Scientist Stories: Henrik Ahlenius, Direct Reprogramming for Studies of Aging and Neurodegeneration

Okay, thank you for the introduction, and thank you, Daniel, for inviting me, and for the organizing committee for allowing me to share a little bit about my views on how we can use directory programming for studying aging and neurodegeneration. Sorry. And I was asked to have a declaration of conflict of interest. This is just to declare that I don't have any commercial disclosures. The only disclosures that I have to make is that I'm not a geneticist, so please don't ask me any difficult genetics questions because I most likely won't be able to answer them. I'm a stem cell biologist that works with the brain. So then we can get started. So I always like to start my talks with this slide, and it's both to remind you guys and also to remind myself why I study aging, because aging is not always the easiest thing to study in the lab. So this is a depiction of a young, healthy brain, and this is how I hope all of our brains in this room look like. Unfortunately, this is where everybody is going. There's no escaping aging. Again, this is how a healthy adult brain looks like in a cross-section. And again, this is what I hope everybody's brains look like today. But we're quite many in this room, and it's quite clear that at least a few of us in this room will develop a neurodegenerative disease like Alzheimer's, Parkinson's, or any other dementia. And as I told you, I'm a stem cell biologist and work with regenerative medicine. So in the context of stem cells and regenerative medicine, why do I study aging? As I see it, any potential therapy, either if it's endogenous, so we're using the stem cells that are already in our bodies, or we transplant stem cells, it will be affected by aging. Either it's the environment where we place the cells or the capacity of the cells themselves. So, this is my reminder to you and to myself why I always keep on studying aging, although it's sometimes very difficult. So, we've heard now about iPS, and you've seen that it has an enormous potential for studying many different diseases. So, my question is then, how can we use reprogramming to study aging? In age-related diseases. And I'd like to name the following slides: the good, the bad, and the compromises, because not everything is as it appears always. So, let's start with the good. So, iPS cells, as Angel mentioned, has been derived from many different types of patients, and both healthy and deceased. And indeed, it has also been able to generate them from centenarians, so very, very old people. This is just one example from one study where they made iPS cells from patients ranging from 92 up to 101 years old. They showed these pluripotency markers that we've been seeing before, and they can also be differentiated into the three germ layers, and they express the genes that they should. At a similar level to young-derived iPS cells. So, that's all good. And it's also been shown that you can make, actually make iPS cells even from senescent fibroblasts. You can make it from these different progeria syndromes, so Hutchinson-Gilford progeria syndrome, Werner syndrome, dyskeratosis congenita, and also from centenarians as I mentioned before.. But what's sort of a little bit forgotten sometimes is that while— when you're doing reprogramming, you actually rejuvenate the cell. So you're taking it back to a very immature state. And that's actually what happens even when you do it from these progeria syndromes. At least the iPS cells in many cases are perfectly normal.. But when you differentiate them back to a somatic cell, they sort of gain these signs of aging. And this is the case for the progeria patients. But when you start with healthy individuals that are just old, and then you do the IPS reprogramming, and then you look at both of the IPS cells and the cells that you generate from the IPS cells, they are resembling more young cells than they are resembling old cells. So, how do you assess this? So, now we're getting into very critical things and very hard things to assess in cells. So, how can we measure aging in a cell? It's not trivial, but there are a couple of sort of hallmarks of aging. And we know on, at least on a global level, there's a decrease in DNA methylation and loss of heterochromatin. You have an increase in DNA damage in these gamma H2AX foci. There's a loss of the proteostasis function, so you have an increased aggregation of protein, and heat shock protein function goes down. And in the mitochondria, you have increased reactive oxygen species and an increased DNA mutations. And when we look at the nucleus, we see nuclear disorganization, lamin A/C goes down, LAP2 goes down, and the telomere length also goes down, and the percentage of cells with short telomeres is increasing. And if we look at these classical senescence markers like senescence-associated beta-gal and p16 levels, they also go up. So what happens then when you take an old fibroblast, or a fibroblast from an old either healthy or deceased patient, reprogram it into iPS, and differentiate it back into, for instance, fibroblast or any other cell type that you are interested in looking at? Are they really rejuvenated? So we look at then these markers again, these hallmarks. And indeed what we see is levels of senescence goes down, the mitochondrial function goes up, goes back up. You have a reduced reactive oxygen species. Telomere length goes up and the amount or the percentage of the cells with short telomeres goes down. You regain heterochromatin and you have less DNA damage and the nucleus again, has a normal organization. And as I said, I put that here also that this is not true when you're working with progeria, when you have a very aggressive progeroid mutation. Then you still see some aging phenotypes in the cells. But when you start with either a healthy old fibroblast or a disease where there's no direct progeroid mutations, for instance, they don't show these signs of aging. So, given this, how do we still use iPS for modeling of age-related diseases? Since we know now that iPS and iPS-derived cells are rejuvenated, and many neurodegenerative iPS disease models, we can detect some findings in these disease models, but they are mainly the genetic forms. And, for instance, I mean, we've heard of something from Parkinson's. You can see some aggregation, and in Alzheimer's you can see sort of different levels of AB42, but in general they're quite mild phenotypes. So then the question comes, can we actually somehow induce aging in iPS-derived cells. And one option is, of course, just letting them grow for a very, very long time or put them into an animal that we let grow old, basically. But there's also a couple of different other variants that we can do. And so, the goal here is to— when we take cells from— this is an example of Parkinson's disease When we take a Parkinson's patient fibroblast, reprogram it to iPS cells, they become sort of young and healthy again. You get young and healthy neurons, except for having the underlying Parkinson's pathology. So, how do we go from here to something back to something that resembles the neuron that was in the patient where we derived the fibroblast to start from? So, we have to induce aging in some way. And there's a couple of different ways of doing that. You can, of course, stress the cells with ROS or pesticides or any other toxin. Toxins. You can activate signaling that we know are increased during aging. But I want to talk today about two different approaches. And one is progerian-induced aging, and the other is altogether a completely different approach where you actually skip the iPS stage and take the fibroblast directly to neuron with bypassing the pluripotent stem cell stage. So, this is work from Loren Studer's lab, which is also working with Parkinson's disease modeling using iPS cells. And when this first came out, I thought it was really a brilliant idea. So, basically what they did was they took their iPS cells, differentiated them into dopaminergic neurons,, and then they transfected them repeatedly with mRNA of the mutant form of lamin A, so the progerian, the mutant progerian basically. And what they observed was quite striking that both from young and old iPS, young and old derived iPS cells, they started to see these nuclear abnormalities shown here both with lamin B staining and in GFP. You see that the nucleus is sort of has a disorganized envelope. And they also started seeing increased DNA damage and also increased mitochondrial ROS. And of course, since they're studying Parkinson's, they of course went to their mutant iPS cells. So they had cells from familial forms of Parkinson's disease, and looking here first at apoptosis. So they couldn't really see any difference in the normal iPS cells or iPS-derived neurons between their control lines and their familial forms of Parkinson's. But when they add then this progerin to these cultures or to these cells, all of a sudden you start seeing these changes between the wild type and the mutated lines. This was also true for when they were looking at neuronal morphology, looking here at the total dendrite length, and also at AKT signaling, which is proposed to be one of the pathogenic mechanisms in Parkinson disease. So, in all of these, they didn't see any changes when comparing sort of normal IPS cells and diseased IPS cell-derived neurons. But when they add progerin to the cells, these phenotypes that were before hidden is now visible. But there's one very important thing here to note, that lamin A or progerin is never expressed in a normal neuron. So this makes the system quite artificial. It's basically not studying Parkinson's alone anymore, it's studying Parkinson's on a progeria background, basically. But it's still very useful to reveal things that we were not able to do before. So, to summarize this first part, it is indeed possible to generate iPS cells from centenarians, and we can also generate iPS cells from progeria patients. However, many hallmarks of aging are abolished during this process., and then progerian reestablishes aging hallmarks and reveals disease phenotypes. So, then next I'm going to talk about— so this is a little bit different from the iPS reprogramming, but it still uses the similar approach. But instead of going through the induced pluripotent stem cell stage, we take the fibroblast directly to a neuron in this case. And I just wanted to give a short introduction to how I came to study direct neuronal reprogramming. So, during my PhD, I was interested in how neurogenesis is regulated in the aged brain. And we were able to show what was already known at that time, that the formation of new neurons— so, neurogenesis, the formation of new neurons from neural stem cells— is severely decreased in old animals. But we were also the first to show that you actually lose neural stem cells from the brain during aging. But when you take these cells into culture and let them proliferate, they are still able to make functional neurons. They don't do it to the same extent as the adult or the embryonic, but they still make functional neurons. So this was when I was finishing my PhD and started looking for places to go for doing a postdoc, and then this paper came out. So Maurice Wernick's group at Stanford took sort of one step back from the IPS reprogramming and indeed took— starting with a very similar approach to IPS disease or IPS generation, they started with fibroblasts, in this case from a reporter mouse where all neurons are green. They threw on the kitchen sink of transcription factors in the lab and waited for a couple of weeks, and indeed they found that you can actually make neurons directly with a semi-strategy as making iPS cells, but they're making them without going through a pluripotent stage. And the factors that was able to do this was Ascl1, Brn2, and Myt1Like, so again a combination of 3, 4 factors that converts cells. So, I must admit that at this point I was quite skeptical to to this direct conversion, as it was termed at that point. But I saw a very huge potential using this to study aging because you would skip this whole rejuvenation you do when you go through an iPS stage. So, I actually ended up doing my postdoc with Marius. And for me, it was quite clear that what I wanted to do is to check, can we make induced neurons from aged fibroblasts? And there's a couple of questions then that arises. First, can we reprogram aged fibroblast neurons? Because it's known from the iPS cells that the older the fibroblast is, the harder it is to make the iPS cells, although it is possible, but it's harder the older the cell is. So, are there any differences in efficiency of reprogramming, or are there any morphological or functional differences? And can we ultimately use this to to model neuronal aging, which is sort of my ultimate goal in life. So, I wouldn't be standing here talking if this didn't work, but it was a long process. But in the end, it worked. So, when I came to the lab, I gathered all the sort of old mice that were still sitting around in the animal house that nobody was using, found a couple of these tau-GFP mice, threw on the transcription factors again, waited, and eventually— so we had to tweak the protocol a little bit, but essentially it's the same protocol as the published one. So we could see these green cells turning on, and they also stain for panneural markers like beta-3-tubulin. They could indeed generate functional neurons as well. So this is looking at action potentials, trains of action potentials, and also the sodium and potassium currents. So, it means that we can actually make the functional neurons. So, the question now came, so are there any differences in sort of the kinetics of the reprogramming? So, we isolated a whole bunch of fibroblasts from everything starting from the embryonic fibroblasts where reprogramming works very well, to postnatal, to adult, to middle-aged, and to very old mice. [COUGH] So we even generated functional iNs from mice that were close to 3 years old, which is in a mouse, a very, very old mouse. And we looked at the efficiency at how this reprogramming occurs, and the MEFS is very efficient with around 20% of the fibroblast starting population turning into neurons. Then it drops dramatically down to postnatal, and a further drop down to adult, but then interestingly it sort of stays the same at the same low efficiency from adult to middle-aged to aged fibroblast. Of course, we wanted to check if the function was different, so we again took all of these different time points and then did electrophysiology again dividing the cells up in cells that fired either multiple action potentials, single action potentials, or failed to form an action potential. Sorry. And as you can see, the embryonic fibroblast is superior. They, all of them, generate functional neurons. Postnatal is pretty good. Then there's a further drop down to the adult. But again, there's no difference with the adult or a very old mouse in the efficiency in making functional neurons. And this is just some quantification of some intrinsic properties, and basically seeing the same embryonic and postnatal does a little bit better, then it's a drop down to postnatal, and then it stays the same throughout age. The really cardinal sort of The primary function of a neuron, though, is to put out and receive information, and the way to test this is to look at synaptic activity. So, we basically generated our induced neurons, put them on a primary neuronal culture, and asked if they can receive synaptic input. And indeed, as you can see, so this is looking at spontaneous AMPA-mediated excitatory postsynaptic currents, and as you can see, the maps are really nice. It's not like primary neuronal cultures, but it's quite close. The postnatal, again, does pretty good. Then the adult is a little bit worse, but they still do it. And again, we have this interesting finding that there's no real difference between an adult and an aged fibroblast-derived iON in the amount of synaptic input they get. Sorry. So, it's been shown before from the iPS reprogramming that senescence and proliferation sort of really affects the efficiency on how you make your iPS cells. So, we looked at our aging fibroblasts both on these aspects, and indeed proliferation, as expected, goes down dramatically. Senescence goes up., and also the— so that's the senescence-associated beta-gal expression, but also expression of these senescence markers p16, p19 goes up with age. And another peculiar thing that we saw is that infection efficiency also drops with age, and also the the way of how easy it is to reculture the cells also goes down with age. So, taking all of this together, although we have a quite low efficiency, we're probably underestimating the efficiency just because it's harder to infect the older cells. They are less proliferative. But nevertheless, we wanted to see if there's any way of sort of improving this because, again, we're getting just a few percent of the starting population turning into functional neurons. So we, at that time, the Anne Bernay lab had just published some quite interesting things on FOXO3, which is one of these pro-longevity factors, and they showed that it was essential for maintaining neural stem cells during aging. And so we started looking at maybe FOXO3 plays a role here. So the first thing we did was to Within our sort of reprogramming cocktail of this Brn2-Ascl1 and Myt1Like, we added FOXO3 to the cocktail, and as you can see, we completely abolish, or almost completely abolish, iN reprogramming. And this is just quantified here, and what you probably can't see in these small images was also that the complexity of the neurons was much less. The few neurons that we got was much less complex, indicating they were less mature as well. And then we got some knockout animals from Ann's lab, started looking at how to generate ions from those. Surprisingly, we couldn't really see any significant difference in the efficiency of making these ions from the knockout compared to the wild-type. But since my focus was always aging, and FOXO3 is a prolongevity factor, I wanted to look, of course, in the adult and the aged as well. And here we started to see something really interesting. And when we look at the conversion efficiency from the knockout, it's actually increased in the knockout only when we look then at the adult and the aged. And this also works when we use a short hairpin against FOXO3. But what was really striking was— so again, when we looked at the the functional aspects of these cells. So, again, there was no difference between the MEFs, between wild-type and knockout. But when we looked at the adult and aged, and as you see, we went up to even 27-month-old, efficiency of generating functional neurons was much, much better. So, not only that we get more neurons, but we also get better neurons when we knock out FOXO3 from this cocktail. So, to really understand that, we have to know a little bit more about the sort of the mechanism behind induced neuron conversion. And the lab was able to show that ASL1, one of these factors, is really the key target here, and it's an on-target pioneering transcription factor that drives the ion conversion. And subsequently, we were also able to show that Ascl1 is actually alone sufficient to form neurons from fibroblasts. So, what does then FOXO3 do? So, let's see if this slide works. I had some problems with this slide this morning, but hopefully it works. So, Ashley in Ann's lab had started looking at what FOXO3 does on a sort of a global genome level, and she found the sort of consensus sequence and this, the binding motif, and saw that it was really this E-box motif was really enriched at other pro-neuronal target sites in the genome. And when doing ChIP-seq comparing Ascl1-bound genes and FOXO3-bound genes, she found a remarkable overlap that FOXO3 binds really at ASL1 locuses. And given that ASL1 is one of the major drivers here, we start seeing some link to the results that we have. She was also able to show that FOXO3 can actually inhibit ASL1-induced expression of downstream targets in the Notch and Wnt signaling pathway, which is very important for neural differentiation. So, just to summarize the second part, yes, we can actually make functional iON cells from aged fibroblasts. And I still believe that with increased efficiency, we can use this as an alternative to iPS cells to model neuronal aging. And we've shown that FOXO3 deletion increases efficiency and also, very interestingly, the function of induced neurons from adult and aged fibroblasts.. And this is likely due to an inhibition of Ascl1-induced transcription by FOXO3. So, for the last part of the talk, I wanted to tell you a little bit— so now we're switching back to the pluripotent stem cells because I wanted to give you my view on how we are using pluripotent stem cells in the lab to model both neurodegenerative disease and also for modeling aging. And my approach is a little bit different. We these days primarily work with human embryonic stem cells, but you can of course do this in iPS cells because, as we heard before, ES and iPS cells are more or less the same, at least when we're comparing human ES and human iPS. So, I use gene targeting. Instead of taking fibroblasts from patients, generating the fibroblasts or the iPS cells, characterizing the mutation, correcting the mutation, I start with the mutation and introduce it directly into the ES cells instead, and sort of skipping the generation of the iPS cell stage. But the idea is the same still. But then I'm not primarily interested in ES cells.. But since I'm studying the brain, I want to study the function of these genes and these mutations in the relevant cells. So, how do we go about doing this? So, in my old lab, we started working a lot with gene targeting, and at that point we were using both conventional targeting, which never works, and we also worked with AAV-mediated gene targeting, which sort of works, but it's very low efficient. But then, as soon as the CRISPR-Cas9 field started, we switched completely over to that, and that's the only means of gene targeting that I do in the lab at the moment. And since the CRISPR technology will be extensively talked about at this meeting, I'm not going to go into detail, just saying that we're using this Cas9 that is guided to the DNA by a guide RNA and then makes either a single or double-stranded cut depending on which version. And then it's either repaired by non-homologous end joining or by homology-directed repair if we provide a repair template. And so this is just a showing our favorite experiment at the moment where we tried to do the very simplest thing. So we basically just induced the double-strand break, let the cells repair the break on its own, and since non-homologous end joining is sort of not foolproof, it causes a lot of insertion and deletions. And this is Surveyor, assay. So, here we have picked 12 clones where we've mutated our— or made these double-strand breaks in our favorite gene. And in the survey assay, so basically you PCR up your gene of interest or the region of interest, denature it, and re-anneal it. And if you have a mismatch, if something happened, you will have a heteroduplex and they will be cut. So every time you see multiple bands here means that we have some form of editing occurring. And as you can see, out of these 12 clones, at least 11 of them had some editing. So this is, of course, an extremely successful experiment. It's not always this efficient, but if you screen a few guide RNAs, you will have something that works for sure. But of course, we want to know exactly what's happening, so we take the PCR fragment clone it into a TOPO vector, pick colonies, and send for sequencing. And this is sort of again looking at our favorite gene. So this is the ATG, and this is the guide RNA that we have introduced, and this is where the cut site is. And here is just a representation of what type of editings we get, and we get everything from from very small deletions to medium-sized deletions to very long deletions, small insertions, medium insertions, and fairly large insertions. And so this is not now 67 different lines, but it's 64— 67 different clones from these 12 initial clones. And in 64 out of these 67, we actually find something. And if you're lucky, in around 50% of the cases, you get a frameshift and a disruption of the gene, basically, and basically a knockout. But, of course, this is all fine if you were studying iPS cells or ES cells, if you're studying mechanisms of pluripotency, but I'm interested in neural cells. So, sort of the classical way these days of doing neural differentiation is the dual SMAD inhibition, which, as you can see, is sort of a quite tedious process. Isaac did this for the project that we've seen presented before, and he's not so happy with this method. It works very well, and it is sort of still the gold standard, but it is time-consuming and yields in the end, a mixed culture of neurons. So this is another thing that we were working with when I was in the US and trying to develop a better method of getting a more homogeneous population of neurons to study from these gene-edited ES cells. And we were able to generate a protocol that is much more simple and faster than the sort of normal, this dual SMAD inhibition protocols. And what we did was basically hijacking the iN technology. So, instead of using a normal differentiation, we select neural or transcription factors that are very potent in inducing neural fate and use similar to the iN experiments, use lentiviruses to express these. Instead of using it in the fibroblasts, we do it in the ES cells.. And this is a typical protocol. And as you can see, we have neurons already at 6 days, and at 2 weeks they are already functional. They receive synaptic input and can also output— also have output. We can even transplant these cells into the neonatal mice and have fairly good engraftment. And at 6 weeks after transplantation, they are actually becoming functional in vivo, which is really, when you're talking about ES cell-derived neurons, really, really fast. So, the goal of this was, of course, always to combine it with the gene editing to reveal some interesting function in neurons. And my old lab has been very successful since in combining this these two methods. Although I should say here that gene editing in these papers have been either conventional targeting or with AAV, but they are also now switching over to CRISPR-Cas9. And they've been able to show gene function in neuropsychiatric diseases, autism, and also just looking at synaptic-related genes. So, this is quite an attractive approach.. But what about glia? It's a famous quote. And white matter matters. So, we know that many neurodegenerative diseases have involvement of glial cells. And it's sort of been a little bit neglected so far. It's starting to come up. But we need, again, better methods to generate the glia. So, this is again two gold standard methods of making both oligodendrocytes and astrocytes from human ES and iPS cells. And if you look here, this is the time it takes to get to the OPC stage. So, 120 and 75 days. We were not so keen to wait that long in my lab to get these cells. So, we thought maybe we can do something similar to what we did with these again hijacking the directory programming, applying it to ES cells, but now doing it with the glial cells instead. And again, we took a very similar approach as before. So this is now data from my lab, starting with ES cells, infecting them again, but instead of using neuronal-specific genes, using glial-specific genes. Select for these and wait and analyze. And indeed it works. We can get these beautiful astrocytes in as little as 10 days. And if you compare that with 120 days, I think this is better. I don't know what you think. I'm pretty happy with this. It is very efficient, as you can see here, staining for astrocytic marker called S100 beta. You can see that almost all our cells in the end are S100 beta positive. You can also have these beautiful cells expressing the cardinal markers, Vimentin, S100 beta, and GFAP. And this is after around 2-3 weeks. So, with that, I'd like to summarize this third part. So, I hope I've showed you that CRISPR-Cas9 gene editing can be very efficient, and that transcription factor-driven reprogramming— so, I tend to call it programming rather than reprogramming because we're starting with a cell that's main job is either to stay undifferentiated or actually differentiate. So, it's not really reprogramming, but we're using a reprogramming approach in ES cells. And I really feel these are superior to normal differentiation. It doesn't exclude that we will use normal differentiation as well. Sometimes reviewers ask us to do normal differentiation as well. Basically, I think we end up at the same product, just it's faster when we do it this way. And, of course, combining these for modeling and neurodegenerative diseases and aging is, in my mind, extremely attractive. I'd like to thank then the people in my lab that has done the work. But of course, I also have a lot of what I presented was also initiated at Stanford together with Marius, Tom, and Ann. And Soham was instrumental in the iN experiments that I showed. And Jingxia was the one who was really the driving behind developing the first iN from ES cells. This is my funding. And since this is also an educational session, I'd like to bring your attention to that we have a very good research school in stem cell biology in Lund, and I happen to be the co-director. And we have programs both for master's student levels, PhD students, and postdocs. So, if you have If you are a master's student, check it out. If you have master's students in your lab, please make them aware of this. And with that, I thank you, and I think both of us are happy to take any questions.

Thank you very much for this nice presentation. Angel, please, as we said, we have run out of time, but We have time for a few questions. We have microphones on both sides of the room. You have microphone over there and we have some on side.

Yes. It's ready?

Yes, please.

Hello.

Is it on?

Oh, hi.

Thank you both for really excellent talks. Since this is an educational session and I don't want to take too much time, just wanted to point out Mike Sheldon at Rutgers University, by the way. To Angel's point about cost and sharing, there are a number of resources available globally for IPS collections that are very well characterized both clinically as well as biologically, and we at Rutgers University maintain two of them for the NIH Institutes for Mental Health and Neurological Diseases and Stroke, and I'm at booth 248 if you'd like any information about how to obtain cells from these collections. They're already made, and they come with extensive clinical information, so I think they might prove to be very valuable resources for some of the people in this room.

Thank you. I have one question over here. Well, first, congratulations for these excellent lectures. Can any neuronal cell type potentially be generated from either fibroblasts or human iPS cells? That's my first question. And my second question is particularly looking at the several neuronal cell types, like granular neurons or Purkinje cells, to study cerebral neurodegeneration.

Thank you. So I think the quick answer is yes. Any cell type, any neuron type can be made, because we see that when we make mice, So they are born with all the neurons in place. Long answer is, yeah, we can, but for most of them we don't know how to make them. It's just a matter of finding the protocol.

This forcing the differentiation by overexpressing in a controlled way key transcription factors is a way to go. If I understand correctly, it means that for Purkinje cells, currently there is not enough knowledge to differentiate them because of the—

no knowledge on transcription factors, no? Am I correct? Yeah, for 3 or 4 types, specific types of neuron that people have been interested in and investing time and effort, there are protocols that work quite well. Some other neuron types, I don't think there's still protocols working, but that doesn't mean that the cells have not the capability to differentiate into those cell types.

It's just that we don't know how to do it in the lab.

Okay, thank you. I can comment on that as well. And I think that what one needs to do is really go back to development and study the development of the neurons that you want to make. And to sort of really find the cues during normal development and how these develop. Because that is how the dopaminergic neurons are the ones that we are best at making from iPS cells that are really bona fide dopamine neurons. And a lot of the protocols are really dependent on the cues that are in them during normal development. So if you find the developmental cues, you will be able to make it.

Okay, thank you very much. You have time for the last question because we are running out of time.

I'm Polata Mikstienia from Vilnius University, Lithuania, and I have a question for Professor Angela Rea. My question is, have you tried your technology or do you use your technology of reprogramming for the functional analysis of variants of unknown significance identified by a new generation sequencing, for example, exome sequencing? Can you repeat the last part of the question? Repeat the question, please. Have you used your technology for functional analysis of variants of unknown significance identified to the patients by new generation technology? If you have a known significant variant and you know to find out if they are pathogenic, and do you use your technology for this purpose?

Yes, we have. I didn't show any. Many labs around are doing this type of approach in combination with the efficiency that is provided in CRISPR gene editing, CRISPR-mediated gene editing. So we are testing using CRISPR to test variants that have been identified in next generation. We don't have any— publicly result so far, any publicly disclosed result.

But yes, it works quite well. Thank you. And another question I'm curious about, do you have any markers or some features or phenotypes of the cell that then you can distinguish that this cell differentiation step is, for example, equivalent to prenatal features of the cell or old cell or young cell or some different cell of human life? Because some pathologists start at the prenatal stage of development and others start later.

And so do you have any features or markers to differentiate this type of the reprogramming cell.

More than reprogramming, if I understand correctly, the differentiation of the iPS cell?

Differentiation of somatic cell from iPS cell. Yes, yes, we do have markers that will tell us, not specifically for neurons but for other cell types, whether this cell that we generate in the lab is representative of more postnatal, early postnatal, or prenatal fetal phase or adult mature cells.

Thank you for your good presentation. I can comment on the last as well. And I think that the really crucial point in when— whether we're generating the cell that we want from an iPS cell or through direct conversion from somatic cell directly to the cell that we want, The crucial thing is to look at function, and that will also tell you how, sort of, at what developmental stage it is.

Okay, thank you very much. I think we have finished the session.

Thank you to the presenters.