Denali Therapeutics Inc. (DNLI) Earnings Call Transcript & Summary

Good afternoon. Thanks, everyone, for joining. Salveen Richter, biotechnology analyst at Goldman Sachs, and really pleased to have the Denali team with us. So with us, we have Ryan Watts, who's the CEO, Co-Founder and President; and Alexander Schuth, who is the COO and CFO and Co-Founder as well. Thank you both for joining us.

To start here, maybe just an outlook question on the company. You've made steady progress with your pipeline here, notably with the positive data that we've seen from the blood-brain barrier with that first data in Hunter's syndrome. And then we've also seen Biogen opt-in recently to the ATV data program. So as we look to the second half of the year, what do you see as the most important steps to further advance these different verticals and programs? And where should investors be most focused?

Salveen, thanks for having us here. It's great to be in the new venue and glad to be with you. I don't think there's ever such an exciting time to be in neuro drug discovery and drug development as now. So much has changed in the last even like 18 months in Alzheimer's and ALS. And when we founded Denali, we felt like the time was right and now we're realizing that across many programs, both at Denali and externally. And so this year for us has been a year of execution. We released data at WORLD. We recently had data at AAN. We're going to go through some of the data. But you're really asking a specific question is, what do we expect for the rest of the year? And probably the 2 near-term data readouts to sort of draw attention to is we'll be presenting at AAIC, our PTV progranulin program, and then again, providing another important data cut for SSIEM, which is for our ATV IDS program for Hunter program. So both of those are transport vehicle-enabled programs. And if you kind of step back and look at Denali, we have focused on small molecules as well as large molecules, all of which are engineered across the blood-brain barrier. But I think a lot of attention is focused in on our transport vehicle technology, but we continue to make steady progress on the small molecule programs as well. And so some near-term data, but more broadly, we now have 7 clinical stage programs and entered into late-stage development last year on a number of programs.

And you've also partnered with companies. You're partnered with Biogen and Sanofi, and given the breadth of the TV platform here, do you anticipate collaborating with other biopharmas on the Ford? Is there a certain point at which you would consider it, like an inflection point? And how do you decide what to keep versus what to partner at? And I know I always ask you this question, but it seems like it's such an evolving space.

I'll hand it to Alex for that.

Yes. And Salveen, you're exactly right. So partnerships, collaboration have been a core part of our strategy, and we're very pleased with the 3 partnerships that we have by Gen, Sanofi and Takeda. Through those partnerships, we're able to operationalize this very large clinical portfolio, 6 molecules in the clinical indications. With respect to new partnerships, there are more opportunities, and we always consider those. Right now, the focus is very much on execution. Execution on the clinical portfolio, execution on the expansion of the TV platform, especially with respect to OTVs. And also, I will say it is important for us that we do have wholly owned assets. Ultimately, the goal is to have a portfolio of commercial stage programs, be the fully integrated company with wholly owned programs. So yes, more opportunities, there are, but we'll keep a high bar and look at the right time and the right partner.

To start, can you describe for us what makes your TV platform different? There are other approaches out there that are being taken by different companies. So maybe compare and contrast why this stands out and why we've seen the data that we've seen so far.

Yes. Great. So just to step back a little bit, the transport vehicle technology takes advantage of a natural transport at the blood-brain barrier. So many of you may know that the vasculature in the brain evolved in such a way to protect the brain from toxins, from unwanted substances crossing the blood-brain barrier. And so it was actually proposed in the late 1980s that one could cross the blood brain barrier using natural transport mechanisms. For example, glucose or iron or essential amino acid you have to get from your diet. All of these have mechanisms to get these molecules across the blood-brain barrier. And so when we founded Denali 8 years ago, the idea was to invent a highly modular platform that allow us to get enzyme antibodies and more recently, actually antisense oligos across the blood-brain barrier using this technology. We had the advantage that there had been several decades of work on blood-brain barrier, but not really a lot of progress where there was clinical data that had really substantiated that transferrin receptor, which is the first target that we're working on, an iron transport that could actually be used to get molecules into the brain. Now it's a highly competitive space where there's a lot of companies that have programs targeting transferrin receptor or other BBB crossing technologies. And so what we decided to do is invent a technology that takes advantage of the natural Fc portion of an antibody. And that Fc portion has fantastic pharmacokinetics. And if we can build in binding to a blood-brain barrier transport or transfer receptor, we could then apply this to antibodies and enzymes and other modalities. And that started roughly 8 years ago. And I think we're very excited in 2020 where we saw for the first time clinical data that not only did we have an effect on the biomarker, we normalized that biomarker. We had a 90% reduction in this key biomarker, in this case, in Hunter's syndrome. And I think that gave us the unique insight that there's a huge capacity for transport of iron in the blood-brain barrier. And if we can hitch a ride without disrupting the actual natural transport, we would be able to take advantage of this big capacity. I think in retrospect, as we now look at our data, we have now 25 plus patients, 27, 28 patients now over 2 years, what we noticed is that actually, there's a relatively minimal dose that's needed to normalize this particular substrate. And that wasn't really predicted from animal models, and I think it reflects the fact that as you scale to humans, you do have this capacity. Now anyone that's working on transferrin receptor could take advantage of this. What's unique about the transport vehicle technology is that we can tune the affinity and we can use it for different modalities. That's not the case for standard antibodies, like standard antibodies with 2 arms that bind to the target. It's much more difficult to do that. And we now are applying this across multiple modalities with 3 clinical stage programs, soon a fourth clinical stage program using this technology. So I think that's a little bit of a background, and we're very excited about the application of this technology across multiple modalities.

Great. And when we look at some of these programs here, Biogen recently opted into your ATV beta program. What should we take note of here? And in terms of, I guess, how to think about this working and then the combinatorial situation that maybe could play out with other approaches, be it TREM2.

Yes. I will take this and Alex, you can comment as well. But I think that the entire landscape of Alzheimer's has completely changed in the last year. We all know it. And pretty soon, I think standard of care is going to be A-beta antibodies that can reduce plaque. Now if you actually look at all the data, what you see is that you either have to have a relatively high dose or frequent dosing, so every other week at 10 mg per kg or 10 mg per kg once a month and you see this plaque reduction. What the transport vehicle technology will allow us to do is likely to go with a lower dose and maximize what's called the Cmax, so you get this pulse of drug in the brain. I think that, that our technology may allow us then to think of different dosing regimens. Maybe we could enable subcu dosing. We wouldn't need a large dose or as frequent of dosing to be able to robustly reduce plaque. You asked a specific question, which is about what about other approaches now? And you have to step back and rethink the combination. And I think that's what we're doing. And we don't actually have an answer for that right now. We don't know what it will mean to combine A-beta with TREM2 or with RIP kinase, but we have to sort of hit a reset and get the preclinical and ultimately clinical rationale to think of those combinations. I think that's the future. And in fact, really, the future is measuring amyloid plaque before you have cognitive deficits and removing that amyloid plaque as a protective mechanism. And so we're excited to enable the next generation of A-beta antibodies, and we're trying to figure out what to do with targets like TREM2 or RIPK, which are going after maybe the, let's say, inflammatory or microglial response, and we don't have an answer for that yet.

And would your ATV or A antibody or Abeta TV asset, would that be LEQEMBI-based? Or is that...

I'll hand that to Alex around the partnership with Biogen and ATV beta.

Right. So as you mentioned, Salveen, Biogen exercised the option to in-license the ATV A-beta molecule, and we had granted that option to Biogen back in 2020 as part of a broader deal, which also included LRRK2. We cannot comment on which fab is being used here, but we'll just say that the clinical development and the execution is fully in Biogen's hands.

You have the -- you also have the TREM2 program in Alzheimer's disease, and we're expecting data from the Phase I by year-end. Help us understand how you think about this target, firstly? And then just what you would expect to see from the safety profile as we go forward, but then what we should be looking for on the efficacy.

Right, so I think I just mentioned the new complexity of combinations with -- Abeta, but let's talk specifically about TREM2. So TREM2 was identified as a genetic risk factor in Alzheimer's disease in a subpopulation, but even more broadly. So in other words, loss of TREM2 function seems to increase risk of Alzheimer's disease. So the idea is basically to enhance TREM2 function. And so if you want insight into how we're approaching this, we published a paper in Nature Neuroscience at the beginning of the year that really went into detail how combining TFR with TREM2 can robustly activate the TREM2 pathway. And what it does is it basically specifically activates TREM2 signaling, which shifts the state of microglia from basically a disease-associated state, or to, I would just say, like a homeostatic state to actually an active state, we see proliferation. It's actually pretty wild. We see formation of new microglial cells in adult animals. And these microglia would be much more responsive, basically. We also see an increase in the metabolic state. So this ability to turn over like cholesterol esters, which accumulate in Alzheimer's disease. And so the actual hypothesis is that gain of function should basically be protective in Alzheimer's disease. I think the unanswered question, which I already highlighted is we don't know what that's going to do when we combine it with A-beta. So that program of plan is to share data by end of year in healthy volunteers. And as I've commented before, it's very potent when you combine TFR with TREM2.

How is your asset differentiated from some of the other ones in development right now?

Yes. So I think the -- if you actually look at the dose that's required to activate the pathway for standard antibodies, it's really high, 30 to 60 mg per kg. And at least preclinically, what we can see is that at much lower dose, we can activate TREM2. We also see this shift in the microglial state that we actually don't see with standard antibodies. And so it's, again, a very robust activation of the pathway.

Great. Your Hunter's program here, so we recently saw data at the WORLD Symposium. Can you just walk us through kind of where your understanding lies today with this asset, next steps here for the regulatory path, and whether there is the ability to use kind of a surrogate endpoint or whether -- given what's playing out on, I guess, the CBR side of everything or whether it's really going to be more functional for you.

Right. I think, again, setting context that a lot has changed in the last, whatever, 12 to 18 months around surrogates as potential approval. You see it in Alzheimer's disease, you see it in ALS. It's kind of interesting because in rare disease, you would think that, that's the first place you would see it. Rare population, huge unmet need and a really linear path. In the case of lysosomal storage disease, you have an enzyme loss of function. You can replace it biochemically. And then you can see that the substrates reduced in the case of, let's say, Hunter disease, it's heparan sulfate, and then that has a linear cascade to improvement of lysosomal function and what we see is improvement of neural network activity and behavior and cognition. So if I step back, what we see with DNL310 is it's the only drug that is normalizing heparin sulfate after short-term dosing. We looked for the first time after 4 weeks, and 4 out of 5 patients had completely normal heparan sulfate. That's essentially -- and that's a relatively low dose. We continue to dose escalate in order to be able to capture all patients, specifically patients that develop antidrug antibodies to Elaprase or to either sulfate, and the idea is that with that higher dose, we can really dose over the standard antidrug antibodies that you see with enzyme replacement therapy. So that was very exciting data. And then since that original piece of data, we now see durability out to 2 years. So imagine your heparan sulfate in brain is normalized and sustained for 2 years. We see improvement in lysosomal function, and then the most recent data is the majority of patients either improve or stabilize on adaptive behavior scale such as the Vinland or in the [ Bailey ] looking at cognition or the global impression of change. And probably most exciting data for the field is that we're seeing patients are having improved hearing measured by auditory brain stem response, but also just actually the number of patients who are no longer, for example, hearing aids. So it's really exciting to see this translate. And these are patients, by the way, who have been on standard of care, which is enzyme replacement therapy, many -- some for many years before they transition on to DNL310. So for me, I think what we're building is this case that heparan sulfate production is leading to clinical benefit, and the question is, how will regulators start to view this. And I think what you're highlighting is that there is a gene therapy program that is basically has recently suggested that they will be filing for accelerated approval based on heparan sulfate. And we think that's fantastic because we normalize heparan sulfate. We don't get a 15% to 60% reduction. We get a 90% reduction in essentially all patients. And we agree that, that should be the biomarker that is sort of predicting clinical benefit. So we continue to engage with regulators. Our Phase III design is based on a number of interactions with the FDA as well as regulators in Europe, and I think there will come a time where, indeed, these biomarkers will be biomarkers for accelerated approval because they will predict clinical benefit, and we're building that case with the data that we're generating now.

And remind me where enrollment stands with your Phase II, III and when we might see data from that program?

So we kicked off the COMPASS study, and just a reminder, this is a 54-patient study where we're comparing against standard of care. This is, again, because of interactions with regulators, the design of the study was pretty carefully crafted to be able to compare to either sulfate. That study began dosing last year. We're enrolling -- we have now 17 sites activated across 9 countries enrolling and dosing in Europe and the U.S. And so it's moving nicely. I think clinicaltrials.gov, is in 2025, it's a 2-year endpoint, again, based on heparan sulfate in CSF as well as adaptive behavior scale, which is the Vinland, and so that's moving along. But again, we continue to generate a lot of data in Phase I/II in parallel with the COMPASS study and continue to engage with FDA, and we hope that it will evolve similar to what we're seeing in the neuro space where these biomarkers are sort of predictive of clinical benefit and you can get accelerated approval.

Remind us about the read-through from this program [indiscernible] you've done to many of your other TV programs and kind of where you stand there with understanding -- kind of just understanding the risk aspect of those verticals?

Yes. So I think this gets to the question of the differentiation across different transferrin receptor platforms. And I think probably the most important point is that architecture matters. The binding affinity, the nature of the interaction with transferrin receptor, and we've done an enormous amount of work on this before Denali and at Denali, and what we show is that monovalent moderate affinity allows for the right exposure. These molecules are effectorless. So if you engage the immune system, there's a risk that basically TFR expressing cells will be attacked by the immune system. And I think that's -- you can avoid that by basically making these effector-less, these molecules. And so there's architecture matters in terms of brain exposure, in terms of dosing, and that's now applied across all of the modalities. And probably the most interesting insight, and this is actually was unexpected is not only can we get enzymes and antibodies across the blood-brain barrier, but we can get antisense oligos across the blood-brain barrier. And so what we learned in Hunter syndrome and in the Hunter mouse model and then ultimately in Hunter patients, we're now translating to additional modalities. And what actually surprised us, and I think we showed this data for the first time about a year or 1.5 years ago is that we can tag an antisense oligo against an antibody that binds the transferrin receptor using the transport vehicle technology, get this ASO across the blood-brain barrier and modulate gene expression. And that particular program, we actually have a paper that people can access on bioarchives, is in revision at a major journal, and I think it's really exciting to see that not only does this work in mouse, but it's working in monkey and the next stage, we'll be applying that to human. And we said it is a broad platform, but I don't -- that's all originated from the work we are doing in Hunter's syndrome is the flagship program. And then we've mentioned TREM2 already in progranulin, but we see a broader platform, including another molecule coming forward for enzyme replacement therapy in Sanfilippo.

What are you most excited about outside of Hunter's with your TV platform?

I think probably just that, that we can open up antisense oligos to the transport vehicle technology. I'm also very excited about building an enzyme replacement therapy franchise and using Hunter as the starting point, but if we can prove the relationship between biomarkers and clinical benefit, we can bring many more enzyme replacement therapies forward and create a franchise. And what's interesting is unlike Alzheimer's disease where the failure rate is extremely high in drug development, it's almost the opposite in enzyme replacement therapy and lysosomal storage diseases where the success rate is very high and we can potentially bring many, many molecules forward. So I think that it's like the base case where we're seeing potential benefit now in patients. So I'm obviously very excited about it. And then the application of that insight into the broader like ASO portfolio.

Great. And you have a ATV program going after Sanfilippo Type A. Maybe help us understand the level of success or outlook for success here, just given what you've seen with Hunter's.

I think it's -- what's interesting about the Sanfilippo data that we have preclinically is it actually is predicting a bigger benefit and a longer duration of pharmacodynamic response. And so we were surprised in Hunter that we were able to normalize heparan sulfate. Based on the preclinical data, our expectation is that SGSH, which we're developing for Sanfilippo, should have a very robust pharmacodynamic response. And I think Hunter is a great example. We're -- and by the way, we're measuring the exact same biomarker, which is heparan sulfate. So this is a big unmet need. Unlike Hunter's syndrome, there isn't a standard of care. And the reason for that is that these patients primarily have neurological deficits and no one has developed an enzyme replacement therapy because it doesn't cross the blood-brain barrier. So unlike comparing against either sulfate, we're just basically going into naive patients with this approach. So we're excited about beginning dosing patients by the end of the year and advancing now our second enzyme replacement therapy program.

And what is the Phase III going to look like there? Have you...

Yes. I think we've learned a lot from Hunter and how to move quickly. Obviously, as soon as we can get CSF data to predict dose, but moving, I think, even quicker into more like pivotal-type studies and kind of transitioning the Phase I/II to Phase II. But beyond that, we're not sharing a lot of data or design. And again, it is competitive. There are others that have BBB crossing platforms that are going after Sanfilippo.

Then with the FTD GRN program here, how is -- I guess how is enrollment continuing in the study? And then what should we -- how should we be thinking about this data that's coming up at AAIC?

Yes. So PTV progranulin is basically taking the full progranulin, using it to the transport vehicle technology. We published a paper in 2021 in CELL that describes the role of progranulin to the lysosomes and how we can rescue -- so it's very similar to the enzyme replacement therapies, right? I think the challenge that we're having with this, I mean is that if you take FTD, which is already relatively rare, it's about 5% to 10% of the population that have the granular mutation. So we start enrolling, we're dosing at FTD granular pace. We're giving no guidance on timing of when we expect the first data, mainly because it's challenging to enroll. There are not a lot of FTD granulin mutation carriers. But our expectation is that the biomarker data will be similar to biomarker we're seeing for enzyme replacement therapies, but there's not such a big signal to noise on some of those lysosomal biomarkers. But I suggest that people take a look at the biomarkers that we published in the CELL paper, and that will give you an insight of what we're looking at.

When do you think there'll be a better understanding about how you can use specific biomarkers in specific diseases, like as you run all these trials, at some point, there's going to be correlations that can be had. So how are you, I guess, internally thinking about that with could Hunter's, in that way, help you not have to run this controlled trial or outcomes trial?

I think you're nailing it. That -- I mean, that's exactly what we imagine is building the case from basically proximal to distal.

So proximal would be your biomarker that the enzyme is targeting or in the case of PTV progranulin. It may be like a few lysosomal biomarkers. And lining that up towards other biomarkers of neurodegeneration, of gliosis and then clinical benefit. Because once you've established that, then you can imagine that you can just do that for disease after disease, including in rare diseases where it's hard to enroll big studies, especially big studies that are either placebo-controlled, which is almost unethical in these rare diseases, but you can use the biomarker to say, "Hey, we are correcting the biochemical deficits. We're correcting the cellular deficits and now we're seeing a correction in neurodegeneration." So that's the goal. And I think Hunter is the flagship program for receptor. And I think it's broadening. It does cover FTD and ALS. And you've already seen that with tofersen and neurofilament has an accelerated approval biomarker in ALS, which if you can have a drug that robustly reduces neurofilament, you can imagine then that you could benefit ALS patients.

You also have the small molecule programs, as you've noted, and the LRRK2 program in Parkinson's disease. You made a recent decision to focus exclusively on the Phase IIb study where recruitment is ongoing. Can you just maybe give us some thoughts about why you decided to refocus on this? And can you provide any details on the recruitment process and how to think about the clinical bar here?

I'm going to hand it to Alex.

Happy to take that. So as you know, back in 2020, we entered into the collaboration with Biogen, and immediately right out of the gates, we started 2 late-stage clinical trials. One trial, which focused on Parkinson's patients with a LRRK2 mutation, and the other in idiopathic Parkinson's disease. Now 2 years later, together with new management at Biogen, we looked at the overall time lines. We look at the overall cost of the program, and we decided to combine those 2 studies. So now the main study is what's called LUMA. That was previously the idiopathic Parkinson's only, which now also will amend the inclusion criteria to allow for the inclusion of Parkinson's patients with a LRRK2 mutation. So what that does is that we will have data in both patient populations much sooner than otherwise. What had previously been estimated and it has always been on clinical trials, it didn't get as much attention, but the LRRK2 positive trial was expected to read out in 2031. So it was a very long trial as designed with 400 patients, so now we have the ability to get to an answer faster.

And I think I'll just add one other point to that is because the neurodegeneration field has evolved so rapidly in the last, again, year to 1.5 years, biomarkers could be the path forward, especially in these like rare subpopulations. So even though it'd be ideal to run a 10-year study, it might not be realistic and/or it would be so challenging that it would be more worthwhile to focus on biomarkers that could predict clinical benefit.

Got it. And then I guess, by merging these populations together, how do you -- I mean, does the trial design change in a certain way or the powering assumptions? I mean, how do they get impacted?

Yes. So the overall trial design remains the same. It's the LUMA trial design. So it's 640 patients. It's 1 year, it's 48 week treatment duration. It has the same endpoints. It's -- the only thing that changes is that now the inclusion criteria also allow for the inclusion of LRRK2 carrier patients. So it should not have any impact on the powering of the study.

The RIPK1 program, maybe just help us understand when you'll move into Alzheimer's disease and then when we could start to see data from the MS and ALS programs.

Yes. I'll take that as well. So RIP kinase inhibitor is in collaboration with Sanofi. There are 2 sets of -- or 2 molecules. There is a central molecule called DNL788. That is currently in 2 Phase II studies in an ALS trial and an MS trial. The ALS trial is called HIMALAYA. It's 260 patients, ALS, FRS, 6-month endpoint and expected to read out next year. The peripheral molecule is also in 2 Phase II studies in cutaneous lupus and in ulcerative colitis and the cutaneous lupus data is expected mid this year. Now the peripheral molecule, DNL758, is completely operationalized by Sanofi. So it's really in their hands of when and how that is disclosed.

With regard to the ulcerative colitis trial here, what type of data will we see when we get it? And what is that clinical bar that you're trying to...

So again, we have -- I mean it's a question for Sanofi and fully in their hand on timing as well in scope.

And then the one other point that you asked, which was around Alzheimer's for RIPK. And I go back to the bar that needs to be -- it's a high bar. And we -- you may remember, historically, we were able to hit 1 of the 2 biomarkers that are really pathway engagement biomarkers for RIPK and Alzheimer's, and we decided to go to the backup molecule. And we've basically been watching these other programs mature. Once you invest in Alzheimer's, it's an enormous investment. You need to be confident in that. And actually, it's interesting now because now we need to ask the question, well, what is RIP kinase inhibition in combination with A-beta antibodies, similar to what we have to ask for TREM2. And so our Alzheimer's strategy is we've essentially are evolving it with these recent, I think, approvals and really how the landscape is going to change.

How confident are you on the some of these biomarker inputs when you're making these portfolio decisions?

Yes. I think we're very confident the more proximal you are. So the closer you are to the target, the more confident you can be that you're seeing the drug effect. It's really the question of that proximal target translating to clinical benefit, but this is light years ahead of where the neuro field has been, where you actually have these biomarkers, if you're confident in your dose, you have the therapeutic index and therefore, you can fundamentally ask is there a clinical benefit? What's evolved is that now those biomarkers such as, let's say just neurofilament as an example, is correlating with clinical benefit across many diseases. And because of that, you can actually predict that there will be a clinical benefit. You don't have to run those massive trials. And I think that there, the confidence is just increasing as disease after disease, we're seeing more examples of these biomarkers translating to benefit.

How do you apply that into the AF2B program given the positive biomarker data?

Yes, that's a great question. So with EIF2B, it's a unique target. It's a translation initiation factor, like a fundamental cell biological pathway. And when cells are in a stressed environment, they activate what's called the integrated stress response and it essentially locks translation. They no longer can translate proteins. This is a big problem. If it's chronically locked, so in other words, if you have chronic integrated stress response or ISR, this is observed in ALS. It's observed in vanish and white matter disease, and there's evidence across other degenerative diseases and more broadly that the ISR pathway is chronically activated and causing ultimately neuronal damage and neuronal cell death. And so what we've shown is that with our biomarkers, they're very proximal. They are the ISR pathway biomarkers, and we can essentially reduce these very robustly with DNL343. Now we're asking the question is, does that translate to clinical benefit? And I think the genetics in ALS is really fascinating because most of the genetic targets are part of these RNA stress granules that form as part of this ISR pathway. So there's genetic evidence, you have now the biomarker showing that we can hit the pathway. And the next thing is to look at downstream biomarkers and clinical benefits. So we kicked off the HELI-platform study, 240 patients, should be able to rapidly enroll, and that study is now actively enrolling and dosing patients. And so we're looking again on that sort of 2025 time frame. So if you think about Denali in the next 2 to 3 years, we have 4 major readouts that's really going to shape Denali's future. The foundation, I believe, is the enzyme transport vehicle and that franchise, but we have these upside potential with 2 ALS readouts and a Parkinson's readout and so it's a very exciting time that these studies are now all off the ground dosing, and we'll have these readouts very soon.

Great. Well, with that, thank you so much, Ryan. Thanks, Alex. Appreciate it.

Great. Thank you.