Voyager Therapeutics, Inc. (VYGR) Earnings Call Transcript & Summary

Hello, everyone, and welcome to H.C. Wainwright at Home Virtual Fireside with Voyager Therapeutics. I'm Patrick Trucchio, senior health care analyst at H.C. Wainwright and your host for today's event. It's my pleasure to be joined by President and CEO, Al Sandrock. Voyager is advancing 2 tau-targeted Alzheimer's programs towards clinical inflection points, including VY7523, an anti-tau antibody with tau PET imaging data expected in the fourth quarter of this year and VY1706, a tau silencing gene therapy that's recently received FDA IND clearance. The pipeline also includes Voyager's TRACER capsid platform and a partner portfolio that features Neurocrine's planned Friedreich's ataxia trial start in the second half of this year. So with that, Al, welcome. It's a pleasure to have you with us today.

Nice to be here, Patrick. Thanks for inviting me.

So just maybe to start, what is Voyager built to do in neurology? And what do you have that others do not?

Yes. So we felt that neurotherapeutics was at a tipping point in the sense that we had -- we understood the targets we wanted to go after. The unmet need is enormous. And -- but the key barrier was the blood-brain barrier, a physical barrier that actually prevents most modalities from getting into the brain. And we felt that if we could solve the delivery issue, we would enable neurotherapeutics not only enabling our own pipeline, but multiple other pipelines as well. And so we have technology where we can actually get AAV capsids across the blood-brain barrier. But we also are now developing shuttles that will get protein therapeutics and oligonucleotides across the blood-brain barrier. So it's really to open up neurotherapeutics by solving the delivery issue into the brain.

Right, right. Great. And so now you now have 3 ways to get into the brain, AAV capsids, there's NeuroShuttle and then systemically dosed antibodies plus a small molecule collaboration. So how are you thinking about prioritizing these different programs?

Yes. So we always choose programs where we can -- since neurotherapeutics is difficult, right? So despite having all these targets, we may have more targets actually, I was thinking about it the other day than oncology, but there's so much more going on in oncology. And I think the delivery issue is the problem. And so -- but we want to choose for our own pipeline programs where we can not only go after validated targets with our novel technology, but where we can de-risk quickly in the clinic. And so we choose programs where we have biomarkers that will allow us to know whether we're on track. So for example, in Alzheimer's disease, we have amyloid PET imaging, tau PET imaging, fluid-based biomarkers, et cetera, for ALS, for example, for our small molecule program. We have neurofilament, serum neurofilament is a blood test that tells you whether or not you're getting the drug to the right place and being active enough to save motor neurons from dying. So that plays the developability and the early de-risking plays into our thinking about our own pipeline, and that weighs heavily in our prioritization.

Yes, that's helpful. And before we dig deeper into the tau programs, maybe you can just give us some background on tau as a target and sort of why it's the right place to be concentrating the focus right now at Voyager.

Yes. So I think we just had recent validation in some ways that in a randomized double-blind, placebo-controlled human clinical trial of hundreds of patients, the Biogen BIIB080 data that tau actually is a validated target. When you look at that data on cognitive measures, it's actually superior to the anti-amyloid treatment effect sizes. On CDR sum of the boxes is comparable. And so -- and then there was a very clear-cut use of PET imaging and fluid-based biomarkers to help understand what's going on in that study. So I think -- so that's why tau was felt like the right place. I'll say this, by the way. We had other programs. We had GBA1, SMA, but we were offered a lot of money by other companies to pursue those targets. And so of course, we have the ability to opt back in for a couple of those programs. So we're also interested in rare diseases, et cetera. But we felt that tau was a very good target because we can know whether our capsids are working, for example. And that right away is of high value to our company to be able to prove with PET imaging and with fluid-based biomarkers that our capsids are safe, getting into the brain, expressing their payload. And then on top of that, these assets could be of enormous value. So that's why we chose to concentrate on tau.

Right. And so moving then to VY7523. This is the anti-tau antibody. Maybe you could describe this for the audience, what it binds, why you went after the C-terminal region? And how is tau it was selected for 700 anti-tau antibodies.

Yes. So that program is actually sort of more of what I would call an old-fashioned program in the sense that we're not doing anything special on the delivery side. We do know that antibodies get in, but it's very, very small quantities. But we felt that this was a very interesting antibody because we found that in an animal model where that expresses human tau where you inject Alzheimer's-derived material into the brain of animals that express human tau. This particular antibody was better than any of the other ones that we tested in terms of blocking the spread of pathological tau in the brain. Moreover, it was selective for pathological forms of tau, and we think that could be incredibly important. In the case of anti-amyloid antibodies that turned out to be critical. So that's -- so we chose it empirically. We had no real preference for which epitope. We wanted the animal model that expresses human tau and that we think recapitulates the way pathological tau spread in the Alzheimer's brain, we thought that this empirical approach will prove to be right. And so far, by the way, we're correct that this animal model is very predictive. The 2 N-terminal antibodies that failed in the human clinical trials failed in this model. Bepranemab, the one antibody that seems to work actually works in this model, and we actually could have chosen an epitope very similar to bepranemab. It just wasn't quite as good as the epitope we chose the C-terminal epitope is the best. Moreover, it correctly predicted that the J&J antibody would fail. So, so far, it's 3 for 3, 4 for 4 in terms of correctly predicting the human clinical trial results. I'm hoping it's 5 for 5 with our antibody later this year when we look at our tau PET results.

Right. Terrific. And so now the Phase I multiple ascending dose trial is fully enrolled the 52 patients, the tau PET imaging data, as you mentioned, is expected later this year. Can you walk us through what that trial is designed to show and maybe what it's not designed to show?

Yes. So it's really based -- it's not powered to show an effect definitively on clinical outcome measures such as CDR sum of boxes or ADAS-Cog. We will look at those measures, of course. But what it's mainly powered to do is to look at tau PET imaging. And we're going to look at 1 year, and we're going to see how we stack up to bepranemab. You'll recall that bepranemab did block the spread of tau and it had actually statistically significant effect on ADAS-Cog in the entire population, not a subgroup, but in the whole population. Also had a trend in CDR sum of boxes that it missed on that endpoint. So we want to be as good or we want to be at least as good, if not better, than bepranemab. So what we're going to anchor on is the tau PET imaging results at 12 months, which we're going to see later this year. And we hope to be at least as good, if not better, than bepranemab on blocking the spread of pathological tau in the brain of humans with Alzheimer's disease.

Yes. That's helpful. And sort of partially answered this next question, but just again, without getting ahead of the data, what sort of magnitude and pattern of tau PET effect would you consider a clear positive? And you mentioned the external benchmark, but -- so maybe you can frame for us what external benchmark makes most sense.

Yes. So we think that lecanemab is the best external benchmark because we have some idea of how tau PET imaging relates to clinical outcomes. We see that we're right on the edge of seeing some effects. And so we want to see greater than 50% slowing. Bepranemab had roughly 50%, depending on the brain region that you look at. So we want to compare region by region, and we want to see that we're at least as good, if not better than bepranemab. And we want to see that at 1 year. And the reason why we chose 1 year, by the way, because I know that's another question you've been asking, which is why not look at 18 months? And yes, you get more of an effect on tau PET imaging at 18 months. Bepranemab did show separation at 12 months. So if we don't show separation at 12 months, we're not at least as good or better, right? So as I said, we have a high bar now. We have to see lecanemab. And we know that at 12 months, bepranemab did start to show separation. So if we don't, we're not as good.

Right. That makes sense. And have you said or can you say which brain regions and which prespecified analysis are going to be part of this readout?

I don't want to get into that kind of detail. We're going to look at multiple regions. There's a region of interest, ROI type assessment. Moreover, there's a composite measure that looks across multiple cortical areas. It's complicated because it depends on where the tau has already spread by the time you enroll a patient. So the starting point on the tau spread is different from patient to patient, depending on how advanced they are. So what we want to do is to look at regions where there's already some tau to see if we can block the spread within that region. But even more importantly, perhaps is to look at whether or not we can slow the spread into regions where there isn't any tau at baseline. And so that's why -- so as you can see, it's region by region and depends on where you start off at baseline, what you want to look at more intensely. But the main thing is that it's mainly cortical areas, it's temporal lobe, frontal lobe, parietal lobe tau in Alzheimer's disease spread to cortical areas. It doesn't spread to subcortical areas, it doesn't spread to the brainstem or the spinal cord.. So it's a cerebral cortex for sure.

Right. And do the fluid tau biomarkers need to move in the same direction for you to believe the imaging result?

We'll look at them. I'm not really -- I want to anchor more on the tau PET imaging because we've seen data actually, Biogen had a trial where their antibody moved the most promising of tau PET biomarkers, the MTBR tau by 40%, clear cut effect by 40% and had no effect on tau PET imaging or clinical outcomes. So for this particular indication -- sorry, for this particular trial, I'm not really looking very hard at the clinical effect of fluid-based biomarkers. It would be nice to see some movement. In fact, the antibodies that failed all moved to fluid-based biomarkers quite well. had either minimal or no effect on tau PET imaging. So I'd rather anchor more on the tau PET.

Right. That's really interesting. And I think in the single ascending dose study, it showed cerebrospinal fluid to serum ratio of about 0.3%. Do we know that that's enough exposure to engage the pathological tau where it matters?

So that shows right away how much of the blood-brain barrier and impediment. 0.3%, okay, which is, by the way, typical of every monoclonal antibody that we've looked at in humans, right? So we throw away 99.7% of the antibody that we inject intravenously. However, we know the concentration we need to achieve in the brain and in the spinal fluid in order to get the blocking effect. We know that from our animal studies. So we have an effective concentration we want to reach in the brain. And we know that we have reached that effective concentration because we use very high doses of intravenous antibody to make sure we have enough in the brain that exceeds that EC50, if you will, of blocking the spread. So we're pretty confident we're in the right dose range with this study.

Most prior tau antibodies have disappointed, as mentioned. And I'm wondering what your explanation is for why they failed and what specifically is different here?

Yes. So the epitope is different. And you might say, well, why does the epitope matter so much? So it turns out that in the extracellular space in the brain, -- and the reason why I'm focusing on the extracellular space is that we're trying to block the spread from one cell to another. And so there must be an extracellular phase where a particular pathological form of tau jumps from one cell to the other and then causes a template-driven misfolding in the receiving cell. When you look at the extracellular space, there's a lot of forms of tau. Many of them are proteolytic fragments. So there's many different fragments and different forms of tau. And one of those, obviously, or maybe some of them, maybe it's more than one, but some subset of those forms of tau are the ones that are responsible for this spreading, which we think is prion-like. And the explanation I have is that the N-terminal antibodies bind to the forms of tau that are not responsible for that pathological spread. And we think that -- and that's what this animal model indicates, and that's what the human trial indicates. We think that the C-terminal ones -- so whatever that pathological form is that's responsible for the spread, what we're saying is we think it contains the C-terminal domain. And that's why we think the C-terminal epitope is critically important. We'll find out if we're right. But as I said, this is all based on empirical data from the animal model. But that's the thinking is that you have to bind to the forms of tau that are responsible for that spread from cell to cell.

Right. That's interesting. And then maybe moving on to VY1706. This is a tau silencing gene therapy. So here we have a onetime IV gene therapy to silence tau inside neurons. So maybe you can tell us why you're pursuing this approach as a potential treatment in Alzheimer's.

Yes. So the antibody relies on blocking the spread by binding to the correct forms of tau that are responsible for that spread. A whole different approach is to just decrease the expression of tau. And there, we don't have to worry about the epitope. We're going to just decrease all forms of tau. And the reason why we like that is that there's been some very good human proof of concept already. I mentioned the BIIB080 data in Alzheimer's disease where there was an effect on cognitive outcome measures. Moreover, Novartis shared earlier this year that they have an antisense that silence the expression of tau, also an antisense oligonucleotide that actually seems to work in progressive supranuclear palsy, another tauopathy -- and in fact, Novartis is now enrolling patients in a Phase III trial of their tau silencing antisense oligonucleotide. So we have 2 different diseases now, one is tauopathy and one Alzheimer's. Well, they're all tauopathies, obviously, one Alzheimer's and one a more rare tauopathy called PSP. And so I like the idea of following an antisense oligonucleotide that has shown proof of concept in the humans, following that up with a gene therapy. Antisense shows that it could work, shows that it's safe after years of treatment, and then we can follow on with a more permanent solution, a onetime IV dose that does essentially the same thing as repeated intrathecal injections of an antisense. This actually happened in SMA. You remember SPINRAZA was a antisense oligonucleotide treatment that was first approved for spinal muscular atrophy. Two years later, Zolgensma comes in and it becomes the standard of care for all infants with SMA. And then now some patients take both Zolgensma and SPINRAZA. If they need a little extra efficacy, they add SPINRAZA. So this paradigm of starting with an oligonucleotide-based treatment that shows proof of concept maybe even gets approved, followed by gene therapy has already happened for another neurological disease, and we want to follow the same path, if you will, for tauopathies.

Right. That's interesting. And you mentioned BIIB080 a few times now. This is Biogen's tau antisense program. Their data showed clinical slowing that was strongest at the lowest dose studied. I'm wondering why does the inverted dose -- what does the inverted dose response tell you about how deep tau lowering needs to be? And does it change any of your own dose thinking?

Yes, it does affect our thinking. And we always said we were going to look at that data as a guide for how much -- what we need to do because that was the first data that was large enough to actually show the correlation between effects on tau, tau silencing and clinical outcomes. So the lowest dose suggests that we may not need to go to the higher doses in our ascending dose first-in-human study. In fact, 50% decrease in tau, CSF total tau seems to be enough to produce a robust clinical effect on cognitive endpoints. So that does affect our thinking as your question implies.

Right. That's interesting. And VY1706 Phase I is open-label dose escalation up to 18 tau PET positive early Alzheimer's patients across the 3 cohorts. I think the dose is 5e13 vector genomes per kilogram. Maybe you could just talk us through kind of why that's the next dose.

Yes, because in our -- both in our mouse and in our nonhuman primate studies in the monkeys we get 75 -- up to 75% reduction in tau expression at that 5e13 dose. We don't think we need to go to even 75%. As I just said, 50% reduction may be enough. So now that was our highest dose that we tested in our toxicology studies. It turned out it was -- that's safe enough to dose humans with. But like I said, why be aggressive why be so aggressive on the dose when Biogen themselves showed that 50% is enough. So yes, we didn't test doses higher than 5e13 because we think we don't want to suppress tau. We don't want to silence tau completely anyway. We knew that. And it looks like the lower doses might be just fine.