Taysha Gene Therapies, Inc. (TSHA) Earnings Call Transcript & Summary

[ Good morning, and welcome to day 2 of Taysha Gene Therapies R&D Day ].[Operator Instructions] Today, the company will present new data and updates for its pipeline candidates in SURF1-associated Leigh syndrome, SLC13A5 deficiency, SLC6A1 haploinsufficiency, adult polyglucasan body disease or APBD, Lafora disease, tauopathies and Angelman syndrome. Joining the call today is RA Session II, Taysha's President, Founder and Chief Executive Officer; Dr. Suyash Prasad, Chief Medical Officer and Head of R&D; and Dr. Steven Gray, Chief Scientific Adviser and Associate Professor in the Department of Pediatrics at UT Southwestern; Dr. Berge Minassian, Chief Medical Adviser and Division Chief of Pediatric Neurology at UT Southwestern; Dr. Rachel Bailey, Associate Professor in the Department of Pediatrics at UT Southwestern; and Dr. Kimberly Goodspeed, Assistant Professor in the Department of Pediatrics, Neurology and Psychiatry at UT Southwestern; Dr. Ryan Butler, Assistant Professor in the Department of Psychiatry and Pediatrics at UT Southwestern will join the Q&A session for Angelman syndrome. Before we begin, please note that this presentation will include forward-looking statements made pursuant to the safe harbor provisions of the Private Securities Litigation Reform Act of 1995. Please see Slide 2 of the accompanying presentation and Taysha's SEC filings for important risk factors that could cause the company's actual performance and results to differ materially from those expressed or implied in these forward-looking statements. Taysha undertakes no obligation to revise or update any forward-looking statements to reflect events or circumstances after the date of this conference call, except as may be required by applicable securities laws. I will now turn the call over to Taysha's President, Founder and Chief Executive Officer, RA Session II.

Thank you, Sarah. Good morning, and welcome, everyone, to day 2 of our R&D Day. Today, you will hear from Dr. Suyash Prasad, Dr. Steven Gray; Dr. Berge Minassian, Dr. Rachel Bailey and Dr. Kimberly Goodspeed on our SURF1-associated Leigh syndrome, SLC13A5, SLC6A1, APBD, Lafora, tauopathies and Angelman syndrome programs. Without further delays, I will now turn the call over to Suyash to kick off the discussions on TSHA-104 and SURF1-associated Leigh syndrome. Suyash?

Thank you, RA. Good morning, everybody, and I'm delighted to be here with you for part 2 for day 2 of our R&D event. We touched on several of our more advanced programs yesterday and over the course of the next 3 hours, we're going to go into some depth on some of the earlier signs on some of our slightly later stage programs. I'm going to start with TSHA-104 for SURF1-associated Leigh syndrome. Next slide, please. Now Leigh syndrome is a disease, which has been well recognized within the pediatric field for a number of years. The official full title is Leigh necrotizing encephalopathy, and it's a disease characterized by a pathological brain. The brain deteriorates over time relatively rapidly from early onset in childhood. And this is associated with developmental abnormalities that are considerable in nature, a multi-systemic disease process, and death usually in late infancy or early childhood. Now as times progressed, we've actually begun to understand the Leigh syndrome and many of the molecular underpinnings of this particular disease. And in fact, there are many different diseases that result in the clinical phenotype of Leigh necrotizing encephalopathy. Now the commonest of which is SURF1 deficiency. Now SURF1 deficiency is a -- SURF1 is a protein, Surfeit locus protein 1, that's what SURF stands for, that's encoded by the SURF1 gene. It's located on the q arm of chromosome 9 and anchors a protein localized a bit in a mitochondrial membrane, involved in the biogenesis of the structural formation of cytochrome C, the cytochrome c oxidase COX1 complex. Now this particular complex is -- the cytochrome c oxidase is a -- is involved in electron transfer as part of the respiratory chain involving the generation of ATP, and involving the generation of usable energy. In the absence of this -- of the appropriate structure of this cytochrome c oxidase COX1 structure, children are unable to aerobically respire i.e., produce ATP or energy aerobically in the appropriate way, so they tend to lean much more on anaerobic respiration, which results in elevated levels of lactate and pyruvate within the body, a disruption in energy metabolism and the clinical phenotype of Leigh syndrome. Now as you can see on the left-hand side of this slide, Leigh syndrome usually presents often initial with gastrointestinal symptoms, unusual within the first year of life. There's a progressive loss of mental and movement capability. The regression is often episodic in nature. So often, the decline -- as opposed to just being a nice gradual straight line, is the decline that it can drop precipitously at times, often associated with significant illness. So for example, the child picks up a viral illness, they will, of course, be possibly in distress of the producing loss of more lactic acid and often there could this episodic regression to that deterioration. And that can occur within 2, 3, 4 years of life. The prevalence number is about 300 to 400 patients in the U.S. and EU. It's not a common disease, but it's definitely one that causes considerable morbidity and mortality. Next slide, please. Now one of the things we do, and we haven't really touched on that much is for every one of our programs, we run, in a very disciplined way, patient-focused groups, where we spend time with between 10 and 12 families of -- and caregivers of children with the various conditions we actually have programs for. And I thought I'd give you an example of what we did for SURF1 deficiency. We pooled together a number of families and we spent 2-3 hours getting into the details of what actually affects the patients and families. And it gives us another perspective on the burden of disease of these children, and it becomes very helpful in our discussions with regulators. Because when we go through to a regulator, we say, "Well, look, we spoke to key opinion leader group and they said this, we're getting advice from you as our regulators," but also, we spoke to patients and families, and we've woven some of that thinking into our clinical trial plans. They're very impressed by that. And they're very impressed by the fact that we sincerely listen and try and weave some of these into our clinical trial endpoints, frankly. So you can see, these are some of the symptoms that patients and families have been -- have said they are most affected by their child's illness. And just as an example, we can see right in the middle of this, feeding nutrition issues. And this is often something that the physicians are less focused on, but clearly, something very important to patients and families. There's an emotional element, of course, to a mother's ability to feed a child and the child's enjoyment associated with that. When you look at the movement of motor spurt skills, of course, we talked about development of regression and issues with milestones. But the way that caregivers talk about it, it's much more personal. They talk about the fact the child -- they didn't talk about the lack of fine motor skills. They talk about the fact that the child can't draw with them. They talk about the fact that with the gross movement that's lacking like to what's affecting the child may not be able to reposition themselves when they're uncomfortable. And so this patient focus and family focus really gives us additional perspective and helps inform our clinical development plans. Next slide, please. In addition to all the patients and family, of course, we look at the literature, and I wanted to share with you some data we haven't really spoken about in too much detail, and that is a very nice, relatively detailed retrospective chart review, natural history study that was performed in SURF1. This was published in 2013 in the Orphanet Journal of rare diseases. It's a review of 44 cases of SURF1 deficiency. And when you think about the fact that there's maybe 300 or 400 patients in the world with this condition, this really is a sizable population and one that we could also use as a comparator for our interventional study. Some of the data includes the fact that, as I've already mentioned, the first 6 months is usually between -- usually below the age of a year so 9.5 months, you can see here. You can see the poor-feeding, vomiting, weight gain. These gastrointestinal issues clearly seem to be more prevalent issues. There is a significant loss of milestones, both cognitive skills and motor skills, as time progresses. Next slide, please. And you can see here some of the clinical features are actually mapped out diagramatically, so there's a visual aspect to this. Children, of course, develop respiratory failure, and that's often a respiratory failure that results in their death. Next slide, please. This is the survival curve. This was -- the 44 patients in the natural history study that I mentioned that was published in 2013, was actually supplemented with another 98 cases from the literature to map out this natural history survival curve. And you can see that the median survival for SURF1 deficiency is about 5 years of age. So it is a progressive, awful disease with significant morbidity, accumulation of disease, issues with development, cognition, seizures, resulting in an early death at 5.5 -- 5.4 years, with some extension for a handful of patients out to mid-teens. Next slide, please. That's a brief overview of the clinical perspective surrounding SURF1 deficiency. I'm now going to ask Steve, our friend and colleague, to talk through the construct and the approach that Taysha in partnership with UTSW are pioneering to treat this condition. Steve?

Thank you very much, Suyash. So the design of the vector to try to treat SURF1 was relatively straightforward. There's no known or theoretical consequence to overexpressing SURF1. It's basically a ubiquitous housekeeping gene as far as our understanding goes. So the strategy here was to use the chicken beta-actin promoter to drive basically high levels of expression in all cell types that receive the transgene. Next slide. As far as our approach to try to treat SURF1 and to measure the effects of treatment, the underlying pathophysiology of the disease is really linked to this deficiency of mitochondrial activity, which is really centered around the formation of COX complex and the resulting activity of COX1. While this is understood, we have taken -- undertaken kind of an exploratory study to just survey fibroblast from various patients that are categorized as typical or severely Leigh syndrome versus patients that have milder presentation of Leigh syndrome, just to see if -- we looked at fibroblast of those patients, if we could see a difference in COX activity, and more formally link this as sort of a relevant biomarker for the disease. And this is just the result of that investigation. This is all preliminary data. But I think, right now, things are trending to show that, basically, mild patients have just slightly higher COX activity than severe patients. And this would be encouraging if it bears out to be true because it would suggest that just a small increase in cytochrome oxidase activity could result in a very large benefit to the patients. So if we want to go to the next slide, this is just looking more granularly at this data. And again, this is still early and kind of preliminary data, but you can see the sort of the spectrum of where heterozygous patients, these would be parents of affected children, fall in terms of COX activity compared to severe patients or mild patients. So next slide. So on the preclinical side, our approach to try to investigate whether gene transfer -- SURF1 gene transfer could be an effective treatment for SURF1 really relied on the use of an experimental knockout mouse model. These knockout mice, unfortunately, have a much milder disease than humans. In fact, they have no measurable behavioral phenotype. They have a normal lifespan, but they do have some of the underlying biochemical hallmarks of the disease that I'll go into. And this includes -- they do have reduced COX activity, and as I'll go into, they do have some differences in lactic acid in the blood. Nonetheless, our approach to try to treat this disease or demonstrate a possible treatment, if you can go back one, was to do -- to treat these mice at approximately 4 weeks old. And we -- our primary route of administration will be intrathecal. But since this is the mitochondrial disease and there could be a component of muscle treatment and some peripheral organ treatments to the degree of rescue, we tested 2 different doses of an intrathecal delivery and also a combination of intrathecal with intravenous. And the combination dose was 2 -- basically 2x the high intrathecal dose. And so our outcomes in this mouse model, because it did have a robust behavioral phenotype were simply biochemical or histological in nature to look at mRNA expression to measure COX activity in various tissues and to actually measure the COX content visually in tissue sections by histology. So we did short-term studies and long-term studies. I'll go through the short term first, where we sacrificed these mice at 4 weeks post injection. Now we can go on to the next slide. So this is -- first, we looked at cytochromes oxidase activity, and this is looking in the cerebrum, the brain, muscle and liver. And so, we can see that there's about a 50% decrease in COX activity between wild-type mice and the knockout mice. The low dose had a very minor effect, but the high dose significantly improved COX activity in the brain, the muscle and the liver. And interestingly enough, when we did this combination of trying to put an intravenous administration on top of this, it really didn't increase the brain rescue or the muscle rescue, but it did show an improvement in the liver. And so, we can move on to the next slide. The -- besides looking at the biochemical rescue, another way to look at this is to stain for MT-CO1. So this is part of the COX1, the assembled COX1 complex. And so, this would not assemble properly in the absence of SURF1. But we can look hypothetically, if we restore SURF1, then it would assemble this complex and so we can stain for this in histological sections. And this is just quantification of this data. And the panels on the right are really showing, in the spinal cord, where, to be honest, in this mouse model, there wasn't -- sorry, if you go back a slide, the right panel, the knockout mice don't have a huge deficit in the spinal cord. But if we look at various regions in the brain, and this is this spider web plot on the left, then we can see the knockout vehicle has a low amount of content in the yellow lines, and then we have this dose-responsive increase in the treated mice that doesn't quite reach wild-type levels but has a significant improvement. So if we go to the next slide, the biochemical data of COX activity and the histological data measuring assembly of the COX complex, these should be telling us the same thing. And so we did a comparison of MT-CO1 -- sorry, if you can go back. So the -- so looking at the correlation between MT-CO1 assembly and the COX activity, we did see a strong correlation in these, which gives us greater confidence in this data. So now if we advance to the next slide. So that's the short-term rescue, just looking at 4 weeks post injection where we had a significant improvement in COX activity. We were able to visualize proper assembly of the COX complex. And then we wanted to see -- well, and that data really honed in on showing that the intrathecal route was -- provided a benefit, and the additional benefit of adding an IV injection on top of that was relatively modest. So to look at long-term rescue, we really focused on the intrathecal route, testing a low dose and a high dose. And in here, the first -- one thing that we could do, looking at a long-term rescue, is look at this change in blood lactate levels, and we can measure this by basically taking a blood sample to look at lactate, running the mice until exhaustion on a treadmill, and then measuring lactate at the end of that period. So we can do this at 10 -- mice at 10 weeks old and also 10 months old. And if we advance to the next slide. At 10 weeks old, the mice don't really show a difference in the lactate levels. They don't show elevated lactate. But this is the advantage of taking these mice out long term, is when we measure this at 10 months old, you can see here, comparing the wild-type mice, the dark blue, with the knockout in green, that these mice have an elevated lactate post exhaustion in the blood. And our gene therapy treatment, either a low dose or a high dose, actually normalizes this. So if we go to the next slide. In summary, we also looked in the knockout mice -- the treated knockout mice long term. And in addition to showing this exercise and this lactic acidosis was normalized. We also confirmed that the COX activity was elevated at 9 months post dosing. And we are actually -- we looked over, I think it was about 18 months post dosing, and we confirmed that the COX complex was assembled just as it was at 4 weeks old. We showed that these tissues had elevated COX activity, just like we showed at 4 weeks old. So -- and we did confirm the mRNA expression of the transgene that was dose-responsive. So if we go back. So overall, this was encouraging for us from an efficacy standpoint, where we were able to show an improvement in every aspect of the disease that we could look at in these mice. In parallel, we conducted a non-GLP toxicology study in mice. We carried that out for 1 year. There were no toxicities observed in that study. And we have an ongoing GLP toxicology in rats. All of those rats have made it to endpoint and have been taken down. They're going through the final histological assessment and bio distribution analysis. With that, I apologize for some of the difficulties with the slides, but we can -- I can let you advance to the next slide and pass this over to Dr. Suyash Prasad to go through the next steps in this.

Great. Thank you, Steve, and thanks for the really nice overview of some really lovely pharmacology data, demonstrating both functional and biochemical improvements in the SURF mouse model. And I'm just going to apologize that we have a couple of technical issues with the slides, but hopefully, that will smooth out. I think, we'll have tech team work on it. I'm now going to talk about the clinical trial plans for SURF1. The intent is to file the IND or IND equivalent by the end of this year and initiate the clinical trial shortly then thereafter. Steve has already talked about the construct, which you can see in the middle part of this slide. The current plan is to go in with a single dose of [ 5e14 ] total VG. But as you've seen from the pharmacology data, we could always increase that dose if need be. We do have some play to do that. It will be a small study, a handful of patients. The drug will be given intrathecally. We're planning to dose patients who are less than 4 years of age, so catch them earlier in the course of disease as opposed to when they're really coming to the end of their life where perhaps the degree of reversibility is limited. The drug will be given intrathecally as the majority of our other programs to give drug. You can see, there's a diagram in the bottom left-hand corner that shows where the needle is inserted. We also keep the patient after infusion of drug in the Trendelenburg position, which means head down at an angle about 15 to 30 degrees to ensure CSF flow around brain and spinal cord and maximize transduction. Once again, the 5e14 dose is a high dose being given intrathecally, but it's actually a relatively low dose in comparison to systemic gene therapies, meaning the risk of an adverse event systemically is reduced. That, in combination with our immunosuppressive regime of prednisolone and sirolimus means that we don't anticipate any safety issues when we start to dose the patients. Next slide, please. Now in terms of endpoints for the clinical trial, we spent a lot of time to keep with -- hopefully, people are still hearing me. I'll keep talking to opinion leaders -- we've -- hopefully, the people are still hearing me. I'll keep talking.

No, Suyash. I think we lost the slides for a second. So let's give our tech team a chance to get the slides back up.

Okay. Sure. I think I heard the recording might have stopped as well.

And the recording stopped as well.

Apologies, all, we'll just -- just give us a minute to get this fixed.

This is life in a pandemic. Maybe if the presenters, Steve, Kim, Suyash, let's go on mute and stop video until we get the slides worked out.

Sounds good.

All right. So again, apologies for the technical difficulties, but Suyash, we'll turn it back over to you. And maybe if we go one slide back. Thank you. Suyash, maybe you want to do this slide again just to make sure that there's continuity.

Absolutely. Hopefully, everyone is able to hear us, and apologies once again for the technical issues. I'm just going to talk now about the clinical trial plans for SURF1, following on from Steve's very nice description of the mass pharmacology data and proof of concept for our gene therapy construct for the use in SURF1. You can see this construct there, it's in the middle of the slide, full-length codon-optimized gene, CBh promote wrapped up in the AAV9 capsid. We'll be delivering a dose of 5e14-total VG. We'll be giving that drug intrathecally. Once again, it's a high dose for a drug that's stuck to the brain and spinal cord, with a relatively low dosing comparison to systemically administered gene therapies. The method of delivery intrathecally, you can see diagramatically in the bottom diagram, where we insert a needle in the L4, L5 space and infuse drug there. We cover with prednisolone for 6 months and rapamycin for a full year. We use this even if suppression ratio are standard across all our intrathecally administered drugs. And based on the results from our Giant Axonal Neuropathy study, we know this regime works very, very well, with minimal adverse events. The intent will be to start this study towards the end of this year. Next slide, please. In terms of endpoints, and I touched on the fact that we always talks to patients and families and caregivers, which help influence the selection of endpoints and goes down very well with the regulators. You can see the endpoints are bucketed into several buckets here. The important ones, of course, are the disease-specific and global assessments and the biomarker activity. So there is a mitochondrial disease scale, the Newcastle Pediatric Mitochondrial Disease Scale, the NPMDS, which has been used in many different mitochondrial diseases, of which SURF1 deficiency is one of them. So there is some data in its use both from a natural history perspective and from a clinical trial perspective. So that will be one of the key endpoints. As for the GMFM developmental progression milestones, we've already seen the milestones and development or any issue in this particular disease. Nutrition and aspects of nutrition, we look at, in particular, it's easier to measure things like the swallowing and dysphagia assessments. We're looking at seizures and a number of other developmental assessments as well as looking at communication, quality of life and EEG and MRI. From a biomarker perspective, which I think is important and were really tied back to some of the preclinical data that Steve showed with a lot of the COX activity, COX expression, lactate levels and the pyruvate levels, as we've already discussed, these patients have a preponderance to producing lactate, which has allowed anaerobic respiration as opposed to aerobic respiration. And so we're looking to see a decrease in CSF lactate and cerebrum lactate in line with some of the preclinical data that Steve showed earlier. Next slide, please. So in terms of next steps, we will be completing the IND-enabling toxicology study, submitting the IND and CTA later this year with plans to initiate the clinical trials shortly thereafter. We're making GMP drug using our commercial process currently, and we'll be completing that soon. And one thing I didn't mention is we actually have a natural history study in plan, and that is due to enroll its first patient shortly. So patients will actually be followed up in the natural history study for a period of time, being administered these particular assessment tools, and then will be rolled over on to the interventional trial once interventional trial is up and running. So we do actually have a situation where patients are acting as their own controls and we're collecting prospective natural history study as part of our clinical development program. Let me stop there, and I'll hand over now to Kim if there's any questions on the specific program.

Great. Thank you, Suyash. So we'll start with the first question from Joon Lee of Truist. Why does some severe patients have more COX activity than some milder patients?

From the -- Steve, do you want to take this? Do you want me to take this?

Yes. Well, for one thing, I want to emphasize that this is something that we're exploring right now to see if we could draw that correlation using patient fibroblast. But all in all, across patients, we would expect some variability, these are humans. And with the clinical presentation of SURF 1, it's also something where environmental components can play a big role. And anecdotally, we hear this where patients might get like a viral infection or a cold or something like that and then have a decline and then be relatively stable for a while. So when we look at patients as mild and severe, probably, the genetics play a large role in that. There's also environmental components that can tie in. So we're expecting this type of analysis to be a little bit messy and variable, but we're looking for overall trends to help guide our understanding.

I think just to add to what Steve just mentioned. I think what's really interesting is when you look at the difference between the heterozygotes, which are the parents of these patients and where these patients are, whether they're severe or mild, you can really see a nice correlation of diminished COX activity between disease patients and the heterozygote or parents. And that's kind of one of the key takeaways for us, is that if you're able to improve COX activity in any way, you should see a significant improvement in clinical phenotype. And I would say the other thing preliminarily that we're taking away is that, really, there's a small difference to in COX activity from a mild patient and a severe patient. And just a small improvement in that, it can significantly improve the patient's clinical phenotype. And this draws a lot of correlations to a lot of the other diseases that we're going after. Again, we talked about GM2 yesterday, taking a patient that is in the infantile form, which has 0.5 HEXA enzyme activity and getting them to the 2% to 4% basically normalized lifespan. You go from an infantile onset, early death of 3, to an adult onset, normal lifespan. And just getting over 5%, you significantly improve the phenotype. Similarly, when you talk about CLN1, getting in that 5% to 10% range of PPT1 activity, enzyme activity, you significantly improve the clinical phenotype. So I think, here, I think it draws a nice correlation between the 2. A small increase in the biomarker could represent a significant improvement in clinical phenotype. Suyash, I'm sorry, you're going to say something.

I think you pretty much covered most things. The only thing I was going to say is that I would just echo the fact there is going to be variability in those individual patient level. And there aren't that many samples, frankly. It's nice preliminary data that's indicative. There are many samples. But as we've talked about, what we want to see as we start to treat patients in terms of trends, is an increase in COX activity and a decrease in lactate. I think if we see that, then we're good.

Yes. And I want to jump in with one more thing. Because we look at the relevance of COX activity, I mean, this is basically whether the mitochondria are functioning properly. This is the underlying cause of this disease. I think that that's pretty clear. When we get hung up on some of these biomarkers, like the COX activity, it's also if we look in patients and we say, is it 30% or 40%, or what's the relevance, that's looking at every cell in the body. But in fact, when we're treating these patients, we're in a different scenario where any given cell that we're treating is going from the deficient level to probably a fully rescued level of COX activity. So it's more about how many cells are targeted, how many cells are rescued. And when we look in a tissue sample, we're looking at the average of that, which is different from the natural, I guess, situation with the disease, where the patients would have COX deficiency in all their cells. But overall, I think it's pretty clear that if we increase the COX activity in any given cell, we're going to rescue it. And that's the way to treat this disease.

Next question, Kim.

Next question comes from Sami Corwin of William Blair. Is the distribution of severe versus mild patients known? Do you plan on enrolling both subgroups in clinical trials?

Suyash, do you want to take that?

I'll take that, yes. So the plan is to have a fairly wide age range in the clinical trial initially. As you probably saw on the slide, we're enrolling patients from 1 year all the way up to 18 years of age, although many of them die a little earlier. And I think that we will include the majority of patients. We'll probably exclude anyone who's really, really mild. There is a very small subgroup who -- of patients who have this mild-ish phenotype that looks a bit like a form of Charcot-Marie-Tooth, CMT type 4K and they will be excluded from the initial clinical trial. But we're actually going to be -- make the study, built broader for the initial study. Yes, there's not that many patients, frankly. And so we want to really explore and understand in a little more detail matters such as COX activity, lactate levels and see what kind of improvement in COX, what kind of decrease in lactate levels actually does clinical improvement. So it's going to be pretty broad this initial study.

Thanks, Suyash. The next question comes from Kristen Kluska of Cantor Fitzgerald. Have you thought about what COX activity would be considered clinically meaningful in the trial, for example, brain or muscles, based off of natural history and findings in the mice, while recognizing the presentation is more mild here?

I mean, I can make a comment and Steve can -- basically, I think I'd probably echo what Steve mentioned earlier that more COX activity is better and less lactate is better. The details are a little indeterminate at the moment, and perhaps it's more related to number of cells transduced rather than actual specific constant expression of COX activity. Steve, do you want to add anything?

Yes. Well, I also want to -- well, we can speculate here because the mice have about 50% COX activity versus a wild-type mouse, and they have essentially no behavioral phenotype. I think the human patients, if we compare -- our comparison was against heterozygous controls, like parents. But if we -- if you look in the literature of COX activity of Leigh syndrome patients, it's actually more around 30% of a normal individual. So that may explain why the mice are so mild because they have 50% activity and patients have about 30% activity. But again, we're really rescuing on a cell-by-cell basis, that any particular cell that we target should have 100% rescue. And in terms of an outcome measure for the clinical trial, it's really going to be focused on exploratory biomarkers, but mostly clinical assessments because we're not going to biopsy the brain to measure COX activity. It's a cellular phenomenon. It's not something -- we can't measure COX activity directly in a fluid like the blood or CSF. But we can measure lactate, which is relevant to this disease.

Great. Thank you. Your next question comes from Joon Lee of Truist. He has 2 questions. One, are there NHP models for this disease and for other diseases or just mouse? And are the UTRs used across different constructs also the same?

I'll jump in here. No, there is no nonhuman primate model for SURF1 that I'm aware of, that has been described. There was a knockout pig model that was attempted by a group in Italy, and there's a publication on that, but those pigs were either embryonic, lethal or they were -- it was actually the reverse of the spectrum, or they died within the first few days of life. Yes. I'm sorry, what was the second part of that question?

I'm sorry, one second here. Let's see.

It was are on the UTRs.

Yes, it was around the UTRs.

Yes. No, they're not all the same between all constructs. We really didn't tailor each of these to the application -- the disease application. So some of them use a relatively, like a larger and stronger [ poly a ] and some of them use shorter minimal poly a, just depends on the situation.

Okay. Great. Your next question comes from Gil Blum of Needham & Company. Is mitochondrial transfer therapy being assessed for Leigh syndrome?

Suyash, do you to take that, mitochondrial transfer therapy?

I don't fully understand what's meant by mitochondrial transfer therapy.

I think that there's some publications that you can -- that mitochondria can be transferred from one cell to another, but I think that, that's still exploratory. I think that the gene replacement therapy should be a, I think, a more potent and effective way to go about this. That's my opinion.

Thank you. Your next question comes from Sue Zhong of BTIG. A small change in COX activity seems to be able to have a large impact on the phenotypes, overall. Do some parts of the body or certain important phenotypes require more COX activities than others when you try to dose to get as high COX activities as possible, well beyond 5%, for example?

Yes. I think this comes back to kind of what we've covered before, where for any individual cell that we target, we're expecting to restore COX activity to normal. We're using a relatively strong promoter. We should get sufficient circulant protein to do that. So it's really -- our efficacy is tied to the number of cells that we transduce, more than an average of some average increase in COX activity.

Great. Thank you. Your next question comes from Kevin DeGeeter of Oppenheimer. In clinical studies, will COX activity and expressible measured in CSF, plasma or both?

We measured in both, CSF and plasma. Yes. I think the CSF is going to be more relevant, but we'll measure in plasma on the side.

Thank you. Your next question comes from Laura Chico of Wedbush. TSHA-104, how do you envision the most appropriate registrational endpoint at this -- at present?

Yes, it's a good question. I think there's a few. As always, with the FDA and other regulators about totality of data and to see a number of different key endpoints moving in the same direction. You can see in this particular disease that survival is significantly compromised. So survival is a key marker here. If we prolong survival, that will be very important. Survival will take a bit longer, though, than some of the other endpoints. Another important endpoint that is related to lethality is the ability to prevent a chugging of a ventilator. You will have seen that respiratory failure is one of the key endpoints in this particular study as well. I think the next level down will be milestones. So if you can either prevent the loss of an acquired milestone to the ability to sit up right to reach out for object and/or you can encourage the acquisition of milestones that have not been acquired that should have been acquired, then I think that will be an important marker, i.e., trying to normalize the child's developmental progression from a gross motor and fine motor perspective as much as possible. And then at another level, you have a number of -- and I mentioned this earlier, a number of specific mitochondrial disease-related endpoints. The one that has most use is the NMPDS, Newcastle Mitichondrial Pediatric Disease scale, and that's going to be one of the key endpoints in our study. So I think those are probably the key endpoints. If they're supported by biomarker changes, that will be particularly helpful. And not just biomarker changes in terms of COX activity and lactate that we've already looked at, but also we'll be looking at MRI brain scans, which demonstrate significant [indiscernible] abnormalities and mitochondrial disease and also EEG basically because these do have quite exaggerated EEG activity, which often manifests as seizures. So I think that group of endpoints will follow a very nice package for registration.

Okay. Thanks, Suyash and thank you, everyone. We have ran out of time for a Q&A session now. I'd now like to turn the call over to Suyash. Back to you, Suyash, for a discussion on SLC13A5. Thank you.

Great. Thank you, Kim. So SLC13A5 deficiency, when you look at our overall portfolio across our 26 programs, they're split into 3 franchises. We have a neurodegenerative disease franchise, which includes programs such as GM2 CLN1, which we talked about yesterday. We have our neuro developmental disorder franchise, which includes conditions such as Rett syndrome, which we talked about yesterday and Angelman, which we'll talk about today. And then we have our genetic epilepsy franchise. And SLC13A5 falls within the genetic epilepsy franchise. And it's the first genetic epilepsies that we'll be talking, and it's essentially the lead candidates at our group for genetic epilepsies. So I'll be speaking about it at a high level clinically, and then Rachel will also be talking about from a preclinical perspective. Next slide, please. So from a clinical perspective and from a molecular perspective, the issue is with the sodium citrate co-transporter, which transports, sodium and citrate from outside the cell into the cell. This -- an abnormality here results in an elevated level of citrate outside the cell and a reduced level of citrate within the cell, and this can have a profound effect on children's development, their growth, their acquisition of milestones, cognitive developments and also seizure activity. And in fact, seizures are often the first sigh of something going wrong in these particular children, and they start having seizures within the first few days of life, which persists, become more severe and then an encephalopathy ensues where the brain develops in an abnormal way, you get delayed motor development, late fine motors development, issues with speech, language, cognition, developmental motor progression and somewhat unusually, there's an abnormality in tooth enamel in this particular condition. Now first-line treatment is always to try and control the seizures. And unfortunately, there isn't -- there's a whole host of different antiseizure medications that are trialed with varying degrees of success. I don't think it'll be fair to say that there is a great deal of success in managing seizures with standard antiseizure medication and a better approach to treatment of disease to try and have some disease-modifying therapies such as a gene therapy approach. We do have orphan drug and a rare pediatric disease designation. And in terms of incidence and prevalence, we're thinking about 1,900 prevalent indications in the U.S. and in the EU. Next slide, please. I just want to dive into the molecular underpinnings of this disease in a bit more detail. I've already mentioned there's an issue with sodium citrate transporter, and you see the NaCT within the actual cell membrane itself. And by this transporter, both sodium and citrate can enter the cell and you end up with a reduced level of citrate when this transporter is malfunctioning. You don't end up with a reduced level of sodium because, actually, sodium has many different types of co-transporters. So sodium can actually enter the cell through a whole host of different ways. But citrate can only enter through this particular transporter. And in fact, this transporter does transport other molecules such as ketoglutarate and malate, that are involved in the fatty acid cycle. Now in the absence of citrate, citrate is a molecule that is a key metabolizer. It plays a role in many, many different intracellular functions, in particular, relating to energy metabolism and also brain development. So you can see, in the absence of citrate, you get a reduction in isocitrate glutamate and, importantly, GABA, gamma-aminobutyric acid, which is an inhibitor neurotransmitter in the brain. So a lack of GABA might be what actually causes the seizures because you don't have this dampening down effect of GABA, resulting maybe to patients having a pro seizure -- a reduced seizure threshold and pro-epileptic activity which results in seizures. GABA is also integrally involved in brain development. And in the absence of GABA, you get abnormal brain development. So it's easy to see our lack of citrate from results in the seizures and developmental delay and regression of this particular cognition. You also have a lack of acetylcholine, lack in neurotransmitter molecule, and you can see our citrate is also involved in cholesterol synthesis, which is needed for cell membrane structure and fatty acid synthesis, which is part of energy metabolism. But ultimately, what you see from a biomarker perspective is a reduced intracellular level of citrate and an increased extracellular level of citrate. And that's one of the parameters we look at in our preclinical studies, and we'll also look at in our clinical studies. This molecular underpinning is what results in the clinical phenotype of SLC13A5, specifically early onset of seizures in the first few days of life, and then developmental delay very soon thereafter. Let's go to the next slide, please. So in terms of groupings of clinical features, you can see therefore, we've already touched on the seizures, they're generally refractory to medication, as I say, and often children succumb to complications of the seizures. There's a movement disorder that's associated. So you see very poor hypotonia, very poor muscle tone associated with dystonic movements and cardiothoracic -- [indiscernible] movements, which are [indiscernible] movements to the body. There's a development of delay. And we've already touched on the biomarker, from a citrate perspective, but also you get quite significant issues with the brain MRIs, which -- and the EEGs. There are no currently approved therapies and patients require constant supervision and care. Next slide, please. So I've touched at a high level on the clinical side of things. I'm now going to pass over to our friend and colleague, Rachel Bailey, who's an assistant professor at UTSW and is really leading the effort on the SLC13A5 program. She's going to talk about the construct and some of the preclinical work that's been ongoing. Rachel?

Well, thank you for the introduction, Suyash. So to treat SLC13A5 deficiency, my group developed a vector carrying a codon-optimized human SLC13A5 gene, the expression of which is controlled by a small ubiquitous promoter. The use of a smaller promoter allows for the more efficient self-complementary packaging. And to target the brain and liver efficiently, which, as Suyash mentioned, are critical targets for this disease, we packaged our vector into an AAV9 capsid in new CSF delivery. Next slide, please. To test vector efficacy, we used SLC13A5 knockout mice. Unlike patients, these mice do not have no phenotype and they have normal outward behavior and lifespans. However, they do have subtle phenotypes specific to the human disease. Like patients, knockout mice display an increase in extracellular citrate and abnormalities and TCA intermediates. They also have spontaneous seizures, although to a much lower extent than patients and have an increased susceptibility to chemically induced seizures. So to test the efficacy of gene replacement therapy, we've performed studies in infant and young adult SLC13A5 knockout mice. We have found equivalent results treating mice early in life and later in life, and here I'll be showing results from the young adult mice that were treated at 3 months of age, well into the disease course. Looking at the graph on the left side. When we look at the plasma in the mice, we verified that at baseline, knockout mice have increased citrate levels as compared to biotype littermates, which is shown in the blue bar compared to the gray bar on the left. When we look at knockout mice treated with our AAV9 SLC13A5 vector, we see that citrate levels significantly decreased 1 month after injection, and that this decrease was sustained to 3 months post injection, as shown in the graph on the right. These results support that gene replacement therapy can restore citrate transport and supports that citrate may be used as a biomarker for the disease and following treatment. Next slide, please. As part of the study, treated mice received untethered telemetry implants to record baseline brain activity. On the right side, the top tracing shows normal EEG activity in a wild-type mouse, while the middle panel is an example of epileptic discharges or spike trains in an untreated knockout mouse. And then the bottom panel shows normal EEG activity in a knockout mouse treated with our AAV9 vector. When we looked at the number and incidence of epileptic discharges, we found that compared to wild-type mice, knockout mice had significantly more spike trains as shown on the graph on the left. Additionally, although rare, we found that spontaneous seizures only occurred in knockout mice and were absent in wild-type mice. When we look at results from our treated knockout mice, we found that the incidence of epileptic activity was restored to normal levels and seizures were absent in vector-treated knockout mice. Next slide, please. So having found that our gene therapy restores baseline EEG activity to normal levels, we tested if we could increase resistance to the induction of seizures. To do this, we used a paradigm where mice received a low dose of pentylenetetrazole, or PTZ, every other day for a total of 8 injections. Following each injection, mice were observed for 30 minutes and seizure severity was scored using a Racine scale, as shown in the graph on the left, where 0 has no effect -- or 0 implies no effect and the numbers increased with immobilization, myoclonus, tonic and tonic-chronic seizures, with a 6 indicating death. As shown in the graph on the left, vehicle-treated SLC13A5 knockout mice are significantly more susceptible to seizures and have more severe seizures than wild-type mice, as shown by a higher score in the knockout mice in blue and compared to the wild-type mice in black. And then as shown by the orange line, we found that CSF delivery of AAV9 SLC13A5 gene therapy completely restored normal levels of seizure resistance in the knockout mice. When we look at the survival of the mice enrolled in our seizure induction study, as shown by the graph on the right, we found that 75% of our vehicle-treated knockout mice died by the last injection as shown by the blue line. This is to be expected as it's known that in humans, uncontrolled and severe seizures can result in death. Importantly, though, we found that prevention of severe seizures knockout mice by treating them with gene replacement therapy protected against seizure-related death and restore the survival wild type levels as shown by the orange line compared to the black line. We have also performed multiple non-GLP toxicology studies in wild-type mice to test the safety of a high dose of our vector, both [indiscernible] and juvenile mice. Looking at more than 1 year post injection, we have not found toxicity clinically or histologically. So overall, we have found that AAV9 SLC13A5 gene replacement therapy is safe, and that even in a developed brain, gene replacement therapy could provide therapeutic benefit as assessed by plasma citrate levels, baseline EEG activity and susceptibility to seizures and seizure-related debt. So with that, I will turn it over to Suyash to discuss what the next steps are.

Great. Thank you, Rachel. As you can see, we have some really nice pharmacology data in the mouse model that demonstrates reduction in seizures and improvements in survival. If we go to the next slide, I'm going to talk a little bit about our clinical trial plans. Now there is actually an ongoing prospective natural history study for this particular disease. It's actually being owned and run by the TESS Foundation who are the patient advocacy group who oversee SLC13A5. We're heavily involved with that particular organization of input -- and have input considerably into the natural history study design. In terms of the actual current clinical trial considerations, you can see that we're separating things out into buckets once again. So biomarkers, we've touched already. Citrate levels in plasma, urine and CSF will be an important guide. We're going to be looking at global assessments, CHOP-INTEND scale, for example, other developmental assessment cells, such as the Peabody and the Bayley. We're looking at seizures, seizure frequency, seizure severity, duration of seizures, what triggers the seizures, what types of medication the patients are on and whether we bring antiseizure medications down or not. We've already touched on the fact that, actually, most of our antiseizure medication is not that effective in this particular condition. And we'll be looking at EEGs also to supplemental look at seizures as well as communication assessments and MRI scouts. Next slide, please. So in summary, we will be completing our IND and CTA-enabling preclinical work. We're going to continue to enroll in the natural history study. As I see that that's overseen by the TESS Foundation. We will be having discussions with the regulators, both in Europe and the U.S. in the latter half of this year. And we'll be completing our GMP manufacturing, using commercial process, with the intent of starting the clinical trial with commercial material. Let me stop with that, and I'll hand over to Kim to take any questions.

Great. Thanks, Suyash. The first question comes from Laura Chico of Wedbush. Could you please speak to the role of partnerships with players like Invitae and how this affects the patient identification strategy in SLC13A5?

Suyash, do you want to take that, and I'm happy to chime in after.

Sure. Yes, I'll make a couple of comments, and please feel free to jump in. Yes, it's a really important point. And we have a very extensive punch with Invitae. To give some historical context here, when I worked in pediatric practice 20 years ago, we would see children who would develop seizures very early on in life. And after doing the basic tests, looking to see if there's problems with blood sugars, often, there'll be this group of kids who have seizures, associated developmental regression. And there was not really much you could do. You could try to an antiepileptic, sometimes steroids, sometimes a ketogenic diet to try and help the seizures. But there was no real way of diagnosing the problem. It's just a bucket of children with seizures that present soon after birth. As times progressed, we've understood the molecular underpinnings of many of these diseases. And many of the genetic epilepsies are only really being -- have only really been described in the past 5, 6 or 7 years. SLC13A5, I think it was only 5 or 6 years ago now, in fact. And now we're understanding, on a molecular basis of many of these genetic epilepsies are. In this particular condition, it's to do with the sodium citrate co-transporter. But what's really happened in terms of diagnosis, because we do know the molecular underpinnings of these diseases, what the neonatal ICU specialists do now is when they see a kid who has seizures in the first few days of life and if the basic blood test are looking for sodium levels, glucose levels abnormal, they send off a panel to Invitae or some of these other organizations. And they screen for the top 20, 25 causes of genetic epilepsies. So actually, a diagnosis is made relatively quickly nowadays. That wasn't the case 4, 5, 6 years ago, but absolutely that's the case now and the partnerships with Invitae are critical for that.

Yes. The only thing that I would add to what Suyash mentioned is, I think when you look across our portfolio, a number of the diseases that we're going after, whether they're genetic epilepsies or developmental disorders or even on the neurodegenerative side, all present with similar -- in the early days, a similar patient phenotype, in the sense, either there's some type of developmental delay, loss of milestone, seizure profile, or whatnot. And so I think what this is our partnership, particularly with the genetic identification franchises really allow a couple of things. One, they help us identify patients for natural histories and, hopefully, to then move into interventional studies. But it also allows us to find out who are the treaters and where these treaters are located, what are the centers of excellence, who are the KOLs we should be partnering with, and particularly allow us to develop more of an effective clinical landscape and profile when we have an interventional trial, ultimately, a commercial program out of it. So again, to Suyash's point, I think a couple of things, patient identification is key now that we understand the molecular underpinning, but also understanding who are the treaters and where those treaters are located.

Great. Thank you. Your next question comes from Gbola Amusa of Chardan. This is a broad question. Would you talk a little bit more about why seizure-related disorders have previously seen less interest from AAV9 gene therapeutic companies? And two, why that seems to be changing with Taysha's program and recent announcements in the space?

So maybe I'll take that on and Suyash, please chime in. It's a little difficult to kind of answer why people haven't gone after genetic epilepsies. What I would probably say, the number of monogenic diseases of the CNS that are out there were in the thousands, or multiple thousands of diseases and disease targets. So I think as the space matured and as the field matured, really, companies were focusing on what we would consider kind of the low-hanging fruit of gene therapy, lysosomal storage disorders, diseases where you had secreted enzymes, diseases where you were able to take advantage of cross correction, or diseases like SMA that you knew the target, you saw kind of transformational clinical data, but the target was relatively accessible using AAV9 within the ability of crossing the blood brain barrier and the ability to be able to transduce motor neurons. And so I think as the field has matured and the technology has matured, I think it really has allowed us to open up a number of different indications that we can now go after, genetic epilepsies being one of them. And I think just some of the profound pharmacology data that Dr. Bailey's lab was able to show, really showed the kind of the transformational possibility of having a disease-modifying therapy for genetic epilepsies, particularly where you understand the underlying biology of the disease. Suyash, please chime, if you have anything to add.

Yes. I think it's a simple matter of a comment I made earlier where only 10 years ago, we would just bucket them as genetic epilepsies. Kids with epilepsies would try and go on top of the seizures. There's nothing we could do for developmental regression. And it's certainly more recently when the molecular diagnosis has occurred more and more. I think the first paper about 13A5 was published in 2014. You'll hear next about SLC6A1, and that was described for the first time in 2015. So they're really quite new diseases, even though they've been around for years. The diagnosis, at a molecular level, is really quite new. And I think I just don't think people thought about it until more recently. I mean, I think -- or having said that Steve and Rachel and others have been working on these for some time, they have the insight to see this and try and work on it. There's only more recently that companies have started to get more interested in these diseases.

You're on mute, Kim.

Thank you. Sorry about that. The last question comes from Sami Corwin of William Blair. Can TSHA-105 rescue brain development in addition to controlling seizures?

Rachel, do you want to take that?

Sure. So unfortunately, with the mouse model that we have to study that disease, it's not a question we can answer at this time because the mouse does not have abnormalities in brain development that would be -- that are often seen in the human disease. So in talking with clinicians that have seen these patients and know this disease very well, one thing that they do comment is that on MRIs, they're relatively normal when they look at the brain. And the architecture and the cells of the brain seem to be actually quite well preserved. So it's our hope that if we can reintroduce this gene, that we can still offer benefit to the different phenotypes that we see in patients. Regarding development, that may be one where earlier treatment is going to be necessary to have a greater impact on development itself. Ever at later ages, we think we can provide benefit potentially to patients.

Great. Thank you. We do have a couple more questions, but let's just get through a couple more. Your next question comes from Joon Lee of Truist. Hi, Dr. Bailey. Is SLC13A5 a ubiquitously expressed gene? Even if it is, would you need to fine-tune its expression to normalized activity not late to excessive or deficient activity of the channel to avoid untoward effects?

So yes, this is a ubiquitously expressed gene. And so the protein is found in both neurons and glia in the brain as well as outside of the brain. Part of what we have done with our safety studies and efficacy studies is actually to look at it highly expressed in the periphery outside the brain as well as highly expressed within the brain. And so far, we have not found any indication that it is not tolerated to overexpress this particular transporter, and that the body deals well with it. So it's bringing in citrate, but we don't think it's having abnormal activity. And so we will also be performing safety studies in nonhuman primates as another measure to try to test for -- if there's potential effects from overexpressing this gene. But so far, we don't have any indication that it's negative to do so.

Great. Your next question comes from Gil Blum of Needham & Company. Is there a direct risk of mortality due to such early onset seizures? Is this is an important aspect in the natural history?

Rachel, do you want to take that? And Suyash -- or Suyash, do you want to start and Rachel can follow up?

Sure. I can make a general comment. Of course, any child who has seizures, there is an associated risk mortality, especially if the seizure is not well controlled. What tends to happen with this condition, though, is the seizures -- with most of these diseases, the seizures get worse as time progresses. So yes, there is an associated risk mortality with this particular condition that is related to the seizures. But as time goes on, it's increasingly related to respiratory failure, secondary to movement disorder and neuromuscular dysfunction. Rachel, anything else you'd like to add?

Yes. I would just add, with what is known with this disease so far is in patients, it appears that life span can be normal, except for when it's not normal, it's [indiscernible], so there's an unexplained death associated with epileptic seizures. So patients will die of -- can die if their seizures are not well controlled. And in addition to death is, as Suyash has mentioned, is regression. So children may have motor abilities or different abilities until they have multiple severe seizures that become uncontrolled, and then they may lose those abilities. So it's also important to try to control this seizure from the standpoint of not having a patient regress.

Great. Thank you so much. And that concludes our section -- or discussion on SLC13A5 deficiency. I'd like to now introduce Dr. Kimberly Goodspeed. She'll be talking about TSHA-103, for SLC6A1 haploinsufficiency disorder.

Hello, everyone. And Kim Goodspeed. I am a child neurologist in neuro developmental disability specialist here at UT Southwestern in an assistant professor level. And glad to talk to you guys about this disease condition. Next slide, please. So this disorder is relatively young, similar to SLC13A5, described in around 2015 initially. And it's caused by a loss of function variant in SLC6A1, which includes a GABA transporter, GAT1, the subtype 1. And it's responsible for reuptake of GABA from the synaptic left, and it's expressed predominantly in GABA-Ergic neurons, but it's also expressed in glia. And what we know with the patient so far is that they typically present with some degree of developmental delay, so that can be relatively broad. And most of them will have seizures. Initially, it was described in the myotonic -- or myoclonic atonic, excuse me, epilepsy conditions. However, on further review, it looks like many of these patients actually also have absence seizures, which are staring spells and they can also have atypical absence seizures, which are prolonged staring spells, typically accompanied by a characteristic EEG abnormality with generalized spikes at the 3 hertz frequency. There's no approved therapies and the epilepsy in some patients can be quite difficult to manage, many patients requiring multiple antiseizure medications to attain adequate seizure control, and some patients never attain control. And when -- many of the patients will also have an autism phenotype, and can have all of the comorbidities that can come along with that, including sleep disorders, GI distress, and maybe behavioral disorders or ADHD, as well as learning disabilities and intellectual disability. Right now, there's an estimated prevalence of 17,000 patients spread across the U.S. and the EU, but that number is a little bit influxed. As some alluded to earlier, it's also included on the NBT panel. And so the primary patient foundation, SLC6A1 Connect is constantly receiving new referrals and connections with new patients that are being identified through those genetic identification efforts. Next slide. We have a couple of videos. We have an ongoing natural history study. And one thing I haven't talked about yet is a particular movement disorder. And for this condition, we're still trying to define this disorder. But what we can see is that the patients are very uncoordinated. So we've got 2 different toddlers here. This first video is going to be of a 3-year-old and trying to capture him running. And if -- not many of you have seen a 3-year-old from before, I can assure you this is a pretty abnormal gate pattern. And you can also see the distress and in that kind of behavioral overlay where he's constantly reaching towards mom and in the midst of a potential tantrum. And I can tell you after this video was completed, that became a full-blown tantrum that was pretty distressing for mom and staff. We can go to the next video. In the next video, we'll see an almost 4-year-old. And again, you see that kind of poorly coordinated gate pattern, very discoordinated run. And he actually also trips. And it's always a question, and that could only be answered if you had them hooked to an EEG, but these patients do have atonic seizures. And so you'd wonder if some of that could have been a slight negative myclonus or an atony that led to that fall. These kids tend to be pretty clumsy. And so what we know from looking at the literature and kind of scouring publicly available databases for patients with this condition is that the developmental profile can be fairly broad. We collected over 100 cases that we had some degree of phenotypic data on. And you can see in the top graph, about half of those kids, we had some degree of developmental assessment. And prior to seizure onset versus after seizure onset, orange bars are before, green bars are after, there is somewhat worsening of their developmental abilities after the seizure onset. But there's still a pretty big spectrum from normal development and intellect, all the way to severe developmental disability and intellectual disability. In terms of seizures semiology, I mentioned that many of these children were initially described in the myoclonic atonic classification of epilepsy. But actually, in this group, the majority of cases had absence or atypical absence seizures. Next slide. And when we look at their EEGs, these are qualitative reports that were available to us, the vast majority were abnormal, but had variable degrees of abnormalities, with most of them having generalized spike in wave discharges, but some having background slowing in focal spike in waves. Next slide. And we have the benefit that this disease is also included on the Simon Searchlight patient registry. And so they're collecting, prospectively, survey data and questionnaires on patients. And so we've scoured the initial 30 to 50 patients who have been enrolled, and we have find an adaptive scale behavior and as well as a couple of questionnaires on autism to start to get a sense of how these patients perform on different standardized measures of development and behavior. And we're using these primarily to inform our natural history study, as well as clinical endpoints that we may use in the trial. I think what's striking, just if we take an overall look at these is that there is progression over time. So they're chronological age on the left-hand graph of the Vineland Scales, is plotted along the X-axis. And their age equivalent is plotted along the Y axis. And what you can see is there is an acquisition of skills over time. However, they are falling well below their age-matched peers in terms of age equivalency. On the right-hand side of the graph, you can see that they are typically presenting with autistic traits and are scoring higher than you would be expected for neurotypical children on autism screeners. Interestingly, between the 2 different questionnaires, the central responses in scale tends to pick up more repetitive behaviors and rigidity, where the social communication questionnaire doesn't quite get that granular on those aspects of the autism phenotype. And clinically, these children do typically present with very strong symptoms in rigidity and stereotype behaviors and repetitive behaviors. And this is also a part of the autism phenotype that we don't have great treatments for. Next slide. And so with this, I'm going to turn it over to Steve to talk more about the preclinical plans.

Yes. Thank you very much. I'll just spend a little bit of time walking through kind of where we are with this program. It's still at a relatively early stage compared to some of the other programs, but we've generated some encouraging data that I'll share with you today. This is a straight gene replacement approach. We're packaging the full-length SLC6A1 gene into a self-complementary genome package in AAV9. So if advance to the next slide, to evaluate the potential for this to treat SLC6A1, we utilized, primarily, SLC6A1 knockout mouse. This is a full knockout mouse. And I want to sort of clarify, the patients are due to a haploinsufficiency, so that patients would all be heterozygous for the disease causing mutation. But we utilize the knockout mouse because it basically has a pretty profound phenotype that we can easily score. It has decreased body weight, it has a noticeable tremor, hind limb clasping, poor motor coordination, anxious behavior, autistic-like behavior which you can score with decreased nest building, and it has a profoundly abnormal EEG. So actually, the knockout mouse model is considerably more severe than any of the patients. But like I said, sometimes it's useful to really challenge the treatment to go after a more difficult model. And we can also do some experiments in the heterozygous mice that would match the genotype of the patients, but those have, say, very mild behavioral symptoms, but they do have EEG and abnormal EEG that we can score. And I'll show you a little bit of data on that. So if we go to the next slide. The primary thing that we've done so far is to investigate sort of the early proof-of-concept work with this, doing neonatal ICV administration of TSHA-103 into knockout or heterozygous mice. And when I say that the knockout mice have a profoundly abnormal EEG, I think you can look here, this -- they actually have abnormal spike trains, about 50% of the time, and treatment, neonatal injection of TSHA-103 into the knockout mice normalizes this EEG phenotype considerably. And so if we get to the next slide, we quantified this by looking at the total duration of these spike trains as a duration of recording time. And so, if we look -- the left panel is the experiment in the knockout mice. And as I said, you can see that these mice have profoundly abnormal EEGs and it's 50% of the recording time. So 50% of the time, they're actually having an abnormal EEG, and the treatment with TSHA-103 significantly reduced that abnormal EEG phenotype in the knockout mice, with some mice actually getting almost a complete absence of this abnormal EEG pattern. Now looking at the right panel, this would be the heterozygous mice, which genotypically would match the patients. And in this case, whereas the knockouts have abnormal EEG 50% of the time, these mice have an abnormal EEG about 8% of the time, and treatment with TSHA-103 considerably reduced these seizures and got them close to a normal EEG pattern. So maybe I'll move on to the next slide and pass on to Dr. Suyash Prasad to kind of talk about what the next steps are for this as we move it forward.

Great. Thank you, Steve and thank you, Kim. So Kim, obviously, Assistant Professor, UTSW and Pediatric Neurologist, talked us through the clinical phenotype of SLC6A1. Whether or not other treatments. And now we're moving forward into some additional preclinical pharmacology work and toxicology work to finalize doses and dose response and establish a therapeutic window before we start the clinical trial. There is a natural history study that is being initiated by the patient advocacy organization, and we're -- we work closely with that particular group, SLC6A1 Connect. And we're in the process of developing our interventional trial currently and giving some thoughts into what we would plan with regard to dosing, design and end points. On that note, let me hand over to Kim -- to Kim Lee to talk about -- to take some questions.

Great. Thanks, Suyash. Your first question comes from Joon Lee of Truist. Hi, Dr. Goodspeed. Isn't GABA a depressant? There are many companies developing GABA analogs for epilepsy. What's the mechanism for excess GABA causing seizures?

That's a really great question. And I think it's ultimately being answered and studied in different labs that are collaborating with the SLC6A1 Connect Foundation. But our hypothesis right now is that it's partly that there could be an over exposure to GABA in that synaptic cleft, and that may have an effect on the density of the GABA receptors on the presynaptic terminal. But it's also possible that GABA is diffusing away from the synapse where it's meant to be and not being reuptake into the presynaptic terminal, which can lead to 2 different things. You can have off-target GABA effects in regions of the brain that wouldn't typically see GABA or wasn't planning to see GABA. And then you also have a deficiency of GABA or depletion of GABA stores in the presynaptic terminal. But those answers ultimately are being explored in labs in electrophysiological studies.

Great. Thank you. And Joon Lee also asked, can you remind me what -- why you think this is a haploinsufficient condition and not a dominant negative?

It's another great question that has been tossed around a lot in the community. And it's difficult to exactly define, but experts have weighed in at multiple SLC Connect-organized scientific meetings. And the discussions have decided that the vast majority of TCC like haploinsufficiency disorders. However, there are a couple of cases and possible mutations that are being looked at in more detail to see if they may effect, for example, how the healthy copy is being transported through the golgi apparatus and expressed at the cell surface. So again, some of those studies are ongoing, but the vast majority of the field believes it's haploinsufficiency for the majority of mutations.

Great. Thank you. Next question comes from Silvan Tuerkcan of JMP Securities. Are you still determining the best promoter? And what are the design choices? To what level would you need to increase expression levels in humans?

So maybe, Kim, I'll take that one, at least the promoter question. And then Kim and Steve, maybe you want to chime in on levels of expression. But from a promoter perspective, what we've done is undisclose the promoter. We have a lead construct, and that's the data that you guys saw today, and that's the actual construct that we're taking forward in our dose-finding pharmacology studies as well as our IND-enabling NHP tox studies. And so we'll be disclosing that full construct at a later date, but we currently have a lead in multiple backups behind it. This is a pretty important program for us. But a lead has been identified that we've hit animal proof-of-concept, and we feel extremely excited about. So more to come. Maybe Kim, Steve, you guys want to chime in on expression?

Yes, maybe I'll weigh in first. So this is something we're honestly taking very seriously. We want to be really careful because obviously, GABA levels up or down can have -- can potentially have a profound impact to patients, so this is one reason why this program is going a little bit slower than some of the others just because we're trying to be very, very careful. We're trying to examine these issues very carefully, seeing what cell -- if we need to tweak some things to get cell-type specific expression or not. Basically, that's why we're leaving it open ended right now. The data that I presented is with the lead candidate that is looking very promising, but we want to really conclude a very detailed analysis before locking anything in.

Great. Thank you very much. Your next question comes from Sami Corwin and Raju Prasad of William Blair. A couple of questions here. What level of Gap1 restoration do you think will result in a meaningful clinical benefit? And are there concerns about Gap1 overexpression? Also, Dr. Goodspeed, what is the best clinical measure of the gait? Is it CHOP INTEND?

So maybe I'll take the one about overexpression. I think Steve just addressed that one. Maybe, Steve, Kim, maybe you could take the first question around what level of expression do you feel would significantly improve the phenotype.

Yes. Well, I mean we're trying to address that experimentally. We have lead construct, we have some other constructs that we've tested. And overall, when we treated mice in kind of a typical safety study design and taking them out long term, the safety has actually been quite good. We haven't seen any significant adverse effect. But theoretically, I could anticipate that if we express this at too high of a level in the wrong cell type that it could generate adverse consequences. So again, this is -- one of these things where we're looking at theoretically and anticipating the possibility of problems, and we're trying to just take this very slow and careful.

Thank you. And RA will you address the best clinical measure for gait and CHOP INTEND?

Kim?

Sure. Sounds good. I think that's become an important question across many disorders, right? It is how do we define their gait. We're looking at a couple of things in the natural history study, including a parent response, developmental coordination questionnaire as well as taking videos of their gaits in clinics so that we can see if there's any patterns that emerge as we see more patients. And then we're also performing the SARA ataxia rating scale in clinic. Some of the younger kids can't perform on everything, but we can at least score pieces of it. And so that's something that we're thinking very critically about. There's also a number of groups that have different technically savvy ways of measuring gait more objectively and quantitatively. So there's still a lot of thought being put into how we can measure the gait. I'm not sure CHOP INTEND is going to be appropriate for this population but -- because I think that they will outperform and probably score at the ceiling on that measure, but it's certainly a consideration we can think about because it is, of course, established in the field, especially in the SMA programs, but there's a breadth of difference between the motor function of these kids and kids with SMA.

Great. Your final question comes from Yun Zhong of BTIG. When would a physician consider genetic testing when treating a kid with seizures?

I can tell you, in our center, we think of it immediately. So our center is very -- is utilizing the Invitae free genetic quite a bit. Part of that is regionally in Texas, access to genetic testing for our patients with publicly funded medical insurance is lacking. But any child with an unprovoked seizure, I'm constantly looking for their Invitae panel, if it hasn't been sent, we're sending it. But in general, the American Academy of Pediatrics as well as Child Neurology Association and -- or Society and the American Association of Neurology have guidelines that any child with global developmental delay, so a delay in 2 domains or more would also qualify for genetic screening tests. So there's a number of kind of safety nets in place that are at least there to recommend patients to be tested.

You're on mute, Kim.

Thanks, RA. Thank you, everyone, for participating. This closes out our section on SLC6A1. And I'm happy to introduce our next speaker, Dr. Berge Minassian, who will be discussing TSHA-112 for adult polyglucosan body disease or APBD. .

Hello, everyone. Thank you. My name is Berge Minassian. I'm a child neurologist and adult neurologists. I'm the Chief of Child Neurology at UT Southwestern. We have a very large program with 30 faculty and 40 trainees, residents and fellows. We're the largest training program in child neurology in the nation. The program is very much geared now towards gene-based therapies. The reason being that much of adult neurology and most of child neurology are genetic conditions. I will discuss today 2 diseases: one in adult onset disease and one on a childhood onset disease. They are connected by the basic mechanism. So they are kind of sister diseases at the molecular level. The first one is this adult disease called adult polyglucosan body disease. If you could go to the next slide, please. I will describe this disease starting with the -- essentially, the pathogenesis of the disease, it's easier that way. It's a disturbance in glycogen metabolism. Normally, glycogen is synthesized by 2 enzymes, 1 called glycogen synthase that elongates the chains of glycogen and glycogen branching enzyme, which places branch points. These 2 enzymes need to work in concerted fashion so that the molecule grows properly into the sphere that it is, which is very much hydrated and keeps it soluble. Whenever the balance between these 2 activities is disturbed, if it is disturbed in favor of glycogen synthase, meaning if glycogen synthase outpaces the branching enzyme, what happens is that glycogen acquires chains that are too long and this leads the glycogen molecule to precipitate. Once it precipitates, it's no longer accessible to enzymes that digest glycogen, and therefore, it will accumulate gradually into structures called polyglucosan bodies, and I'll show you pictures of those in the next talk. And over time, as this polyglucosan bodies accumulate, they drive a neuroinflammatory and a neurodegenerative process with resulting phenotypes. In this case, we're talking about the deficiency of branching enzymes or branching enzyme deficiency, leading to glycogen synthase activity outpacing it and creating glycogen molecules with chains that are too long. This disease presents around age 50 and proceeds basically very similar to amyotrophic lateral sclerosis with upper and lower motor neuron signs, bladder control problems, ataxia. And these patients are wheelchair-bound within some years and eventually even acquire some dementia and die prematurely. So this is the disease. And what we're trying to do in this project is to restore the balance between glycogen synthase and glycogen branching enzyme by reducing glycogen synthase activity so that it matches that of branching enzyme and corrects the problem. We've shown that this works very well with all sorts of experiments using transgenic mice. And in this project, we're basically developing an intervention, which Suyash will describe in a minute. And I'm happy to take questions after. Dr. Prasad?

Great. Thank you, Berge. If we go to the next slide, please. And it's been really wonderful having Berge join us for this session. He's actually traveling. He's not at UTSW at the moment, so pleased he could make it. He's been a wonderful partner for us, him and his team over at UTSW. He leads the pediatric neurology department, and he's a world expert on conditions such as Lafora and APBD. So as Berge has already talked about biochemistry, he's talked about restoring the balance between glycogen synthase and the going, which produces glycogen and the destruction of glycogen metabolism of glycogen on the other end. So the approach that we're using, and this approach has really been led by one of Berge's bright young researchers, Emra [indiscernible], and he has been running a number of proof-of-concept studies in the animal. And what you see here in the construct is the -- is a knockdown approach to treating glycogen synthase. So this is actually -- it's a small interfering RNA on a micro RNA backbone wrapped up in a self-complementary capsid and driven by the CBh promoter. And the intent is to bring down, to reduce the production of glycogen, therefore, trying to help restore the balance in the biochemical pathway that Berge just outlined. If we go to the next slide, you can see some of the work from the APBD knockout model. And you can see the western blot loss on the left and the expression of glycogen synthase on the right. And what you see here is that -- this was a preliminary construct that Emra was working on, which included the knockdown approach to treating a glycogen synthase. But this was also associated with GFP, so you could actually see expression of protein. And the -- you can see on the left-hand side, you have the PS-treated mice on the right-hand side of the western blot, you've got the true mice treated with a specific construct And you can see here that as GFP levels increase, IGFPs being expressed, you see a reduction in GYS1. If you look at the miRNA bands and compare it to the PBS bands, you can see they are lighter. And those -- the lines of the band actually corresponds to the degree of expression that have seen in the GFP, i.e., we are seeing a reduction in glycogen synthase activity, which you can see reflected in the graph on the right-hand side of the screen with the mRNA-treated, construct-treated animal models. Of course, this is what happens at the cellular level, what happens in terms of histologically. If we go to the next slide, you can see -- Berge has already talked about the polyglycosan bodies and how they impair clinical functioning and contribute to the phenotype. And you can see in this particular slide a decrease in the polyglycosan body formation, specifically in the hippocampus of the brain of these mice. Go to the next slide, you can see, therefore, that we've got animal proof-of-concept achieved with this knockdown approach. We're now moving forward to dose response, more complex pharmacology work and then their actual performance of toxicology work as we start to think about the clinical trial. And we're already giving some thought to interventional clinical trial protocol currently. So it's one of our earlier programs. We're very excited about it, and I'll hand over now to Kim to facilitate any questions we might have.

Great. Thank you. I think we should -- let's move on to Lafora disease as Berge will also be talking about that topic as well. And then we can combine questions. So Berge, I will turn it back over to you. Thank you.

Sure. Can you go to the next slide, please. So here, again, the pathology is that glycogen acquires chains that are too long. The enzymes in question here are called Lafora and Malin, which my lab discovered. One is a glycogen phosphatase, the other is a ubiquitin E3 ligase. It's a very long story and complex story as to how these enzymes are involved in the structure of glycogen. And a lot of it is still quite speculative. We don't really know. This whole area of control of glycogen architecture have been just neglected and unknown. It was assumed that simple concerted activities of chain elongation and chain branching is sufficient to produce perfect glycogen molecules. Without this rare disease, we wouldn't know that there's this added level of regulation. Unfortunately, there's no time to go into what we know about what these enzymes do. But the bottom line is the end product is the same as the stride for APBD. Glycogen now has chains that are too long, making it prone to precipitate, accumulate, form polyglucosan bodies, which, in this case, are called Lafora bodies, which drive the disease. Next slide, please. These are Lafora bodies. On the left, you see Lafora body completely occupying the cell body of that particular cell right in the middle of the image. And then several other Lafora bodies, which are these round dense brownish structures in this image. In B, you see a Lafora body that is so big that it outgrew the cell it was in and has destroyed the cell it was in and it's freely floating in the neuro pile. And C, the pink structure smack in the middle of the picture is a juxtanuclear Lafora body. And in D, again, right in the middle, at the electron microscopy level, is a large -- well, really small, but in the picture, it's large Lafora body, which is occupying a food process of an astrocyte near a synapse. Next slide, please. This is Chelsea. We showed this with permission from her family. I want to make a very important point by showing you this picture, and this point applies to the APBD disease. You see Chelsea here, at age 14, she's already been diagnosed with Lafora disease based on the fact that she had some seizures and some myoclonus. But you see how bright she is, how great she looks and how you wouldn't tell there's anything wrong with her. I mention this because in these diseases, which are neurodegenerative diseases, there is a lot to save. Imagine preventing this kid from deteriorating to what I will show you next. And the same applies to APBD. Here is Chelsea at age 21. And through the years, from that first picture to this one, basically what happened is she had increasing amounts of seizures, which soon became intractable. She developed visual hallucinations, which are invariably scary. As if there's not enough in these patients' lives, the hallucinations are not pleasant, always scary. She developed a gradual dementing process, which is of a frontal lobe major resulting in this inhibition and eventually reaching a vegetative state and you see the blank phase that she has in this picture, if the team can show the short video in the next, please? This is just a -- yes, there you go. So that's the state she was in with the constant myoclonus basically all day long. And then she died soon after that. So again, restoring the balance or glycogen synthesis and branching enzyme, meaning downregulating glycogen synthase to lead to proper length of glycogen branches basically rescues the animal models as we've shown in many a transgenic experiment. And now we have data with the intervention with this drug TSHA-111 that we are proposing to advance. And this drug, obviously, will apply to both these diseases and a number of other conditions. And again, these are neurodegenerative diseases where the child or the adult is healthy at the beginning, and we can save them from what they would otherwise experience, they and their families. I'll turn it back to Suyash.

Thank you, Berge. Thanks for the very nice overview of Lafora disease. And yes, Berge is actually correct. We're using the same construct or the same construct to treat APBD, Laforin and the two forms of Lafora disease, Laforin deficiency and Malin deficiency. Once again, it's a small interfering RNA on an mRNA backbone that is wrapped up, it's driven by the CBh promoter and wrapped in AAV9 capsid, and the intent is to knockdown GYS1 glycogen synthase activity and production of glycogen therefore. Next slide, please. This is a similar slide to the one we saw previously for APBD. So we looked at 3 separate mouse models in APBD model and Laforin knockout model and the Malin knockout model. And what you see in all of them, and as you see from Laforin and the Malin knockout model, this is the same construct as I showed previously, where you are knocking down the GYS1 activity. You do have a GFP associated with this particular construct, you look for expression of the construct. And you can see both in the top western blot and the bottom western blot, a reduction in the darkness on the -- of expression with the mRNA-treated Laforin knockout model and the Malin knockout model, which corresponds to a reduction in glycogen synthesis. You can see the activity is reflected on the right-hand side, so you see the relative expression of GYS1 is reduced. When you look at things qualitatively, on the right on side of this image, and go through the next slide, what you see here is a reduction in the formation of Lafora bodies, which is the actual histopathological characteristics that then accumulate and can then go on and result in some of the clinical features of both -- of Lafora disease. And as Berge rightly pointed out, this approach actually is something of a platform approach. There are many glycogen storage disorders where this approach might be deemed suitable for, i.e., any kind of disease where you're trying to reduce the production of glycogen, i.e., anything where there is a glycogen storage that results in a disease, this approach might well be useful for. Of course, what we're trying at the moment is APBD and Lafora disease, 2 different forms of Lafora disease. Next slide, please. So we've got another proof-of-concept achievement, knockout mice model of TSHA-111-LAFORIN, TSHA-111-MALIN and of course, APBD. This is an early program, so we're still in the process of evaluating dose response and age response. And once we finalize the dose from our pharmacology studies, we'll then go on and do toxicology and think about the interventional clinical trial. We've already given some consideration to endpoints and study design, but this is a little way off for the time being. So let me stop here and hand over to Kim for any questions on APBD and Lafora disease.

Great. Thank you, Suyash. Your first question comes from Joon Lee of Truist. MicroRNAs typically have many targets. What have you done to ensure the target specificity?

Berge, do you want to take that question around off-target?

Yes. So what you saw is the construct with the GFP. The final construct will not have the GFP. That's important to mention. And currently, bioinformatics has been used for off-targets, but we are doing RNA seek experiments to make sure that our final microRNA is not downregulating anything else to any significance.

Great. Thank you, Berge. Next question comes from Yun Zhong of BTIG. Are there any natural history data available for polyglycosan body disease and Lafora disease? Or are there any ongoing efforts to conduct such studies?

It's a great question. There is extensive natural history study for adult polyglycosan body disease. All the patients have -- not all, the large majority of patients have 1 of 2 mutations, and they are in a very tightly knit group. It affects predominantly patients with Ashkenazi Jewish background. And so that foundation has been very strong, and there's a lot of data, a lot of it published. Lafora disease is spread out across the world in all ethnicities. We are presently doing in the midst of completing a natural history study of Lafora. Now Lafora disease has been known for 100 years. We have abundant data on its natural history, but obviously, you all here understand the importance of doing a formal natural history study, which we're in the midst of completing as we speak.

The only thing that I would add to what Berge was saying is, particularly APBD, this has been studied extensively in the Ashkenazi Jewish population, as Berge mentioned. And what we found out through some of that exploration and investigation is that a number of ABPD patients are actually misdiagnosed with multiple sclerosis. And what you actually see is, and as we all know, multiple sclerosis is somewhat of a clinical -- kind of day clinical diagnosis and collective of a number of different under -- clinical underpinnings that go into that. And it was thought that a large percentage of patients that are actually presenting, particularly of the Ashkenazi Jewish descent, that actually have been diagnosed with MS actually have a genetic underpinning and a mutation in APBD. And as a result, would be appropriate for treatment here. So this is ultimately how we get to our prevalence estimate of around 10,000 patients in the U.S. and in Europe because of the large percentage of patients that are actually misdiagnosed with MS and that should be appropriately genetically screened for APBD.

Yes. I would also add that it's not just MS actually, but there's now increasing reports of patients who have dialysis, ALS actually diagnosed, and they have APBD as well. And to the comment about the natural history, Berge is quite correct that there is quite of a natural history data already out there. There's actually a really nice papers published. I don't know if you saw it first by the Italian group, recently on Lafora's disease, which I thought was quite extensive. And for those of you who are interested in learning more, the lead author is Federica Pondrelli. We can send the paper out, but it came out just 3 or 4 weeks ago, I believe.

Great. Thank you. Your next question comes from Silvan Tuerkcan of JMP Securities. What is the heterogeneity and severity in these diseases? Is the aim to arrest progression or reverse?

Berge, do you want to take that?

Sure. That's a great question. So again, that's why I emphasized the large window we have in both cases between diagnosis and neurodegeneration. The Lafora patients, when they're diagnosed, they're just like any other epileptic patients and the occasional seizures. The APBD patients are having some minor difficulties by the time they're diagnosed. So indeed, for the time being, the treatment aims to prevent progression, which is tantamount to a cure, if you start early. To what degree these treatments will help advanced cases? We do not know. There's a suspicion that if we stop more polyglycosans from forming and accumulating, that there may be systems themselves to clear those to some degree, such as autophagy. But that, in these diseases, the amount produced exceeds the amount that can be cleared, and there's a net accumulation. If that is the case, then stopping formation may lead to actual clearance of what's there. But this, we will only find out when we actually try. Certainly, we should be able to stop progression of the disease.

And I think to just add to what Berge said, it's extremely important. Because the intervention window is so wide, particularly, you saw the picture of Chelsea at 14 years of age. You saw most -- the APBD patients, it's a prime of life disease is how it's described. So most of these patients present at around 40 to 50 years of age. So the interventional window is actually pretty wide if we're able to identify these patients through genetic screening and genetic testing early on in life. And to Berge's point, you may be able to stop the disease where it is before they actually -- patients actually present any type of system -- symptoms, which would be extremely important. I think also something that both Berge and Suyash mentioned is the fact that this program is somewhat of a pipeline and a program, as you would think about it, and the fact that we've already achieved animal proof-of-concept in 3 genetic models of the disease, 1 in APBD, 1 in Laforin and 1 in Lafora Malin. And so the thought is that if we're able to prevent polyglycosan body formation across these 3 models, we may be able to expand this at some point to other glycogen storage diseases like Pompe and a bunch of others. And so really, I think the opportunity with this particular program is quite unique and one that's very different from kind of the more gene replacement programs that we have within our portfolio, the fact that we could take a single construct and perform IND-enabling and CTA-enabling toxicity studies across a number of different -- that would be attributable to a number of different diseases, and then just need to do kind of the very specific -- disease-specific pharmacology studies to get dose response, I think, can really open up a really large landscape of indications that we could potentially go after with a single construct. I think this is really important from a value creation perspective.

Great. Thank you, RA. Your next question comes from Gil Blum of Needham & Company. Is there potential for systemic downregulation of glycogen synthesis by targeting just one? If so, are there risks associated with activity outside of the CNS?

Berge, do you want to take that? I think we may have -- I think Berge may be frozen. Suyash, do you want to take that in Berge's absence?

Sure. So could you repeat the question, please, Kim?

Sure. Is there potential for systemic downregulation of glycogen synthesis by targeting just one? And if so, are there risks associated with activity outside of the CNS?

Yes, it's a good question. We're going to be giving the drug intrathecally. And so the target is the brain and the spinal cord. So yes, we do want to reduce glycogen synthase in the brain or spinal cord. You will, of course, get some leakage in the systemic circulation. So there is a risk of reducing the production of glycogen outside the CNS. And the -- I guess, the theoretical risk is that of hypoglycemia, i.e., glycogen, which is used as an energy substrate, gets broken down to glucose. But the reality is the liver has so much glycogen, it has many methods for making glycogen. I think it's going to be inconsequential outside of the CNS and the PNS. So no, I don't think there will be any risk of, of course, the only kind of off-target fact outside the CNS. But we will need to look for it. We'll look for it in the preclinical studies. We'll look for it in the clinical studies as well. But I think it's highly unlikely.

The only thing that I would add to that is, theoretically, you would need to -- in order to basically eliminate glycogen, fully knockdown GYS1, you would basically need to transduce every cell. And we understand that from just the technology we have today from an AAV perspective, particularly AAV9, we're not -- we're going to give good transduction in the CNS particularly using CSF delivery, but you're not going to transduce every cell in the CNS, i.e., even when you have CSF turnover, you're going to have extremely limited exposure to how many cells that you transduce. So I think the risk here is relatively low. Obviously, we'll need to look for it in our IND-enabling toxicology studies. But the risk is relatively low. And this is something that's been talked about pretty extensively with Berge's lab.

Yes. You know, we're not turning off glycogen. You can see, it's reduced glycogen synthase. And that's in the pretty high doses, 70-11 in the mouse data that we've just shown. So you will not be turning off glycogen synthase. And RA is quite right. You won't trust this every cell over.

Great. Thank you for that. Your next question comes from Sami Corwin of William Blair. Could knocking down glycogen activity have any adverse effect on other biological processes? Do you think additional preclinical or clinical data from either TSHA-111 or 112 will derisk the other programs?

Yes. Kim, I think we just addressed -- kind of just addressed it in the last question.

Yes. Great. Next question comes from [indiscernible] of BTIG. Proteins accumulate over time, early diagnosis seems to be important. Below what point or threshold that therapy can deliver the best outcome?

Suyash, do you want to take?

I can take that. Yes, so it's -- this is an important concept for our whole portfolio of diseases. How well do you need to treat to really make a considerable clinical impact. Or the other way of looking at the question is how late is it. How late can you go before treating doesn't actually have any meaningful change, i.e. at what point does this disease become reversible. And it's a disease-specific discussion we need to have for every program. So for example, if you joined us yesterday, you will have heard that with GM2, there's a very rapid decline. There's an ongoing progressive rapid loss of neurons, once you've lost a neuron, it doesn't come back. So once a child is not responsive, needing a ventilator, not moving in GM2, I think that's likely to be a nonrecoverable patients. So the earlier you treat, the better the outcome and if you can treat at birth or soon after birth for that particular disease, you will see a much better outcome. Now for Lafora and for APBD, it's a little different. Once again, there is accumulation of a particular aggregate in Lafora disease is for Laforin bodies and APBD is polyglycosan bodies. There's actually not really much difference between the 2 histologically. It's just accumulation of glycogen in different cell types. With APBD, as you've already heard, the progression is longer. It takes longer to accumulate these bodies and actually patients remain relatively asymptomatic until the prime of their life, their 30s or 40s at which point diagnosed -- symptoms emerge and diagnosis becomes more relevant and more important. That's less the case for Lafora disease where diagnosis is often made at an earlier stage. Berge talked about the case of the young lady who you saw who had symptoms in their teenage years and succumbed before the age of 20. So I think with APBD, you've got a very, very wide opportunity. If you can treat it before even symptoms are acknowledged, i.e., you can divest the patient presymptomatically and treat, that's probably the best thing. Even the early stages because it's so slowly progressive, you have a very, very wide window opportunity to treat. Maybe the very late stages it becomes -- yes, the prospects of any kind of benefit becomes more limited, but I still think for APBD you've got a very wide window. It's not quite so true for Lafora. You saw the case earlier, the young girl died in her late teens, and symptoms were present for about the age of 10 or 11 or 12 onwards. The earlier you treat there, the better. So I can't quantify it with a specific number or specific age. But conceptually, it's a really important point to consider across our portfolio, and it's very disease-specific. What I will say is with APBD, a very wide window. Lafora, probably a slightly smaller window or definitely a smaller window. And then for some of our other more severe diseases and more rapid and aggressive, GM2CL1, either are well.

Great. Thank you, Suyash. And this concludes our discussion on Lafora disease. I'm happy to turn it over back to you, Suyash, to talk about TSHA-113 for tauopathies. And we want to -- we also -- I'm sorry, we also want to thank Dr. Minassian for joining us as he's international as well.

Correct. I think sadly, his connection dropped and we won't be able to say goodbye in person. But yes, as I say, Berge is a great partner and collaborator to work with and very happy he could join us. Okay. Let's move on to tauopathies. So I'm just going to give a very brief overview, then I'll hand back over to Rachel to go into some more detail on the construct and some of the signs. So if we go to the next slide, please. So tauopathies and tau is often in the news, and all kind of mechanism, modalities that are thought to contribute to Alzheimer's are very prevalent in the biotechnology news currently. Tau specifically is a microtubule associated protein, MAPT, that's expressed primarily in neurons. One terminal of tau is often associated with microtubules, the other domains bind to plasma membranes. And essentially, they're involved in helping stabilize the structure of neurons. Now tau can degrade, it can phosphorylate in an abnormal way. And this results in the self assembly of tangles, which are -- tangles are paired helical or straight filaments, which are involved in the pathogenesis of Alzheimer's disease and several other neurodegenerative diseases that we list here. So frontotemporal dementia, progressive supranuclear palsy, corticobasal degeneration and several other conditions. You do get tau isoforms expressed both in the central and the peripheral nervous system, so it's throughout the CNS and the PNS. And the prevalence of these disorders, if you sum them, is about 13,000 patients. So that's specifically these tau-associated disease. And of course, many, many more patients exist with Alzheimer's and there is some thought that tau or abnormal tau can contribute to the development of Alzheimer's. Now one of the key features or the classic features that you see with these tauopathy light diseases are a deterioration in the MRI appearance in association with many of the clinical features. So in this particular slide, what you see here is 3 different images. The -- well, 6 images in total, but the 3 images at the top are all normal brains and the 3 images at the bottom are abnormal brains. And you can see -- if you look at the middle diagram, to start off with, you can see that there is a loss of matter in the -- on the right side of the brain, there is some deterioration there. So this is -- you can see the loss of white matter and brain structure. On the extreme right-hand side of the image, it's a slightly separate disease. Sorry, B is the corticobasal degeneration type form of tauopathy. C is known as Pick's disease. And what you see here is once you have loss of white matter in the pattern across the cortex and the cerebrum. And the image on the extreme right A is -- it's thought to be reflective of frontotemporal dementia. And it's a bit of a subtle feature, but in the midbrain, which is bang in the center of the image, you see a decrease in brain tissue, which results in the expansion of the third ventricle, just a breadth of the -- above the midbrain, which is the kind of dark pocket that you see. And below that, with the eye of faith, you can see -- this is known as a hummingbird sign. We see an image of a hummingbird in the brainstem. Very difficult to see. I can't quite point it out to you at the moment. But all these MRI features demonstrate the structural abnormalities associated with tauopathies that relate very strongly to the clinical features of tauopathies. They do cause significant disease. There's a big burden disease, with a particular for admission. If we go to the next slide, I'm now going to hand over to Rachel Bailey. Reintroduce Rachel, Assistant Professor at UTSW, who's been leading the tau program, she's going to talk to you about the approach to treating tauopathies.

Right. Thank you, Suyash. So to treat tauopathies, my group developed a vector carrying an artificial microRNA shuttle to deliver lifelong reduction of tauprotein and neurons in glia. We used the U6 promoter to drive ubiquitous expression and packages in the more efficient self-complementary AAV9 vector. To target the brain, we packaged -- to target the brain, we delivered this through CSF. So this project is currently in preclinical development. But today, I will discuss our vector development and early results. Next slide, please. So here, I'm showing the science of the treatment approach. You can see the AAV9 viral vector binds and enters a host cell where it then traffics to the nucleus and delivers the vector carrying our therapeutic gene. Once in the target cell nucleus, primary microRNA constructs are transcribed and then processed by the RNases, Drosha and Dicer and the nuclear exporting factor export the 5. The mature antisense tram that then is in the cytoplasm, incorporates into the RNA-induced silencing complex, a risk to elicit sequence-specific degradation of messenger RNA from all 6 tau isoforms found in the human brain and/or in the mouse brain. Next slide, please. To identify the lead microRNA shuttle, we cloned 10 different human tau artificial microRNA shuttles into a U6 promoter-driven expression cassette. These were used in an initial in-vitro dulcite for a screening assay, in which we clone the human tau target sequence into the 3 prime UTR of a CPAD luciferase and use firefly's luciferase as a standard. Knockdown efficiency and disease specificity were tested following co-transfection of the reporter plasmid in each tau shuttle or scrambled control vector in HEK293 cells, and that's what's shown in the graph on the left. The most effective constructs were then used in a secondary screen. Next slide, please. For the secondary screen, plasmids with our 3 top candidates were transfected into HEK293 cells expressed in human tau, and we measured reduction of tau messenger RNA as shown in the graph in the left using QPCR analysis. We then looked at tau protein levels using western blot analysis as shown in the middle and confirmed reduction of human tau proteins without altering mouse tau protein levels. Using a HEK293 cell line that stably expresses GFP tech in human tau, we found that our most effective candidate, now referred to as Tau5i, decreased tau protein levels as showed by the decreased green GFP signal as compared to the control scrambled microRNA. hTau5i was then cloned into self-complementary AAV9 vector that also coded a separate GFP reporter for in-vivo delivery. Next slide, please. To assess distribution and target engagement, we used a tau mouse model that expresses mutant human tau at a 7x higher level than the endogenous mouse tau. In these mice, human tau forms fibrils and aggregates with age, which results in neurodegeneration and early death. In this pilot study, mice were treated close to 3 months of age and were then assessed 2 months post injection. This is a time frame that is early in the disease progression and tau pathology is mild. To look at distribution, we stained for GFP as this will be expressed in cells transfused by our viral vector. And here I show representative images of the entorhinal cortex. Looking at the top-left panel, we see no GFP staining. And in the bottom-left panel, we see that GFP expression, which is the dark brown, is present in both neurons and glia. We then stained serial sections with a tau antibody and show normal expression levels in a vehicle-treated mouse, which is in the top right panel as well as we found reduction of tau in an AAV9-treated mouse, which is shown in the bottom right panel. We then measured human tau messenger RNA levels in these mice and found out vector-treated mice had approximately 25% less human tau messenger RNA levels. As a reminder, this mouse model overexpresses human tau sevenfold more than the normal amount, so this is an agreement with studies that have used ASOs to reduce tau levels that 25% to 50% of human tau reduction is likely the best we will be able to achieve in this particular mouse model. We have also designed mouse tau-specific microRNA shuttles and cloned our lead candidate into a self-complementary AAV vector to test for the level of tau reduction in wild type mice that have normal tau levels as well as to test the safety of reducing the endogenous tau protein. So overall, we have shown that we can target the tau protein and lead to reduction of tau messenger RNA and protein levels. So now I'll hand it over to you, Suyash, for the anticipated next steps.

Thanks very much, Rachel, and really very nice proof-of-concept work showing that -- this tau knockdown approach is potentially going to be clinically meaningful. So we have an own proof-of-concept, as you've seen in the knockdown mouse model. We are now moving forward into mass pharmacology studies, limited dose response, age response and we'll finalize the dose, the presumed dose for clinical trials from the pharmacology studies. And following on from that, we will be performing some toxicology studies using our usual approach of 3 species, mouse, rats and NHPs to design and develop a therapeutic window within which we can dose our clinical trial. And the clinical trial protocol is just starting currently. Let me hand over to Kim for any questions on this early program.

Thanks, Suyash. Your first question comes from Yun Zhong of BTIG. There have been a lot of failures in targeting tau. What lessons do you think you've learned or have learned? And what differences do you think that TSHA-113 has -- have a better chance?

Rachel, do you want to take that?

Sure. So in some of the earlier treatments to target tau, they're actually targeting the tau protein. And so many of those approaches may rely on having a drug that first needs to be able to cross the blood-brain barrier. And then second needs to be able to get into the cells to target the tau protein, which oftentimes that is not efficient. So you don't get -- you don't get good interaction of the drug with the tau protein. There has also been vaccines to try to target tau that's actually outside of the cell as tau has been showed to spread almost in a prion-like form. And again, that relies on getting enough of a drug into the brain spread widely to target the drug that it doesn't then interact with the tau inside of cells. So we feel that by vectorizing and providing a copy of the gene for the cells to then continuously make this drug, that we will be able to efficiently get our drug inside of cells and that the cells will be able to have a constant level of this reduction to have more efficient downregulation of tau than what's been seen with other treatments.

Thank you for that. Your next question comes from Joon Lee of Truist. Dr. Bailey, is there a general goal of thumb as to what targets you use mRNA? And what targets are better suited for shRNA approaches?

So I don't know if there's a general rule of thumb of strictly shRNA versus microRNA, but something to keep in mind is there have been studies that show that with shRNA because it is expressed at a higher level typically than microRNA that gets processed down by the endogenous machinery, that you can get some toxicity associated with expressing too much RNAi. So we feel by using the microRNA shuttle for this disease in using the endogenous processing system that we will be able to avoid potentially toxicity that's associated with that using shRNAs.

Thank you for that.

Kim, the only thing that I would add to what Rachel just mentioned is I think it really depends on the indication that we're going after, which is going to be what's the most appropriate way to knockdown, whether that's going to be an mRNA approach or an shRNA approach to interfere with translation. So really, I think we take this kind of on a disease-specific basis.

Thanks for that add. Your next question comes from Joon Lee of Truist. Let's see. For tauopathies, is your mRNA specific from mutant tau? Or wild-type tau? You target wild-type tau, what's the therapeutic window?

Rachel, do you want to take that?

Sure. So this is targeting all tau, whether it is a mutant or not, so it would not need to be changed between patients. I think it would just need to be checked to see if they have a mutation to make sure it does not fall within where the guide RNA is finding. So we chose this approach because the majority of tauopathies are actually not associated with mutations and they occur just with the normal wild-type tau. So this would allow us take the same treatment approach that we can use for genetic forms and potentially use for sporadic forms as well. In terms of therapeutic window being a neurodegenerative disease, somewhat as is mentioned, for the GM2 study, there's going to be a point where you won't be able to rescue a patient because if the cell is degenerated and is no longer present or is no longer functional, we cannot treat that cell with the virus because the virus needs the cell to go inside. So this, of course, would be another disease like any degenerative disease that the earlier you treat, the better the therapeutic outcome. However, we will be also performing efficacy studies in our mouse models have different ages to assess at which point we feel is our window for the best treatment.

Thank you. Your next question and final question comes from Joon Lee of Truist. Now that a dulum is approved, could this be done as an add-on common issue there?

Yes, I believe it could be because they would not overlap. And so it's something that can be used in combination potentially.

Right. Thank you very much for that, and that concludes our session on tauopathies. I'd like to now hand it over to Suyash to talk about TSHA-106 for Angelman syndrome.

Great. Thank you, Kim. Let's move on to the next slide. So yes, TSHA-106 for Angelman syndrome. Next slide, please. Angelman syndrome is it seems to be a rare disease. It's one of the commoner rare diseases. So the estimated prevalence is about 55,000 patients in the EU and the U.S. It is a neurogenic disorder developmental sort of due to genomic imprinting. Specifically, what happens from a genetic perspective is that the UBE3A allele is exists on Crimson-15 on both maternally derived allele and the paternally derived allele. In Angelman syndrome, maternal allele is silenced by a strip of long, noncoding RNA and the maternally derived UBE3 allele is meant to express the UBE3A protein but due to a mutation, it doesn't express the protein. So there's a lack of UBE3A, which results in a whole host of -- we'll touch on the functions of UBE3A in the moment, but it results in the clinical phenotype of Angelman syndrome. Now these kids generally look fine in the first few weeks after birth, but subtle features of developmental delay emerge early on in life. This translates into more severe impairments and behavior and motor function and communication and sleep and usually the disease is diagnosed after at the age of 6 months, but usually before the age of a year. The symptoms progress and there are some debilitating seizures and ataxia in the form part of this disease. So it's really an unpleasant disease to manage because -- frankly, because a lot of the behavioral issues associated with this particular disease. It was always known historically in the pediatric context as this happy puppet syndrome because these children often have a lot of smiling and laughter -- not necessarily a pleasant laughter, but sometimes a maniacal-type laughter. And they walk with a fairly kind of rigid posture due to some of the ataxic features that they have with this associated. They have feeding issues, limited ability to communicate in their speech, abnormal sleep cycles. And sleep is a really, really big problem in these particular children. And there are a number of characteristic EEG findings associated, of course, with the distraction of sleep. And as time progresses, as the disease gets worse, they have seizures, and they have more and more seizures as time progress. And initially, they're somewhat controlled with antiseizure medication. But as time progresses, the seizures get worse. Oftentimes, they have a normal lifespan, but they can't live independently. Next slide, please. There's no cure for Angelman. The current treatment focuses on managing medical and developmental issues. It's all supportive care. We've touched on antiseizure medication, which is something really helpful, physical therapy to assist with walking, movement problems, fine motor skills, trying to ask the person to write with a pen, special schools are often needed. Speech and language therapies needs to help with the articulation of words and the bulbar function. And very significant behavior therapy is often needed to overcome hyperactivity, short attention span, some of these laughing features. And sometimes neuropsychiatric agents can be used to treat some of the behavioral issues associated with Angelman. But that's really a -- from a family perspective, from a quality of life perspective, often, it's the behavioral issues are the hardest thing to deal with. Next slide, please. So UBE3A is a ligase that's involved in regulating the expression and function in the number of proteins and a deletion or loss of function in the maternally inherited allele of UBE3A in association with the normal silencing of the paternal chromosome with this long noncoding strip of RNA is what results in Angelman syndrome. And in particular, UBE3A, if we go to the next slide, is heavily involved in many, many aspects of neuronal development. And you can see there's many, many functions. This slide summarizes some of the key functions of UBE3A, especially in the context of development of the brain. And I won't go through all of this in detail, I'll pull out a few issues. So for example, UBE3A, ubiquitin-protein ligases involved in the activity-regulatory, cytoskeleton-associated protein, which, if it's increased, reduction in synaptic plasticity. It has impacts on alpha-synuclein, which also has an impact on synaptic plasticity, even has some effect on mTOR. So as a whole, there's a global degeneration in neuronal development because UBE3A targets many, many, many of these molecules in these pathways. And there's several that are not even fully understood yet. Next slide, please. Now we've touched on already the underlying genetic defect. And I think the way to think about it, once again, is if you look at the top diagram, when you look at the middle chromosome, this is the commonest form of Angelman. We have a maternal deletion. So the blue is the allele that's derived from the father, the orange is the allele that's derived from the mother. Don't forget, in the normal circumstances, the blue allele is silenced by a strip of noncoding RNA. So you don't get any UBE3A from the paternal chromosome. On the yellow, which would normally produce the UBE3A, there is a deletion in the commonest form of Angelman, which results in no UBE3A being produced. And that's what results in the clinical phenotype of Angelman syndrome. So from a genetic medicine gene therapy approach, there are 2 clear approaches to how you might treat this genetic abnormality. The first is just simply replace the maternal chromosome, the maternal allele, i.e., you've got this deletion that's causing the stock code on -- causing an issue with UBE3A production. So if you replace that gene that's missing, you should be able to then produce the UBE3A that's missing and, therefore, ameliorate some of the signs and symptoms that we've already mentioned in the earlier slides. That's one approach. The other approach is to use a silencing mechanism or rather a small interfering RNA where what we do is we silence the noncoding strip of RNA that binds the paternal allele and causes silencing of the gene. I.e., we are unsilencing the production of UBE3A from the paternal chromosome, okay? So there's 2 approaches, a simple gene replacement approach and there's also a way of silencing RNA that's causing silencing of the paternal chromosome. I.e., we therefore unsilence the paternal chromosome. Hopefully that's clear. We're looking at both approaches with the impression that's talked about is the unsilencing approach. And you can see this in the bottom screen where you've got the intact paternal allele that's silenced by the UBE3A-ATS, which is a noncoding RNA that silences, and we're going to unsilence that allele by interfering with the silencing allele, the silencing strip of RNA. Now the other approach that's been trialed by several other companies is ASO approach to skip the exon of the long noncoding RNA and, therefore, allow the paternal chromosome to keep producing UBE3A. And that I think from our perspective really derisks our approach because we've seen some very nice data from that approach, which I'll mention in a couple of slides' time. Next slide, please. So this is the construct. And you can see here, we've got the anti-UBE3A-ATS shRNA. Once again, it's an shRNA that binds to the UBE3A-ATS. This is noncoding RNA that silences some allele wrapped up in AAV9 capsid. So you unsilence the paternal allele, and that paternal allele produces the UBE3A that's missing. Now Ultragenyx have actually run a study looking at an ASO approach to this particular disease, which, once again, unsilences the paternal allele. And you may or may not be familiar with that data, but I shared it a few months ago. They treated 5 patients with this antisense oligonucleotide. And they were able to encourage the expression of UBE3A from the paternal allele, which resulted in really quite clinically significant improvements in these 5 patients. There was a significant improvement in the Clinical Global Impression of severity of improvement scale, specifically related to Angelman by at least 2 or 3 points out of 7-point scale. There was clear improvement in the EEG findings and also improvement in milestone acquisition and also improvement in sleep. So many of the clinical features of Angelman that friends and families talk about is causing a great deal of compromise. The ASO is generally well tolerated, but the problem was -- with this approach is that of the 5 patients -- and don't forget, from an ASO approach, you have to dose on a relatively frequent basis. So these children receive an intrathecal injection of an ASO on a monthly basis for several months. And what happened in 3 of the 5 patients is that there's some inflammation that occurred at the base of the spine really as a consequence of administering the drug intrathecally in a repetitive manner that has cause quite considerable lower extremity weakness to the point where 3 of the children had lost the ability to walk. This is only temporary for a period of time. There was a significant side effect. So that's the only issue with this ASO approach, but the efficacy data actually look very, very favorable. And we feel that using gene therapy to unsilence the paternal allele, as opposed to ASO, offers a unique set of advantages. It's onetime dosing, of course. There's widespread transduction through CS using an AAV9 vector. And we have a lot of experience now in clinical trials where a drug had been given intrathecally. We talked about GAN in detail yesterday, CLN3, CLN6 of the intrathecally dosed soldiers in the study. So this is our [ thecal ] approach to silencing the -- to unsilencing the paternal allele using a gene therapy siRNA approach seems to make a lot of sense. Next slide, please. So I have 2 slides that talk through some of the early scientific data prior to preclinical work. And this work is being led by an assistant professor at UTSW, Ryan Butler, who will be on hand to answer some questions, if indeed you need your questions answered. And he's really led this work for Angelman over the past 2, 3 or 4 years. And there's been a number of candidates of shRNA candidates that are being trialed to see if you can increase the production of UBE3A in conjunction with reducing UBE3A, the actual sola is in transcript. And what you see here with 26 different lead candidates that Ryan has worked on, you do see increases in UBE3A. So this gives us preliminary scientific proof-of-concept that this approach will work to allow the expression of UBE3A. This was in neuroblast cell line and demonstrates consistent knockdown of UBE3A-ATS and an increase in UBE3A across these 26 distinct shRNA candidates. Now this is the first step in the process. Ryan is repeating the experiment currently with a neuro cell line and actually a larger number of constructs, about 45 or 46 shRNA candidates. And from that larger pool where we're going to -- where we expect to see higher fold elevation of UBE3A, we will actually select the construct. Next slide, please. And what we see here is preliminary expression of UBE3A in the Purkinje cells of the cerebellum, the Angelman mouse model, following intrathecal delivery of one of these experimental shRNA AAV9s. So of course, as we've talked about already, shRNA targets a [ regional upstream ] of UBE3A-ATS, which induces paternal expression of UBE3A. We've talked about it being similar to the ASO approach. And you can see the beginnings of expression of UBE3A in the Purkinje fibers of the cerebellum. So once again, very early preclinical proof of concept. Next slide, please. So what was a wonderful work here, you can see we still have a number of candidates to screen through and look at. And once we've selected candidates of the many canvas we have, we will be looking at expression of safety data from confirmatory -- for confirmatory NHP studies. For this kind of silencing approach, the preclinical package tends to be a bit lighter based on the ASO approach. But of course, that remains to be seen as we have our discussions with the regulators, which will be occurring sometime during the course of next year. We'll be evaluating dose response, age responses, finalizing the dose from a pharmacology studies. And we're already giving some thought to the interventional trial. I think we'll probably base a lot of the learnings from the Ultragenyx's 5-patient study. They had a very nice selection of end points, and we'll probably pool some of that through into our interventional trial development. Let me stop there, and I'll hand over to Kim to facilitate any questions you might have.

Great. Thank you, Suyash. Our first question comes from Eun Yang of Jefferies. A general question here for not only your commercial stage products, but also your subsequent programs here. Are you planning to use commercial-grade materials for clinical trials? And can you comment on your commercial manufacturing process?

So maybe I'll start off and, Suyash, please chime in Fred's absence. So we are absolutely planning to use commercial-grade material in both our interventional trials, but also in the toxicity studies leading up to those interventional trials. We think it's actually particularly important as the FDA has added a lot more rigor earlier on around understanding the characterization, the full characterization of the product, the process in which you're manufacturing under and how that translation works from preclinical pharmacology studies and translational studies into the clinic. So this is something that's extremely important to us that -- and is the reason why we derisked our manufacturing platform through kind of our 3-pillar approach, and we'll highlight that approach and kind of what we're doing in July at our manufacturing day. But the short answer is yes, we are using commercial-grade material for our interventional studies. Suyash, do you want to add anything to that?

Sure. I think there was some guidance published by the FDA I think in January or February that really gave a lot of good guidance on gene therapy approaches in neurodegenerative diseases, looking at 3 elements, so clinical trials, preclinical work and CMC manufacturing. But what is clear, what is crystal clear just from that guidance, which we found very reassuring because it's actually checked all the boxes of our approach anyway, what's clear from that guidance, what's clear from what we're hearing from the FDA in our discussions with the regulators and what you're seeing in terms of clinical trials is there's much, much more focus on getting manufacturing right very early. And from a scientific perspective, what this means is we just have to make sure we characterize the product well and earlier move to commercial processes well and early. So you've heard us talk before about this three-pronged approach. We've got GMP facility and UTSW. We've got a strategic collaboration with Catalent who we're building as our own GMP facility. We have the groundbreaking ceremony a couple of months ago, and that will be up and running some GMP products in 2023. That gives us lots and lots of flexibility. But for now, we need to characterize the product we're making very, very -- in a very disciplined manner from a purity perspective, looking up whole cell protein, whole cell DNA from an empty, full capsid ratio perspective and from another part as the FDA are really making us pay very close attention to. Fred is not on the call. He's our Chief Technology Officer. He is extremely well versed in this and he's -- he and I have all lived through situations where manufacturing has not gone quite as well as it could have done. So we've used those learnings to really make sure we've dotted the I's and crossed the T's from a CMC perspective for all our programs. But specifically, all top studies will be done using commercial-grade manufacturing and all the pivotal part of clinical trials will be done with commercial-grade manufacturing.

Thanks, Suyash. Your next question comes from Joon Lee of Truist. For large, rare diseases like Angelmans with larger players like Roche and Ionis and Biogen are starting to enter here, what's your competitive strategy? And any plans to partner at a certain point? Or will you take it all the way to commercialization on your own? And can you share your thoughts on some of the larger indications that this may be applicable to as well like the others?

No. It's a really good question. And I think as we think about partnering, we're pretty fortunate that we haven't necessarily had to make any hard prioritization decisions. I think we mentioned this yesterday, we kind of moved at the speed of biology or the speed of science. And when we brought in this portfolio from our partners at UT Southwestern, the programs were already appropriately staged. You had some programs. They were -- that we're right about to go into the clinic. You have a number that had already hit animal proof of concept that we're just needing to do some additional IND-enabling toxicity work before they went into the clinic. And then you had some programs that were a little bit earlier on, they were right on the cusp of hitting animal proof of concept or just even actually starting at the discovery point of view. And we've mentioned, and we've actually shown some great progress on the entire portfolio over the last 2 days. And so I think as we think about programs like Angelman, particularly, or Fragile X or indications that are still within what we consider rare diseases but maybe on the larger side of rare diseases, I think these are programs that as we see them today would be ideal for us to take them all the way into the commercial setting. I think what we're fortunate to be able to do is achieve significant economies of scale both across the clinical development aspects of our program, but also manufacturing. But lastly, commercialization because the majority of the call points here are going to be either pediatric neurologists -- pediatric neurology centers of excellence or neurological centers of excellence. So the commercial infrastructure that you need to actually support these particular programs is relatively limited and kind of building out that foundational commercialization franchise around giant axonal neuropathy, GM2, CLN1, SURF1 and kind of bridging you to Rett really sets you up nicely for then taking on programs like Angelman, Fragile X and a bunch of others. So for the time being and as we see for the foreseeable future, these are programs that we feel pretty strongly that this -- that we're able to kind of take forward from translational all the way through the clinic and into the commercial setting with the infrastructure that we're building out. And this was kind of a central thesis around the company and why we focus solely on monogenic diseases of the CNS, particularly pediatric diseases, while we focus on the use of the single capsid, route of administration and manufacturing system because we want to make sure we're able to get and achieve significant economies of scale. The second part of the question was around some of the larger indications that kind of fall outside of the rare diseases, tauopathies being one of them. I think realistically, for us to be able to take a program like tauopathies from today and fully exploit that to -- through the Alzheimer's population, if successful, I think we would need to have more of a established infrastructure around us to be able to kind of pull off both clinical development and then commercialization to really be able to achieve -- and really to be able to achieve kind of a full appreciation for what that program could be. So I think we're going to be opportunistic as it -- as we consider partnering opportunities as some of the larger indications that kind of fall outside of rare disease. But particularly for the rare disease indications and the indications that affect pediatric populations, I think these are going to be programs that we look forward to taking ourselves.

Thanks, RA, for that. Your next question comes from Gil Blum of Needham & Company. Are imprinting or silencing mechanisms all similar that is involving expression of long interfering noncoding RNA? This is potentially a new drug class.

Yes. So maybe we'll have Ryan. Ryan, if you're on the call. I can see you're on the call, but you may be on mute and off video. Maybe you want to take that question?

Thank you very much. No. They don't all work the same. However, siRNAs and miRNAs are quite similar. For ASOs, they generally are on more nuclear-specific. They use a mechanism of RNAs, RNAh. For siRNAs and miRNAs, they typically are nuclear and cytoplasmic and a recruit complex known as risk RNA interfering silencing complex. And so they aren't the same. Was there another part to that question?

Yes. We'll see. That is -- is this potentially a new drug class?

I would say yes, but I'll let RA or Suyash.

So to kind of -- to answer that second part and I think Ryan is spot on, and it's a great question. We'll be looking at this as similar approach, and we're already investigating a similar approach across a number of different indications. The second one that is somewhat associated with Angelman is Prader-Willi where we're taking an almost identical approach around -- and it's just the reverse. We're looking at unsilencing the maternal allele in this case. And then there's a number of other indications that we're exploring with Dr. Butler and his lab that is going to be using similar technology or even a combination of gene replacement and upregulation using shRNA or microRNA. So more to come. But yes, we kind of see this as, again, another kind of novel-targeted approach depending on the indication. So more to come.

Great. Thank you. And our last question comes from Eun Yang of Jefferies. Thanks for the comprehensive overview of the programs. With a program like this, would you rate this as a high probability of success besides the GAM program? And what other ones would you consider high probabilities of success and also high risk?

Yes. So it's an interesting program and it's an interesting question. So we appreciate the question, and maybe I'll start and Suyash, please feel free to chime in. I think when you look at some of the indications, and we're fortunate that some of our earlier indications that we've either achieved clinical data and have really good long-term clinical data are on the cusp of achieving clinical data or getting into the clinic. These are some of the diseases where you could kind of say the proof of concept or the probability of success is somewhat higher. And this is really just due to the underlying biology of the disease. For example, in GM2, it's really around expression of HEXA enzyme and that enzyme activity. And the natural history informs that, that a very little enzyme goes a long way, and we're fortunate to have the experience of the enzyme replacement therapies coming before us, and Suyash can talk about some of that experience. But really, just in GM2, taking a patient from almost a knockout in the infantile form to 2% to 4% in the adult onset form significantly prolongs lifespan and improves the clinical phenotype. Again, as I mentioned in CLN1, taking that patient up to somewhere between 5% to 10% levels with significantly prolong survival and significantly improved clinical phenotype. And so -- and then in giant axonal neuropathy, again, it's a disease where you have a buildup of some substrate that can ultimately be cleared once the gene is replaced. And so again, there's a lot of similarities to these programs and really give us the opportunity and the confidence that there is -- by addressing the underlying biology of the disease, we should be able to really have a profound improvement in the clinical phenotype. I think this then allows us to bridge to some of the diseases where the biology is a little bit more nuanced, but we feel very strongly that our targeted approach should have an effect, but ultimately, the experiment is going to need to happen in the clinical trial. I think when you talk about diseases like Rett Syndrome, again, the underlying cause of the disease is this loss of MECP2. But from a biological perspective, the fact that these patients are mosaics and they had half of their cells are wild type and half of their cells are null. And the fact that if you increase MECP2 production in wild-type cells, you make them sick. And if you don't get enough production in the null cells, you don't actually get a therapeutic level. That goldilocks or very tight dosing window is extremely important. But in order to address that, I think we've kind of innovated around this notion of the self-regulatory feedback loop in the MI rare platform. And in the case of Angelman, I think what's nice is we have nice human proof of concept from the Ultragenyx trial in the ASO approach. That gives us confidence that unsilencing the paternal allele to knock down UBE3A-ATS is an approach that can restore UBE3A production and that would ultimately translate into the clinic. And we think the vectorized approach or the gene therapy approach has a number of -- has a number of benefits over the years. So I'll stop there. Suyash, do you have anything to add to that?

I would just echo the fact that I think it's a little bit disease-specific.

Yes.

And I think you're quite right, RA. The more we understand the molecular biology, the better we are able to quantify the probability of success. And the more we can learn from precedent, and it doesn't need to be gene therapy precedent, enzyme replacement therapy precedent clearly applicable to GM2 and CLN1. Well, a little bit of enzyme on goes a long way. And so we only need to make a little bit of enzyme from our gene therapy constructs in GM2 and CLN1 to know that's going to result in a meaningful clinical outcome. In addition, the precedent with the ASOs, specifically for Angelman, really gives us a lot of confidence that with a single administration of silencing -- sorry, unsilencing construct, which transduces itself permanently, we should see persisting, ongoing efficacy. I think that, for me, the one group of disease that's a little harder to quantify as less precedent and probably genetic epilepsies. There's less precedent there, the disease is newer, a little bit less is known. But it clearly works in the animal model. So we should be able to translate that in humans as well. So I guess I'd make those general comments for the group.

Great. Thank you. Just one final question while we're on this topic is from Sami Corwin of William Blair. How are you thinking about prioritizing the development of these programs you just discussed?

Yes. We're fortunate that, again, when we brought these programs in, they were already fairly appropriately staged. So we have a number of programs that will be in the clinic by the end of this year. We'll have a total of 5 giant axonal neuropathy, GM2, CLN1, Rett Syndrome, SURF1. And then we have a number of programs that we'll be looking to move into the clinic next year, first, from our genetic epilepsies franchise, SLC13A5 program. We haven't mentioned today, but our GM2 AB variant program, which is a collaboration with Queen's University and UT Southwestern. And then a number of programs kind of behind that, that we'll be looking to get into the clinic towards the end of next year, APBD, adult polyglucosan body disorder being one of them. And so again, I think with the infrastructure that we're building out, we've taken the company from a handful of employees this time last year, we're roughly about 135 employees. We're building out internal manufacturing capabilities. We had good capacity at UT Southwestern and Catalent. And we're bringing on the right people in order to be able to move the portfolio forward. So when we talk about kind of from the next 18 months or so, we're really focused on about 10 to 11 programs. And then ultimately, that will be foundational for building upon the rest of the portfolio that we'll be looking to move forward. But really, today and yesterday was really about showing you guys the breadth of the portfolio and the research engine at our partnership at UT Southwestern.

Thanks, RA, for that. That closes our section -- our discussion on Angelman syndrome.

Awesome.

I'd like to turn it over to Suyash for our final discussion on the redosing platform. Suyash?

Great. Thanks, Kim. And we're close -- we're just at time. I'm just going to spend a few minutes talking about our redosing platform. And this is important because we started off the whole 2-day event with a deep discussion on giant axonal neuropathy, which is our lead program. And we're kind of coming full circle because much of the preclinical work for our redosing platform has actually been done for giant axonal neuropathy. And as you know, redosing is clearly an issue in the world of AAV gene therapy and many, many companies are looking at approaches specifically for this. Let's go to the next slide, please. Now there's actually 2 things we're trying to accomplish with our redosing platform. And specifically, it -- what happens with this is we give a direct injection of gene therapy, of vector, AAV9 vector into the vagus nerve. Now the vagus nerve is the tenth cranial nerve. It comes off the base of the brainstem. It runs through the shoulder, splits into 2. One branch goes down the center of the chest, innervates the lungs. The other doors down the left side of the body, innervates the heart. The branch that goes down the center innervates restriction muscles and then goes on to the gastrointestinal tract and feeds up into different parts of the body. And you can see some of the parameters in the right-hand side of this picture, which are innervated by the vagus nerve. And specifically, this is the autonomic nervous system. So autonomic is the functions of the body that happen automatically without you thinking about them. So for example, breathing. You don't think about breathing, but the rate is set by autonomic nervous system, the same with the heart, the same with gastrointestinal motility, the same with a pupillary response to light or to dilate, the same with bladder contraction and salivary production. So in all of these neurological diseases, we have to talk about CNS, which is more the brain. We talk about the PNS, which is more the spinal cord and peripheral nervous system. But often, we don't talk about the autonomic nervous system. And its deficits in the autonomic nervous system that actually give patients and families sometimes the most concern because they cause symptoms that specifically have a major, major impact on the quality of life, in particular, gastrointestinal symptoms. So severe abdominal pain, severe constipation, the kind of containing gas causing severe wind within the GI tract, the inability to get appropriate nutrition, these all cause a lot of issues from a patient perspective. And if we can ameliorate some of those signs or symptoms, it will be very, very helpful to patients and families. You will also recollect from our giant axonal neuropathy data that I presented yesterday and data that was published in the recent journal by Carsten Bönnemann and his team, in brain, this publication came out about 4 weeks ago, there is some specific information that's been captured on autonomic nervous system features of giant axonal neuropathy. And you may recollect that on the table I showed yesterday showed that using this particular scanner, the COMPASS 31, and using other specific measures of autonomic function, i.e., measures of sweat, measures of lacrimal tear production, you can see that this is significantly compromised in these patients. Not only is it comprising down, but there's also a lot of discussion around the autonomic nervous system dysfunction in Rett Syndrome and indeed many other diseases. So being able to target the ANS with a direct injection of the vagus nerve, that'd be critically important. And importantly, you can target ANS with a direct injection to the vagus nerve even in the presence of AAV antibodies and even if the patient has already received an intrathecal or a systemic dose of AAV drug, which I'll show you in the next slide, which is the mouse data looking at redosing with a direct injection to the vagus nerve. So you see 2 studies here, Study 1 and Study 2. These mice received an injection intrathecally of the GAN construct. So this is the giant axonal neuropathy construct. The full-length GAN gene wrapped up in AAV9 capsid. They were given that drug, and then they were given the drug several weeks later in study when it's 4 weeks later, it's 12 to 16 weeks later of another AAV9 construct, but this time associated with GFP, expressing GFP. So let's look at the tissue on the microscope. But the key thing here is that we were looking to see if we could dose a second dose of AAV9 construct after the mice had already received AAV9 construct intrathecally. And in lots of those, we were successful in being able to do that. So you dose the Study 1 at 4 weeks, you dose the Study 2 at 16 weeks. And what you see is nice, robust expression of GFP both in the vagus nerve and the nodose ganglia. The next slide, please. This is looking at staining in different parts of the brain, in particular, looking at the brain stem and the medulla where a lot of the autonomic nervous system features are controlled. So the nucleus ambiguus is involved in bulbar function and articulation of vocal cords. The pre-Bötzinger Complex is involved in respiratory rhythm management. These are autonomic features, and you see nice transduction and nice production of GFP in those parts of the brainstem. Next slide, please. And you also see in these preimmunized AAV9 rats, that's the diagram on the right side, the nonimmunized, naive rats on the left-hand side. You do see a little bit more expression in the nonimmunized rats' side without AAV9, but you still see significant transduction in the AAV9 pretreated rats. And specifically, if you look at the top left-hand corner of the diagram on the right-hand side in the area postrema. Now the area postrema is part of the brain that is the most closely associated with the circulation. So it's a part of the brain that's most vulnerable to the presence of antibodies. And the reason it's very closely associated with the blood system is it actually picks up toxins in the blood system and encourages vomiting. So if you've taken some toxin, somehow either accidentally or you've been poisoned in some way or you've ingested something, that's detected in the area postrema of the brain because of its very close alignment with the blood circulation. And then that causes -- induces a vomiting effect of trying to rid yourself of vomiting that toxin. So the area postrema is the part of the brain that's most closely associated with blood flow. So being able to see transduction in this particular part of the brain when the rat has been exposed to AAV9 already is a very nice thing. The rat had been able to demonstrate transduction in the presence of antibodies to AAV9. So this is a really big thing from a preclinical perspective. Next slide, please. And in terms of moving forward, we don't have a summary slide I'm afraid. We've got that rat study and the mice work done. We're moving forward to larger animal studies. And we will, at some point in the not-too-distant future, think about an approach to IND-enabling studies to move this redosing approach and vagus nerve targeting approach into clinic. Let me stop there. It's been a wonderful R&D Day event, 6 hours of great discussion and great questions. I'm going to hand over back to Kim and see if there's any -- Kim and RA to see if there's any final questions before we close.

Great. Thank you, Suyash. Since we're out of time here, let me turn it over to RA for his closing remarks.

Awesome. So really, we want to thank people for their attention and interest in Taysha over the last 2 days. As Suyash mentioned, it's been a wonderful 6 hours of discussion and really what we wanted to do is just highlight the transformative nature of what we're trying to accomplish through our collaboration with our partners, our great partners, which are literally across the street from our headquarters here in Dallas at UT Southwestern. The second half of 2021 will be a busy time for us. We have numerous value-creating milestones remaining for the year. For TSHA-120 and GAN, we anticipate data from our high-dose cohort, which is 3.5 V to the 14th total VG. We expect to have that data here in the second half of this year along with a regulatory update based on our conversations with various regulators around the world. In GM2, we expect to have preliminary biomarker data, which will be Hex A enzyme activity in the second half of this year from TSHA-101. In CLN1, we anticipate dosing the first patient under our currently open IND and expect to have a number of additional INDs filed in clinical trials initiated by the end of this year, including Rett Syndrome and SURF1-associated Leigh Syndrome. So in total, by the end of 2021, we expect to have 5 clinical programs undergoing clinical investigation. And so again, this is a very -- what we consider a very kind of fast-moving opportunity for Taysha being, in fact, that the company has been discovered and seeded about 1.5 years ago, and we've built up really a robust portfolio pipeline, built out a team, infrastructure to support this ongoing clinical translation. On the investor front, we'll be hosting additional Investor Days, including our Manufacturing Day, which will take place on July 27. This will be followed by a CLN1 Investor Day in August and a Rett Syndrome Investor Day in September. We'll do a deep dive on the natural history, the pharmacology and translational studies and as well as an extensive deep dive on the interventional trials that will be planned for later this year and end points. And we look forward to kind of seeing you guys at those upcoming events and providing further updates on progress as the year goes on. Lastly, I just want to really give a special thanks to our thought leaders who participated in our event over the last 2 days. Dr. Steven Gray, Berge Minassian, Rachel Bailey, Kim Goodspeed, Ryan Butler or more broadly, the entire team over at the UT Southwestern gene therapy program. There are about 70 employees stretched across multiple functional areas, including discovery science, translational development, GMP manufacturing and clinical care. And they just do a fantastic job, and we're so honored to be their collaborators and really trying to do something extremely special. And that's eradicating monogenic disease of the CNS. So with that being said, I just really want to thank everyone for their attention over the last 2 days and their continued interest and support of Taysha and wish you guys all the best and more to come. Have a great day.