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

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. Great. I'm wondering if you can start us off with a brief overview of the company's initiatives.

Sure. So we have -- at the very beginning of the year, we put out a little shareholder letter that said there's 3 major things going on this year. One was related to tau. In fact, I called it the year of tau. And 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. And 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 VY-7523, 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. So very similar in terms of mechanism of action to BIIB080 and to the tau silencing ASO that Novartis has. And that's why I call those 2 programs out, except that what we're doing is we're using AAV to express a vectorized siRNA to silence tau throughout the central nervous system. That's tau. The second piece is capsid POC. So after many years, approximately 5 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. And we're entering the clinic for -- so this is the first test in humans now. We have 2 programs that are testing these novel capsids in humans. One is the tau silencing gene therapy program that I already mentioned that we call 1706. We have announced that we have FDA clearance and Canadian regulatory clearance. We're 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. And also in the second half of next year, we expect to share data that says that we're expressing the gene of interest, in this case, the tau silencing siRNA in the second half of next year. So that's the capsid POC. I should also say that our partners at Neurocrine have said that they're entering the clinic with also a blood-brain barrier penetrant capsid derived from our platform. For Friedreich's ataxia, they're entering the clinic this year, and they said they expect to share data on that program next year. So 2 different programs for -- that should provide proof of concept that our novel capsids work in humans. And the third thing and the final thing is we expected -- we said that we would be showing more data on the NeuroShuttle platform. So this is not AAV gene therapy, but it's basically leveraging the receptors that we discovered from our novel capsids. So we reason that if these capsids get into the brain through the blood-brain barrier, they must be leveraging receptors to get there. We've discovered a handful of these receptors. The first one we've talked about publicly is ALPL. And 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. And we have said that we're going to share data. In fact, we have a plan that we have an abstract accepted for a meeting in December, where we're 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. So those are the 3 -- that's a high-level overview of the 3 major, I would say, catalysts for Voyager.

Great. Great. That's super helpful. If we can dive into the first area, which is tau efforts. As you said, there have been a few important readouts this year. Could you comment on BIIB080's finding? That's the siRNA, maybe more related to...

ASO.

Sorry, ASO. More related to your other tau program, like vector tau siRNA program. But maybe we can start there and also talk about some of the antibody data this year.

Yes. So 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 the 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% to 45%, 50% relative to placebo. Also on Mini-Mental Status Exam, 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 percent 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. I mean there are some questions that were -- that came out of that study as well, like why is there what seems like an inverted dose response curve. And they themselves, Biogen themselves say that they're moving into Phase III with the lowest dose, which they think is in the sweet spot, if you will. It's hard to understand fully why the dose response was apparently inverted. I think it could -- I think it relates more to off-target effects, perhaps associated with intrathecal ASO because we've 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's 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, which obviously is not great for cognitive measurements in Alzheimer's patients. So -- 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, I wonder if that's due to some of the intrathecal ASO issues. But anyway, so 2 different diseases where tau silencing -- and by the way, Novartis is also entering Phase III. In fact, I believe they're already enrolling patients in Phase III with their tau silencing ASO. So it's 2 separate companies moving into Phase III with their tau silencing approaches. I think that means that -- and then there's a lot of other programs, I can tell you in tau that -- so I think it's a pretty exciting time for tau as a target.

Got it. Got it. I mean in terms of that confusional state AE, in your opinion, is that on target or not on target? Or could you...

Yes, I don't think it's on target because if you look at, for example, tau lowering and BIIB080, all 3 doses lowered tau, roughly the same levels. At the lowest level -- at the lowest dose, it was 50% lowering, at the highest level, 60%. If you look at tau PET imaging too, not much separate. So 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. Yes, that was what looked like an inverted dose response. That's one of the reasons why I don't think it's on target. And then yes. So 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. And they speculate in those published papers that it's due to either effects on ion channels or on maybe neurotransmitter receptors. I mean we saw it in the old days with SPINRAZA. You get temporary hind limb paralysis in the animals. And so you put these highly negatively charged antisense oligonucleotides in at high doses, 100 milligrams plus into the small space of the lumbar cistern and you get this steep gradient, it's not surprising perhaps that you might get these. And then the acute confusional state began a week after injection and lasted for about a week, at least what was apparent to the patients lasted for about a week. So anyway.

Got it. Got it. Very helpful. And that's also -- so 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?

Not demonstrated. We don't have any drugs that are approved yet. So I think the preliminary data suggests that it's very promising. I'm old-fashioned. Once it's approved, then I'm more comfortable using words like demonstrated.

Right. Yes, sorry about using that too early. Yes. So that's exactly what I meant in terms of read-through to tau program, right? Let's now talk about the antibody approach. I think J&J reported data this year.

That was another readout, and that was a negative readout. The antibody against the mid-domain did not work. And 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. So we would have predicted that, that antibody probably would not block the spread of tau. And it had like, I think, a 10% effect or so. But it was much less than bepranemab, put it that way.

Got it. Great. So let's talk about your upcoming data from your antibody program. So can you set us up in terms of the study design, what data we're going to see? And what's the bar for success?

Yes. So this is a multiple ascending dose study where we're going to show data from the third cohort. And 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. And we know that with the doses that we're giving in this multiple ascending dose study, we are achieving exposures in the brain that should block the spread of tau. And as I said, we relied heavily on this animal model where these are transgenic mice that express human tau, P301S tau 301S tau. And 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're saying is what we're 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 had been 4 for 4 in predicting the human results. So the 2 N-terminal antibodies, the one from Biogen and the one from Lilly fails 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, positive predictive value, 2 negative. So I'm hoping it's 5 for 5 that was 7523 because in our hands, in that model, we have the best effect, the most robust efficacy on blocking the spread of tau. And that is, by the way, what we're hoping to see in the humans. We want to be as good, at least as good, if not better than bepranemab. Why? Because bepranemab did have an effect on ADAS-Cog on cognition. You'll 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. So we saw that Roche exited that partnership with UCB. But in their last earnings call, I believe USB said -- UCB said that they are moving forward to Phase III with bepranemab. So that's a nice benchmark. We want to be at least as good, if not better, than bepranemab. And so we're going to look carefully at our tau PET imaging data later this year and benchmark it to bepranemab. And if we're not as good as bepranemab, we're going to terminate our program. We don't need anything less good than bepranemab. But if we're at least as good or better, now I'm interested, and we're 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'll 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. So we have some options with our anti-tau antibody. But the first step is to get the readout that we're planning for in Q4.

Right. For bepranemab, the benchmark, is that tau PET data for tau reduction? Is that like 33% and 58%?

Yes, it was in the -- I would say, 40% to 50%. I mean they had 2 separate measures based on what -- it's a composite measure based on a region of interest, right? And so there's this -- what's called the Jack named after Clifford Jack temporal lobe tau measurement. And then there's more of a cortical composite where you look at all the cortical areas. So there's 2 different ways of looking at tau PET imaging. And relative -- so it didn't really decrease tau. It slowed the spread of pathological tau by about 40%, 50% based on those measures. So that's our benchmark, and we want to slow the spread of tau more than that.

I see.

Relative to placebo. And by the way, our multiple ascending dose study does have placebo. Each dose -- each cohort had placebo patients that were randomized as well. So we can -- so we're going to compare to placebo relative to placebo and the benchmark is bepranemab.

Yes. Right, right, right. But what's your thoughts on, let's say, you did reach that benchmark or you exceeded it, right? But how do you think about -- or how should we think about the performance of the benchmark antibody on CDR-SB versus ADAS-Cog, right? It is not...

Yes. No, that's -- and by the way, are we seeing a pattern here that the anti-tau -- the tau targeting approaches have a bigger effect on cognitive measures than functional maybe? I don't know. It's an N of 2. But I would say that that's why we say it has to be at least as good. And I would also say that UCB has produced some tantalizing data that if you start off with people who are either carriers or noncarriers of APOE4 and particularly if you have low tau burden, you may have a bigger effect. Those are post-hoc analyses, and we have to take them with a little bit of caution. But I -- yes, so I think that we want to be at least as good or better on tau PET imaging, and we're going to seek a partner. And as I said, we may shuttle it.

Okay. Got it. Great. Yes, let's talk about VY-1706, vectorized tau siRNA. Can you help us understand the data you generated in NHP so far?

So what we have these very potent BBB penetrating capsids. We're very excited about the capsid we're using here a Gen 2 VCAP capsid. It's very potent. We do not want to be anywhere near 1E14 VG per kg because that's where -- that's typically where people have safety issues, right, the Sarepta and other drugs. So we want to be in the E13 dose range. And the maximum dose that we're going to be testing is 5E13 vg per kg. And with that, we have shown that we can get up to 60%, 70% knockdown across the brain at the highest dose. But BIIB080 has shown us that maybe you don't need more than 50% knockdown. So we may not need to go to that highest dose. So this will be a study where we do a onetime intravenously delivered AAV vectorized siRNA against tau. And we're going to be looking at CSF tau as a measurement, do we actually -- so first thing is safety. We're going to say, in Q1, we're going to be able to say whether or not the acute safety is there. And when you have problems with capsids, it generally occurs early. And so Q1 acute safety. Later in the year, we hope to show data on spinal fluid tau levels because if we're seeing a decrease in tau, CSF tau, we must be getting expression of the vectorized siRNA. So by the end of next year, we should be able to say that it's safe and that we're 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. And there, we want to benchmark it against BIIB080, which we just talked about. So that's how I see that program. And it's going to be an ascending -- now with gene therapy, you have to -- 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. And so yes, 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 detargets the liver 30-fold relative to AAV9. So in addition to the lower dose, we're using the capsid that basically detargets the liver. So that's why we're pretty excited about the capsid because it's very potent, we can use lower doses and because it detargets the liver. And we have the biomarkers. We have CSF tau, we have tau PET imaging that we -- 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 that whether or not we've lowered tau, we can benchmark it against the BIIB080 too. So -- 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. And 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. I see. The acute safety -- so that's the first -- of course, the first milestone, right? For that milestone, what is the signal that you will be monitoring?

Well, we're going to be carefully monitoring the patients. We have a very. With the FDA, we designed a study where we're going to be dosing a patient. We're going to be observing very carefully. And then we're going to dose the next patient after a careful. And then after all 3 patients are dosed in the first cohort, we're going to have an external safety monitoring committee that will say, yes, you can go to the next cohort. So that's a pretty -- 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. Got it.

And we'll be looking at all -- look, I mean, when you look at AAV, I mentioned liver, right? The common toxicities associated with AAV are liver, something called TMA, thrombotic microangiopathy. And those are the 2 main things. And so we'll get in -- 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.

And -- like how low are you -- is your starting dose going to be?

We haven't said that. And just is telling me I can't say that because I haven't said it yet. Not supposed to.