants raised their hands. So Shruti, if you want to ask a question.

Yeah, hi, uh, thanks for the seminar, Sonia. Um, I was just wondering about the effect in the brain during the time period when there is no microglia, like when you deplete and then engraftment. Between that, there is no microglia in brain, right? So how long does it take and what's the effect in the brain?

Sorry, what was the last little bit of that question? How long does what?

How long does it take, like, for the— between the depletion and the engraftment, what's the time period?

Yeah.

And, um, what is the effect on the brain, if there is any?

Right. Yeah, so it's a great question. Um, so in our genetic models, as I mentioned, we're depleting at P1 and P2, which obviously you're depleting microglia in a really important time in development. We're also doing that with this PLEX model. We're depleting at early development time points while we're doing these transplants. We haven't fully studied the effect on the brain. What I will say is from the Green Lab has done studies in which they have depleted microglia using PLEX and shown that they're able to repopulate after 7 days. At least endogenous microglia are able to fully repopulate the brain after 7 days. In terms of our model, in terms of repopulation, the earliest we looked here was 14 days after transplant. We see the engraftment that I showed, which is pretty significant engraftment. What's the effect of not having microglia during development? Probably something. That being said, the really interesting and useful part of this tool is that you can actually do this in adult animals. You can mitigate the effects of not having microglia on development by PLEX treating adult animals and doing transplants then. Thank you.

Thank you so much, Sonia. So Jerry also posted here, and post about the seminar happening on January 31st. If someone wants to be a moderator, it's highly appreciated. Thank you one more time, Sonia. That was a great meeting, great talk. And thank you everyone for joining.