Thanks very much. It's my pleasure to be moderating this chat with Ryan Watts, Founder and CEO of Denali Therapeutics. I'm sure most folks know the company well, but Denali just provided a great update on a number of programs. So, Ryan, maybe you can kind of go through that and set the stage, and then we can do Q&A.
Yeah, great. Paul, great to be here in New York with you at this conference. It's a very exciting time for Denali. I think, as some of you may know, we founded Denali to do two things. One was to cross the blood-brain barrier and to develop technologies to get medicines in the brain, and the second was to defeat degeneration. So, setting that long-term vision. And I think this last quarter is an example of a lot of progress across our pipeline, including our late-stage programs and the several new programs, including new modalities entering the clinic with regulatory filings. And so, just as a bit of background, our technology is known as the transport vehicle technology, which is part of a broader class of brain shuttles. And the key here is using the natural mechanisms at the blood-brain barrier to get basically molecules into the brain.
And in our case, at least the lead programs are using transferrin receptor, which I think has now been validated as a path to the central nervous system. We have three franchises. We have the enzyme transport vehicle franchise, where now I'd say our fourth medicine is entering clinical trials in Pompe. The lead program is in Hunter, where we have a BLA that's under review right now. We have a Sanfilippo program, which we plan to file for accelerated approval. That's important. And then we have a progranulin program, which we put categorically in our enzyme franchise because it's a protein replacement therapy. The approach is very similar, and that's for FTD-GRN. And as I mentioned before, the fourth one is in Pompe. I think very excitingly, we have our first oligonucleotide molecule now entering the clinic.
I know, Paul, you and I have been discussing this for four years. I believe, although I can't say definitively, that we're the first ones to show that you can get an oligonucleotide across the blood-brain barrier using transferrin receptor, using the transport vehicle. This is a really exciting class of medicines. Now, I wish we were the only ones doing it. It became highly competitive within the last couple of years, others using their technologies. Our first program is targeting MAPT, which is the gene that codes for tau. We're excited to have filed that program. I'll just comment that we're close to filing our second Alzheimer's program for amyloid beta.
So, I think ultimately you see a span between rare diseases like Hunter and common diseases like Alzheimer's, but the uniform relationship is using the transport vehicle technology to get distribution throughout the whole body and the brain. I maybe make one last point. What's unique about the Pompe program is that it's muscle targeting in addition to brain targeting. So, in terms of Pompe, there is severe Pompe or infantile onset, but there's also LOPD, which is basically the disease that really focuses mainly on muscle. And what you see with these transferrin receptor-enabled technologies is also improved distribution to muscle, bone, and brain.
Great. Sounds good. So, let's talk about 310 briefly. Any updates you can give on the regulatory review? And there was this major amendment recently. Anything to that kind of suggests the review is going in a different direction, or do you feel like it's more or less administrative?
Yeah, I think we were very surprised to get the major amendment based on a clerical error around the molecular weight of TIVI, but it was done in our experience a very collaborative way with the FDA. They've been engaged from the beginning since we got priority review. Obviously, that shifts the PDUFA date from January 5th to April 5th. I think we immediately identified where this clerical error originated from and how it affects specifically just one portion of our BLA, which is the population PK calculation based on this theoretical molecular weight. That's very easy to correct. That being said, we've continued very positive engagement with the FDA, including beginning label negotiations and finished the late cycle meetings. So, we see a lot of momentum with the FDA regardless of this shift in timing.
I think the way we're viewing it right now is that we're planning to April 5th as our PDUFA date, but we're prepared.
But you're in the middle of discussions.
That's right.
That would suggest things could actually close out a lot sooner.
That's possible. I think so. And I think if I look at the overall engagement from the beginning of the review, the FDA has been highly engaged, very relevant questions. And so, we've been impressed with that. It's the same team we've worked with now for multiple years in the rare disease and medical genetics. It's the same team we work with on the START program for Sanfilippo and now for Pompe. So, that consistency has been important for us. And obviously, we were disappointed. And I think technically the FDA has that ability to delay the review based on having to recalculate the population PK. I should just comment that that theoretical molecular weight shift doesn't change anything in terms of the conclusions of the data package, which I think is really important.
Right, right. Okay. Makes sense. So, maybe just on the launch preparation side, how are you building out that infrastructure in the U.S.? And I've asked you this many times, but what % of the ex-U.S. market is unlocked by accelerated approval in the U.S.?
Yeah. So, I think I'll frame a couple of sort of asking two questions there. So, in terms of launch preparedness, we have a small but mighty team, low double digits, ready to launch in terms of the field team, all in place, ready to go. That's the same team that will launch not only Hunter, but Sanfilippo. And then obviously, by the time we get to the point of launching in Pompe, we'll expand that team. But that's the value of having a franchise like the enzyme franchise. In fact, all the physicians, treating physicians are the same across Hunter and Sanfilippo. In terms of the US ex-US markets, the way we think about it today is a third, a third, a third.
So, one third of the market is the U.S., and then a third is a portion of the market that we think we can access with the accelerated approval data package, okay, rather than the COMPASS data package. That will represent another third. And then the final third is likely Europe generally, where it may require the clinical endpoints. Although we have robust clinical data, it's not active comparator controlled, which is essentially what.
But it sounds like areas like Brazil and South America.
That's exactly right. MENA, other areas that we're looking at, U.K., Japan, these are areas where we will take the data that we have and then advance. So, rather than being like two binary events, it's going to be a stream of launch events now with US accelerated approval, hopefully.
Right. Okay, okay. And can you update us on Sanfilippo data you've generated to date and how kind of close you are to figuring out what you're going to need to file there?
Yeah. So, I think exciting, again, I mentioned this quarter a lot of advances in the portfolio, so we completed enrollment of the phase 1/2 portion of the study that's required for accelerated approval, so that's roughly 20 patients. And so, as perspective, when we enrolled the Tivi trial, we have 47 patients in that filing package, and so our goal is to have learned from that, be able to do things ideally half the size, twice as fast, and certainly in the case of Sanfilippo, we're looking at 20 patients. That enrolled in September, that's 49-week endpoints, so 49 weeks from September, that data package will be complete, and then we'll start preparing the filing package for the BLA.
What's a little unique about Sanfilippo or this particular program is that unlike the Hunter program where we're manufacturing out of Lonza, we're going to be manufacturing ourselves the Sanfilippo program. So, we have our own site. We've now onshored most of our manufacturing, but we've begun manufacturing that particular program. So, that means we'll also do the commercial manufacturing for that. I think that will end up determining our timeline for BLA filing. We just want to be really ready to go because it will be the first time launching a product out of our own facility.
Yep, yep. And any clinical data you can speak to there?
Yeah. So, what we had mentioned before, we had top line that we had 24 weeks of data. The first, I think roughly eight patients where we had robust biomarker effects, and we've now seen up to 49 weeks of data. We plan to have our first actual data presentation, an interim look. Obviously, it won't be the entire data package that's in the filing that will be in the filing because that actually comes out in September. But this upcoming WORLD, we plan to present that data at that medical conference.
Okay, okay. And then I guess for each of these indications, obviously with 310, you're well underway, right, with like a confirmatory trial. And from my seat, right, it's like the biomarker data proves 100% that the drug works, but there's still just the risk of going into a placebo-controlled study, like not fully understanding placebo effects, like variability. So, maybe start with 310 and just kind of your confidence that you've designed the pivotal study conservatively and then how that's going to follow through to Sanfilippo.
Yeah. Well, I appreciate that you feel the way we feel, which is the data is very compelling that the drug works.
The heparan sulfate and neurofilament data are like unequivocal. It's just neurocognition, how do you, yeah, I'm sure you guys have thought tons of things.
Yeah. And I think what's been nice for us and I think very valuable is seeing the neurocognitive data, the behavioral data, the hearing data mature in our own data set. And I think it's becoming very clear, unequivocal that we have a medicine that has really robust pharmacological activity and efficacy. Now the question is, how do you design the trial? And it's actually interesting because we're putting this in the context of Hunter, but it's also important to think about it from Sanfilippo and from regulators' perspective there. So, in the case of Hunter, this is an active comparator control. So, we're comparing to idursulfase. Now, interestingly, the vast majority of our patients in the phase 1/2, we're on idursulfase. So, they're switching.
We look at hearing improvement after, let's say, four or five years of idursulfase and hearing's declining, then they go on to TIVI and we're seeing an improvement in hearing as an example. Same with cognition and behavior. I think as that data matures, it becomes more and more clear. What we're looking at in the Compass trial is a two-to-one randomization. The co-primary endpoint is CSF heparan sulfate. We know where that's going to play out. The other is basically the VABS, which is the Vineland, which is a behavioral scale. There, if you look at our phase 1/2 data, basically all patients either stabilize or improve, which is very different than what you see in the natural history data. I think we're confident in that. Now, it's always about powering and scale.
These are rare diseases, so it's hard to enroll 200, 300 patient trials. This is 42 patients in cohort A and the neuropathic. That being said, it's a two-year endpoint. That trial will complete enrollment essentially at the end of this year. You're looking at 2027 to complete that trial. Now, in Sanfilippo, it's a different story. Their regulators, both in the U.S. and in Europe, are very open to using natural history as your comparator, so rather than having to do a placebo control.
Because there's no standard of care.
That's exactly right. There's no, so it's really, I think in the case of Hunter, the argument was you have Elaprase that's providing some clinical benefit. So, that's fine. But in the case of Sanfilippo, there's no standard of care. And so, it's likely, although we're finalizing that now, this will be part of our accelerated approval sort of package for Sanfilippo that will use natural history as the comparator and will pre-specify what we expect to see in terms of benefit across endpoints for these patients. Again, similar endpoints, behavior, cognition that we're testing in Hunter.
Yep. Okay, okay. And to round out the enzyme conversation, and then I want to talk about Alzheimer's, but on the Pompe program, so like obviously logical application with the transferrin platform, the one question I have is like, do we have good evidence that the existing drugs for Pompe are inadequately clearing toxic substrates like within the muscle? Because I feel like usually the biomarkers in these indications are not actually like in the proximal tissues.
Yeah, that's a great question. So, I think the answer is there's definitely still an unmet need, both in sort of categorically two ways. One is an IOPD, anything neuro related, there's definitely an unmet need. And so, that is something that could be addressed with ETV: GAA, or DNL95.
Is that a big component of Pompe?
It's a small, very small component, and it would be like the obvious place for some additional proof of concept. But actually, we didn't develop ETV:GAA for that. We developed it for the entire population, and there we also believe there's an unmet need around muscle. But I think what you do is you are looking at distal biomarkers initially to show activity, and then you look at proximal endpoints. And there it's really around muscle strength, clinical endpoints, right? So, I would say stay tuned. We have an analyst day on December 4th. We're going to talk more about our strategy for Pompe. At least it'll be the first glimpse looking at the strategy. We'll also talk, answer your questions around the addressable markets, US, ex-US broadly for Hunter. But that'll be a great time for us to talk about Pompe.
I think notably, as we have key opinion leaders who work in this area and physicians, they work in the Hunter, Sanfilippo, and Pompe area and very much understand where the unmet need is. Therefore we tailor our trials addressing that.
Yep, yep. Okay, okay. Anything you want to kind of close out this ERT discussion or enzyme replacement discussion with, with FTD and maybe one, where things stand with your program and two, like any learnings or implications from the Alector failure?
Yeah, so I think we've now correctly put ATV progranulin into our enzyme franchise because the mechanism is essentially protein.
It's a lysosomal mechanism.
It's a lysosomal enzyme replacement therapy. And I think what we've learned from, I mean, a couple of things we've learned from Alector. At the time that we first started our trial, it was actually difficult to enroll because there were others like Prevail and Alector who were enrolling. And our enrollment has very much accelerated as those programs have stopped enrolling. So, B2 enrollment is complete. We're now enrolling very rapidly cohort B3. We'll talk about that clinical trial design, I think for the first time at this analyst day as well. And I think what to think about is that this is an entirely different mechanism than blocking a natural receptor. So, the approach that Alector took was let's block sortilin, which is one of the natural receptors, and then you get an increase in progranulin measured extracellularly.
You can measure it in blood, CSF, wherever you want to measure it. In our case, we're replacing progranulin because these are hemizygous patients. They have a loss of at least one copy, and basically what we can do is replace the protein that's missing, and I think the best example to look at is our Hunter program or the Sanfilippo program. Our focus will be on biomarkers that are both proximal and distal to progranulin, so the proximal biomarkers would be like glucosylsphingosine and anything related to lysosomal function, and distal would be like GFAP and NfL, and I think NfL would be, obviously that would be ultimately the goal.
Yeah, and those patients have very high NfL levels, right?
That's right. They have elevated NfL levels.
Okay. Got it. Great. Switching gears on the OTV side, maybe walk through with Tao just the preclinical data you've generated. I guess I think the key question is really that with delivering an oligo to the brain via the transferrin receptor, do you get the same kind of biodistribution and knockdown in the key areas that are going to matter for Alzheimer's?
Yeah. I think that's a great, great question. And I would say that what implied in that question is today the standard of care for oligonucleotide delivery to the brain is intrathecal delivery, right? So, you inject it into the spinal cord and you have, and we show this I think pretty clearly in others as well, you have an uneven distribution. You get very high concentrations of oligonucleotide in the spinal cord and you get decent concentrations in regions that are adjacent to cerebrospinal fluid, but not so much in deeper brain regions. And so, what the transport vehicle enables, and this I think is the key point, is an even distribution throughout the central nervous system.
And in a disease like Alzheimer's, which affects many regions of the brain, not the least of which is the hippocampus and then obviously cortical regions, we would expect to see a robust biodistribution. And that's the advantage of the transport vehicle, which is much more challenging with intrathecal delivery. So, the preclinical data that we have is that, and this is why it took us a bit of time, we had to do further engineering on the oligonucleotides. We've since expanded and are using all types of oligonucleotides, si, ASO. And the reality is any type of oligonucleotide we can transport across the blood-brain barrier. But this particular program, we specifically engineered an ASO, took additional time to really fortify the chemistry around that particular molecule. And what we see now is a very sustained reduction of MAPT, which then is a sustained reduction of tau.
When you think about it from a clinical perspective, we're going directly into patients and you're looking at two things. CSF reduction in tau protein, which takes some time because tau has a decently long half-life. And then the second is ultimately tau PET imaging where you'd want to see a reduction in the PET signal.
Yeah, and are these both going to be endpoints of the phase one?
They will, definitely. Yeah.
Okay. How long? Maybe just walk through that design. How long are you treating or following patients, and what's your expectation for like the time points at which you should see these PD markers move?
Yeah. So, I think here we're lucky. We're not flying blind. I think there's been intrathecal delivery of ASOs for MAPT. So, you can get an idea at least what would be today's like expected kinetics based on that data. And I think we would model it based on that in terms of timing. This will also be something we'll lay out at our analyst day on December 4th. But in terms of we'll need a certain dose escalation and then from there we'll have to expect a certain duration of treatment before you see a reduction both in Tau protein and in Tau PET.
Your point on the biodistribution with intrathecal, I guess I thought the Tau PET data for BIIB080 is like pretty good. Do you feel like that drug is not fully testing this Tau hypothesis or do you think it's probably good enough that we're going to get like a real idea?
Yeah. I thought it was really good as well. Like pretty good or really good. It's always hard with PET because I don't know exactly what I should believe is baseline or not. But in general, it taught us something really interesting, which is it wasn't entirely clear if neurofibrillary tangles could.
can be cleared by the brain if you knocked down an enzyme.
That's right, and so I think that's actually really interesting, and so that should be part of the standard. In terms of distribution, there was a recent paper published on some of the anti-amyloid antibodies, and these are some post-mortem analyses, and what was really fascinating is that there is an uneven removal of amyloid plaque. Now, this is a systemic delivery with an antibody, but my guess is that with intrathecal, it's going to be the same thing we see that we showed in our non-human primates, which is an uneven distribution of the oligonucleotides, so will there be enough distribution with intrathecal to really test the hypothesis? I hope so, and then you could replace it with a systemic IV or subQ therapy.
Yep, yep. Okay. Maybe taking a step back, like how should a Denali investor think about just like safety risk in general for transferrin targeting? Like how many transferrin clinical programs do you think there have been total?
So, there's.
25 in the industry?
Yeah, there's at least over a dozen who have like longer-term data now, right? And I think that's important.
It's like maybe, I don't know, maybe 10 have had some on-target safety then, maybe more. If you just count like mild anemia or like the stroke or the vitiligo or the Trontinemab, like I guess like what I'm trying to ask is like, I think some investors who look at this space think it's got tremendous potential, but there's still some smoke around safety. It's hard for people, I think, to kind of get in the weeds on each program or each platform and really make sense of like who's cleverly designed out some of the safety risk. Obviously, you've been in the space like longer than anybody. So, how do you think about that with your platform and program to program at Denali?
Maybe just step back in time. We were the first ones to observe a hematological effect. Now, the thing that's interesting about the hematological system is that it's evolved to be biologically robust. So, if you lose blood, you very quickly start to retic and you regain red blood cells. It moves very quickly. We were the first to show that transferrin receptor targeting molecules have the ability to reduce reticulocytes and then subsequently cause anemia and also how to engineer around that. So, then in 2014, this was a year before founding Denali, we actually had a paper that basically showed that if your molecule is immune inert, you can basically preserve reticulocytes. Now, it's been very interesting to observe the mechanisms by which reticulocytes are lost. The primary mechanism is when you have full effector function.
This is our strongest argument around differentiation from Trontinemab is that our A beta antibody has an on-off switch. When it's bound to transferrin receptor because of a mutation in the Fc, it's off. You cannot bind Fc gamma. It's called LALA. It's basically based on the architecture of the Fc So, in other words, our antibody will bind to transferrin receptor on reticulocytes or immature reticulocytes and will not engage the immune system. However, when it's bound to plaque, it then can engage the immune system and then microglia can engulf plaque. And I think what you see with the data is maybe 20% anemia. Now, we saw it with our TREM2 molecule. We reintroduced immune function with TREM2, which made it very robust at depleting reticulocytes.
We still had a window, but we took a more conservative view about not developing that molecule in the Alzheimer's population with that particular risk, right? So, all of the hematological pieces, I think, are monitorable, reversible, easy to understand. Like that actually, and maybe easy as an understanding. We've been working on it for a decade, so maybe it feels easy now. But the principles around engineering, you can get there. And I think that's a huge differentiator for our ATV: Abeta program. I think these other points, these one-off safety events, I can't categorically say those are related to transferrin receptor. And I think you point out the Trontinemab example, that is a full effector function antibody, right? It can engage the immune system fully with the brain shuttle.
And then you look at some of these other programs, and if they're related to transferrin receptor, they're giving 100 mg per kg, more of a very high dosage.
We talked about this. How they're describing dose for the payload, not the antibody.
Yeah, exactly. So, I actually don't know if I can even say they're transferrin receptor related or if they're dose related or both or neither. I don't know. But like, and so, I think definitely there's a lot of experience, which there we can figure out how to engineer over time.
Yep, yep. But as we look across your portfolio of clinical candidates, like what are you seeing in terms of any on-target transferrin agents?
Yeah. So, I think the one that we've described before, which again, we don't know if it is on target. This is how complicated it is. So, in Hunter syndrome, there's anemia of chronic disease. So, 20% of patients have anemia at baseline before we even start treating them, right? We see that when we're doing blood draws, but I think importantly, these patients, they retic. So, the retic numbers aren't dropping. We know from what has been previously reported from Trontinemab that they actually deplete reticulocytes, right? So, beyond that, if you look across our portfolio, that's like basically what we see.
Okay. Makes sense. So, as we think about like building the company for the next phase, you may have two clinical candidates in Alzheimer's. How much of this do you envision yourself funding independently versus partnering? Like do you think you could take one all the way? What's the thought process there?
Yeah. I mean, that's the great debate in Alzheimer's, which is it feels inevitable that at some point you'll partner with a bigger partner. Obviously, the further we take them on our own, the more value we can generate. And frankly, probably the faster we can move, at least initially. So, at this point, we're very excited to advance both our MAPT program and our A beta program as far as we can. But I think there'll be a point in time where it makes a lot of sense to seek a partner in an area that could use additional firepower and can use a global footprint. I would love to be able to launch our own Alzheimer's medicine at some point. Now, it may not be these first two, but it could be a second or third generation.
Yep, yep. Makes sense. How is your A beta brain shuttle similar and different to what Aliada had?
Yeah. So, I actually put like Aliada on one end of the spectrum and Trontinemab on the other end of the spectrum.
From an effector function perspective.
It's basically how much do you engage effector function, right? That's the biggest difference. The other is that what's unique about the transport vehicle is because of where we bind transferrin receptor, it has a unique architecture that allows us to make it immune silent or immune active. I referenced that in sort of the safety question that you asked, Paul. So, I think that's the other piece. We have the added advantage of now we have decades' worth of clinical data to pick the A beta binding profile, right? That's important. Pick the right A beta binding profile, have the engineering that's specific to the transferrin receptors.
Are you picking a more lecanemab-like binding profile or donanemab?
It's a great question. I have a preference to be as unbiased as possible. So, the one thing I would be biased against is binding monomeric A beta, right? And so, I don't know.
Which takes away the rest of your.
Then it becomes a question of does pyroglue actually capture everything or would you prefer to capture something more broadly?
The potential.
Yeah. So, maybe the latter would be. I'd be more interested in.
The ProtoFi.
Yeah, a broader profile. Now.
Is that the Crenezumab still in your heart?
I don't know if it's Crenezumab. Not really. I think it's interesting because you have Bapineuzumab, Crenezumab.
Crenezumab was like the oligomer one.
Yeah.
You probably worked on that one, right?
I did. And then Gantenerumab and then Ponezumab and Solanezumab. I think what we've learned is that the molecule needs to remove plaque.
Right.
It likely needs effector function.
Yep.
Ideally, the profile isn't one where you're binding monomers, which would be irrelevant to the efficacy. So, if you can like exclude the monomers and then and I think I don't want to overthink it beyond that. Because we can talk about ADDLs or protofibrils or oligomers. All of those are either fictitious or reality, but they're all stochastic in what is driving toxicity.
Yeah. Okay. One of the coolest things you showed me, and I remember it was at whatever hotel you guys post up at, at J.P. Morgan. And you were showing me the biodistribution data for the A beta brain shuttle and how using transferrin changes the degree to which the antibody will actually pool in certain areas that you hypothesize might be the driver of ARIA. Since we've seen some Trontinemab data that like completely decouples amyloid lowering in ARIA. Can you like speak to this? And I guess to me with Trontinemab in phase three, like that's the biggest question. Like is this safety profile replicable? Because if it is, it completely upends, I think, the category. But what's your thought?
Yeah. So, we published a paper in August in Science around that exact hypothesis. And I think the data is very consistent. It's always hard to prove a hypothesis, but the theory is this: with transferrin receptor, you have an even distribution throughout the brain, but you also have reduced blood concentration. And the reason for that is that when you bind transferrin receptor, you get this rapid distribution through muscle, bone, brain. Like you actually see in a whole animal this tissue distribution. Your blood levels go down. And as a result, you don't accumulate concentrations of antibody in the perivascular space that you do with a standard antibody.
So, if we take a standard anti-amyloid antibody or any antibody and you look at where it is in the brain, it's at very high concentrations in the perivascular space, which is where you have arteries and arterioles. That's where you have cerebral amyloid angiopathy. That's where we believe you initiate ARIA. And so, you have a shift in distribution. So, not only do you have better distribution in the brain parenchyma, you have less concentrated in the perivascular space. And I think it is actually that second point that is underappreciated. The fact that you don't have very high concentrations in areas where you have high cerebral amyloid angiopathy means that you shift the safety profile. And exactly your point. This is actually mind-blowing. Everyone thought plaque reduction would correlate with ARIA.
It makes sense. It's like it's on target.
It doesn't. In fact, Trontinemab is three times better plaque reduction, three times faster, one-fifth the dose, and substantially less ARIA, right? Then we also have another case where you activate an immune receptor and you have ARIA and no plaque reduction, right? That's in the case of TREM2. So, I think we're realizing that part of it could be geography, where you're localized, or biodistribution, if I'm using the biological term. The other could be the mechanisms of clearance. So, I think for us, that's a huge and I think that combined with a better sort of, let's call it a reticulocyte profile, that's where we think the ATV Abeta will have sort of the best in class. So, I think you asked the question of Trontinemab versus the Aliada approach.
It's like if you're completely effectorless, you will most certainly have less ARIA, but you may have less plaque reduction. We tried that with Crenezumab. So, Crenezumab was an IgG4.
Yep, yep. Okay. We're out of time, man, but that was great. Thank you. Appreciate it.
Thanks, Paul.
Always a pleasure. Thanks.