Alexander Disease Research Update
Discussion of recent publications from the research literature on Alexander Disease, hosted by Albee Messing from the University of Wisconsin-Madison.
Alexander Disease Research Update
Alexander Disease Research Update - Episode #3: crossing the blood-brain barrier
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Albee Messing, Berit Powers and Rachel Battaglia discuss the following recent publication:
[4:11] Nagata T et al. (2021). Cholesterol-functionalized DNA/RNA heteroduplexes cross the blood-brain barrier and knock down genes in the rodent CNS. Nat Biotechnol. (in press)
PMID: 34385691
[26:30] Questions from email
Send your questions and feedback to: axdrupodcast@waisman.wisc.edu
Help support research on Alexander Disease at the University of Wisconsin-Madison's Waisman Center:
https://alexander-disease.waisman.wisc.edu/donate/
Hello and welcome to Alexander Disease Research Update, episode number three, recorded on September 13th, 2021. I'm Alby Messing from the Wasteman Center at the University of Wisconsin, and with me today are Barrett Powers from Ionis Pharmaceuticals and Rachel Bataglia from Brigham and Women's Hospital, Harvard Medical School. Barrett and I have worked together since 2016 on the anti-sense approach for suppressing GFAP expression that is the basis for the current clinical trial. Barrett, since this is your first time on the podcast, can you say a little bit about your background and your current position?
SPEAKER_02Sure. First of all, thank you so much for having me, Alvi. It's great to see you and to be here with you today. I did a PhD in neurobiology and behavior at the University of Washington in Seattle, focusing on spinal cord injury and remyelination following injury. And then I did a postdoc at Harvard in the Rubens lab, focusing on stem cell models of disease, particularly ALS or the Garrix disease, and then screening to see if we could find small molecules that would, or drugs that would improve survival and other phenotypes in disease models. And then I came to Ionis and been here about eight years. And I'm on the preclinical research side in our neuroscience drug discovery department. So that means what I do is I try to come up with interesting disease areas to go after where there's considerable unmet patient need and where we think that our ASO platform technology could be applied to sort of directly get at the root cause of disease rather than just treating symptoms of disease. So I work with models of disease, and then once we sort of understand that we have a burden of proof showing that we could show some benefit for patients, then we work toward uh getting a clinical candidate uh identified and then bringing that to trials. So that's the process that we've gone through together on the Alexander Disease program.
SPEAKER_01And now, Rachel, your turn.
SPEAKER_00Hi, thanks, Albi, for also having me. Um so I'm more on the basic side of things research, but it's fun to talk together with someone who's more involved with the clinical aspects of it. So it's a great opportunity. Um I did my undergraduate work at Rutgers University, where I studied uh fertility in the fruit fly. And from there I went to pursue a PhD at UNC Chapel Hill in Natasha Snyder's lab, where I studied Alexander disease and also giant exil neuropathy, which is a somewhat related neurodegenerative disorder. And I worked to develop different human cell models for both of these diseases using stem cells. Oh, and now sorry.
SPEAKER_01Last but not least.
SPEAKER_00Yes, last but not least, uh I am currently a postdoctoral research fellow in the lab of Melfini, uh researching molecular mechanisms of neurodegeneration. Um, and Mel has a model for Alexander disease and the fruit fly. Um, but I'm also looking at other things like uh an Alzheimer's disease model and Parkinson's disease.
SPEAKER_01So All right. Well, thank you both for being here
Nagata et al.
SPEAKER_01today. So today our topic is a paper that is currently in press in Nature Biotechnology by Tetsuya Nagata and Chris Adwyer, along with a long list of co-authors, including Barrett. The title is Cholesterol Functionalized DNA RNA Heteroduplexes Cross the Blood Brain Barrier and Knock Down Genes in the Rodent CNS. But before we get started on this, I just want to make a few comments on the feedback we've received so far about the first two episodes of the podcasts. And while everyone is very happy that we're doing them and listens to at least some or all of the podcasts, uh, several people have commented that they're a little bit too technical for them to follow. Uh and uh and so we don't want to um sort of indulge ourselves too much by um talking about lots of details that we may enjoy. Uh we want to make sure that the main messages uh get through. So I will keep asking for feedback on an ongoing basis. So at the end of the podcast, I'll remind everyone about the email address that I'd like you to send that to. But up front, I'll I'll say why I thought this paper was of interest. The antisense oligonucleotides were ASOs are in already in clinical use and in a number of clinical trials, but the ones that are being used don't pass the blood-brain barrier and so require direct administration into the cerebrospinal fluid. Uh, and so I think there is a high priority for everyone in finding a less invasive way to administer the ASOs. And to me, this paper describes at least one step towards that goal. Uh, Barrett, uh, perhaps you could begin by providing a short and simplified description of what ASOs are, how they work, and also some comments about the blood-grain barrier, and then walk us through the key points of the paper.
SPEAKER_02Sure, Alvi. ASO stands for antisense all the nucleotide. So it's a short strand of synthetic nucleotides that have been chemically modified to confer drug-like properties, such as being stable once they get into cells. So, to understand how ASOs work, I'll describe the central dogma of molecular biology. This describes the flow of genetic information from DNA or our genes to RNA and then to a functional protein. So you can think of DNA as an instruction manual, RNA as a copy of instructions for an individual protein, and then protein as the functional workhouse that performs tasks in a cell that need that are needed to make an organism live. So we know that disease, such as Alexander disease, can be caused by gene mutations that lead to the production of abnormal proteins or too much of a protein. So ASOs can be designed to cause the RNA to be degraded or knocked down. Uh, so then less of the protein is made. And this is similar to our approach that we're we're using for Alexander disease. So uh, as Alby mentioned, our current ASOs don't cross the blood-brain barrier. So, blood vessels in your body are leaky everywhere but in the brain, where only the smallest molecules can uh get across the BBB in order to prevent toxins from getting into your brain and things like that. So, for today's paper, we'll talk about a strategy that we can use to get ASOs across the BBB so that they may be injected into the bloodstream or under the skin instead of directly into the spinal fluid, which is how we deliver ASO drugs today. So, sort of the main takeaways of the paper that Albi selected to talk about today are that it is possible we can get the ASOs over the blood-brain barrier and achieve the target RNA reduction, leading to then protein reduction of clinically relevant genes. And this can be done across the brain and spinal cord. Um, and it can be done in a dose-responsive manner. So um if you give a little bit of the drug, you'll get a little effect, and if you get a lot of the drug, you'll get a lot of effect. Um, it's also fairly long-lasting. So once you give an injection, it lasts quite a long time, after a couple months after a single injection. Um, and then we did find some toxicities while we are doing this work, and it looks like some of those can be mitigated by adjusting uh the dosing parameters that are used, whether it's sort of minimizing the amount that you give at a single dose or um trying different routes of administration instead of intravenous, trying subcutaneous, which is into the skin. So I wanted to sort of explain what the paper is about in terms of the technology that's used. So I spoke about ASOs, antisense, all the nucleotides, they're these strings of uh uh of nucleotides that can bind to RNA. And typically when we administer these, they're administered as a single strand. So there's no complementary strand. Um, but many years ago, our our chemists, as well as um Nori, who's the the senior author on the paper, uh, were talking about some ideas they'd had um sort of around um sirna technology. Um, this is a different RNA targeting technology that's used therapeutically. I won't talk about that today, but but typically it is delivered as a double-stranded molecule. So you have one strand that's sort of the active strand, and then another that's just a um a complementary strand to to what you've you've what your drug is. Um so they thought, well, we'll try it out with ASOs, and and Nori published a paper demonstrating that these um these duplexes, so an ASO with its complementary strand attached, can enter cells in the liver and can be functional drugs. So also at that same time, um Nori and our med chemists decided to sort of work together on these projects. They had met at a conference and um were interested in conjugating, so the attaching different um molecules to the ASOs to see if it would change where the ASOs are able to go. And so attaching a lipid um stands to reason that that might help the ASO get into the brain because we know that there are transport systems that are present within the blood brain barrier cells that can transport specific molecules across the blood brain barrier. Um, and lipids are some of these molecules that get transported across. So the big question was: can we tether our ASO, our drug, to something that would pull it across the blood brain barrier as a cargo? Um, and as it turns out, the structure that worked best was not tethering it to an individual single-stranded ASO, but instead uh tethering it to the heteroduplex, the the ASO plus its complementary strand. And so if you look at the paper, if you're able to access it, and you can see figure one A shows a sort of a schematic of what these drugs look like. But basically, what you need to know is that the ASO is there, and then it's um its its complementary strand is carrying also a cholesterol on the other end. Um, and so what we did uh initially was try a bunch of different lipids. So you can see in figure one B that um we tried a lot of different fatty acids. It's another way to say lipid. And the best two candidates that seemed to get uh ASOs across into the blood brain barrier and through into the tissue were tacopherol and cholesterol. Um so my role in this paper was um sort of in the early days, I haven't actually done work on this in many years, um, just testing to see how well these things got over the blood brain barrier. Um and so we did these initial studies with a lot of different lipids, and then we we we worked with um with Nori's lab to then um test um multiple doses and um different designs. And ultimately, uh as I mentioned earlier, we found that you you didn't get really great activity if you just attached it to an ASO itself. It's the heteroduplex, the HDO you'll see in the paper, that was really functional.
SPEAKER_01Let me just um interject for a minute. Most of this paper um involves, it doesn't involve GFAP, it involves a different gene target that I can't quite pronounce. Is it mallet one?
SPEAKER_02Yep, mallet one.
SPEAKER_01Yeah, and so why did you why did you pick mallet one?
SPEAKER_02That's a good question. So oftentimes you'll see in in papers that IONIS publishes, we use quote-unquote tool compounds. So we use compounds um that have been identified and are used routinely in our experiments, you know, across different studies when we're targeting different organs or targeting um, you know, we're trying to conjugate it to something new or trying out new chemistries to see if it increases stability or activity. Um, we use uh uh mallet one oftentimes because it is uh ubiquitously expressed, which means it's expressed, it's expressed by almost every cell in the body. And we know that we have a lot of um ASOs in our toolbox that are functional and are good at knocking this particular target down. So we we will often use um this as just sort of a first pass. And then um, you know, to to expand upon what was first found with the the um the malet one RNA. Um I should say also there seem to be no detrimental effects from really plobbering this this malet one RNA. Um and although it's its biological role is not really well understood, it seems if you if you knock it down for prolonged periods of time, there there really aren't um detrimental effects in an organism. So then we went to in other parts of the paper, knocking down um clinically relevant targets and looking at the protein level and making sure that when we were we were knocking down that middleman, the RNA, that that then resulted in a reduction of of the protein. So we looked at DMPK, which is um uh a gene target for myotonic dystrophine one. Uh we also looked at GFAP, which is of course the the gene target for Alexander disease.
SPEAKER_01And SOD1.
SPEAKER_02And SOD1, that's right. And that is a gene target for um SOD1 related ALS.
SPEAKER_00Um, if there is a mutation in your gene of interest that you want to knock down, is it important to design the ASO not around that mutation and avoid it? Or can it recognize small small differences?
SPEAKER_02So it really depends on your intended purpose. So in in some cases, a mutation could be used to specifically target um the mutated allele. So we typically have two copies of a gene, one from our mom and one from our dad. And if you have a new carry mutation where it's just mutated in one copy and you want to maintain the other copy because you don't want to knock down every bit, you could you can maybe use um mutations to specifically target, and we can use single nucleotide polymorphism SNPs to specifically knock down uh alleles with a mutation that carry mutation is disease-causing. Um, but in a disease like Alexander disease, because we have so many different mutations across the gene that can cause disease, we want to target our ASO. And because it um it's not necessarily uh shown to be a bad thing so far to just really knock down uh GFAP to almost zero. It's it seems to be real tolerated. So we're not worried about trying to um avoid the non-mutant uh allele. In this case, we we do want to avoid those those regions so that we can make sure that our our drug would be functional for for everyone um who cares in the tissue.
SPEAKER_00So I was really interested in the choice of lipids for conjugating. I understand you tested um a whole array of different lipids, but you landed on cholesterol having the biggest effect. Um, and I know that you know cholesterol is very important in the brain, it's a key molecule, it's part of the cell membrane, and that lipid metabolism can be altered in neurodegenerative diseases. So I was wondering if you've looked at all at whether adding excess cholesterol could be a problem potentially.
SPEAKER_02Yeah, that's a really good question. Um, it's certainly something that we would need to explore really thoroughly if this were to ever become something that we would want to try in a human patient. Um but I think uh, and I could be wrong, um, the the the amount of cholesterol that we would be delivering, I think would would pale in comparison to the amount that's um present in the brain normally. I think it's like 22% of the dry weight of a brain is cholesterol, as because as you mentioned, you know, it's it's really critical in in cell membranes and uh myelination, for example. Um so so it may be that that the amount we would deliver is negligible, but we we would really have to carefully investigate that.
SPEAKER_00And then back to cholesterol for a second. Um I was thinking about how you looked at different cell types and how well they took up the ASO, how well um each gene was knocked down in the specific cell type. Um and then astrocytes, um, I think the gene mallet one was only knocked down to 40%. So I was wondering if um, you know, different cell types, so astrocytes typically are generating cholesterol and not necessarily taking in cholesterol. So could that be an explanation as to why the knockdown is not as good in astrocytes?
SPEAKER_02Yeah, you know, it's a really good point. Um, we published a paper, I think, earlier this year, um, where we looked at uh uh ASO activity, so efficacy of knockdown using mallet-1 ASOs that were injected directly into the spinal fluid in mice rats and and and non-human primates. Um and we we found that generally speaking, neurons and mycoglia um were not the best at uptaking ASO and the activity was actually, it was almost flip of what was seen in this cholesterol targeting paper. So the astrocytes and the oligodendrocytes were really best in terms of the level of knockdown that was achieved at a particular dose that was given. And that's not to say that we couldn't achieve great knockdown at any dose. It was just that the amount that it took to knock it down by 80% was much less in an astrocyte and a molovidendrocyte versus a neuron and microglial cell. Now, this cholesterol paper, on the other hand, showed the opposite. Um, and that was my first thought as well is you know, astrocytes are really the main generator of cholesterol within the central nervous system. And so maybe the mechanisms that would allow that a cell to uptake cholesterol really aren't super present in. In an extracte.
SPEAKER_01So that means that depending on your cell that's being targeted, in the case of Alexander disease being the astrocytes, you are going to need different conjugates than this. And you're really going to have to tailor it to what the particular cellular uptake mechanism is in each case. Is that the idea?
SPEAKER_02Well, that could be. I mean, it could be just that you know slightly higher doses are required than if you were targeting a neuron. You know, with the I think in the extended figures, they show a mallet um one ASO, a GFAP ASO, with a particular chemistry that's that's not as potent as what's used in the rest of the paper. And I think you can see with the GFAP, you still got pretty good reductions using that that particular chemistry of the ASO. So it could be that it's it's fine to use something like this, but um, but certainly there are other uh mechanisms in uh state different stages of investigation that could be either better tolerated or better functional than the target targeting weight that that uses cholesterol.
SPEAKER_01Of course, a major question about GFAP is has always been what role it plays outside of the central nervous system. And we know there's a long list of cells that express GFAP, especially non-myelinating Schwann cells in the peripheral nervous system and pteraclea and the gut and a number of others. And nobody's ever really come up with an idea of what it means for those cell types by giving it intrathecally into the CSF. We know from your other studies that not much gets out. And so we don't worry about knocking it down in those other cells we also know from the knockout animals that have no GFAP anywhere that there don't seem to be any phenotypes in those other organs. But in this case, if you're giving high doses of a conjugated ASO systemically, it's going to be at highest concentration outside the central nervous system first before it gets in. And so it'll be um, I guess, interesting or important to see what's going on in all those other cell types and organ systems where GFAP might be expressed.
SPEAKER_02Yeah, and and you bring up an important point, which is oftentimes the the sort of genes that we're targeting um with our centrally administered with our lumbar puncture administered ASOs are things that you maybe wouldn't want to knock down in the rest of the body. So you you might feel that an anti-sense uh approach where it is centrally administered by lumbar puncture is actually advantageous because you you don't want to affect the the rest of the body. There are many examples of that sort of thing where um we feel more comfortable going after um treatment of a disease that's affecting the central nervous system because we wouldn't be affecting the rest of the body very much. So um sort of two sides of the same coin.
SPEAKER_01What more would you like to see before considering this for human use?
SPEAKER_02Yeah, I think the the toxicities that we identified, which are um sort of outlined in the extended data, are not encouraging, although they seem to be mitigated by uh by these alternate dosing strategies that I talked about earlier. So, you know, I think we'd really want to be careful about the safety. Um and then also, you know, you'd want to make sure doing studies like like the ones you described, Albi, that it's okay to reduce this uh this protein everywhere in the body, not just in the brain.
SPEAKER_01All right, now
Questions from Email
SPEAKER_01for some email. And I'll remind everyone you can send questions to axdru podcast at Wasteman. That's W-A-I-S-M-A-N dot Wisc.edu. And we'll try our best to address them in a future podcast. Here's the first from Gil. I would love to hear what you think about mini-brain research. Rachel, do you want to start off on this? What are mini brains?
SPEAKER_00Definitely. Um, so mini brains is actually kind of a deceptive term uh in the fields. We prefer the term brain organoid because um mini brain gives the idea that this is the same thing as the brain organ, just scaled down to a smaller size. And it's really not true. It's like that organ. Um, you know, it has similar properties, so it has the same cell types that you would see in the brain, although not all of them. It's missing some cell types, notably endothelial cells, which make up the blood vessels. And these brain organoids have similar functions to the brain. So the neurons that are in those brain organoids can make connections to each other, they can form synapses, um, and they can send electrical signals. So it's the same cell types and the same functions, but not necessarily all the same structures. So you won't see, you know, in the brain, you have large structures like the hippocampus, and you can generate hippocampal-like cells, but it's not the same as that structure, um, and it's not connected to all the structures that you would see in the brain.
SPEAKER_01How have they been used so far in disease models?
SPEAKER_00So typically these brain organoids will be generated from a stem cell source. So you can take those from embryonic stem cells, or now it's kind of more popular to take them from induced pluripotent stem cells. So these don't come from embryos, they come from a skin sample or a blood sample from a patient. And those are really advantageous because you get that that patient's exact genetics. And we know this is really important for Alexander disease, uh, where genetic modifiers are so important. Um, so people have used them to model diseases by generating those stem cells from patient scanner blood samples and making these brain organoids. Um, so you can do that with a monogenic disorder like Alexander disease, where we kind of know the already the causative disorder. Um, it's also very useful for diseases such as Alzheimer's disease, where we know some genes that can definitely cause it, but um in some cases it's it's can be sporadic, so we don't know what genes are causing it. And um yes, people have done all sorts of things with these brain organoids. Um, so once you grow them, uh you can collect all sorts of samples and collect the proteins that Barrett was talking about. Um, you can look at the gene expression, uh, that RNA to see which recipes are active in the brain organoids. So then you can compare them to brain organoids that you generate from neurotypical individuals that don't have this disease, and you can see um what are the differences. If there are some recipes that are more active in the disease organoids, for instance, then you may be onto something.
SPEAKER_01Of course, the brain doesn't exist in isolation from the rest of the body. So I can imagine that organoids that are restricted to central nervous system cell types like neurons, astrocytes, algodendroglia are good for some things, but to the extent that the disease reflects an interaction between the brain and liver or gut, then it gets much more complicated. And you're not going to model that with a I was about to say a simple brain organoid, but they're obviously not so simple.
SPEAKER_00Yeah, they take a long time to grow too. I didn't mention that, but um so you can keep these in culture over a year, uh a very long time. And to get the target cell that we're um we're looking for in Alexander disease research, the astrocytes, um, that can take past six months. Um, so it's a very long time. But yeah, they they are not necessarily connected to these other um important organ systems, one of which I mentioned was the endothelial cells, so that that kind of restricts how big these um brain organoids can get, and that's why people call them mini-brains because they I think at the biggest they grow up to four millimeters, so it's a little smaller than the size of a pea. Um just because um these brain organoids are bathed in growth media that has all sorts of important factors for them and that can't get to the inside without blood vessels, so the cells in the end in the inside become um necrotic and they start to die.
SPEAKER_01So there are no publications yet about using organoids in Alexander disease research, but perhaps some of them are forthcoming. You know, the idea of culturing something for a year just brings me back to my early days as a graduate student as a postdoc, and it would have been a nightmare to think about trying to keep things free of contamination for that long a period of time. Um, so I I admire the people who are manage to do it.
SPEAKER_00Yeah, it's a commitment. They can take uh daily care at some points in the in the protocol.
SPEAKER_01Right, and a lot of luck.
SPEAKER_02Yes. I wanted to jump in to say that um brain organoids can be very useful in the case where you want to try to um uh come up with a drug to treat a patient. Um, and unfortunately, there's no model of disease that's been made yet. There's no mouse or fly um that carries a genetic mutation or whatever it is that you're trying to model. And so the best um and fastest way to get to the creation of a disease model is actually to use the patient's own cells and generate these membranes and see if you can um sort of recapitulate any of those uh disease phenotypes and then you know treat them uh with a drug and see if you can have a benefit. And so we are using these now with with our ASOs, which is great. They get in there, they penetrate, and they do what they're supposed to do.
SPEAKER_01All right, now for the second email. Chiara writes, I was wondering if you have an idea of how many people are currently affected by this disease. Adults are difficult to recognize, I guess, but perhaps there is a number for younger people affected. I'm reading so many different things about it. 500 since 1949, 100 worldwide, one in a million. Um what's the correct number? And this is something I I addressed in my in my book from a few years ago. And honestly, we don't know. There's a widely cited figure of 500 or 550 that popped up in Wikipedia several years ago and has been cited ever since. I don't think it's based on any data whatsoever, uh, but it seems to have taken on a life of its own. Um, there's only one population-based study that was ever done in Japan. It was published in 2011 by Yoshida et al. And they did that by surveying all the pediatrics and neurology clinics they could identify throughout the entire country, asked how many Alexander patients they had they had seen in their practice. They tried to eliminate the the overlaps, the patients that had been seen, same patient who had been seen in more than one clinic. They got a response rate of around 73%. So they adjusted the number upward by that factor, and they ended up with a a final number of two one in 2.7 million based on the population of Japan at that time. It was uh probably an underestimate. They they did not consider patients who were very severely affected. Oh, and I should say this was um a five-year prevalence rate, because they asked the question they asked was, have you seen any patients with Alexander disease over the past five years? So it was sort of a complicated epidemiological type question. And uh and they didn't count patients who are very severely affected, um, so I'm not which they didn't really define. So I think it's an underestimate on that count. Uh, we think that there's probably an underestimate of adult patients generally, because they're often they're often misdiagnosed as multiple sclerosis or Parkinson's disease somewhat earlier in their course, and it takes several years before they eventually get to a diagnosis of Alexander disease. So we don't really know. And having said that, 500 may not be that far off. I forget how many people there are in the world right now, maybe five or six billion. If you say it's one in 2.7 million, then you'd end up with maybe a couple of thousand worldwide. Um, so it is uh it's definitely a rare disease, um, perhaps an ultra-rare disease, depending on what your definition of that term is. And there are large areas of the world that really don't report it very well. So uh we have very little information about uh Africa and parts of Russia, certainly. That's all for today's episode of Alexander Disease Research Update. Thanks for listening, and thank you to Barrett and Rachel for joining me today. Our theme song was written by Charlie Allenson, special technical assistance from my daughters, Zoe and Rebecca, and from Clark Kellogg at the UW's Wastement Center. And thanks to our donors for these podcasts, the Barrett Riddle family. I'm Albie Messing. See you next time.
