 Well thank you all for coming and welcome to the NHGRI Intramural Research Program Research Seminar Series today. I should tell you that this is being broadcast via a Zoom webinar and during this presentation you can ask for questions by going online to the Zoom Q&A part and Dr. Malik Dan will be conveying some of those questions at the end for us to give to Dr. Vernig and you see on the screen that Dr. Vernig is going to be talking about next generation cell therapies for the brain. My name is Dr. Gaul. I'm a senior investigator here in NHGRI and I thought I should just tell you a little bit about our welcome guest today Dr. Vernig did physics in graduate school so that's good background congratulations on that and then went to medical school at the Medical University of Vienna. He did his graduate school work in physics at Vienna as well then he did a residency in neuropathology and general pathology at the University of Bonn and did a postdoctoral fellowship with Rudy Janish at the Whitehead Institute. Then went to Stanford in 2008 as an assistant professor became an associate professor in 2014 and a full professor in 2019 in the Department of Pathology and the Institute for Stem Cell Biology which he is currently the co-director of at Stanford. Dr. Vernig has garnered many different awards. I'm just going to mention a few of them. The Clotterelli Prize for Outstanding Scientific Excellence by the National Academy of Sciences in the United States. The Assina Award from the Ministry of Science and Research in the Republic of Austria. The New York Stem Cell Foundation Robertson Stem Cell Prize and the Howard Hughes Medical Institute Faculty Scholar Award and there are others on here I haven't really mentioned. He's published over 130 articles and reviews and many of them are in nature journals and some of them are in science translational medicine and cell and PNAS and development really good journals. He's delivered over 150 invited lectures around the world and he's extremely well-funded and you'll see what an expert he is in this field. So I'm going to now pass the podium on to Dr. Vernig. Vernig and Hilt's talk is Next Generation Cell Therapies for the Brain. So take it away sir. Thank you so much for this very kind introduction. These long lists of awards and so forth just means I'm getting older and older. That's not much. That's what happens when you're a scientist in business for a while. But of course it's great to be a scientist and I hope I can share with you a little bit of our excitement and what gets up in the morning every day to work on. As probably many of you know, the concept of hemorrhoid development and cell differentiation is really much governed by the epigenetic mechanisms that people have in mind. So beautifully illustrated by Conrad Waddington in the 1950s with this landscape where balls run down valleys and once a ball has found its valley then there's no way to go back anymore. However, there's an alternative possibility that people have come up with thought maybe cell lineage specification is really rather dictated by a specific set of transcription factors. We just tried to illustrate with these letters. The idea was that if you just knew the identity of the key lineage determination factors such as ABC for pre-reported cells all the way on top of this landscape here then you should be able to really reprogram the cells into any direction that you like. So as you probably know, these type of experiments worked and led to the discovery of IPS cells which was rewarded with a Nobel Prize in 2010 I believe to Shina Yamanaka and our contribution to this field was that we wanted to extend this idea and ask whether this reprogramming also works between two somatic lineages that are really distantly only distantly related and we were actually able to convert fire blasts directly into cells that really looked and behaved like neurons so we called them in an analogy to IPS cells induced neuronal cells and we found these three transcription factors A is the one with one like in print two to do that job and these cells remarkably not only look like neurons they actually had all the functional properties that you would like to see from a neuron namely they were able to fire action potentials and form functional synapses when we were co-cotching together. We're very fortunate to work with Tom Suthoff an expert in synaptic biology and neuroscientist who helped us characterize these cells. So ever since people got very excited about this idea because it really suggested that maybe everything is possible and if you just knew the right factors you really should be able to move cells along the entire developmental potential room right into into every corner and there's a lot of example this is an image from a review we wrote several years ago and it's already way outdated. There's a lot of amazing examples out there now where you can actually really use this transcription factor based approach to program cells into a specific lineage of your design and we have really worked over the years to generate neurons from also different cell types from different donor cell types always with the same goal in mind to generate neurons and we were actually were also able to reprogram even endodermal cells from completely different germ lineage hepatocytes as well as which I'm quite proud of lymphocytes which to me is really one of the most differentiated sort of specialized cell in the body they can think of you know some some small round cell that flows around in our blood right even that we could manage to reprogram to to neurons I just wanted to show you some pictures how we can slowly but surely morph these round cells over several weeks to this beautifully looking neurons that have these nice dendritic arborizations we got help with some small molecules in addition to the transcription factors that improved their reprogram efficiency quite well and even those lymphocyte adult human lymphocyte derived induced neural cells or in cells we're able to receive synaptic input from primary neurons so they were also functional based on electrophysiological recordings I briefly showed this already earlier but we also explored what these transcription factors would do to pluripotent cells themselves and we actually started out using those more like as a positive control because a pluripotent cell is poised to differentiate so you assume if you add a transcription factor to to to their portfolio it would have even stronger effects than a somatic cell would have and that worked much better than we had even thought it would and we found that actually it has single transcription factor NGN2 and many people actually around the globe are using this transcription factor now to just generate neurons from ES or IPS cells because it really produces this very homogeneously looking population of of it seems like pure neurons that that also develop in in very rapid time like of a timeframe of two to three weeks this beautiful synapses that you can see are stained with with this red dots here synapses and when we do patch clamp recordings they are beautifully functional have synapses and importantly they have excuse their exclusively excitatory so they're only used to neurotransmitter glutamate we explored them a little more in more detail by doing a single cell sequencing RNA sequencing analysis and what we noticed quite early on actually already is that compared to undifferentiated cells these mature you know synapse forming NGN2 cells derived from pluripotent cells are quite heterogeneous even though from a morphological point of view and from a functional point of view they are homogeneous they're all excitatory neurons but from a transcription point of view there is quite some spreads on this single cell Disney plot as you can see so when we looked a little bit closer how to explain this obviously they were all neurons that expressed panneural markers and they were all excitatory which are expressed excitatory markers such as V-glute 1 and V-glute 2 which is exactly what we had seen from our functional characterization but when you look at other neurotransmitter markers such as cholinergic markers we saw some of them being at least partially induced and it seems to explain these different populations so in particular this population down here expresses you know it's this choline transporter as well as the vesicular choline transporter and some portion of this upper cluster is positive for these genes as well and when we then look at transcription factors that are sort of associated with this cholinergic program we noticed islet 1 as well as Fox2B and Fox2A which I don't show here quite nicely also characterize these clusters and it seems we have like based on these transcription factors we have really three groups they are all excitatory but one group expresses islet 1 and Fox2B which is this cluster down here and another group expresses only islet 1 but it's negative for Fox2B and then there's a set which is negative for both of these these transcription factors we have to sort of try to you know illustrate these three populations over there so it became quite clear that even though we from this function perspective we thought they were homogeneous but when you look more deeper in the transcriptional response or through transcriptional characterization there are actually some heterogeneity and in particular we saw cholinergic as well as other neurotransmitter programs being somewhat partly partially induced in at least some subpopulations of these cells and that by the way in itself is an interesting finding we think and we still don't really understand the answer because we start with a very homogeneous population of ES cells and put just a single transcription factor into these cells and we select for the expression of this factor so there's not much variability in the expression levels really and still we have a reproducible sort of heter... three population outcome right how this mechanistically works is is still a mystery to us so we next asked whether these transcription factors that be used as labels right for these different clusters actually also play functional roles and just added those in combination with NGN2 to these cultures and they all made beautiful neurons somewhat differently looking though and when we then look for these couple key cholinergic sort of reporter genes again we actually did see that both ILAT1 and Fox2B and to some degree also in their combination were actually responsible for this cholinergic program so it seemed that these are true key transcription factors that are mediating this sort of other neurotransmitter phenotype that causes some heterogeneity and of course one goal was to make this protocol even more homogeneous and maybe try to eliminate these subpopulations that would co-induce such a cholinergic program and we thought of two approaches the first approach we thought of was ESLs are super plastic and they're quite easily to moved and patterned in in various neural progenitor cells it's particularly straightforward to differentiate them in primitive neural ectodermal cells that can then be patterned with wind or without wind into an anterior or posterior population and we thought let's pre-specify so to say these neural cells first and then add our NGN2 transcription factor because we know NGN2 this famous pro neural BHLH transcription factor is very good at just turning everything into neurons and if we have a pre specification perhaps we can limit the developmental potential first and then just induce the neural differentiation with the transcription factor so that's what we did and the was quite remarkable that this patterning for one worked really well but it actually was maintained throughout the differentiation so let me quickly guide you through this principle component plot here and so here you have the three different populations undifferentiated H9 ES cells the anterior and the posteriorized neural ectoderm cells that were exposed to NGN2 for either two days so very early in these light colors and in these open circles for 28 days so for long time so those are the really mature functional neurons here and what was really striking is that we see a difference right this principle component to really is dominating the difference between these three different donor cells if you want and that is maintained beautifully throughout the differentiation protocol so that the patterning works and then the differentiation that is mediated by NGN2 doesn't really change much of this original differential patterning that we have induced into these cells and on the other hand their maturation stage which seems to be corresponding to this principle one component here is also quite similar between those so it's this very similar maturation stage and the type of patterning is maintained throughout the process so that already indicated to us that presumably we didn't really achieve our goal here right because what we actually wanted is to end up with a more homogeneous cell population right in the end of our you know these four week old enzymes but we probably only have accomplished is we have sort of still maintained the original patterning that we have induced into the cells right so that was that was the initial conclusion for this and but it was more solidified when we then looked at the specific genes that are you know underlying these these principle component plots and just as Saturday checks you know when you look for posterior genes like hawks genes they are beautifully induced right in this in its posture populations and anterior genes like OTX genes are beautifully highly expressed in anterior populations and repressed in the posterior populations so it's this patterning worked exactly as advertised and but importantly it's really kept through the differentiation because in fact all of this data plotted here are from these 28-day old cells which are the mature neurons but unfortunately as we had already sort of concluded from the principle component analysis we have not really accomplished our goal so as you can see eyelid one is in fact repressed in this posterior population but it is really not affected at all in this under population indicated in the red color here and compared to the non-patterned population in blue so we do seem some effect but it actually was the exact opposite that what we had expected because you know there's a motor neurons are a famous cholinergic neurons in the spinal cord so if anything we probably would have expected that cholinergic problem should be rather induced in the posterior component as opposed to the anterior component and when we look at the sort of marker genes for a variety of neurotransmitters sub-specification markers we see that there's really not much change right we did I mean there is change but but not there's not a significance of the suppression of the resulting neurons so we're also interested in mechanisms and asked where is the transcription factor actually landing in these different cell populations right the patterning into under and posterior population also changes the epigenetic landscape and transcription factors are thought to be quite dictated right in their ability to access the genome by the chromatin by open and closed chromatin where these transcription factors can can go and what we actually saw is that the engine to distribution throughout the genome is actually not all that much changed in this different combination in these different cell populations so here again are the ESLs the anterior and the posterior populations and here a chip sequencing experiment putting out all the sites where engine 2 is bound and you see that there's some sites that might be lost but there is certainly no sites that are newly gained for engine 2 so overall quite similar pattern and in blue you see an attack seek signal which is a measurement to measure the accessibility that the openness so to say of the chromatin and there's a typically a very good correlation between where engine 2 binds and where a chromatin is open so overall not so much happening there on that under the front what we did make the surprising observation we made is that actually in this context engine 2 we found to be directly bound at the islet 1 promoter sites as well as one of the at least one of the cholinergic genes presumably directly regulating such a cholinergic program which doesn't make too much sense actually because engine 2 is expressed in a variety of different you know neural precursor cells and responsible for the differentiation of many different types of neurons certainly not a specific cholinergic inducer so that is interesting that in this particular context of an ESL engine 2 would do that so since our first approach didn't really pan out as we had hoped we came back to you know our love to transcription factors to begin with and thought why not you know combine transcription factors right from the start it also makes things easier I don't have to worry about you know media exchanges and and variability how efficient cells differentiate along a different path and just start with the undifferentiated cells and partner engine 2 with other transcription factors which by the way this is not this is nothing new right developmental biologists such as François Guillemot have already you know discovered many transcription factors that are thought to be responsible more for subspecification right really not as long stretch and we should have thought of this much earlier because even earlier we actually were very successful to generate pure cultures of inhibitory neurons, GABA-ergic neurons, by combining another BHLH transcription factor called ACL1 with such a homeobox factor which is more restricted and thought to be more instructive for a sublinage for the GABA-ergic lineage and that worked really really well so we assume ACL1 is really there to induce a sort of a pan-neuronal differentiation program and DLX2 is its job is really to sub you know specify the lineage and produce this beautiful cultures of purely inhibitory neurons. So back to our problem to sort of get rid of this cholinergic program in our engine 2 cells which are otherwise excitatory and thought like what transcription factors could potentially do the job and we since we wanted to generate a cortical sort of forebrain type neurons we picked some transcription factors that are good candidates for that and there's this class of homeobox transcription factors EMX1 and 2 as well as OTX1 and 2 that are beautifully expressed in this sort of extended pattern starting from the from the very unterritid tip of the telencephalon going all the way at the at the midbrain hindbrain boundary and we also included FoxG1 which is expressed about similarly as EMX2 from from the middle of the telencephalon and when we combine these transcription factors as two factor combinations with NGN2 and then do an RNA sequencing characterization of these cells we actually very you know very relieved to see that we actually generate a very different set of different neuronal populations so you're kind of cool you know with a different combination of transcription factors you can really sense that the cells in sort of different lineages of neurons and those are all neurons right by by functional as well as by sort of morphological criteria but they obviously have a very different RNA sequencing pattern but our main goal was to see where they can sort of purify the population right and and hone in the the glutamilagic program and rather repress all the other programs so we looked specifically at that and we're very pleased to see that in particular the EMX genes and FoxG1 FoxG1 was very efficient to to block ILAT1 as well as this Fox2V transcription factor which we thought are really the sort of the key nodes right responsible for for cholinerative programs so they were really good at that and also quite good at repressing choline atrial transferase which is another marker as well as this vesicular marker and when we then plot more of these lineage indicators these lineage marker genes here we actually did see a much better suppression than in our previous attempts with pre-s, you know, differentiations of this pre-patterning here in particular we liked the combination NGN2 and EMX1 which did the best job of repressing these non-glutamilagic programs and again we wanted to see can we explain this on a molecular level and in quite contrast to what I've shown you previously where NGN2 is bound in these pre-pattern cells now when we add EMX1 or FoxG1 now we see a very different binding pattern of NGN2 so that the partnering of homobox factors or this forkhead factor FoxG1 has a much stronger effect on how NGN2 can access and does actually access the the chromatin and now we actually have quite a substantial set of loci where the binding of NGN2 is not only lost as in the previous example but also really gained and for FoxG1 we couldn't do the experiment but for EMX1 we were able to to find good antibodies and actually in this case you used a flag tag to do that and asked where is EMX1 going and the quality as you can see is not quite as good as for the NGN2 but still we saw I think I've circled us here for those sites that are where NGN2 is gaining new access to the chromatin we actually see a quite clear enrichment also of EMX1 suggesting that EMX1 may actually drag NGN2 to this to these new sites and to investigate this also for the for the FoxG1 sites where we couldn't really do the chip seek experiment but we could look for the for the motifs behind it right that was also actually quite striking where we look when we look for the sites where NGN2 is bound exclusively in the presence of FoxG1 we see a clear enrichment of the FoxG1 motif presumably FoxG1 would also be binding there and the same is of course also true for the sites that are newly bound by NGN2 dependent on EMX1 we see a nice EMX1 motif being enriched there and for that case we also know that EMX1 protein itself is present there as well so a lot more protein-protein interaction actually than we had imagined before a little bit closer look at EMX1 itself it seems that to be rather have a transcription or repressive role than an activating role when we look at the overall gene expression so plot the sort of average gene expression globally as well as the EMX4 tar genes they're much more they're significantly repressed rather than induced and when we look at the differentially regulated genes and do a geothermal analysis the genes that are in fact repressed by EMX1 have you know geo terms that are associated with several neuronal specification programs that are not excitatory right you see spinal cord and sympathetic neuron and so forth cranial neuro formation and so forth so things make a lot of sense all right I am now pretty much at the half of my time which is a perfect timing because I would like to change some gears now and tell you a little bit more about our a bit more translational efforts how to really go about sort of new types of therapies I know this is really a little bit of a pipe dream you know people in particular we had over lunch with the students discussed it's a little in particular investors are very scared of cell therapies in general very complicated you know manufacturing and so forth but we think as academics we have the freedom to you know pave new ground and try things out without the pressure of having to make money right we I always say my job is to burn money and gain knowledge and whereas investor has to make money right so so luckily we are in this luxury position and throughout my you know scientific career I've really thought you know how can we get the cell therapy to work in the brain this is taken from one of my papers that I did with Rudolf Ennis as a postdoc which by the way yielded this concert really price from the national academy of sciences so it's thanks for mentioning this this by the way was also very nice trip to washington I remember probably not too far from it or was it them all this this right the academy is there yeah anyway so so this was actually one of the first times that IPS cells were shown to have a therapeutic effect and this is all mouse at this point you see on the data is early on so we differentiated mouse IPS cells at the point to dopamine urgent neurons and and showed that it has a therapeutic effect in a red model of of the disease of of a Parkinson model of reds but what you also see so I should oh I should point out that the slide is stained for th so the the black substance here is is the transplanted dopamine neurons presumably and this is the normal side right this this is how the dopamine neurons normally project into the into the stratum obviously very homogeneously and you know throughout the entire brain structure or this area at least whereas our grafted cells even though they had a beautiful therapeutic effect they were really you know stuck exactly where you put them and there was really not much migration and for Parkinson people pursue this I mean there's clinical trials going on now multiple action of them you know for Parkinson's maybe okay to to have that level of innovation this local you know problem and at least for the motoric symptoms you you may have a good benefit but it's still really not satisfactory from a more general point of view we really would like to have something that has you know better incorporation pattern into the brain and there's at least some migration um so just take this this as an example of you know many of my during during my training as well as in my own lab attempts to get some their cell therapy really to work so at some point we thought about my cochlea those are hematobic cells they're not you know the cells that usually come to our mind that we would want to replace or you know they're not degenerating in diseases and so forth but they're still important those brain cell types and there are there is this body of literature that claims that when you do a bone marrow transplantation that some cells can get into the brain at that point when we started this work about five years ago so and it was not that clear how robust this is um it was even some ever some you know old literature from the like 80s um that even in human right in um Y chromosome mismatched donor samples they would see some Y chromosome positive cells in the brain so it was really exciting and but even mice at that point was not so clear at least to us you know how robust is this phenomenon so we wanted just to repeat this in our own in our own hands and uh Yohei Shibuya the postdoc who started all this work just did a conventional bone marrow transportation we used a buzelfan um also in mice because that is what mostly is done in the clinic in people injected bone marrow into this into this mice and then just followed these mice and looked in the in the peripheral blood as well as in the brain what happened to this GFP label cells and as of course expected there was a very high rapid chimerism being formed in the peripheral blood and somewhat intriguingly um there was also a chimerism formed in the brain which is actually quite substantial you know on average maybe 40 percent or so but it was there was a long delay so at this early time point of let's say four weeks where essentially the entire blood is already GFP positive there's really not much yet going on in the in the brain it takes another good you know two months before anything really shows up in the brain so that's one thing you know why why is this this delay why what if it was why does it take so much time for the cells to find the way from the circulation into the brain but the second issue was that yes we we on average got a relatively you know nice chimerism 40 percent is substantial but it really was highly variable and this is from this you know the same batch of cells of course different mice but the same hands you know the same post-doc injecting exactly the same way and and highly variable results that didn't really sort of hone in into into into a less variability situation after longer time points so sometimes we would get like beautiful chimerism in this in this in this brains but sometimes really you know it's hard to find any cell so not really something good to work with so we thought there's really room for improvement and read more papers and stumbled upon this what what you know i would like to call the microlele niche conundrum because there are some experimental conditions where apparently it's very easy for a circulation derived cells for peripheral blood cells to get into the brain and there's other conditions where it all seems almost impossible for the cells to enter the brain and one example which i'm shown here from from as just as an example but as many papers like that from benedal 2018 where he took a sees if one receptor knockout mouse so i should point out that sees if one receptor is a key signaling pathway for microglue survival so without that receptor without that intact signaling pathway microglue cannot survive so when you knock out that receptor these these mice don't have any microglue in the brains and when when you inject bone marrow actually just IP into this newborn mice you can see that entire brain is full of these GFP positive cells so in that context the brain is almost like a sponge and you know takes up you know all these these cells from the circulation and and they become microglue like cells if you do a very similar type of experiment and rather block the same pathway pharmacologically then genetically which which you can do very well these that there's some some drugs developed that are super potent and you can and we could rep reproduce this data readily and in like 10 days you can deplete more than 90 percent of these cells beautifully it's a beautiful reagent so you also deplete microglue and when you withdraw this drug you also see a recovery also a repopulation somewhat similar to this but in this case all the cells that have repopulated the brain are from within the brain and not from the circulation not a single cell has been found to be recruited from the circulation so we assume that the same signals are involved right because in both cases you you have a transient period where the microglue are depleted so what's going on there we wanted to simplify things a little bit because you know there's it's quite a complex thing right going from the blood into the brain is the blood brain barrier that needs to be some attachment first and then once they're on the other side of the of the of the vessel they need to differentiate right so you wanted to simplify the situation a little bit and take the blood brain there out of the out of the equation and just put these bone marrow cells right behind the blood brain right into the brain so it's a little bit of a silly experiment you would think and some people have done this actually before when they were dreaming of trans-differentiation some of you might remember these times so we injected bone marrow into the brain into the brain and what happened was not much it what did happen is though that these cells did survive for at least two weeks but they were hanging out there they still have round sort of hematopoietic pericasal cell morphology they don't express iber 1 which is a microglue marker so they are not sure what they do there they are hanging out there not knowing what to do and and that's it however when you do the exact same experiment into a mouse brain that was previously depleted for microglue so into a mouse brain that has no microglue the situation the outcome was completely different first of all we saw many more cells and there were much more spread around the injection site and literally all of these gfp positive cells that injected bone marrow cells had changed their morphology and had these beautiful ramifications and all of them expressed iber 1 there were none of these pericasal cells round iber 1 negative cells left at least everything that was left was was microglue like cells so that experiment told us that these these niche factors that that is that these these factors that apparently are produced in a microglue free brain are a very strong environment for hematopoietic cells to adopt a microglue like phenotype so it's as if the brain would call for help right if there's no microglue around i'd grab anything that comes around and turn you into a microglue maybe anything hematopoietic so without that's great that seems a really strong signal let's take advantage of this and let's combine this with with the bone marrow transportation protocol that we had done before and let me just come right to the chase when we do this we do first the bone marrow transportation and then let the mice recover a little bit give this plexigon drug well you know this is a C1 receptor inhibitor from this company plexigon which by the way doesn't exist anymore anyway the drug is still very good and then let the mice recover from from that drug and then analyze these brains we have a very consistent you know over 90 percent gfp positive hematobic population based on facts and i'm a trained neuropathologist so i always have to do sections and look you know how these these dots on the fax board actually look like and when you do that we see the entire brain is full of these transport cells and when you look a little bit more more in the detail they look beautiful and we actually when we have a good drug and a good depletion because we had bad batches of the drug but when we have a good drug we have really hard times finding it in dodgers microglue in this in this context so in i don't see any micro i have one positive cell which is red that is not gfp positive and if anybody can spot one please let me know i couldn't so it's a very very efficient the entire microglue at least parenchymal you know part of the microglue population is replaced with these grafted cells so let me just go back to to this image how different does that look like than this this pns paper i just showed you where we transferred these dopamine neurons where the cells were just stuck there exactly where we put right now we have access to the entire brain and it turns out also the spinal cord so the the entire cns with a relatively simple already you know clinically you know almost you know the variation we just did a variation of the clinically you know approved procedure so that that was really exciting and of course one of the first things we asked is can you know this is useful can we can we you know see some therapeutic benefit that we can take advantage of the system for some some disease models and one of the things that come to mind are lysosomal storage diseases and i had the chance to talk about it earlier but i don't show this here because it was already published earlier but another thing that also came quite early to mind is i'm sure everybody in this room is aware of this strong you know genetics link between microglue genes and isomers disease in particular trim 2 variants confer they're relatively rare compared to apoE but but the risk they confer is i think on par with the apoE4 a little so there's also a syndrome called nasohochol disease that is where people are deficient for this trim 2 this microglue receptor which has a very rapid neurogenital phenotype so that was sort of you know screaming at us literally right we can replace or we can fix the problem right we can change the microglue that have the deficient receptor with wild up cells right as easy as that so we got a hold of an isomal disease model which is a plaque model so they express a beta and form these beautiful plugs which are well i'm sure they're beautiful from a you know from a neuropathologist point of view not not for the mouse and not for the patient of course but they form these these a beta plugs and in a normal mouse the microglue have a strong reaction to these plugs so you see in in black again the eyeball stain you see how these microglue react to them and and sort of circle around these plugs and you know try to do something with them they are not able to quite fuck as i told them but they certainly do something to these plugs and are reactive so when you knock out trim 2 on top of that these microglue couldn't care less they seem to be completely insensitive to this pathology as if they are you know blind you know so it seems and that's what i guess this these data are sort of the basis for why people believe that this trim 2 signaling is really important you know for measuring disease and and then responsible for for a microglue activation which is apparently beneficial right you also did the discussion earlier whether you know those are bad or good microglue so that is something quite measurable and we were wondering whether we can you know use this as a as a readout to see whether such a cell therapy could work so we did the exact same experiment with with our transplant cells now on the other hand so when we take a white type bone marrow cells and replace them in these alzama models in these plug models as expected the the cells again nicely react to the to this plug series like two examples you will notice and i sort of forgot to point this out these bone marrow the rough cells they're a little different than microglue so that's why we don't call them we cannot call them microglue they are different morphologically as you see they're a little bit more simplified they are ramified and they have all the functional properties that you want to see they can focus at those they respond to to inflammatory stimuli with with secretion of cytokines and so forth but they do this slightly differently than than than real microglue more more important to also have the proper control experiments done here but importantly when we take knockout cells as as a control right so for that reason this is an important control also these bone marrow cells are you know insensitive seemingly to to to these plugs so that is a great result because we can now in principle restore this dysfunctional trend to receptor in at least in this mouse model and importantly you know these these cells are not activated but the microglue function apparently has also an effect on the overall plug load in this model so with tram 2 typically you have fewer of these of these plugs which is shown here right in the control conditions tram 2 y type here has fewer plugs as well as compared to the to the knockout and we see a beautiful you know similar effect or rescue if you want of the of this particular mutant when we transplant the the wild type bone marrow and replace the tram 2 mutant microglue in these brains with with these with these circulation derived cells all right so i have a lot of clinical fellows in in the lab so the the potential applications of this hematopoietic you know replacement therapy is is quite quite big and one of my fellows got really interested in neuro inflammatory disease in particular ms and in particular because there is evidence that bone marrow transplantation may have a long lasting beneficial effect on a progressive and chronic relapsing multiple sclerosis so this is taken from a from a relatively recent review that has collected you know metadata from for for how many patients but but quite a lot as you can see and compared to historical data is my understanding that it's quite an impressive finding that after 10 years of of bone marrow transplanted ms patients you still have a 60 percent disease free or relapse free period and so that particular review also looked at different conditioning there seems to be some stratification depending on how you do the bone marrow transplantation seems to have more or less better effects but there seems quite a convincing effect on the bone marrow transplantation on that disease but the mechanisms are completely unclear i think most people think of T cells and lymphocytes you know this is an autoimmune disease and there's you know specific antibodies and specific epitopes that T cells attack on on myelin and then oligone resides of course so i guess the overall idea in this bone marrow transplantation is to rather have a lymph ablation right and sort of kill lots lots of the lymphopoietic system and replace that with a sort of a reset immune system if you will but nobody thinks so much i think about myelid cells so we thought we have now a nice way to assess really the contribution of brain myelid cells or microglia in particular and what the contribution of those cells would be because we have now this way to really completely exchange those cells as well so that's why we got a hold of an EAE model there's a bunch of them and we picked on purpose like the chronic form of EAE i should say this is you know a very common neuroinflammatory model not sure the experts would be happy to call it a good MS model but this is the model people use and we have it stands for experimental autoimmune encephalitis and several myelitis and also in our hands we get this to work quite nicely we see this chronic lesions here you know a dappy enrichment of cells and infiltration of of immune cells including ibogon you know myelid cells and a clear demyelination here when there's this in here for flora myelin so the model is really well established and was readily working in in our hands as well so we did a couple cohorts of many many mice with the with various different conditions to assess whether these added microglia replacement has any you know clinically functional effect on the course of the of the disease so let me quickly walk you through this through this data in red you see the normal course of EAE in this model so you typically have a very consistent peak of severity and that then over time sort of becomes more more chronic goes down a little bit but then sort of really stays up sorry my stays sort of on the same level right there's other models which are more chronic relapsing so that go more in waves but this is the one which is really more sort of chronic stable and in the other colors are the various control groups so first we also of course tested whether this this plexi contract itself may have an effect because of course you know it's quite a big thing to eliminate microglia for like a week or so and then have to come back that alone could have could have an effect but in you know this seems this transient depletion of plexicon during these 10 days here and didn't really affect the the course of the disease at all however the bone marrow transcription groups did so that that is in line you know with some papers that have been done in in rodent models as well as you know in some open label clinical studies that I showed it as proposed that bone marrow transcription may have may have a benefit and indeed a conventional bone marrow transportation right around the time you know after the cell transportation these curves diverge here and the the clinical scores here are getting a consistently better in these bone marrow transcription groups and quite nice to see for us is that when we then add our plexicon trick here which then you know boosts the microglia replacement right around the time where you actually expect an effect which is at the end of this plexicon treatment we see a further improvement of the of this treatment group here so it seems that the replacement of the of the mild cells actually does have a functional effect and the difference between green and blue line cannot be explained with the difference in the lymphoid compartment so it turns out that this chronic phase of EA is actually not that well characterized most studies like are around here whereas our you know cell therapy is of course takes some time to to reach the brain so we are rather interested in the chronic phase and so to characterize what's what is going on on the both on the recovery side as well as on the on the molecular side we wanted to single cell experience single cell characterization single cell RNA sequencing characterization we actually did a nuclear sequencing here and that experiment actually worked really really well so I can highly recommend this is somewhat expensive but but this experiment actually was was worth the money I must say so we saw a beautiful representation of the many different cell types and first just just characterized control with with the EA and as you would expect there is a huge increase in particular in in here in green you see that the mild compartment and and you do see relatively speaking also a huge increase of of course of lymphocytes here as well as expected and we look at the what populations changed their gene expression pattern was also actually the mild cells which had the strongest dysregulation of genes much more than the cells that are actually affected which are oligonucleides and astrocytes to some degree so can we how can we explain the the functional benefits right you would hope that there is more myelination going on and that's exactly what we see so when we when we stain and quantify the amount of myelin with this with this myelin stain I just show you some some representative examples but we have quantified this and there is indeed increased myelination in the in the treatment groups both in the bone marrow and the bone marrow with plexigon group and when we look at specifically the gene expression in the oligonucleides lineage which is composed of you know the the entire path of an undifferentiated oligosand precursor cell all the way to a mature cell I didn't quite appreciate the the heterogeneity of of this lineage in in in the mouse's brain and we see actually that the program that is induced by EAA sort of the disease state is rather reversed in the bone marrow treatment groups both conventional as well as with plexigons so I apologize for it for a little bit complicated way to to show this data but this is essentially a correlation between genes that are induced or repressed in in the control versus EE which is proton on this axis and the genes that are then changed if you want in either the bone marrow transportation group versus EE or the bone marrow this is our modified version with with with the plexigon group and as you can see in both cases there is a negative correlation in those genes saying that you know the genes that are induced by EE are rather repressed in in both treatment groups and the genes that are down-regulated in response to EE are rather up-regulated in other words these both treatment groups seem to rather normalize the gene expression pattern of the oligonucleides cells and which is of course nice and consistent with with a better myelination and can explain why in the overall there's a healthier you know brain and a healthier mouse so the exact same thing was true for astrocytes again remarkably heterogeneous population in our brains it's it's it's a lot of people think about the heterogeneity of of the cells in the in the brain now and again when we do the same type of analysis we see a rather a more normalization of the gene expression pattern in the astrocytes specifically however when we look now at the myelid cells specifically which i told you has actually the largest changes overall and also a very very heterogeneous we actually see the exact opposite so in contrast to the other glial populations the sort of microglial response to EE which is again is plotted here on the left on the on the y-axis is rather enhanced in our treatment groups which want to remind you are of course clinically better so the microglial response due to EE is not normalized or going to a normal microglial state in in these in these treatment groups it's rather makes the microglial reactive state more pronounced right so it's a stronger reaction so to say that we see in this in this microglial which is i think highly suggestive of the conclusion that this particular state this particular microglial response or reaction to the disease is actually beneficial because when we enhance this program we see a better clinical effect and we have a functional perturbation here because that is exactly the cells that we have boosted right through our through our transportation approach all right that's all i wanted to share with you today um this is a picture of the lab and you can see it's full of happy people i will point out that maybe it was a little cheating um the fact that really everybody smiles so well might be due because my 12 year old boy took this picture so it probably was a very funny thing to look at i should also briefly mention the names i think i mentioned yohei shibuya who has now recently left the lab who he really brought the microglial project into the lab young jinn was almost significantly responsible for this term two project and maris made up for the EE project and the first part of my talk was really spirited by a former student at shin who is now a poster in post thank you so much well that was really wonderful presentation marius thank you so much great basic research with profound clinical implications and we'll take some questions so let's take a couple from here if there are any and please go to the microphone and tell us what your question is i see shan going there that's great uh i was wondering in both the the induced neuronal cells and in your microglia from the bone marrow when you profile those do you see ghosts of their past in the in the transcripts are there are there parts of their identity that haven't been converted and do they fall into sort of regulatory aspects that maybe are intractable to the to the inductions yes excellent question i i didn't go into this but we looked early on at this very question very carefully and uh there was actually one of the reasons why for that purpose we we used a more defined donor cell type we used hepatocytes rather than fibroblasts which are not really well defined you know but hepatocytes you know pretty clear what it is and look very specifically at remnants of hepatocyte genes and we did see more expression of sort of a hepatocyte signature than you would expect but it was rather well repressed i must say and and we you know we we didn't drag it out too long so it it it goes down over time and i would assume you know the longer you wait the less the less you have but it was was not very prominent in that context with other people's work um what seems to be conserved though is what people call it some sort of aging signature so when you take a fibroblast from an old person and reprogram that to IPS cells and then turn these cells into neurons much of the sort of cell biological markers and signatures and transcriptional signatures of aging has disappeared so it seems when you go through an IPS cell reprogramming step you actually rejuvenate the cells and from our Mausburg we can also confirm because you know these IPS cells they make they make mouse babies right and they look very young even when you take them from adult tail tip fibroblasters or right so there is a sort of a cellular rejuvenation process going on in IPS cells it seems seems quite plausible however when you directly convert somatic cells into neurons with with our transcription factor methods not going through IPS cells then these features seem to be maintained other question was that your question other questions about the microwave microwave is a little different because these cells actually are not they're not identical as I briefly alluded to from a transcriptional point of view and a morphological point of view and some functional properties that they're distinct they're mild cells there are sort of other types of tissue resident macrophages not exactly microglia but you know something very similar but but distinct other questions from this audience may do we have any questions online she's a question though okay so thank you thank you for the talk so I was just wondering you know when I read the papers that you've presented here I was quite excited because we were working on some neurodegenerative phenotype but also the fact that the the CSF1R inhibitor that you used is actually FDA approved for something else it is I forgot to mention that yeah yes and so would it be possible to think about like inhibiting other cells in the brain and would you expect the same I guess competitive advantage of bone marrow cells you mean other cells yes like replacing oligonary sites with the same trick right huh interesting thought um potentially we um we probably picked the um sort of the lowest hanging fruit or the easiest population because microglia are known to repopulate really rapidly rapidly I'm not sure it's known in the human brain but in mouse brain literally in in days from a completely deplete microglia brain you have a fully repopulated brain and what we think is happening is that the transplanted cells and the endogenous remaining microglia are in a competitive situation in this repopulation essay or in this repopulation situation and the transplant cells they win the race and that's why the entire brain is full of these these grafted cells so that is a little harder to recapital it with cells that are much slower with a much slower turnover but in principle the same trick could work too yeah good point so apparently we need microglia to feed and interact some of the neurons in the brain is that interaction more for some types of neurons than for others or do all the neurons in the brain need microglia to the same extent that is a great question that is a great question and I don't know the answer to this um I was shocked to learn that uh short-term effects of microimplementation don't have much of an effect actually so when you deplete microglia with this drug for example and then do behavioral analysis um people cannot really find differences we know from human genetics in particular that long-term damage to microglia genetic damage right like term 2 causes problems over decades what these microglia do to nurture you know the environment in this chronic phase is I think a very exciting field of study that I would encourage everybody to work on yeah I can sort of imagine if you do enough bone marrow transplants which is essentially microglial replacement with different human diseases we may find that out eventually yeah good point Donna yeah I recently learned something interesting that I wondered how it plays into some of your work and that is that um microglia uh copper cells um some other uh cells or um actually uh embryologically derived from the yolk sac that's rather than uh the bone marrow and so and as such you know they are under perhaps genetic different transcriptional programming and it's only so they are derived from the yolk sac go to the liver and then from there uh we'll populate uh you know where the brain microglia will be and also in the liver where copper cells will be and dendritic cells in the skin for instance and then only later so I don't know how long after embryonic what embryonic day later you've eventually developed a bone marrow and then you begin to create some of those cells that are hematopoietic stem cells that will ultimately go to some of these same places and differentiate into those same types of cells but they're they're they're they're coming from different populations and and as such they're they're they interact we know that they interact but they're different uh and perhaps under different transcriptional programs and I wonder as you were talking about um the microglia that were that once you did the the transplant that they weren't um you weren't having as many chimera right in the brain in microglia I wonder if that's because um it's not normal for those cells to be coming so much from the bone marrow rather they're they're established there from this prior population in the yolk sac is that yeah something that's yeah everything you said is 100 correct as far as I know at least in mice there is some intriguing circumstantial evidence that perhaps in human brain the situation is very different um and even there is some some reports in mice that are not properly lineage traced unfortunately or transplantation based with with you know things like irradiation and so forth so they're quite perturbed um but there might be a bit more influx from peripheral cells into the brain that we think um at the moment because just uh you know the the studies on the yolk sector it is all through which it's very well characterized in mice and there's in the normal conditions you know not not not convincing evidence that in mice physiologically cells would go from the circulation into the brain but it's a it's a very interesting question and we actually thought since it's not it's not published we have to bite the bullet and do a proper lineage tracing experiment and wait for two years until the mouse gets old and see whether maybe in in in aged mice there is maybe more influx from from the periphery to the to the brain which may also be sort of regulated by aging or causing aging or something so may is there one last question from online yeah so from Yvonne actually so um wonderful talk with over expression of TREM2 have similar effect on the clearance of beta amyloid in the in the brain sorry the TREM2 exertion has over expression of TREM2 yeah would it have and the same effect on the clearance of amyloid in the brain oh rather than using like white type levels trying to overexpress TREM2 it's hard to know I think there's one study out there that tried that they're supercharged TREM2 and they used a humanized model and it's it's that paper claimed that over activating the pathway could could be beneficial even in like a TREM2 white type situation right so in a normal you know population when you just increase TREM2 function that could have a therapeutic effect but it's not our work all right this was spectacular thank you so much and really appreciate you're coming here and also talking to all of us individually so we really appreciate that next work thank you