Scientist Stories: Fred Gage, DNA damage and Repair in the Brain

Hello everyone, my name is Rick Henson and I am the incoming President of the BNA. I will be introducing our closing plenary lecture by Fred Gage in a moment, but before I do so I'd like to say a few words of thanks to some of the people who have made this fantastic festival such a success. Indeed, I'd like to invite some of the key people to join me briefly online, so hopefully Anne and Annette and Tara will briefly join me. Foremost to thank is Anne Cook and her team at the BNA: Louise, Sophie, Alex, Hannah, and Joe, without whom none of this would be possible. Anne and her team have worked tirelessly over the last few months planning how best to host an online conference. And liaising with our providers, Outsourced Events, who've been very responsive and supportive. I'd also like to thank again all our partner societies, the UK DRI, and our sponsors who've contributed to the content of this festival. Another person I'd like to thank is our outgoing president, Annette Dolphin, who has worked hard to promote the BNA over the last 4 years, 2 years as president-elect and the last 2 years as president. Particularly coping with the challenges that have been brought about by the pandemic. I know how much Annette was looking forward to seeing everyone in Brighton, but I hope that can happen in 2 years' time. I'm beginning to realize what a big task being President is, and I look forward to Annette's wise advice over the next 2 years as our President passed. I've heard that Annette is a very keen theatergoer, and I hope you've received some small token of our appreciation by post. And I think you'd like to say a few words, Annette.

Yes, thank you so much for the lovely flowers and the theatre tokens. I really look forward to spending these theatre tokens when theatres open. So as outgoing— that's not my dog— as outgoing President, I'd just like to very quickly take this opportunity to add my thanks to all the people that made this possible. The BNA team are wonderful, CEO Anne Cook and all the staff who've worked so hard since last July when we made that fateful decision to go online, which turned out to be absolutely the right decision. I also want to thank the programme committee chaired by Sarah Guthrie and Hugh Piggins who've spearheaded, I think what we can say, is a wonderful programme that we've all put together, and it's been a great success I think given the circumstances in which we find ourselves. So I'd like to thank Rick for his good wishes and wish him all the best, best of luck for his future as President and our future meeting in Brighton, and look forward to seeing everybody in Brighton in 2 years' time. Thank you.

Thanks, Annette. Looking forward, I'd also like to announce our new President-Elect, who is Tara Spires-Jones. Who is Professor of Neurodegeneration at the University of Edinburgh. She's on the screen now. I'm very excited to work with her too and learn from her wealth of experience in neuroscience and more general science-related issues. Finally, I'd like to thank you lot for attending the festival. I hope you've enjoyed it despite the problems of virtuality and the lack of Guinness. I hope to see you at our next festival in 2 years' time. Which we hope will be in person in Brighton, and also to see you at the FENS 2022 meeting, hopefully in Paris next year. Indeed, please stay online after Fred's talk if you want to help shape the next FENS meeting. Finally, if you're not already a member of the BNA but have enjoyed the festival, then please consider joining our society. Details are on our website, and it's not a lot of money to contribute to your community and support the team that put on great events like this. Right, to return to the main feature today, I'm delighted to introduce Dr. Fred Gage, also known as Rusty. I've already encroached too long on his time, so I'll only briefly highlight some of his many contributions. Rusty is the president of the Salk Institute for Biological Studies and the Adler Professor in the Laboratory of Genetics at the Salk Institute, and his work is focused on the plasticity and adaptability that remains throughout the life of all mammals. He has made many important discoveries, including, together with Erikson in 1998, the finding that the human brain produces new nerve cells even in adulthood, contrary to the textbook dogma. He's won numerous awards, which you can find on his Wikipedia page, but the entry there that I like most was the rumor that Rusty is a descendant of Phineas Gage, the famous clinical case whose prefrontal cortex was destroyed by an iron tapping bar, because that entry is qualified with the addendum that this proposition faces considerable difficulties, chief of which being that Phineas Gage had no known children. As with all our plenaries, please type questions at any point during the talk into the questions tab of your live discussion box. Rusty has kindly pre-recorded this talk, but he'll be back online at the end to take any questions. Thank you again, Rusty, for taking out— taking time to talk to us today, and I look forward to hearing your talk.

Well, thanks for having me, and I look forward to answering questions at the end of the day.

Hello, I'm glad to be here, happy to give this lecture for you, although it's, uh, obviously more complicated doing it through virtual presentation, but I'm— hopefully I'll be able to communicate these ideas to you well. Title is Age-Related Degenerative Disease and Genome Instability. I'm going to thank Jerome Mertens and Dylan Reed, who led parts of these projects, ahead of time.

As I said, I'll divide it up in two parts.

One is modeling aging with human cells, and the second part is genome instability. These are two new, yet to be published papers in press.

So aging itself is not fully understood—

why we age. There are hallmarks: genome instability, epigenetic alterations, loss of proteostasis, mitochondrial dysfunction, and more than likely some combination of those reflects what aging really is rather than it being a single thing. What we do know is that aging is the major risk factor for many diseases including many cancers, heart disease, and certainly Alzheimer's disease and other neurodegenerative diseases. There are a lot of hallmarks of Alzheimer's which have been focus of a lot of research, including tau and tangles pathologically, as well as loss of brain tissue, loss of neurons. People have modeled this disease in a variety of experimental animals, usually by overexpressing mutations in the familial forms of Alzheimer's disease in cells, in C. elegans, Drosophila, mouse, but 95% of the patients are sporadic. So we're still not sure how predictive the familial forms of Alzheimer's are of sporadic disease. Chose to use the reprogramming technology in human cells from patients. In this case, we used old and young patients, took biopsies, isolated and purified their fibroblasts, and differentiated them into iPS cells, a technology that we know is useful now for differentiating cells into a a variety of tissues, including neurons. We have done this with a cohort of young and old individuals, starting 1 day old through 80 years old. And with fibroblasts, we can see clear differences, statistical differences in gene expression between old and young. However, when we convert the cells to iPS cells, we lose that ability to distinguish between old and young. Patients through their cells, and this is completely lost in iPS cells. We've implored— employed a second methodology of programming called direct reprogramming, where we take fibroblasts and overexpress pioneer factors in GN2 and Ascl1, and within a period of 3 weeks, the cells can be induced into neurons. This is presumably because these factors are pioneer factors, which directly direct the cells to become a specific lineage, in this case neurons. When we implore this technology, we find not only does it work well, but the cells now represent old and young quite effectively. So we pulled out genes that are uniquely expressed within the elderly population. One of the genes, for example, that we pulled out was RanBP17, which is a nuclear pore protein, and its decrease results in a decrease in nuclear integrity. We explored this and showed that there's an age-related increase in nuclear pore leakage over time. As you age, which can be reversed and you don't see when you use iPS cells, suggesting that this is actually a target for modification. So now we have used this technology for directing iNs and making iNs and exploring a variety of age-related changes. We've made these vectors available through Addgene and through sharing, and there have been several publications over the last 4 or 5 years showing changes in epigenetic methylation, stress survival, heterochromatin, and proteostasis changing with age using this methodology. One thing that we have shown using sites associated with epigenetic clocks, so methylation sites, that the INs, and the fibroblasts from aged individuals are correlated with age, whereas older patients, our patients here, however, if we look at direct iPS-derived cells, they are more correlated with early prenatal time. So even with regards to methylation as a genetic marker of age, epigenetic marker of age, there is— it's captured by these IN cells. And using the BrainSpan Atlas from Alan, where you take all the transcription factors that are associated most with early brain development all the way through aging, so you can map your genomes to this transcriptome map. And when you do that, you can find out where your transcriptome ends. We've done this, and what we see is that for the iNs during early development, there's— our gene expression patterns are not— or have a low similarity score with early genomics, whereas the iPS iNs are highly correlated with this early time period. The INs are more similar to age-related genomes, whereas the iPS are not similar to the INs— sorry, to the age-related genes in the Allen Atlas. So let's go to Alzheimer's disease. We've used 19 non-demented and 13 sporadic and a few familial cells, patients' cells, for comparison. We've differentiated all these cells into INs, and they all convert— both the AD and the controls convert nicely into neurons. They are highly represented by NeuN, VGLUT, and GABA, pretty much the same proportion in all cases. The only population that shows an A-beta secretion, 4240, is those that have mutations, the familial forms of the disease. We don't see this, as have not others seen in sporadic forms. So let's look at the differential expression patterns comparing controls and ADs. In the fibroblasts, we really don't see any difference using volcano plot for differential expression between the controls and the ADs. However, when we look at the INs, we see a significant number of genes that are upregulated and downregulated. Upregulated and downregulated in this population comparison. They fall out nicely in a t-SNE plot analysis, and you can see that the expression patterns also are different using this diagnostic plot as well. Now, as a way of determining whether or not the changes that we're seeing are similar to what are seen in post-mortem RNA-seq data, we've gone back to data that's already been published from post-mortem tissue comparing AD and controls. And the color code here is yellow being very similar, and, and the blue going down is also similar. So what we see here is that there's a high correlation between the RNA-seq from the post-mortem tissue in both the upregulated and the downregulated genes for comparing between post-mortem tissue and the INs. So it's some level of validity. We've also taken a recent paper by Grubman from Nature Neuroscience where they did the single-cell sequencing from frontal cortex postmortem tissue and reanalyzed their single-cell tissue and then compared those with our INs. And what we see here is that there is a high correlation between the new IN group of cells here that they've defined as differentially expressed in AD and those that are overlapping with our IN cells. We find very little relationship with the other cells in their pool. So we think that this represents again that our INs are recapitulating well the changes that are occurring in vivo and postmortem tissue. Look at this a little more clearly, we can look at the Alzheimer's disease genes that are downregulated in GO terms. We see synaptic transmission, nervous system development, rhythmic processes, cell-cell adhesion. And really what emerges is this idea that a basic identity of neuronal fate is downregulated in AD compared to the controls— axonal guidance, neural differentiation. So neuronal activity and synaptic function are all declining in AD. We confirmed this by doing double labeling with SNAPs and PSD95, and we see a significant decrease in these markers in AD compared to controls. We've done morphological complexity analysis of the cells, and the cells from the AD brains patients are significantly lower. And finally, we've done calcium imaging, so we can do live imaging of these cells And we see a lower frequency of spontaneous network activity with the calcium imaging in the IANS compared to controls. If we look at the upregulated genes, interestingly, inflammatory genes, cell cycle genes, stress, and apoptosis-type genes, in addition to evidence for cell dedifferentiation. So when we look at these again using an alternative analysis, many of the genes that are upregulated appear to be involved in cellular stress, apoptosis, cancer, and some stem cell signaling pathways. So we're trying to figure out what's the connection between these upregulated genes and Alzheimer's disease. It appears that there is some evidence for the fact that the INs, the Alzheimer's INs, are dedifferentiated and have lost their— losing or losing their fate of neuronal fate. So one of the things we do see associated with this is an increase in ROS signaling. So the ROS levels are higher, reactive oxygen species, And if we look a little bit more closely at that, looking at H2AX, which is a DNA repair-associated protein analysis, we see an accumulation in the nuclei of the AD but not in the controls. You can see this in quantitation here by nuclear fluorescence. And in addition to this, we see an increase in the amount of DNA content within the AD population, suggesting an attempt at cell cycle regulation. So we looked at that a little bit more closely and saw that cyclin B1 is upregulated in AD. But there is some evidence of cell cycle entry, but it's not completed. So when we label the cells with EDU. We got no double labeling, so the cells were not completing their cycle of cell cycle. The cell fate seems to be unstable, as I've shown you. And what about the epigenetic identity? We used ATAC-seq to look for open chromatin as a marker of immature progenitors. We know that Immature cells have much more open chromatin relative to differentiated cell types. And here are the plots we see for AD ATAC-seq peaks near the promoter region of genes, and we see an elevated increase in the accessibility of chromatin in this region. Shown here as well. So there's more open chromatin. If we look more specifically, what we find is that this promoter region is significantly more accessible. And if you look at the sites there that are open, we find proteoglycans in cancer, metabolic pathways, apoptosis, all enriched for being open. So open promoter target— promoter sites are associated with cancer, metabolic, and apoptosis genes. More specifically, when we do DNA motif binding factors for INs, we find these genes that are associated with glycolysis, for example, LDHA, elevated in AD. HK2 is also elevated, suggesting that the cells are dedifferentiating from their oxidative phosphorylation utilization or generation of ATP to a glycolytic form. And then we see increase in cell stress and stem cells. So DNA binding motifs for stem cells and cancer transcription are more accessible. Genes that promote dedifferentiation transformation become expressed. Is this phenotype age-dependent? So the question is, we have these aged individuals that have— all of them are aged and they all have Alzheimer's, or half of them have Alzheimer's disease. Can we capture that Alzheimer's disease in without aging. And the value is we have two methods from the same patients. We can either do this direct conversion, or we can make iPS cells. So we decided we would make iPS cells from the same cohort that we'd shown in direct to ask specifically whether or not the iPS, the reprogramming of cells, was a— would have a negative impact on identifying the phenotypes. Or put another way, could iPS reprogramming rescue phenotype that we've identified in direct reprogramming. So we compared both the iPS-derived neurons and the induced neurons from the same patients. We use the same vector to convert them into neurons. It's only that this cohort went directly from fibroblasts to neurons, and this cohort was induced into iPS, but they both became induced neurons, and then we compared between the two. Quite strikingly, in the iPS cells, we see no differences between Alzheimer's disease and controls. Likewise, in the iPS-derived neurons, unlike with the iNs, we see no differences, no significant differences between the two. So iPS reprogramming of patients erases the differences between them. Highlighting the idea that there is really an age-dependent requirement to be able to see these phenotypes in the aging brain. So neural differentiation allows for neuronal connectivity and brain activity and cognitive functioning. Sporadic Alzheimer's disease, which is what we're studying here, is a sporadic form of Alzheimer's disease, is an age-dependent pathological program which promotes a neuronal cell fate loss, a loss of neuronal functionality and circuitry, epigenetic destabilization, and activation of cell death mechanisms, as well as a partial but not complete cell cycle re-entry and an increase in the utilization of glycolysis at the expense of oxidative phosphorylation. Interestingly, many of these pathogenic programs are consistent with the transformation of cells from a somatic form to a cancer form in a process known as the Warburg phenomenon. I think this opens up a new way of thinking about Alzheimer's disease and its relationship to aging. I just want to highlight for a second that we did find in our RNA-seq in the aging INs compared to the age-matched controls, an increase in reactive oxygen species, increase in DNA, and an increase in DNA damage pathways. So these are transcription factors associated with DNA repair, and they're differentially expressed in young versus old. So there's a genomic instability in the total DNA content. So now I'd like to switch to talk a little bit more about the mechanisms of DNA instability in neurons specifically. And this, of course, is just to remind you that in Alzheimer's disease there's a loss of neuronal identity associated with synaptic transmission, plasticity, differentiation, but a gain in these factors associated with neuronal stress. So what about what role does genome instability play in aging? There's a long-standing history of DNA damage theory of aging. We know that mutations to various forms of repair proteins result in severe neurological damage. Muki and colleagues have shown that there's an early neuronal accumulation of DNA double-strand breaks in Alzheimer's disease. And interestingly, when you, when you think about chronic DNA damage response, it triggers altered metabolism, proteostasis, downregulation of mitochondrial biogenesis. So this genomic instability is associated with a variety of things that we've seen in the previous presentation that I just gave you. Now neurons preferentially repair certain types of DNA. There is some evidence supporting the fact that terminally differentiated neurons repair transcribed genes, so activity dependency in some sense, and there's also evidence that neurons can incorporate radioactive nucleosides following DNA damage by unscheduled DNA synthesis without any damaging agents, suggesting that there's ongoing DNA damage. In fact, the estimates are somewhere around 1,000 DNA damaging events occur every day in every cell. So there are hundreds of thousands, if not millions, of DNA damaging events that are occurring in the course of a lifetime. Most, of course, are normally repaired. But where does that repair occur in the genome? Is it sporadic, is one of the questions that we would like to ask. So to do this, we developed a methodology we're calling RepairSeq. We incorporate— allow EDU to be incorporated into the genomes of post-mitotic neurons. In this case, we're using iPS-derived neurons. Ones I described to you previously. These are not IN neurons yet. And we can label DNA with this. So the way we do the quantitation is— this is just qualitatively showing that they do incorporate into nucleus as well as into cytoplasm mitochondria. We take post-mitotic neurons, we label with EDU, we allow time for the EDU to incorporate into DNA damaged sites, naturally occurring DNA damaged sites, and then we use click chemistry, including the biotin azide, make a library, pull down with the biotin, and then do PCR amplification of the beads, and then we do next-genome sequencing. Here you can see an example in this synuclein gene. These are hotspots of repair, so this is what we call hotspots of repair. This is repair occurring all along the gene, but they're only in sub-regions where we get these hotspots, which accounts for about 1.3% of the genome. It's quite reliable, so in replicates of the same line, we find high reproducibility of the hotspots.. And even between different cell lines, iPS lines, we find very high correlation between them. When we look at the peaks, they have a general size of about 500 base pairs, so they're pretty sharp band around the site where the repair peaks occur. And generally speaking, we find a higher amount of peaks found in the 5' UTR of genome. To look more specifically at where these repair sites are occurring, whether or not it's in open chromatin, we did ATAC-seq, which you can see here, and we did ChIP-seq for H3K27 acetylation. You can see that there is an association, some level of association between ATAC-seq. Most hotspots are not associated with them, so more than 50% are not associated with ATAC-seq, but most of the heavily repaired sites tend to be associated with ATAC-seq and H3K27 acetylation. There are regions of the genome generally around the promoter area called quadruplex, and we used ATAC-seq— that are associated with ATAC-seq sites, and they're a folded DNA that had some structure presumably important for transcription, and what we find is that those areas are highly repaired, meaning that they were undergoing damage and repaired preferentially at a very, very high level, and this is relative to sort of random site within the genome So strongly enriched in promoters, these quadruplexes, and a lot of repair is occurring at those sites. So what about transcription? So if you look at all areas of repair, not just the peaks, just the repair, there is a correlation between transcription RNA-seq and repair-seq. But if you look just for the peaks, there's really no association with transcription. So it's not just transcription alone. But if you look at where those peaks are occurring, and look at their GO terms, they appear to be in neuronal— primarily in neuronal areas. So what's happening is the genome is repairing primarily areas of identity, of neuronal identity, We've only done this in neurons, so it'll be obviously important to do this for muscle and other tissues as well to see whether or not each cell type has a repair program specific to its cell type. We used a method to pull down proteins and do mass spec on proteins associated with chromatin that had been repaired. And when we do this, we find 79 proteins that are enriched. And using a variety of databases of conformity, we can find that the genes— these proteins are associated with histones, H2A isoforms, and we've confirmed this by Western blots, as well as RNA-binding proteins. So the importance obviously of these cells and cell identity, and these proteins and cell identity and DNA repair, are interestingly linked to our hotspots, part of the chromatin. When we also looked at databases, we find that the consensus is that 21 of the 79 are proteins that show abundance in those proteins that have been demonstrated to change with Alzheimer's disease and other neurodegenerative diseases. And many of these hot spot proteins were found independently in a protein analysis of postmortem tissue that were associated with late stages of Alzheimer's disease. So here's our cluster analysis here, and then here's this mapping WTCA from Dan Geschwind's group. Cell reports. When we looked at epigenetic drift, and I spoke to you about this a little bit before, the epigenetic clock can be used to measure age. And when we looked at the signal for aging that's been identified for the epigenetic clock, we find that those areas of the genome in young cells are highly repaired. So there seems to be a preference for repairing sites associated with the epigenetic clock. So in general, we think that DNA repair— and others have shown this, and this is in a review by Karl Herrup some years ago— showing that there is a decline, and they're trying to be repaired, but with some events— Alzheimer's disease, DNA damage— there is a breakdown in the ability of the cells to repair the amount of imbalance between the amount that can be repaired and the amount of damage that occurs, and that's when we shift from the sort of normal decline with aging to a progressive loss. So we mimic this a bit by using DNA damaging agents, and what we find is that there is both a loss and a gain of peaks, so the location of where the gain and loss of peaks occur. So one can see that there is a baseline healthy amount of DNA damage and repair occurring that preferentially repairs specific locations in, in the genome. But when that is lost, then there's a shift and some areas are preserved and others aren't. Another way of looking at this is to compare the repaired sites to those sites that have been identified as being lost with age. So several investigators— go to Chris Walsh, a very nice paper in Science showing that there's an increase in SNVs, single nucleotide variants, with age and accumulation with age. So we mapped these sites that have been identified by Chris using a relative distance marker to identify distance between two genomic coordinates. What we see is that there's, first of all, there's very little overlap between the sites that are repaired and those that are, that show up pathologically as SNVs. And in fact, we see very little relationship to them. Sorry. Um, we've also used another methodology to look for genome evolutionarily conserved sites. So these are areas of the genome that are very similar between many, many species, evidencing the fact that they are conserved. So they're evolutionarily conserved sites. And while there's no relationship between or no, no commonality or similarity between ATAC-seq and repair for this, for these SNVs, we find that there's a very high, they're very, very, very close to these conserved evolutionary areas, specifically the repair areas. So you see that repair the seek or these repair mechanisms are really trying to perpetuate, preserve these evolutionarily conserved regions. So in conclusion, I'd like to say this last study we've shown that DNA repair in neurons is not random. There are hotspots of DNA repair, accumulation of DNA repair occurs. Promoters, splice sites, DNA binding proteins, histones, are targets of repair. In neurons, neuronal identity genes are peripherally repaired. It's interesting, from the previous study we showed that there's a loss of identity in Alzheimer's disease, and then there's evolutionarily conserved regions of the genome that are also preserved. And finally, the biological clock, which is so valuable in predicting age, is also protected by this repair process. So we hypothesize that DNA repair declines with age, and this might be the driver. This decline in DNA repair might be a driver in disease, and that any perturbation that occurs to disrupt the hierarchy of repair could be detrimental to areas depending upon the disease type and where the damage occurs. So when thinking about DNA damage, using the analogy of the bullet hole misconception, the damage to the hotspots—that's where the repair is—is probably the most detrimental to the organism. With that, I'd like to thank you for your attention. Thank Dylan Reed and Jerome Mertens in particular for this, and the core facilities at the Salk, as well as those funding the research. Thank you for your time.

Thank you so much, Rusty. That was an inspiring breadth of neurobiology that you covered there. While we're waiting for questions to come in from the audience, let me start with a few. My first is actually a request for clarification. So in the first part of your talk, when you're talking about induced neurons from patients with AD showing downregulations associated with synaptic density and activity and upregulations associated with cell stress, I wasn't clear whether these were the sporadic or familial AD or whether you'd collapsed them, because one of your early slides you showed that amyloid secretion was only present in the familial AD cells.

So what do you see? Yeah, so we only have 2 or 3 familial forms in our analysis, and we did include those in the synaptic analysis. So we do see a decrease in synaptic function both in the familial form as well as in sporadic form, but that analysis is dominated by the sporadic. There's something like 16 to 3 in those populations. When we separate them, and it's not really fair 'cause the n is so small for the familial, but we do see a decrease in the synaptic function as well. We're increasing our cohort of familial forms and different types of familial forms in future analysis. But our real focus here is to try to take a look at the sporadic forms, which are so hard to analyze in most of the model systems we look at.

And then my, my question was that how many of the differences that you observed in these, in these ions do you think would be specific to Alzheimer's compared to other neurodegenerative disorders? So I don't know whether you've tried this or what picture you'd have if you took induced neurons from FTD or other tauopathies.

Yeah, so we are, we are doing that. Right now. I can say that we have done this for other diseases like Parkinson's disease, and there's very little overlap. One of the— one of the— in the— in— there's—

there is—

I should say from the positive side, there is an overlap in issues related in particular to metabolism. So there's a core phenomena that seems to happen in these age-related diseases where they shift their metabolic activity from oxidative phosphorylation to glycolysis. And this cell cycle entry is something that seems to be ubiquitous. It deserves a lot more attention, but it is something that is emerging for us as a metabolic—

there's a metabolic component to this whole thing that is really quite interesting.

Yeah. Right.

We've got a lot of questions coming in. Let me start with a couple from Annette. So, well, two questions here. Either are the synapses in induced neurons from AD patients different in morphology as well as having fewer synapses? And if you can hold that in mind, do you find a difference related to ApoE status? Yeah, there's not—

we don't have analysis on the ApoE. There's no— in the small sample we have, we don't We don't find a relationship there. The caveat for the first part of that question is that we do find synapses defined by these immunomarkers, but when you go to the ultrastructural level, they're not as developed as you'd like to see. Well, you'd like to have to be able to make the kind of analysis that Annette is asking for. So there's enough variability there. We're using these really simplistic markers of synaptic formation. So they're very early on. We have now begun to use organoids where we can get much more complete synaptic structures, and we're doing ultrastructure on those. And this observation at the 2D level has inspired us to make a more complete story, and hopefully we'll have a better answer to that in the future.

Okay, and a question here from Tara: do your results have implications for amyloid and tau-directed therapeutic strategies?

Um, let's see, from a— I would say not a, not a lot of, um Phrase that question again because I want to give my answer in the context of the way the question was phrased.

Well, as written here, it says, do your results have implications for amyloid and tau-directed therapeutic strategies?

No, I would say they don't. And since we're not really looking at those features, we don't get a lot of that in these, in these assays. I don't think this really informs it very much one way or another.

Okay, uh, some more questions now from the audience. So Susan Chalmers asks, or says, first, fascinating talk, thank you. Is there any opportunity in the iN protocol to renucleate with nuclei from young iPS or to introduce young mitochondria that maybe haven't built up damage?

Yeah, no, I think this is a brilliant idea and one that we are looking for collaborators to work on. The repopulation with mitochondria, with young mitochondria is one thing, or other kinds of pharmaceutical interventions that increase mitochondrial biogenesis, some way of rejuvenating the mitochondrial function, I think is a really good target. And we're very excited about that. The nucleation part is interesting, The fact that we lose most of the phenotypes when we reprogram to iPS cells heightens our interest in reprogramming itself as a way of rejuvenating cells. And there are some strategies that people at our institute, in fact, have been pioneering to try to do this in an in vivo setting. It's very interesting.

Okay, there's another question here from, uh, Kislain Katani— please forgive me if I mispronounce people's names— great talk. What are the genes that regulate the repair process and hierarchy of repair?

Yeah, so there are hundreds of genes that are involved in the repair process, and interestingly, neurons that are post-mitotic, at least these post-mitotic ones that we're examining, only use a fraction of the DNA repair capacity of the complete machinery. So, there's some sorts of DNA repair processes that are not functional in our iNs or in neurons at all. So, it's a subset of these repair processes.

We—

and I don't think I can tell you a really clear case that there's one repair gene that's involved. There's sort of a dampening. And interestingly, here's even makes it a little bit more confusing. There's, I term it a misregulation of repair, 'cause some of the repair processes are upregulated, some of the genes are upregulated. So it's really a dysregulation or a misregulation, if you will, of the process. And it deserves a lot more attention to understand what that machinery is how that machinery is disrupted selectively in these neurons, which already have a minimal form of repair compared to dividing cells.

Thank you. So Maya Hanspel asks, so very interesting that AD is a cellular program into loss of neuronal fate identity. Do you anticipate any similarities between AD-induced neurons and quiescent neural stem cells?

Um, that's a good question. Um, the answer at the surface level is yes, and mainly because quiescent neural stem cells have more open chromatin. Um, so, you know, that surface level, you were— we're seeing this move back in that direction. I'm not sure that I think that that's the way to, you know, the most effective way to think about it, because we also see this loss of neuronal fate or loss of cellular fate in cancer, which is, you know, that's the one that's most intriguing to us is this idea of a Warburg effect where you get this dysregulation of an existing cellular somatic state back into a more immature state, where cancer cells will go on to divide, we think that what's happening with the neurons as they dedifferentiate is they more than likely go into a quiescent state or senescent state. So I would say the quiescent stem cell is more like the senescent neuron than it is like a cancer cell. So there's the analogy is, and I can say that we see a subfraction of these cells taking on characteristics of markers of quiescence. So within this larger population, there's a subpopulation of cells that have quiescence. And that's where maybe the correlation is best. But we're in the middle of that— well, actually more toward the end of that sort of comparison. And we'll be submitting that soon. Very interesting.

Kevin Marinus, asks, um, could you speculate on why age-related neurodegenerative diseases shift towards glycolysis rather than oxidative phosphorylation?

Yeah, um, this is the focus. This is a real focus of our attention at this point. And let's say, let's think about it in a couple ways. One is What's the difference between them? Well, in glycolysis, you often use that when you really need rapid production of ATP, that there's, you know, something is requiring a more rapid— the slow turnover of mitochondrial and generation of lots more ATP is good for a quiescent cell who's busy doing lots of activities. When a cell is undergoing bombardment or undergoing stresses, there's an evolutionary machine built into our cells which drives them to generate quick ATP. So that could be part of it. That's not really a— I mean, it's more of a global explanation than a mechanistic understanding. I should say that— so another way to look at it is very mechanistically and say I showed you in some of the figures that you may not have seen that some of the genes that are involved critically involved initial phases of glycolysis are upregulated. And the sites around those genes that are— the chromatin area is open in these areas for some. So, some mechanism is driving the opening of regions which are associated with the formation of, or the function of glycolysis, are opening up as the cell is decreasing. Or dedifferentiating. So there are mechanisms in there for doing that. We need to look and understand that a lot better. But I think what Kevin is asking is a deeper question, is why is this happening? And it may be a red herring to talk about this, and it may be for a completely different reason that this is happening. It's just part of a machinery that is initiated as a function of stress and that these programmed machines— we were fascinated by the idea that increase in DNA accumulation and DNA repair and mitochondria, all these things, see, we think of them and study them almost in independent isolation, and yet they're very much linked to each other. And what these kinds of studies show us is just how linked inflammation, DNA repair, epigenetic mechanisms are to each other. And it's at the interface of those that I think that I will be able to give a better answer to Kevin's really, really important question.

Thank you. Another question here from Annette. So are the BRCA mutations a risk factor for AD?

Braca? You mean cancer? Oh, you know, there is a whole interesting story in the cancer field and the Alzheimer's field about the relationship between them. And I think the story is that people with cancer have a lower frequency of Alzheimer's or one, one way or another. And I, I should have been prepared to, to answer that part of it, but there is, yes, there is an association between And I don't know about BRCA specifically. I should know, but I don't have it on the top of my head. But there is a literature on this that I haven't contributed to yet, but it's out there. And it's a fascinating link between, I think it's an inverse link between Alzheimer's disease and some forms of cancer. So yes, in general.

Question here, a very general one. Uh, what are the mechanisms that regulate DNA repair?

I guess that's quite a broad question. Yeah, and just— I'll give it a— it depends on the kind of damage that occurs. In— there's different types of damage. You have double-strand damage, and so there are repair mechanisms which will do either very selectively a clean repair of that double strand, or you can do blunt end repair. Other times there's single point mutations that occur, and that's a different sort of machine. And what it does, a set of proteins will grab a nucleotide, or in some cases a ribonucleotide, and put it into the DNA as best as possible. So you can imagine there are a lot of proteins that are sitting there identifying and sensing the chromatin and then chaperones around looking for the environment for the nucleotide that would best fit into that location. It's a— it's a— obviously it's a whole field of investigation and one that's really quite deep. And machinery underlying it is evolutionarily— it's one of these things where it's amazing that over millions of years of evolution there is evolved this machine. But you can understand that having a machine that repairs DNA is crucial for the survival of the species. And so this is a, you know, it's just a fabulous area of investigation. So if you're interested, there's introductory levels and as you get into it, more senior advanced levels that you can learn lots about that.

We ought to wind up soon, and I hope we haven't exhausted you too much, but there is another question here from Kevin. What are your thoughts on the contribution of de novo mutations to sporadic neurodegenerative diseases?

Yeah, so again, our data were We were surprised because we looked at the, the, the, the published data on sporadic, I would say, de novo mutations that have been published. And they're not that many. These are all at the single cell level being able to identify. And then we looked at our hotspots to see whether or not there was any relationship. And it turns out that where the neuron is directing repair to occur is uncorrelated with where those sites of damage are. It's, and this is where we use that analogy with the airplane, is that the repair, you know, you can look at it this way, this is really sort of a sardonic way of looking at it, is that those spots, de novo spots that you detect are the ones that the cell decided that they're not worth repairing. Or less concerned about. Um, on the other hand, when we did our manipulations to look at, um, what happens to the hotspots when you challenge the cells, they do shift to do repair other places. But when they shift to other places, they free up spaces where they were spending a lot of time repairing. So it's a whack-a-mole sort of game of trying to, you know, and, you know, the other part of it is that hopefully linking back to Kevin's question, one of the things that we learned early on in our studies was one of the things independent of disease is that mitochondrial function declines with age. And we're not the only ones who have seen that. We just see it here. So here you have a decrease in mitochondrial function, decrease in ATP, You got these highly efficient machines that are busy repairing, costing lots of energy. So, they're already compromised. And then, you know, on top of that, you add a disease. I think this is where we begin to reveal where are the sites that become open. And that might be where specificity of disease comes in, where you have some disease, you know, certain areas of the chromatin become more vulnerable to damage in one type of disease versus another, and that might have something to do with phenotypic outcome.

Great, thank you very much, Rusty. It's been brilliant. Thank you for giving up your early morning slot to talk to us today. I hope to see you in person at some point in the future.

Right, and congratulations for a very successful meeting and for you coming in as president. Thank you.

Just to remind everyone who's listening that this is the last session today, but please stay on if you want to learn about the FENS meeting next year and contribute to the program there. Thanks again to everyone who has helped make this festival such a success, and again, look forward to seeing everyone in real life at some point soon.

So, bye Rusty, bye everyone. Thank you.