Great. Thanks everyone for being here. I'm Yanan Zhu, one of the biotech analysts here. It's my great pleasure to be joined by Al Sandrock, CEO of Voyager Therapeutics. Thanks, Al, for being with us. Thanks for having me, Yanan. Great. I'm wondering if you can start us off with a brief overview of the company, the initiatives. Sure. At the very beginning of the year, we put out a little shareholder letter that said there's three major things going on this year. One was related to tau. In fact, I called it the year of tau. That was in part because there's a lot of data coming from multiple companies on tau. We've seen that so far already. BIIB080, the tau silencing ASO from Biogen. Novartis had some data on PSP with a tau silencing ASO as well. We ourselves are about to have data from our multiple ascending dose study with an antibody directed at the C terminal of tau. This is called VY7523, and we expect to see data in Q4 of this year, tau PET imaging data. In addition to that, we have a second program in tau, which is a gene therapy tau silencing approach where we're vectorizing an siRNA. Very similar in terms of mechanism of action to BIIB080 and to the tau silencing ASO that Novartis has. That is why I call those two programs out, except that what we are doing is we are using AAV to express a vectorized siRNA to silence tau throughout the central nervous system. That is tau. The second piece is capsid POC. After many years, approximately five years of work, we have discovered these blood-brain barrier penetrant capsids that can get into the brain after IV delivery, get broad distribution throughout the central nervous system. We are entering the clinic, so this is the first test in humans now. We have two programs that are testing these novel capsids in humans. One is the tau silencing gene therapy program that I already mentioned, that we call VY1706. We have announced that we have FDA clearance and Canadian regulatory clearance. We are activating sites as we speak. We expect to enroll the first patients this year, and we have said that we will expect to share acute safety data in Q1 of next year. Also in the second half of next year, we expect to share data that says that we are expressing the gene of interest, in this case, the tau silencing siRNA, in the second half of next year. That is the capsid POC. I should also say that our partners at Neurocrine have said that they are entering the clinic with also a blood-brain barrier penetrant capsid derived from our platform for Friedreich's ataxia. They are entering the clinic this year, and they said they expect to share data on that program next year. Two different programs that should provide proof of concept that our novel capsids work in humans. The third thing, and the final thing, is we said that we would be showing more data on the NeuroShuttle platform. This is not AAV gene therapy, but it is basically leveraging the receptors that we discovered from our novel capsids. We reasoned that if these capsids get into the brain through the blood-brain barrier, they must be leveraging receptors to get there. We have discovered a handful of these receptors. The first one we have talked about publicly, it is ALPL. Much the same way that companies are leveraging transferrin receptor and CD98, we have these other receptors that we are leveraging as a shuttle for all sorts of modalities, antibodies, enzymes, peptides, oligonucleotides. We have said that we are going to share data. In fact, we have a plan now. We have an abstract accepted for a meeting in December, where we are going to show a lot of animal pharmacology data, as well as with ALPL, the first of these receptors, as well as some of the safety data associated with ALPL shuttles. That is a high level overview of the three major, I would say, catalysts for Voyager. Great. That's super helpful. If we can dive into the first area, which is tau effort. As you said, there have been a few important readouts this year. Could you comment on BIIB080's finding? That's a siRNA. Maybe more related to— It's an ASO. Oh, sorry. Yeah, ASO. More related to your other tau program, like the vector— Right. —with the tau siRNA program. Maybe we can start there. Sure. Also talk about some of the antibody data this year. Yeah. First I want to congratulate my former colleagues at Biogen for executing on that, on BIIB080, and for showing the results and telling us. We learned a lot by listening to what they shared with us at the scientific meetings this year, as well as their publications on the earlier phase trial. But the way I look at it, at a high level, I think it does validate tau as a target for Alzheimer's disease. I think the data on the cognitive measures, such as ADAS-Cog with BIIB080, were unprecedented in terms of efficacy, 50% roughly, 40-45%, 50% relative to placebo. Also on Mini-Mental State Examination, again, unprecedented levels of efficacy. On CDR-Sum of Boxes, it was sort of in the same range as the anti-amyloids, in the 20%-ish range, 25% maybe. Curiously, not so much of an effect on activities of daily living, which the anti-amyloid antibodies actually have no problem showing effects on ADLs, but for some reason, the tau-silencing BIIB080 did not. There were some questions that came out of that study as well, like why is there what seems like an inverted dose response curve? Biogen themselves say that they are moving into phase III with the lowest dose, which they think is in the sweet spot, if you will. It is hard to understand fully why the dose response was apparently inverted. I think it relates more to off-target effects, perhaps associated with intrathecal ASO, because we have seen things like this with other ASOs, like tominersen, for example, at the highest dose, showed worsening of clinical scores and ventricular enlargement. You see with other ASOs, things like inflammation in the spinal fluid. I think there is some untoward effects from intrathecal ASOs that may have played a role here. In fact, BIIB080 itself showed a steadily increasing incidence of acute confusional state. Right. Which obviously is not great for cognitive measurements in Alzheimer's patients. But overall, as I said, I think it does validate tau as a target in Alzheimer's. In addition, Novartis has been sharing data with the tau-silencing ASO for PSP, and they showed a very nice separation in neurofilament relative to placebo. They also showed, actually, what they call encephalopathy in a few patients. Again, wonder if that is due to some of the intrathecal ASO issues. Anyway, so two different diseases where tau silencing. By the way, Novartis is also entering phase III. In fact, I believe they are already enrolling patients in phase III with their tau-silencing ASO. So it is two separate companies moving into phase III with their tau-silencing approaches. I think that means that. Then there is a lot of other programs I can tell you in tau. So I think it is a pretty exciting time for tau as a target. Got it. In terms of that confusional state, AE, in your opinion, is that on target or not on target? Or could that be addressed? I don't think it's on target because if you look at, for example, tau lowering in BIIB080, all three doses lowered tau at roughly the same levels. At the lowest dose, it was 50% lowering. At the highest level, 60%. You look at tau PET imaging too, not much separate. I think we're on sort of the plateau in terms of the biology, if you will. The effects on tau lowering are sort of on the plateau of the dose response, yet there was what looked like an inverted dose response. That's one of the reasons why I don't think it's on target. Ionis has published, actually earlier this year, there's some untoward effects of intrathecal ASOs on neuronal function. For example, firing of action potential, synaptic transmission. They speculate in those published papers that it's due to either effects on ion channels or on maybe neurotransmitter receptors. We saw it in the old days with SPINRAZA. You get temporary hindlimb paralysis in the animals. You put these highly negatively charged antisense oligonucleotides at high doses, 100 mg plus, into the small space of the lumbar cistern, and then you get this steep gradient. It's not surprising, perhaps, that you might get these. The acute confusional state began a week after injection and lasted for about a week. Mm-hmm. Yeah. At least what was apparent to the patients lasted for about a week. Yeah. Anyway. Got it. Very helpful. It sounds like you think the tau antisense or tau direct reduction approach intracellularly, in your mind, has demonstrated clinical benefit in Alzheimer's, right? Well, demonstrated, we don't have any drugs that are approved yet. Right. I think the preliminary data suggests that it's very promising. Promising. I am old-fashioned. Once it is approved, then I am more comfortable using words like demonstrated. Right. Yeah. Sorry about using that too liberally. Yeah. So that is exactly what I meant in terms of read-through to— Yeah. —tau programs. Yeah. Right? Yeah. Let us now talk about the antibody approach. I think J&J reported data this year. Yeah. That was another readout, and it was a negative readout. Right. The antibody against the mid domain did not work. By the way, we actually had an antibody in an overlapping epitope, which we chose not to pursue because it failed to block the spread of pathological tau in the P301S mouse that we use to determine which is the best antibody to move forward into the clinic with. We would've predicted that that antibody probably would not block the spread of tau. Okay. It had, I think, a 10% effect or so. But it was much less than bepranemab, put it that way. Got it. Great. Let's talk about your upcoming data from your antibody program. Okay. Can you set us up in terms of— Yeah. —the study design, what data we are going to see, and what the bar for success is? Yeah. This is a multiple ascending dose study where we are going to show data from the third cohort. What we did was we chose doses based on the single ascending dose study that we had already done in normal healthy volunteers, where we measured CSF and plasma exposure. We know that with the doses that we are giving in this multiple ascending dose study, we are achieving exposures in the brain that should block the spread of tau. As I said, we relied heavily on this animal model where these are transgenic mice that express human tau, P301S tau. In that model, you inject paired helical filaments from human Alzheimer's brain into one side of the brain, into a particular region, and we look at the spread of pathological tau across the brain. And what we are saying is, what we are hoping is that whatever biology is responsible for that spread from cell to cell is recapitulated in that mouse model expressing human tau. And we chose our antibody because it was the best, it was the most robust at blocking that spread. So far, that animal model has been four for four in predicting the human results. The two N-terminal antibodies, the one from Biogen and the one from Lilly, failed to block the spread of tau in that model, pathological tau. And sure enough, they failed in the clinic. Bepranemab, it did block the spread of tau in that model. And sure enough, bepranemab does block the spread of tau in the human, as shown by tau PET imaging very clearly. And it also predicted that the J&J antibody would not block the spread of tau. So far, one positive predictive value, two negative. I am hoping it is five for five. That was VY7523 because in our hands, in that model, we had the best effect, the most robust efficacy on blocking the spread of tau. And that is, by the way, what we are hoping to see in the humans. We want to be at least as good, if not better than bepranemab. Why? Because bepranemab did have an effect on ADAS-Cog, on cognition. You will remember that even in the overall population, the P value was less than 0.05 on ADAS-Cog. But the P value on the CDR-Sum of Boxes was not less than 0.05. We saw that Roche exited that partnership with UCB. But in their last earnings call, I believe UCB said that they are moving forward to phase III with bepranemab. So that is a nice benchmark. We want to be at least as good, if not better than bepranemab. And so we are going to look carefully at our tau PET imaging data later this year and benchmark it to bepranemab. And if we are not as good as bepranemab, we are going to terminate our program. We do not need anything less good than bepranemab. But if we are at least as good or better, now I am interested, and we are going to look for a partner for that program because Alzheimer's disease is too big for a little company like Voyager to pursue on its own. By the way, one thing we could do is to shuttle it. One thing we showed in our phase I single ascending dose study is that we get 0.3% into the brain. The brain to plasma ratio is 0.3%, so we literally throw away 99.7% of the antibody. If we shuttle it, maybe we can make it even better. We certainly can get more into the brain. And you will recall that gantenerumab had a modest efficacy as an anti-amyloid antibody, but when they shuttled it and made it trontinemab, they improved the efficacy and the safety. We have some options with our anti-tau antibody, but the first step is to get the readout that we are planning for in Q4. For bepranemab, the benchmark, is that tau PET data for tau reduction, is that 33% and 58%? Yeah. It was in the, I would say 40%-50%. They had two separate measures based on what it is. It is a composite measure based on a region of interest, right? And so there is this, what is called the Jack, named after Clifford Jack's temporal lobe tau measurements. And then there is more of a cortical composite where you look at all the cortical areas. So there are two different ways of looking at tau PET imaging and relative. So it did not really decrease tau, it slowed the spread of pathological tau by about 40%, 50%. Right Based on those measures. So that is our benchmark, and we want to slow the spread of tau more than that. I see. Right. Thanks. That's— Relative to placebo. By the way, our multiple ascending dose study does have placebo. All right. Each cohort had placebo patients that were randomized as well. We are going to compare to placebo, relative to placebo, and the benchmark is bepranemab. Yeah. Right. What are your thoughts on, let's say, you did reach that benchmark or you exceeded it, right? How do you think about, or how should we think about the performance of the benchmark antibody on CDR-SB versus— Yeah. —ADAS-Cog, right? It is not tau. Yeah. No. By the way, are we seeing a pattern here? That the tau targeting approaches have a bigger effect on cognitive measures than functional maybe? I do not know. It is an N of two. I would say that that is why we say it has to be at least as good. I would also say that UCB has produced some tantalizing data that if you start off with people who are either carriers or non-carriers of APOE4, and particularly if you have low tau burden, you may have a bigger effect. Those are post-hoc analyses, so you have to take them with a little bit of caution. I think that we want to be at least as good or better on tau PET imaging, and we are going to seek a partner. As I said, we may shuttle it. Okay. Got it. Great. Yeah, let's talk about VY1706. Vectorized tau siRNA. Can you help us understand the data you generated in NHP so far? We have these now very potent BBB-penetrant capsids. We are very excited about the capsid we are using here. It is a Gen 2 VCAP capsid. It is very potent. We did not want to be anywhere near 1E14 vg per kg because that is typically where people have safety issues, right? The Sarepta and other drugs. We wanted to be in the E13 dose range, and the maximum dose that we are going to be testing is 5E13 vg per kg. With that, we have shown that we can get up to 60%, 70% knockdown across the brain at the highest dose. BIIB080 has shown us that maybe you do not need more than 50% knockdown. We may not need to go to that highest dose. This will be a study where we do a one-time intravenously delivered AAV vectorized siRNA against tau, and we are going to be looking at CSF tau as a measurement. First thing is safety. In Q1, we are going to be able to say whether or not the acute safety is there. When you have problems with capsids, it generally occurs early. Q1, acute safety. Later in the year, we hope to show data on spinal fluid tau levels. Because if we are seeing a decrease in CSF tau, we must be getting expression of the vectorized siRNA. By the end of next year, we should be able to say that it is safe and that we are getting gene expression in the brain. Now, whether or not that asset, as a tau silencing asset, has legs in Alzheimer's disease, we're going to need to wait for the tau PET imaging data, which will be after next year. Then there, we want to benchmark it against BIIB080, which we just talked about. That's how I see that program, and it's going to be an ascending. Now, with gene therapy, even your lowest dose has to have a chance of helping patients, otherwise it's not ethical. So our lowest dose will have some tau-lowering capabilities. So, yeah, so there'll be several cohorts, the highest dose being 5E13 vg per kg. By the way, in addition to the lower dose, our capsid de-targets the liver 30-fold relative to AAV9. So in addition to the lower dose, we're using a capsid that basically de-targets the liver. That's why we're pretty excited about the capsid, because it's very potent, we can use lower doses, and because it de-targets the liver. And we have the biomarkers. We have CSF tau, we have tau PET imaging where we can determine whether or not we're getting it into the brain, into the cortex, because we have the PET imaging, and we can measure CSF tau to know whether or not we've lowered tau, and we can benchmark it against BIIB080 too. And so I think that the program, if it shows that our capsids are safe and effective, I think that's pretty big for our field. Then if later on we show that we have tau PET imaging data that's in the range required for an asset to be approved for Alzheimer's, that's also exciting. I see. The acute safety, that's the first, of course, the first milestone, right? Yeah. For that milestone, what is the signal that you will be monitoring? Well, we're going to be carefully monitoring the patients. With the FDA, we design a study where we're going to be dosing a patient, we're going to be observing very carefully, then we're going to dose the next patient after a careful. Then after all three patients are dosed in the first cohort, we're going to have an external safety monitoring committee that will say, "Yep, you can go to the next cohort." So it's an external safety monitoring committee that says we can go to the next cohort, and they won't say that unless they think it's safe. Got it. And we'll be looking at all. Look, when you look at AAV, I mentioned liver, right? The common toxicities associated with AAV are liver, something called TMA, thrombotic microangiopathy. Mm-hmm. Yeah. Those are the two main things. We'll get informa— You can imagine all the measurements we're making in the clinic to make sure that we have neither TMA nor liver. There are some other issues such as insertional mutagenesis and things, and that we won't know for years. How low is your starting dose going to be? We haven't said that. Okay. Trista's telling me I can't say that because I haven't said it yet. Right. I'm not supposed to. But we could have data early 2027. Is that going to be multiple dose cohort data or just the starting dose? Well, all we've said is that in 2027, we'll know whether the acute safety is there and whether or not we have evidence of gene expression based on tau lowering in the spinal fluid. Right. Okay. Right. Gene expression, it's a CSF tau, and the PET imaging tau data will have to come later because that takes Takes more time. It takes longer to reduce tau. That's right. Okay. Typically people show 12-month or 18-month tau PET data. Okay. Yeah. Got it. Yeah. That's very hopeful. Okay. I guess you kind of answered my next question, which is how you think about your, or how we should think about your development strategy for the antibody versus the siRNA— Yeah. —or vectorized. But sounds like the antibody will be partnering material, right? Yeah. Right. Well, look, we've said that Alzheimer's disease is too big for Voyager to take on all by itself, so we're going to look for a partner for both assets. For both. Okay. I wanted to generate proof of concept data in humans before we look for a partner, and hopefully we'll achieve that. Got it. Now, how they're going to be used in the clinic, it depends on the data. I would say that if you look at other diseases like spinal muscular atrophy, almost all infants in this country now gets Zolgensma as infants, the SMA gene therapy that replaces SMN protein, right? And many patients also then, if they're not fully treated by that gene therapy, they take SPINRAZA or they take risdiplam. So there are situations that we've already seen where you don't have to solve all the problems just with the gene therapy. That maybe you use two drugs to really maximize the efficacy and safety for patients. And so hopefully we'll have choices here. If both work, maybe we'll see ourselves using both one day. I don't know. Because a lot depends on what we see in terms of the data. But the precedence in SMA is that people get gene therapy and often one additional treatment in addition that also increases SMN protein. Got it. I forgot or I missed this. For the CSF tau data, is that after the early 2027 acute safety update? Well, we expect to see the CSF data next year. Next year. Sometime next year. Yeah. Right. Okay. I was trying to think— Because we have already shown that we can see data in three to six months in the non-human primate on tau lowering. You do not need 12 months. Got it. Three to six months is enough. I see. I was wondering which program will produce the first proof of concept for the BBB crossing capsid? Yeah. It's going to be a race between us and Neurocrine. Right. I don't care who wins. Okay. Yeah. I'm rooting for both of us. Right. Neurocrine says they are going to enter the clinic this year and show data next year. We are saying we are entering the clinic this year, so maybe, but at least simultaneous. I do not know. Right. Got it. Let us talk about the NeuroShuttle program. Okay. Can you remind us the data you have generated so far? Yeah. What are the key development efforts there? Yeah. We reason that if these capsids get into the brain by crossing the blood-brain barrier, they must be leveraging receptors on the blood-brain barrier to get into the brain. We have discovered a handful of novel receptors that I can tell you would never have guessed could be used as shuttles. The first of these is ALPL, and we have already shown data that it is differentiated from transferrin receptor, which everybody else seems to be using. We have data coming up at a scientific meeting in December, where we are going to show lots of animal pharmacology in more than one species, on what sorts of drugs you can get into the brain, proteins, peptides, oligonucleotides, perhaps. Also, we are going to be sure that it is safe. As you know, transferrin receptor shuttles sometimes cause hematologic adverse events. We are not going to have that, I don't think, because we have no change in reticulocyte count because we are not affecting transferrin. But humans with loss of function mutations in ALPL can have a different disease called hypophosphatasia, which is a decreased bone and teeth mineralization. Now, you have to have pretty severe loss of function. One of the things we have been looking very carefully is, can we shuttle drugs in while producing no problems on the bone? In other words, can we shuttle things with ALPL without causing a defect in bone mineralization? We will be showing some of that data too. Okay. Got it. But you do have multiple additional targets. We have several other, and we're not telling anybody. They don't even tell me the name of the receptors. Not that they don't trust me or anything, but we want to keep it secret. Right. For now. When can we start to see potential getting into clinic and all that kind of thing? Wow. Shoot, that's a bit of a ways off. Yeah. But do you think there might be partnership activity— Yeah. —before that happens? We're actively talking to partners. You'll remember in the early days of the novel capsids, we did a lot of business development. Yeah. Right. I am a big fan of business development because it brings in non-dilutive revenue. Also there is so much to do in the CNS that we cannot possibly do it all at Voyager anyway. We have done some very nice deals and you can bet that I am open for business when it comes to partnering also on the NeuroShuttles. Got it. Yeah. Great. I wonder, for the partner programs, obviously, you touched on Neurocrine. Yeah. Nine partner programs. Five at Neurocrine, three at Novartis, one at Alexion, which was originally at Pfizer. I think, boy, if we show that our capsids are safe and effective, man, I think that our partners are going to be very interested. I am very flattered that Novartis chose our capsids for their next generation SMA gene therapy. Yeah. As you know, they are working on Huntington's too. We are very excited about the potential for gene therapy because I think it can provide transformative treatments for children and adults. Got it. Great. I think with that, we are out of time for this session. Thank you so much, Al. You are welcome. Thank you for having me. Appreciate it. Yeah. Our privilege. Thank you.