Reprogramming Stars #2: Reprogramming Towards Neural Lineages—An Interview with Dr. Henrik Ahlenius

Introduction by Dr. Carlos-Filipe Pereira (Editor-in-Chief, CELLULAR REPROGRAMMING)

Dr. Pereira: Good afternoon. My name is Filipe Pereira, associate professor at Lund University and editor-in-chief of Cellular Reprogramming. I'm very happy to bring you the second interview of Reprogramming Stars, our flagship series capturing the findings, projects, and ideas of the leaders in cellular reprogramming. Today we have Dr. Henrik Ahlenius, group leader at Lund University in Sweden. His lab aims to understand how aging and neurodegenerative diseases affect the brain, neural cells, and neural stem cells.

Henrik did his PhD at Lund University, where he studied adult neural stem cells, and postdoc at Stanford University in the US, at Marius Wernig Laboratory. In the Wernig Lab, Henrik was part in pioneering the field of reprogramming towards neural lineages. He was then recruited back to Sweden, where he has been awarded prestigious startup grants from the Swedish Research Council and the Swedish Society for Medical Research. This allowed him to further develop and use the reprogramming technologies to study aging and neurological disorders.

A quick reprogramming note: Henrik just finished editing a book on neural reprogramming, an effort that we all find critical to transfer knowledge to the next generation of reprogramming scientists. Dr. Ahlenius, thank you so much for joining me today.

Dr. Ahlenius: Thanks for having me. It's my pleasure.

Dr. Pereira: It is a pleasure to have you as our second Reprogramming Star. So, your lab has great expertise in neural reprogramming. Not so long ago, you published a very nice paper in Nature Methods entitled “Rapid and Efficient Induction of Functional Astrocytes from Human Pluripotent Stem Cells” (Canals et al., 2018). I find the concept of forward reprogramming very interesting, and I wonder if you could tell us how your journey started in the cellular reprogramming field?

OPEN IN VIEWER

 Dr. Henrik Ahlenius

Reprogramming Star : Dr. Henrik Ahlenius is a principal investigator at the Department of Clinical Sciences and Lund Stem Cell Center at Lund University in Sweden. The Ahlenius lab uses novel stem cell, genome engineering and reprogramming technologies to study how aging and neurological disorders affect the formation and function of neural cells. The lab has recently been focused on reprogramming towards the astroglial lineage for investigating the role of astrocytes in neurodegenerative disease.

Dr. Ahlenius: It was right about when I was finishing my PhD when the whole reprogramming field really exploded. I have to say that I was quite skeptical towards reprogramming in the early days. I couldn't really see the value of it. I guess I thought it was too artificial. But I soon realized the potential of reprogramming, especially for the research questions that I have, towards understanding what is going wrong in the adult and especially aging brain with the different diseases that comes with aging. So, at the end of my PhD, I realized that I had to learn more about epigenetics and reprogramming. It was becoming so important in the field and I felt that this was something you cannot miss.

It ended up that I went to Stanford, at Marius Wernig's lab, for my post-doc to learn more about reprogramming. This was when the first direct conversion from fibroblast to neurons papers were coming out of his lab. In Marius's lab, I was studying neuronal conversion of aged fibroblasts, because aging is one of my main interests. First of all, we established reprogramming from aged fibroblast, but we also found out how one of the longevity genes, FOX03, regulates this process (Ahlenius et al., 2016). But we also did a lot of other interesting stuff in Marius's lab. We developed direct conversion to neural stem cells and oligodendrocyte progenitors. We studied the transcription factors that are involved in the conversion process, how they function in induced and normal neurogenesis (Mall et al., 2017). And of course, we also worked with induced pluripotent stem cells (iPSCs) and disease modeling as well.

So, I guess that was the start of my journey. I was skeptical in the beginning, and it's kind of funny that I now have a lab that's more or less completely dedicated to cellular reprogramming.

Dr. Pereira: I completely understand the feeling. I still remember when that first neural reprogramming paper came out. Looking back, I don't think no one would anticipate that it would ignite a whole field - neural reprogramming - where people try to generate different types of neurons or progenitors and define their identities with this direct reprogramming approach. This lead us to disease modeling, and many other interesting applications that I am sure we will learn more about today. But it would be interesting to hear more about your Nature Methods paper. What were your main findings there, and what was the main contribution to the field?

Dr. Ahlenius: Another project we were working on in Marius's and together with Tom Südhof's lab was to get better protocols for differentiating human pluripotent stem cells. To do that, we applied the logic of direct conversion, but instead of using fibroblasts, we used the same transcription factors that turned fibroblast into neurons, and applied them to human pluripotent stem cells. To our surprise this approach very efficiently drove differentiation towards neurons (Zhang et al., 2013). I was part of developing two of these protocols, one to excitatory and another to inhibitory neurons. These protocols were also the ones that I started using, when I was starting my own lab, for modeling neurological disorders.

But we fairly quickly understood that we needed to be able to make human astrocytes as well. In fact, we had a disease modeling project where we were expecting to see quite striking phenotypes, but we couldn't detect anything! We started to think that this could be explained by species-specific differences, we were using mouse astrocytes, not human astrocytes; this was the gold standard at the time, before we had efficient protocols to generate human astrocytes.

So this paper started off as a side project, and my postdoc at the time, Isaac Canals, jumped on this and really made this happen. We were a bit lucky as at the time, Vania Broccoli's Lab, had just identified factors that were able to convert mouse fibroblast to astrocytes. However, this was not really efficient in human cells. I then thought that if we can drive neural differentiation from pluripotent stem cells into neurons so efficiently with transcription factors, it should be possible to make also astrocytes, right? So we tested the factors identified in Vania's lab and a couple of others and applied them to human pluripotent stem cells instead. It turned out that this was really, really efficient in driving astroglial fate. We then went on to define the minimal combination of transcription factors that was required to have an efficient induction of astrocytes from pluripotent stem cells.

Dr. Pereira: What were those transcription factors?

Dr. Ahlenius: We initially started with NFIA, NFIB, and SOX9. And then we came down to that NFIB alone, or NFIB together with SOX9, which is the combination we prefer, as sufficient to drive astrocytic differentiation. And it's quite interesting, because almost at the same time as our paper was published, there was another paper that came out from Lorenz Studer's lab, showing that NFIA is important in the gliogenic switch, but not necessarily in mature astrocytes (Tchieu et al., 2019). So, this is probably the reason why SOX9 and NFIB, or NFIB alone, are the best astrocytic reprogramming factors.

After having established these factors and looking closer at the process, we could see that after just a couple of days the cells changing morphology, and they started expressing the phenotype of astrocytes. But then, by taking this a bit longer, we could really see that these induced astrocytes, were really behaving phenotypically, transcriptionally and, most importantly functionally, as bona fide human adult astrocytes. This is quite remarkable, because in normal differentiation, if you use developmental cues and use standard differentiation protocols, it takes months to achieve mature functional astrocytes.

Then we finally used the protocol that we had developed and combined it with CRISPR genome engineering, to model a rare disease called Alexander disease, where mutant GFAP, which is an astrocyte-specific protein, causes the disease. We introduced one of these mutations in human embryonic stem cells using CRISPR, and then combined it with our transcription factor programming protocol to make astrocytes. Using these together, we could detect pathological changes in astrocytes, which to us proved that our protocol was good for disease modeling. Alexander disease is a good model disease to study astrocytes, it's not a disease we intended to start with, but now it has become a focus in the lab.

Dr. Pereira: Do you envision that the main use of these new astrocytes you can now generate is for disease modeling? Or does it also open therapeutic avenues?

Dr. Ahlenius: Well, we are exploring the use of our protocol for therapeutic purposes, but the main interest for us was to be able to have an efficient way of producing human astrocytes. Because again, it's very difficult to get them from patients. You're limited to postmortem material or resection material, which is not really an ideal cell source. So, our primary goal was to use these for disease modeling but given that it's so hard to get hold of human astrocytes, I think they will be also very interesting to study basic astrocyte biology.

Dr. Pereira: Oh yes, this will definitively be interesting to explore. Now, I'm curious about the development of the whole subfield of forward reprogramming, or forced differentiation, as it has been named. So, if we zoom out and think about other lineages… could you give us an overview of the technologies and lineages specified with this idea of forcing transcription factor expression in iPSCs? the main advantage of this approach would be to drive iPSCs specification into lineages, speeding up development and resulting in high degree of homogeneity, right?

Dr. Ahlenius: So this is something that has been around for a while. People have used it, not so much in pluripotent stem cells, that's fairly new, but for instance in neural stem cells, to drive differentiation with various factors. But the field really kicked off after establishment of direct conversion protocols, because it's often similar factors that drive direct conversion that also works in forward programming. This is now pursued along many different lineages beside the neural, such as in the hematopoietic, muscle, pancreatic and liver systems. There have also been recent high-throughput studies were large amounts of transcription factors have been screened for their ability to induce different fates in iPSCs (Ng et al., 2021).

One has to bear in mind that forward programming is somewhat artificial. A colleague at Stanford actually called this cheating, if you compare it to normal differentiation, but it is a much easier and faster way of generating functional cells. The way I see it is that you give the cell no other option than to differentiate in the direction that you want, and then you also push to speed.

Dr. Pereira: Taking the example of astrocytes, how many days do you save with forward programming? comparing a regular morphogen-based differentiation versus the forward reprogramming approach with the NFIB and SOX9.

Dr. Ahlenius: It depends on what level you put it, but we see that our cells are more or less functional after a couple of weeks. They start showing the immunophenotype of astrocytes, already after a couple of days, and after two weeks they really look like mature astrocytes. After a couple of weeks, we think that they are already mature, but then of course some aspects will improve with time. But it's a matter of weeks. And if you look at the sort of traditional protocols, that varies from around a month, two months, up to extreme cases where people have had beautiful studies in organoids, for instance, but then it can take up to a year to get fully mature cells in organoids. I think it's a matter of switching from months to weeks. Let's put it that way.

Dr. Pereira: That's very impressive! forward reprogramming could then be explored to generate human cells at scale for therapeutic application, right?

Dr. Ahlenius: One has to be aware that this again is a bit artificial, what we and others typically do is we're overexpressing these factors with lentiviruses. So, I still think that for instance, for cell replacement therapies, probably the standard or more normal differentiation with developmental cues might be better, or at least they will be easier, because you won't have always this sort of fear of where does these lentiviral vectors land, and will you have some insertional mutagenesis, and so on. But for us, it's really an efficient tool for disease modeling. But of course, having identified which transcriptional regulators that are important for driving the fate, one can go back to more standard differentiation paradigms and make them better.

Dr. Pereira: Of course, if you understand the rules of that particular differentiation path, then you can think about small molecules or external cues to improve efficiency or to drive astrocytic specification without viral vectors, right?

Dr. Ahlenius: I mean, if you would find out, then small molecules that would efficiently activate NFIB and SOX9, for instance, that would be a way to go forward.

Dr. Pereira: Yes, that's a very interesting thought. I'm wondering if you could tell us a bit more about what you're doing now in your lab, and what are your project and avenues that you're exploring?

Dr. Ahlenius: We are of course working with our induced astrocytes and using them in our disease modeling efforts. And I mean, that's why we developed this in the first place. And at the moment we're focused on frontotemporal dementia and astrocytopathies, demyelinating and neurodegenerative disorders that starts and the mutations actually sits in the astrocytes, and not in the neurons. And in fact, we now actually see a phenotype in the disease model that I talked about earlier, where we initially used mouse astrocytes and couldn't see anything. All of a sudden, now when we're using these human induced astrocytes, we're starting to see very interesting phenotypes. But we're also of course testing reprogramming into astrocytes from different cell sources, like human fibroblasts. And also, as you were asking, we're thinking about how we can use this for potential therapeutic strategies.

Dr. Pereira: You mentioned a disease phenotype for the Alexander disease model, is there another disorder you are currently exploring?

Dr. Ahlenius: The Alexander disease is quite clear, because it's known that the phenotype arises from astrocytes. Then it's not really known how that affects myelination, which is the main phenotype in patients. So this was a project where we were modeling frontotemporal lobe dementia, actually. And there has been studies, and that's also why we started developing this protocol, that in more neurodegenerative disorders, the role of astrocytes is becoming more and more recognized. So that was the start of it. And now we see, when we co-culture neurons and astrocytes from our mutated pluripotent stem cells, we start seeing phenotypes that we didn't see before when using mouse astrocytes.

Dr. Pereira: You also mentioned several times the dichotomy of mouse and human. How come astrocytes are so different across species?

Dr. Ahlenius: Well, I think this is specifically, but not exclusively, very evident in, astrocytes, because mouse astrocytes and human astrocytes, they're really vastly different. One thing is just the size. It's just enormous difference in size. And given the size, it also connects with much more neurons than a mouse astrocyte would. So it has a bigger territory and a larger interaction with neurons compared to mouse.

Dr. Pereira: For the non-specialists, can you summarize the main functions of astrocytes?

Dr. Ahlenius: The most well-known function is of course that they take up excessive neurotransmitters. So they regulate excitotoxicity, and then. they have these endfeet on blood vessels, and take part in the blood-brain barrier. They are key players in immune functions in the brain, so they get activated upon injury or disease, and are inflammatory cells in the brain. And they of course form the glial scars around the injuries. They are also involved in metabolism in the brain as well. So as you can see, they have really diverse functions, right? they can interact at many different levels. That's why we also think that they are very interesting to study in neurological disorders.

Dr. Pereira: I'm very happy to see that cellular reprogramming has opened so many possibilities in the astrocyte field. So, can you highlight what your main challenges in utilizing reprogramming for disease modeling?

Dr. Ahlenius: Well, in terms of disease modeling I think we really need to continue to improve the direct conversion from adult human cells. We know we can make very good cells from pluripotent stem cells, but it would be very important to be able to make equally good cells from somatic cells. And this will be crucial to study adult onset disorders, as well as sporadic diseases that are not monogenetic. And it also will be important to capture epigenetic components to disease, because it's believed that the epigenetic landscape is not as remodeled when you do direct conversion as compared to when you do iPSC reprogramming. So that I think this is one of the important challenges that we have before us. Another aspect is to be able to make subtype-specific and regional-specific cells, because that's also becoming very clear, it's needed to really dissect disease mechanisms.

Dr. Pereira: Do you see you can learn from other direct reprogramming processes that have been described in the literature? It's a body of evidence that has been building regarding the identity of the transcription factors, ways to deliver these transcription factors, stoichiometry, and also the supplementation with small molecules. So what are the things you can envision that will be more critical to achieve neural reprogramming with high fidelity?

Dr. Ahlenius: I really believe that we can turn any cell into any other cell. It's just a matter of finding that combination, as you say, right? Which are the transcription factors and what are the external cues that we need in order to do this? I think delivery we are really good at now. It's fairly easy these days with the very advanced lentiviral constructs and the inducible systems, to get really high level of transcription factor, which is really essential to get good conversion.

So the key, I think, is in finding the right transcription factors, and as we've shown, and others have shown as well, more is not always better. Sometimes you actually have to cut down the transcription factors, and sometimes you might need to play around with at what time point they are turned on. Because there might be something that's really important in the initiation but not in the maturation of cell fate change. And then you have to find the best combination of transcription factors, and with the small molecules or instructive factors that you add to your cocktail. There are now companies and people actually working a lot with bioinformatics to tease this apart. There's a couple of recent papers on this, really looking globally at what the transcription factor combinations that would drive certain fates.

At the same time, I should say that in some cases it seems to be very difficult. And there, I don't think the silver bullet lies in your obvious candidates. Then you can probably not rely on what is published. You have to really generate your own data sets on expression in both the cells you're starting with and the cell you want to make, right, to really try to figure out what's missing from your cocktail.

Dr. Pereira: More and more we can take advantage of single cell information, so that may also help us, improving the conversions that we already have established. In a general sense, can you share with us your ambitions and vision for the future of cellular reprogramming?

Dr. Ahlenius: I think we will see major developments in improving the direct conversion from adult human cells and also making more subtype-specific cells. But also in vivo conversion as a sort of a potential cell replacement strategy, and even for rejuvenation strategies. I think these are areas on the strong uprise.

Dr. Pereira: Picking up on rejuvenation, how are you utilizing that concept in your lab? you mentioned you also want to model aging. Are you using the astrocyte system to model aging?

Dr. Ahlenius: I've been a little bit hesitant going in to study aging with direct conversion. I think you can get some useful information, but I still think that we need better protocols. But once we have better protocols, then of course it's a matter of taking fibroblast of different ages, and then generate astrocytes from those, potentially even making multiple different neural cells from differently aged fibroblast, and then compare how these function alone and together. Similar to what we do in mouse studies, we transcriptionally profile cells from different ages and see what are the transcriptional and epigenetic changes that comes with aging. So, I mean, that is one approach.

Then when it comes to cells derived from pluripotent stem cells, I think we have to take a different approach, because pluripotent stem cells, it's quite well known now, that they represent a more fetal stage or early developmental stage, even if you differentiate them well. But, we have the opportunity now, since we can make neurons and astrocytes and other neural cells very efficiently. We can take sort of a genomic approach and study longevity genes, for instance. What are the actual function of longevity genes, or even these different types of mutations or SNPs that are associated with longevity? How do they affect function of neural cells? So I see it as complimentary approaches.

Dr. Pereira: There has been literature on direct reprogramming of neurons and the maintenance of aging signatures. With direct reprogramming, the epigenetic, and some hallmarks of aging are maintained, right? through iPSCs there is a completely erasure, right?

Dr. Ahlenius: We know that reprogramming to iPSCs involves major epigenetic remodeling and that it rejuvenates cells, which makes them not so good for studying normal aging. However, you can still use them to study genetic components of aging. Direct conversion on the other hand is believed to retain aging signatures of the starting population. It's clear from transcriptomic studies that directly converted neurons retain the aging gene expression profile of the starting cells.

Dr. Pereira: Henrik, do you have any advice for the younger reprogramming scientists starting their career now in the field?

Dr. Ahlenius: I think it's important to let what excites you also lead the way forward. And you need to find the question you want to answer. And also, if you have an idea that you really believe in, then you should stick with it and really fight for it. It might be an uphill battle, but I think it will pay off in the end to really stick with what you believe in.

Dr. Pereira: Completely. It's really about these eureka moments we have, right? And that keeps you fueled for a couple of years. And then some other eureka moment comes. Do you have an example of one of these moments you could share?

Dr. Ahlenius: Sure. I mean, I'm still really excited every time we get new data in the lab. I mean, this never stops to excite me. But then of course, I think there was these moments during my training, when I made my first iPSCs and my first directly converted neurons. And for instance, in our study to make astrocytes, I was really amazed the first time we started getting data from this. I mean, how efficient it was, and how fast it was. Because although we had the indication, and I was convinced that it should work since making neurons from IPS cells even with the traditional protocols although time-consuming, it's quite well-established. But with the astrocytes, you have these sort of daunting, very, very, long protocols. And maybe I was a little bit scared that it wouldn't work. So that was really our eureka moment, when we saw how efficient it really was. The same strategy that works for neurons worked equally well, if not even better, for astrocytes.

Dr. Pereira: Papers take several years to develop, but the core of the contribution resumes to a couple of moments!

Dr. Ahlenius: Yeah. It was never really getting this to work that was the problem. The problem was really when we started thinking about what defines an astrocyte, because then, turning to the literature, there was no really one thing that defines an astrocyte. There were many things, and we basically had to round up every single functional assay we could think about and start looking for diseases where we could actually apply this. And so that's where sort of the bulk of the work came in.

Dr. Pereira: I'd like to close the interview with questions that are not strictly related to science. If you could answer any single scientific question, regardless of your expertise or chosen field, what would that be?

Dr. Ahlenius: I guess it would be why do we age? I worked in aging throughout my entire career, and it's one of my biggest interests. And it might seem like we know already a lot, but I feel that there's still so much to learn. As I said, we really don't know, still, what's the phenotype of an aged human neuron? I mean, before it was believed that we lost a lot of neurons from the brain when we were aging, but we know now that's not the truth. It's more sort of subtle and discreet changes that occurs in the brain. So I think this is still a fundamental question that we have to answer.

Dr. Pereira: Do you have any hobby that you never had the time to pursue?

Dr. Ahlenius: I think I have too many hobbies. I love being out in nature, fishing, surfing, and diving, anything close to water. But when it comes to sport, I think it would be fencing. I fenced for probably 15 years, but I started late and didn't really have the support or the time to really commit. I think I could have been pretty good. I was competing, but I think if I would have started earlier and really stuck to it, then I could have become pretty good.

Dr. Pereira: And if you were not a scientist, what would you be?

Dr. Ahlenius: Probably a chef. I know this is quite common answer among scientists, but for me it was really a serious decision. In Sweden, when you go to senior high school, you have to choose sort of direction, right, and I chose eventually natural science, but I was seriously considering going into the direction of hotel and restaurant. And it's quite funny, because my parents always reminds me that as a kid, I always told them that I wanted to open a wilderness hotel in the northern part of Sweden. So maybe that's my fallback.

Dr. Pereira: I think this is a very, very interesting plan B, right? Dr. Ahlenius, thank you so much for joining me today. I thank you for your time. It was great to learn more about your interests and your science.

Dr. Ahlenius: Thank you.