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

Good afternoon. I'm Salveen Richter Biotechnology analyst at Goldman Sachs, and we're pleased to have with us Taysha. And with us from the company, we have R.A. Session, CEO; Suyash Prasad, Chief Medical Officer and Head of R&D; and Kim Lee, Head of IR. And with that, all right, thanks for joining us.

And maybe to level set and start here. Could you remind us of the catalyst path for you. And in 2021, you're clearly focused on neurology diseases within -- with gene therapy and have taken quite a few drugs into the pipeline, and we're looking for first proof-of-concepts. So would love to hear.

Yes, absolutely. And Salveen, thank you guys for having us over at Goldman. I would probably say you guys have the best intro music of the pandemic, just so you know, quite dramatic. So thank you for that. But the company is a spinout of UT Southwestern, the gene therapy program there. We closed the seed round last year, the crossover in the summer. So we're about a year from the crossover round and went public in September and since then have hit a number of milestones. Taking the company from preclinical into the clinic with our first program, GM2 gangliosidosis. We've recently just brought on another clinical stage program, really at the regulatory stage. At this point, this is TSHA-120 in giant axonal neuropathy. This is the real kind of key catalyst for the company this year. We presented earlier when we announced the acquisition of this program, that the program has 5 years' worth of clinical data, multiple trial participants with more than 3 years' worth of durability. We've seen clear arrest of disease progression at the second highest dose. And we still have the high dose data later to come this year. We're going to be approaching regulators around our approval pathway for this program in the second half of this year, along with the high dose data will be released into the Street. And so that's going to be a key catalyst for the company as we discussed what that regulatory pathway looks like and that high dose data. What I would also say is, for our initial clinical stage program in GM2 gangliosidosis, TSHA-101 we'll have biomarker data. So HEXA enzyme activity data for that program later this year and extremely excited about that. For our CLN1 program, we currently have an open IND. And we'll be initiating first patient dosing on that program here in the next few months. We anticipate having biomarker data, which is enzyme expression in the first half of next year. And then our Rett syndrome program, which we're extremely excited about TSHA-102. We just had data published in brain, our pharmacology animal proof-of-concept data published in brain. They basically demonstrated a genotypic dependent expression of MECP2, which is essentially thread in the needle in that Goldilocks scenario essentially minimizing gene expression in wild-type cells and getting back up to wild-type levels of gene expression in the disease cells. We plan to be taking that program into the clinic later this year, along with our SURF1-associated Leigh syndrome program that will be entering clinical development later this year as well. So in total, we anticipate having 5 clinical stage programs by the end of this year, and a number of programs moving into IND-enabling studies. And we're fortunate to be able to highlight a number of these programs at our R&D Day on June 28 and 29.

Great. And you're hosting 2 R&D events later in June. Maybe help us explain what you're looking to cover with these events?

Really, we have such a robust pipeline. We always find it difficult to get to everything and give everything kind of its just deserved. And so what we're going to do on the first day is really highlight a number of our clinical stage programs and programs moving into the clinic this year. So we're going to go through the giant axonal neuropathy program in depth. We'll talk about some of the primary efficacy endpoints that we discussed on our previous investor event. But we'll also talk about some of the secondary endpoint that we haven't talked about. And there's just a breadth of data to be presented in that program and really some key learnings that provide nice read through and validation for our scientific approach for the rest of our portfolio, HEK293, AAV9, endothelial delivery. And so we'll talk about some of the commonalities as well as a natural history. We're fortunate in giant axonal neuropathy to have a robust natural history study that's been ongoing for 8 years at the NIH. And so we'll talk about some of those key outcome measures that really help validate the clinical data that we've gotten in the interventional study. We're also going to talk about the CLN1 program. The preclinical studies, the translational studies that led to the open IND, but also the natural history as well as the proposed clinical study design, and we'll talk about that a little bit more in depth. We'll do the same thing on the first day for Rett syndrome. We'll talk about some of the pharmacology data that were just recently published in brain. We'll also talk about the natural history and the proposed clinical study design. And we'll do the same thing for our GM2 program, which is currently being investigated in the clinical trial in Canada at Queen's University. We'll also talk about what we consider registrational program, which will be our U.S. IND, which will initiate later this year. On the second day, we're going to do the same thing with our SURF-1 program. We're going to talk about the natural history study. We're going to highlight this really nice correlation of COX activity to disease progression. And some natural history work that we've just recently done with our collaborators over a UT Southwestern as well as with the advocacy group [ pier ] tier SURF-1. And then we're going to go through some of our IND-enabling for our programs currently in IND-enabling studies. Most notably APDD or Adult Polyglycosan lupus and Body Disorder; Lafora disease, this is our GYS1 knockdown approach, a vectorized GYS1 knockdown approach. We're going to go through our SLC13A5 program and our SLC6A1 program. These are genetic epilepsies, where we've seen some really nice animal proof-of-concept data that are currently in IND-enabling study that's undergoing IND-enabling tox. And then we're going to talk about some of our earlier preclinical programs that have just hit animal proof-of-concept. Our tauopathies program, which obviously, Tauopathies and Alzheimer is kind of the topic of the day with the approval of the Biogen drug. We're going to talk about our approach tackling MAPT-associated Tau production and knocking that down using a vectorized microRNA approach. And obviously, the ability to be able to not only go after Alzheimer's, but also being able to go after MAPT-associated FTD and MAPT-associated PSP. And then we're going to talk about our approach in Angelman, which is basically a vectorized short hairpin RNA targeting UBE3A-ATS, which is a silencing mechanism for the paternal allele. So this is very similar to what Ultragenyx is doing and some of the ASO approaches are doing, but just using an ASO approach. And basically, what we're doing is just vectorizing that approach. With the hopes of having more potent expression, more stable expression and a onetime dose, obviously. And so we're pretty excited about that day and the ability to be able to get to multiple programs over a 2- day event. So it's going to be a 3-hour event on the first day and a 3-hour event on the second day.

So maybe jumping into the pipeline portfolio here. And you mentioned you acquired recently in AAV9 gene therapy TSHA-120, which is in GAN and in Phase I trial. Just help us understand the inhuman data that you saw that basically led you to bring it on board? And what you're looking for -- or what we should be looking for from the highest dose cohort and then what that regulatory update might look like?

Sure. Sure. I'll start and give an overview of kind of what was our key conviction that led us to bring it in the program, and then I'll let Suyash go through some of the clinical data. But really, this is the program that we've known about for quite some time. I've had a relationship with UT Southwestern and Steve for many years. This dates back to our old days at AveXis and his days back at UNC. And this was a program that -- it's a groundbreaking program. It's the first intrathecal dose gene therapy program in history. And I think really laid the foundation for a number of programs that came behind it. Most notably the Amicus CLN6 studies and CLN3 studies. The SMA intrathecal dose study for Zolgensma, a strong trial. And a number of other programs that are using CSF delivered AAV9 and really kind of helped provide a nice framework for that. It also provided a nice foundation for the immunosuppression regimen that we're using for the entire portfolio. And so as we went through building out the portfolio for Taysha, this program kept coming up. Steve would kept saying, this is the reason that he had a high conviction around this approach for GM2. Or this was the reason where he had a high conviction for this approach in CLN1. And we finally just said, Steve, can we sign a CDA and see the data. And ultimately, we did -- we signed a CDA with the NIH, signed the CDA with Hannah's Hope, which was the advocacy group that control the IP. And Suyash and I got to look at the data, and we both look -- and we've both looked at each other and said, this to be approved today. And so ultimately, that's what gave us the conviction. But Suyash, maybe you want to comment further.

Yes, happy to. I mean, I was very excited to bring this program on. I've known about this for quite some time. I'm linked in with Carsten Bonnemann, who's the principal investigator at the NIH, he's been running the study. And the study, as R.A. says, the first ever intrathecal dose study was started in 2015, so a good 6 or 7 years ago. And 14 patients are now dosed with the AAV9 intrathecally delivered gene therapy product for giant axonal neuropathy. Now you may might want to be familiar with the disease, but giant axonal neuropathy is a very severe, progressive neurodegenerative disease that affects principally children. These children start off in life, they look fine at birth, but around the age of 2 or 3 develop an imbalance, a problem with their gate, this progresses to a motor dysfunction. These kids underpin wheel chairs at the age of 10, on ventilators at the age of 15 and usually death in the late teens or the early 20s. So it's a relentlessly progressive fatal disease. Now what we've seen in the interventional trial is that the ongoing trajectory of natural history is a decline over time of 8 points per year with the MFM32 score, which is a scale that is a little bit like a chop intend for the younger infants, but it's for the older children. It's well accepted by regulators, it's well accepted by clinicians. And there's this 8-point decline in natural history on an annual basis. And 4 doses have been administered in the gene therapy interventional trial. We've shared data on the first 3 doses, the high dose is yet to come. But in the 2 medium doses, the medium low and the medium high. With the medium high dose in particularly you see clear stabilization of the disease progression, i.e., this 8-point decline every year translates just flattens out, it immediately flattens out to in the print of 8 points over a year, 16 points over 2 years and so on. And don't forget a 4-point change that MFM is deemed to be clinically relevant, clinically meaningful. So as we approach our discussions with the regulators on this particular program, as R.A and I already said that we think it should be improved now. We see -- it checks all the boxes the regulators need it to check. We've got dose response across 4 doses. We've got a great, very in-depth natural history study of 45 patients. Clear arrest of disease progression at the medium high dose with the high dose yet to come. Got a long-term safety, long term efficacy, the 6 patients dosed now over 3 years. And importantly, long-term durability of effect. So there's no diminution of the effect over time. So that I think will set us up nicely for the regulated discussions we're planing to have towards the end of this year.

Perfect. And we're also expecting from your -- the rest of your portfolio, I guess, prior to this acquisition. But first data on early biomarker data in GM2 in the second half of this year. And help us understand what the clinical bar for success looks like with this first read and what that means for proof-of-concept for your platform?

Suyash, you want to go ahead and -- well, I could give an overview, and you could go ahead and go through a little bit more depth. Salveen, I think we're pretty fortunate here. There's a couple of things about GM2 that makes us feel pretty confident in this approach. The first thing is it's a lysosomal storage disorder. There's no fear around over-expression of the enzyme and like a lot of lysosomal storage disorders, a little bit of enzyme goes a long way. What we're also fortunate of is the fact that the natural history for GM2 is really well-elucidated. And it's really -- there's this really nice correlation around enzyme activity level and disease progression. And so I'll stop there and let the doctor speak, but really, we feel pretty confident about this approach, and it gives us a lot of confidence going into CLN1 as well, where you have a very similar disease where it's a secreted enzyme, no fear around overexpression and a little bit of enzyme goes a long way. But to answer your question around GM2. Suyash?

So yes, sure. So R.A. is quite right. There's 3 forms of GM2 gangliosidosis, the commonest form is the infantile form, about 80% of these patients have a rapidly progressive disease, which results in death usually by about 3 years of age. They have less than 0.1% enzyme activity. The juvenile form, also an awful disease resulting in death by the teenage years they're running around about 0.5-some levels of enzyme activity. The other line enzyme that's missing is HEXA. And then the adult form they have usually a normal lifespan of some significant neuropsychiatric cognitive motor dysfunction, but they're running about 2% to 4% of enzyme levels. So we actually think that a 5% level would probably be enough to normalize lifespan and ameliorate clinical size and symptoms very dramatically. Now the reason we know this is because the enzyme is a secreted one. So what that means is that when you think about what will happen when you inject the gene therapy intrathecally, the drug will travel to the brain, the capsid enters the brain, the DNA starts being produced, with the alpha subunit and the beta subunits, which then go on to join the function of heterodimer. It can start to break down GM2 gangliosidosis within each cell and can then leave that particular cell and enter another cell through the Mannose phosphate receptor and work in that now to breaking out additional GM2 gangliosidosis. You don't need much, only a little bit of it goes a long way. For those reasons, we think that it's a really good target. And as we've already mentioned, we expect our biomarker data from this program towards the end of this year.

You could think about the approach, just to add to what Suyash was saying, when you could think about the approach for both GM2 and CLN1, the goal is to basically transduce as many cells as you can at a high rate, essentially turn them into biofactories. And we know that there is a pretty high bar around the fear of overexpression, that there's really no fear around over expression. So the goal is just transduce as many cells as you can, turn them into biofactories and allow them to secrete out its enzymatic activity out and to move from cell to cell. That's really the goal here.

And what about the translation, I guess, to functional benefit here. And I know you said there's a lot of good natural history data. But one thing we've been seeing recently is that the natural history data, that when you design a trial can just result in -- I mean, there's so much variability. So how are you guys thinking about those 2 aspects?

No, it's a good question. Suyash, you want to take that?

Sure. So R.A. is quite correct. The natural history data is well understood for this particular condition. Part of the reason for that is we've known Batten disease for many, many years. It's first discovered in the late 1800 by William Tay, a British ophthalmologists. And HEXA was identified in 1969. And because of the preponderance of the Ashkenazi Jewish population, genetic screening, mutation analysis has been going on for a long time. There are several good publications on the natural history. And true, there is some variability. But the key thing, once again, is that the clinical outcome is predictive -- well, is predicted somewhat with the baseline enzyme level. So that's really our key with this biomarker that can really help in that regard. And I think when you look at the preclinical data, we've actually shared some very nice preclinical data on our construct. In the Sandhoff mouse model that looks at survival. It looks at neuromuscular, neurophysiological outcomes, and it looks at reduction of GM2 gangliosidosis in a dose response manner. We see a really nice improvement as dose-dependent in survival, in neuromuscular outcomes and also in reduction of GM2 gangliosidosis. They all trend in the right direction and they all mirror each other. So a very powerful compelling data. But with regard to outcomes specifically, I think the key thing with this disease, and in fact, the key thing with CLN1 is to diagnose the patients early and treat them early. That, I think is, regardless of level of enzyme that's being produced, I think that's going to be the key determinator of how all the patients do. So in parallel to that, we're actually working very hard on speaking to the relevant bodies about newborn screening programs for this program and for CLN1. The soon you diagnose a patient and get them on treatment, the better the outcomes are going to be because this is a neurodegenerative disease while neurons are being lost on an ongoing basis. And when you're losing neuron, you don't get it back. So if you can treat before there's too much of an accumulated loss of neurons, you should see a really nice functional and clinical outcome.

Right. And maybe you could touch base on the other data sets that we're supposed to see thereafter, the CLN1 data in particular? And what -- how we should think about the bar there for clinical successes?

Suyash, do you want to tackle that?

With some of the other programs?

It's about CLN1 and then we can talk about...

Yes. So CLN1, once again, is a lysosomal storage disorder similar to GM2 gangliosidosis. The disease is also an awful disease but slightly less rapidly progressive. So usually, the kids develop, they lose their milestones, they start having seizures, they lose their vision, usually die around the age of 5 or 6 years of age. It's still really awful, but not quite as severely progressive as GM2 gangliosidosis. Lots of learning there, once again. This is lysosomal storage disorder, the enzyme missing is PPT1. In the absence of PPT1 you get a build-up of palmitoyl-weighted substrate, which then causes neuronal degeneration, neuronal loss. We will be -- we're actually including the full-length CLN1 gene within AAV9 capsid, we'll deliver the drug intrathecally in the clinical trial. The full-length human gene, we'll enter the brain cell, start producing enzyme, which would start to break down the accumulated substrate of palmitoyl substrate and then once again, it's a secreted enzyme. So the enzyme who leave the cell -- enter the cell through the mannose phosphate receptors and breakdown substrate there. We do have an open IND for the program. Just through the acquisition, we are in the process of making drug for the program. And we're making a few tweaks to the protocol. But we have plans lined up to start the clinical trial later on during the course of this year. We do have our sites identified, and there's a nice pool of natural history data that are available for the CLN1 program. Both in a site in the U.S. and also a site in Europe. So we're working closer with those individuals. And we're looking forward to moving forward. And once again, I think the cadence of events for this particular program will be, we inject the drug intrathecally. So the AAV9 capsid enters the cell, DNA pops that, enzyme starts being produced. You get maximal transgene expression around about 3 weeks of age -- 3 weeks after dosing, sorry. So the 1-month time point, we should see a nice jump up in CSF levels of enzyme, it should go even [ 5-uptime ] when hit the 3-month CSF time point. I would hope to see the beginnings of some clinical stabilization around about the 3-month time point and maybe beginnings of some clinical improvement. So that's how I imagine the cadence of events occurring after dosing.

Salveen. So to answer your question, what should you guys expect? We expect to have data on this program sometime in the first half of 2022. A win here is really enzyme activity around that 10% level of normal. That's really kind of what we've seen from some of the enzyme replacement therapies. But also what's interesting is and I'd point to as a comparator program, the Amicus CLN6 program, which is a little different because it's a membrane-bound protein. So it doesn't have the benefit of secretion. But this is also AAV9, this is also intrathecally dosed. And if you recall, I was fortunate to be a part of that program early on, that's a program that came out of nationwide children's, Brian Kaspar's lab, and I help transact that program to Amicus after we sold AveXis. So I was doing that in my free time. But once we saw that program, that gave us a lot of conviction to bring on this program into the fold once we had an opportunity to do that because the bar was a lot lower here because this is a secreted enzyme versus a membrane-bound protein like CLN6. So it gave us a lot of conviction here that the bar is a lot lower. So what we would be hoping to see is somewhere around 10% of normal enzyme expression.

Got it. And then I think the Rett program, yes, the Rett syndrome program, just remind us where that stands? And also, it's a little different. You need to thread the needle carefully with how you manage the outcome. Maybe walk us through that and what makes you comfortable for the preclinical data?

Yes. This program, we're quite excited about. The clinical trial is expected to start before the end of the year. You're absolutely right, Salveen, in a sense this is what's called the Goldilocks disease in a sense of too little protein expression. You don't cure the phenotype too much. You get another disease, MECP2 duplication, which is very similar to Rett, but it can be a little bit more severe then Rett. And this is really the culmination of about 15 years of work out of Dr. Steven Gray's lab. And Steve will joke with you. He'll say -- he was a little young and naive when he took this on and if he would have known it would have taken this long, he would have never done it. But fortunate for us, he did and fortunate for patients, he did. And the preclinical pharmacology data and animal proof-of-concept data was actually just published actually on Cinco de Mayo because I had a nice margarita to celebrate, I remember that. But that data was just published in brain, and it was just really compelling data. The first data that actually showed genotypic dependent expression of MECP2 protein, which means basically minimal expression in wild-type cells but allowing for expression in disease cells and quantifying that in certain parts of the brain, quite compelling data. And I encourage the people who are watching the fireside chat to take a look at that paper. It's a quite compelling paper. I'll pause there and allow Suyash to kind of go through the biology of the disease and what makes us so confident in the preclinical data that we've achieved so far translating to humans.

Yes. Thanks, R.A. Salveen, you're absolutely right? Too little MECP2, which is a protein missing in Rett results, in Rett syndrome too much and you get overexpression toxicity. And the challenge here is that the children with Rett are girls and this an excellent dominant disease. And what that means is that due to random X inactivation or lyonization as it's known, half of the cells of the female patients have normal levels of MECP2, the other half of the cells have abnormal levels of MECP2. And it says 50-50 balance between normal and abnormal cells that results in the phenotype of Rett. Now the challenge, therefore, the gene therapy approach is that in the cells that have no MECP2, you want to bring up levels to the point at which is effective, but the cells already have plenty of MECP2 or wild-type cells with a minimally express MECP2 or not express MECP2 at all, which means you want to regulate the amount of expression of MECP2 on a cell by cell basis. And Steven, as R.A. mentioned, has been working o this for 15 years and he seems to have cracked it and got some great preclinical data in how this works. And the way it works is that Steven has incorporated what's known as the miRARE platform, microRNA auto regulatory element. And what that is a strip of microRNA binding sites that bind to down regulatory microRNAs that are endogenous cell. So in the normal healthy cell, these micro -- you have microRNAs and those are upregulated in the presence of MECP2 that bring down endogenous production of MECP2. So you have this feedback loop keeping MECP2 at the normal level. And there are many of these feedback loop systems within the body. Now by incorporating this strip of miRARE in our construct, what happens is this? We inject the drug intrathecally, it travels to the brain, the capsid enters the brain cell, the DNA pops out, MECP2 produce starts being produced. As the MECP2 goes up, endogenous down regulatory microRNAs are triggered, they then bind the downregulatory binding sites on the construct in the untranslated region of the construct, this brings down the output of MECP2 from the construct. So what happen's you end up reaching the steady-state level of MECP2 within the appropriate physiological premise, and not too much, as we've said, and definitely not too little. It all sounds great in theory, but we've also got a lot of really nice animal data to show that it works. By what -- we have dosed wild-type mice to show the lack of any kind of expression toxicity, and we've dosed knockout mice, i.e. null cells to show that there is lot of efficacy. So I would encourage, as R.A. said to read the paper that's published in brain very recently. It's a very long paper, a very detailed paper goes to this data beautifully. And in addition to the functional animal data, you've got quantitative expression of this particular protein, showing upregulation in the null cell and downregulation in the wild-type cells in different regions of the brain.

And what I would say just to finish what -- on to what Suyash mentioned is, it was a nice -- not only was a nice proof-of-concept for this approach in Rett syndrome, but it's a nice group of concept for this approach in other diseases where you may have this dose sensitive type of gene. There's a number of diseases where they are dose sensitive, Angelman being one of them. But a number of others Pitt-Hopkins and there's a couple of others where you really need to tightly regulate the production of that protein. And so what we're going to be doing is actually moving this platform forward in a number of other new programs that we're going to be going after. The most notably, the one that we've disclosed is FOXG1, which was always thought to be on the Rett spectrum. It was always start to be a part of Rett until the gene mutation was discovered, not too long ago. And so this will be the second program that will be utilizing the miRARE platform.

Maybe 2 last questions here. One is you have a really, really broad pipeline that's in the clinic already with a lot going on in optimization and new assets. How are you prioritizing and managing your resources, your time and the infrastructure? And I guess, secondly, you've talked about a 3-pillar manufacturing strategy. Just remind us where you kind of stand with kind of scaling up the HEK293 platform there?

No, absolutely. And both really good questions. Fortunately, we haven't had to necessarily make any type of prioritization decisions. When you think about gene therapy, drug development, it's much different than kind of classic modality drug development because you're able to do these things at a fraction of the cost because we're not doing any high throughput screening, we're not trying to identify the target. We know the target. The target is a mutated gene. So by holding a bunch of things constant, and this is really our scientific approach and kind of central to our thesis. We use validated gene therapy technology, coupled with very novel payload design. So basically, it enables us to iterate on one key aspect of a program, which is the payload. One key aspect. We hold constant AAV9. We hold constant the intrathecal route of delivery. We hold constant the manufacturing process. So we're able to really focus and be very thoughtful and novel and targeted as it pertains to the payload design. So in some cases, it's bicistronic vectors, 2 genes packaged into a single construct. That's what we're doing in GM2. In some cases, it's a knockdown approach, a -- vectorized knockdown approach. That's what we're doing. In our Tauopathies program, where we're targeting MECP-associated tauopathies. In some cases, we're unsilencing allele. In the case of Angelman, we're using a vectorized short hairpin RNA to silence UBE3A-ATS to restore function from the paternal allele. And in some cases, just classic gene replacement. And we've already talked about our approach, our regulated approach in Rett syndrome. And so again, by holding a bunch of things constant, we're able to gain significant economies of scale because we're not iterating across the board. So that saves a lot of time, and that also saves a lot of resources. What I would also say is our strategic collaboration with UT Southwestern allows us to do a lot with a little. We would never be able to do target identification, construct design and early translational animal preclinical proof-of-concept studies as efficiently with the resources that we have, better than UT Southwestern does. So from the time it takes us to say we're going to be going after a certain target until we actually get an animal proof-of-concept, actually get a definitive go, no-go decision with animal proof-of-concept data in hand. It's about $1.50 million to $2 million because we're able to do it at scale in UT Southwestern. If we brought that in-house, we would never be able to do it for that price. It would be double the cost it would take us probably twice as long. And so by leading on UT Southwestern scale and expertise, we're able to do a lot with, again, very little resources. And again, we do what we do best. We focus on kind of everything post-animal proof-of-concept. So IND-enabling studies, GMP manufacturing, clinical development and then, obviously, commercialization, where UT Southwestern focuses on target identification, construct design and early preclinical. That really allows us to do a lot with very little. And because where we in-licensed the programs from UT Southwestern, it was already appropriately kind of prioritized. You had some programs going into the clinic. You had some that just hit animal proof-of-concept they were going into IND-enabling studies. And then you had a bunch that were just actually at the discovery stage just embarking on their animal proof-of-concept assays. And so we haven't necessarily had to make those trade-offs today. Not to say we won't have to do that sometime in the future. But today, we've been able to keep everything relatively moving in parallel because they're putting pressure on different parts of the organization. So that's -- hopefully, that answers your prioritization question. I don't -- Suyash, do you have anything to add to that?

No. I think you covered it all. I think we let the science lead the way more than anything. And the partnership with UTSW. I can just speak from personal experience, works wonderfully well. They do the discovery, lead optimization, early work. We then get involved, we move it more towards clinical and regulatory, but there's ongoing close collaboration all along the way. There's 60 of them there at UTSW, coupled with 120 people [indiscernible]. So it really is a significant discovery and lead optimization organization that's really under our auspices.

What I would say, Salveen, I'm in our new office. We just opened this week, it's actually pretty cool to be back in the office. But UT Southwestern is literally across the street. So we're pretty close with our collaborators there, and we work almost as a single unit versus kind of 2 distinct organizations. Your second question was around our manufacturing strategy, and we have this 3-pillar platform. And really, the first it starts with UT Southwestern and the GMP facility that they have at UT Southwestern. They have 500 liters worth of capability within the GMP facility, but also had 200 liters worth of non-GMP capability that we use to run the early pharmacology studies, translational studies and do some of the early tox work, and which is nice that we're able to design a construct, do early pharmacology work and also manufacture GMP material, all within kind of a single site, which is really, really cool, and it cuts down on a lot of what we would say capacity constraints that you have within the rest of the industry. What we also supplement that is with our strategic collaborations with CDMO partners. Most notably with Catalent. So you remember our long experience going back to our AveXis days with Catalent. We basically designed the AAV9 manufacturing process together. At the same time, we were building out the liberty build facility. And you probably know this, but Catalent is the only license -- commercially licensed producer or manufacturer of AAV9. They're actually commercial license to produce Zolgensma. So they're really, really good at this. And we actually have 4 ongoing GMP runs right now at Catalent. So we're always able to kind of slot in -- a slot over a Catalent. And because we're not bearing the process from one program to the next, all we're doing is replacing a single plasma. So we use triple plasma transfection and suspension. Essentially, we're using the same AAV9 plasma, the same helper plasma. And the only thing we're swapping out from 1 program to the next is the transgene plasma. So it basically makes it a lot easier from going from one program to the next. And then the third pillar of that is obviously building out our own internal manufacturing platform and capability. And we disclosed late last year that we had signed a lease on our very own 187,000 square foot facility located in Durham, North Carolina. We've broken ground on that facility. We had actually a great groundbreaking ceremony about 2 months ago is actually where it was the first time that a lot of that team actually met not over Zoom, so it was actually cool to get everybody together, had a patient organization out to help us break ground on the facility, and construction is ongoing. That facility at scale will have 2,000 liters worth of capacity with the ability to double that capacity of 4,000 liters. It will be able to supply not only preclinical material, but all the way to commercial scale material. And again, it will be our own facility, and it will have to -- at scale, it will have 200 employees at that facility. So we're really excited about that to bring that online. That will come online in 2023. Now when we start talking about some of our near-term opportunities, particularly giant axonal neuropathy, that's a program that manufacturing, in particular, commercial manufacturing is going to be really important now. And so what we've decided to do is to actually manufacture that virus at the same-facility, with the same CDMO partner that manufactures the clinical material because we want a like-for-like process. So we're going to be using the same cell line, the same manufacturing system, which we're fortunate to have, which is HEK293 suspension. And we just signed an agreement for that manufacturing slot. So we're going to be bringing that online here in the near future and hope to have product here to be able to do analytical comparability on in order to support the BLA.

Great. Well with that, thank you so much, R.A., and Suyash, you can really appreciate the time today.

Thank you for having us. Always a pleasure.

Bye. Salveen.