Denali Therapeutics Inc. (DNLI) Q3 2026 Earnings Report, Transcript and Summary

Good day, and thank you for standing by. Welcome to Denali Therapeutics Third Quarter 2026 Earnings Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded. I would now like to hand the conference over to your first speaker today, Laura Hansen, VP, Investor Relations.

Good afternoon, everyone, and thank you for joining us to discuss Denali Therapeutics Third Quarter 2026 Financial Results and Business Highlights. Earlier today, we issued our press release and filed our quarterly report. The press release and today's presentation are available in the Investor Relations section of our website. Before we begin, on Slide 2, I'd like to remind everyone that today's discussion includes forward-looking statements, including our expectations for fourth quarter AVLAYAH net product revenue and the timing and availability of clinical data and any regulatory submissions. These statements reflect our current expectations and are subject to risks and uncertainties that could cause actual results to differ materially. Please refer to our SEC filings and to the cautionary language in today's release and presentation. Denali undertakes no obligation to update these statements, except as required by law. Joining me today are Ryan Watts, our Chief Executive Officer; Katie Peng, our Chief Commercial Officer; Peter Chin, our Chief Medical Officer and Head of Development; and Alexander Schuth, our Chief Operating and Financial Officer. Ryan will open with our progress this quarter. Katie will cover the AVLAYAH launch. Peter will review the clinical programs, and Alex will discuss our financial results before Ryan opens the Q&A session. Ryan?

Slide 4. Thanks, Laura, and thank you all for joining us today. This was a defining quarter for Denali, and we are excited to share it with you. I want to start where we always do with our purpose: transforming life for people living with serious disease. We are focused on making a real difference for children, adults and families by delivering medicines across the blood-brain barrier and into the brain. In recent weeks, we have spent time with Hunter syndrome families, hearing directly what this disease asks of their children and what families hope AVLAYAH can offer, brings our purpose into focus. Their trust is a responsibility we carry into the launch and into every program we advance. Slide 5. It starts with AVLAYAH, but it is bigger than one medicine. For the Hunter syndrome community, AVLAYAH is the first new FDA-approved therapy in nearly 20 years. It is the first FDA-approved medicine of its kind designed to cross the blood-brain barrier through the transferrin receptor and the first to come from our TransportVehicle platform. That is a milestone for patients and a milestone for our field. AVLAYAH gives Denali something we have worked toward for a long time, a commercial product that shows what our TransportVehicle platform can do. That same delivery capability underpins a broad pipeline across rare and common diseases. Our ambition now is to repeat what we have achieved in Hunter syndrome program after program. Slide 6. You can see the breadth of that pipeline here. On the left are rare genetic diseases, including Hunter syndrome, Sanfilippo syndrome type A, FTD granulin and Pompe disease. Together, we estimate these diseases affect more than 30,000 people worldwide. On the right are Alzheimer's and Parkinson's disease, which we estimate affects more than 40 million. Different modalities, different diseases, one shared advantage: our ability to deliver medicines across the blood-brain barrier. Today, we will focus on our most advanced enzyme replacement therapies and our Alzheimer's disease programs. Slide 7. Three things defined this quarter and each one moves us forward. First, AVLAYAH delivered $17.8 million in its second full commercial quarter, well ahead of our expectations. Katie will explain the adoption and access behind this result. Second, the Phase I/II study of zafinofusp alfa in Sanfilippo syndrome type A has completed the prespecified treatment duration for the primary endpoint. We are analyzing the data now and working toward a BLA submission in 2027 as we have planned. Third, both Alzheimer's programs are in the clinic with initial data expected in 2027. Together with readouts in Sanfilippo, FTD granulin and Pompe disease, those studies make next year a landmark year for the breadth of our platform. Katie, over to you.

Thank you, Ryan. We are now 2 full quarters into the AVLAYAH launch. Patient interest, physician engagement and access are all moving faster than we expected. This has accelerated revenue growth. Let me walk you through what is driving that momentum and why we are so excited about the opportunity ahead. Slide 9. Families, physicians and payers each have a role in getting a patient on to treatment. This quarter, we made substantial progress with all 3. AVLAYAH generated $17.8 million in net product revenue, up almost fivefold from $3.6 million in our first full quarter and well above the $10 million to $12 million we guided to in August. It is fantastic to hear from families and physicians who tell us that AVLAYAH is making a difference. We are also seeing payer recognition of AVLAYAH's value, which has translated into an increase from 50% to 70% of commercial lives covered since last quarter. Looking ahead, we expect $30 million to $32 million in the fourth quarter, which would be more than 65% growth over this quarter. Behind those numbers is our priority: helping more eligible families reach treatment. Slide 10. Let me start with families. We estimate about 500 people living with Hunter syndrome in the United States, of whom roughly 375 are eligible under the current pediatric label. Interest is broad from families of newly diagnosed children and those considering a transition from standard enzyme replacement therapy. One parent and advocate, Jeanette, described what she has seen in her son: I've seen meaningful changes in our everyday life, including more expressive language. We have heard similar stories from multiple families about their positive experience with AVLAYAH. This reflects why families are seeking a therapy designed to reach the brain as well as the body. We believe that AVLAYAH is on its way to becoming the standard of care for Hunter syndrome. Our focus remains on supporting families through their treatment experience and ultimately reaching all eligible families. Slide 11. Turning to physicians. We reached approximately 90% of our targeted health care organizations since launch. Before launch, in a survey of 35 physicians, 90% found the AVLAYAH data motivating to prescribe. What we've seen now is action, and treatment centers are leaning in. They are working through prior authorizations, appeals and infusion readiness alongside families, and that commitment reflects the value they see in treating neurologic manifestations. Slide 12. Access is expanding fast. Commercial coverage increased from more than 50% to more than 70% of lives during this quarter. 20 state Medicaid programs now list AVLAYAH as covered, up from 14 last quarter and a permanent J-code took effect on October 1, making billing and reimbursement more straightforward for treatment centers. Having policies in place is an important step, and our team is there for every family along the path from start form to the first infusion and beyond. We work alongside physicians and families at each step. And as those steps become more familiar, we expect that path to get shorter and smoother. Slide 13. Let me tell you why we are so optimistic about what comes next. The launch is ahead of our plan, and we see a long runway for continued growth. Families are engaged across the eligible pediatric population, physicians are activated and working with their treatment centers to move patients towards infusion, payer access continues to expand, and we now have a J-code to simplify reimbursement. Together, that progress took net product revenue from $3.6 million in the first full quarter to nearly $18 million in the second. This step-up was larger than we anticipated. Let me explain why. Two factors contributed to our progress this quarter. First, payer coverage expanded rapidly. As I mentioned, AVLAYAH is now covered for more than 70% of commercial lives and 20 state Medicaid plans now list AVLAYAH as covered. Second, the majority of eligible patients from our Phase I/II study moved commercial treatment faster than anticipated. These patients had already reached their target dose, so they contributed meaningfully to revenue right away. That puts fourth quarter at $30 million to $32 million, and we expect the patient base to keep building from there as new patients start treatment, access processes improve and doses escalate. I also want to put the broader launch dynamics into context as well. We've described our revenue launch trajectory as an S-shaped curve with the steep part coming in 2027. Based on what we're seeing today, that inflection is arriving sooner due to the strong launch execution. Some of the growth we expected later has been pulled forward, including the Phase I/II patient transitions. So the step-up between these 2 quarters shouldn't be read as a run rate. Over the next several quarters, we expect revenue to keep growing at a steadier pace as we reach additional families one by one and help each patient through the steps to treatment. The upside of this patient-by-patient approach is persistence. MPS II is a lifelong condition and AVLAYAH is a chronic therapy. So each patient who starts treatment becomes part of that base that builds on itself. Slide 14. Three drivers support continued U.S. growth: reaching more eligible patients currently on standard therapy, treating newly diagnosed children and reducing the time from start form to infusion. The patient interest, the physician commitment and access progress we are seeing gives us great confidence that AVLAYAH will become the leading treatment for Hunter syndrome in the United States. The capabilities we are building will also power our broader enzyme franchise. Our ambition extends well beyond the United States. We are working with distributor partners in Latin America, the GCC and select Asia Pacific markets, determining the right regulatory path for each. COMPASS is expected to support broader approvals and European reimbursement. Over time, we want to reach approximately 2,000 people living with MPS II in commercially accessible geographies. Peter will now tell you how COMPASS moves that goal forward. Peter?

Thanks, Katie. Slide 15. The launch is built on the strength of clinical evidence and our ongoing studies are extending it. I'll cover COMPASS, zafinofusp alfa and our 2 Alzheimer's programs. Slide 16. Our Phase I/II study of 47 participants supported U.S. accelerated approval of AVLAYAH. COMPASS is how we build on that foundation, and I'm delighted to report that we have completed enrollment in Cohort B for a total of 65 participants across cohorts A and B in this global randomized, double-blind, active controlled Phase II/III study. I want to thank every participant in family as well as the study investigators who made that possible. COMPASS is designed to provide the evidence for conversion to full U.S. approval for potential expansion of the label to adults and for global regulatory and reimbursement decisions. COMPASS has co-primary endpoints, CSF heparan sulfate and adaptive behavior. As a reminder, CSF heparan sulfate is a CNS biomarker that serves as the basis for AVLAYAH's accelerated approval. And in our Phase I/II study, we observed skill gains in most participants on measures of adaptive behavior and cognition. We expect the study to complete around the end of 2027, and we look forward to sharing the data in the first half of 2028. Slide 17. Turning to zafinofusp alfa. Sanfilippo syndrome type A is a devastating disease and families need treatment options that address the brain and body. In the third quarter, the Phase I/II study completed the prespecified treatment duration for the primary endpoint. We expect the resulting data package could support a BLA submission in 2027, seeking accelerated approval. The package includes CSF heparan sulfate reduction in the key efficacy cohorts, support of central and peripheral efficacy endpoints and up to about 2.5 years of safety follow-up. As we have previously said, we reached alignment with the FDA that CSF HS may be considered a reasonably likely surrogate endpoint to predict clinical benefit and may therefore be used to support an accelerated approval path for zafinofusp alfa or MPS IIIA. We are analyzing the full data set and plan to present it at a medical congress in the first quarter of 2027. In parallel, we are moving quickly on manufacturing readiness and start-up of the Phase III confirmatory study designed with extensive input from FDA and EMA. Slide 18. As a reminder, this slide shows preliminary results from the first 2 dose-finding cohorts presented at WORLDSymposium earlier this year, and they were encouraging. At week 49, mean CSF heparan sulfate was reduced by 80% with 3 of 7 assessed participants reaching the normal range. We believe these are the strongest biomarker data presented for MPS IIIA. We also saw a 61% mean reduction in CSF GM3, an 83% mean reduction in urine heparan sulfate and normalization of liver volume in all participants assessed, reflecting zafinofusp alfa's potential to address disease manifestations in the brain and body. Infusion-related reactions were the most common adverse events, and the safety profile has been generally consistent with established enzyme replacement therapies. The next data set will include all 20 participants and a broader view of efficacy and safety. Slide 19. Finally, Alzheimer's disease, where both of our programs are now in the clinic. DNL628 is an intravenously delivered antisense oligonucleotide targeting tau. Its Phase Ib study is randomized and placebo-controlled in people with biomarker-confirmed early Alzheimer's disease, and we expect initial clinical biomarker data in the first half of 2027. DNL921 is our blood-brain barrier-enabled anti-amyloid antibody. Its Phase I/Ib study is ongoing, and we expect initial safety and clinical proof-of-concept data in 2027. The field has made real progress in understanding Alzheimer's biology and measuring it with biomarkers. We believe the next advances depend on delivering therapies safely and more effectively throughout the brain. And these 2 studies will give us our first clinical readouts on that approach. In 2027, we expect data from 5 TransportVehicle programs: zafinofusp alfa in the first quarter; DNL593 in FTD granulin; and DNL628 in Alzheimer's in the first half; and initial data from DNL921 in Alzheimer's; and DNL952 in Pompe disease. Alex, over to you.

Thank you, Peter. I will briefly review our portfolio strategy and pipeline priorities and then get to the financial results. Slide 21. The TransportVehicle technology has broad potential for a wide range of indications where improved distribution of biologic medicines to the brain and other tissues can be beneficial. We are focusing on 2 main fields or franchises, which complement each other in our long-term growth strategy for Denali. The first franchise are brain-penetrant enzymes with AVLAYAH as the commercial foundation, and we're excited about the strong recognition of the value for patients as seen in the successful launch. Beyond AVLAYAH, we are developing a portfolio of additional ERTs with zafinofusp alfa for MPS IIIA, the most advanced. Enzyme replacement therapies are a sizable market opportunity as conventional enzyme replacement therapies, which generally don't reach the brain, currently represent a market of more than $9 billion. We believe that the next-generation products with TfR-enabled brain and tissue delivery can significantly expand this market. In addition to market size, the ERT field is attractive as these programs historically have a high probability of success and relatively fast development time lines. We also expect that most of these indications would qualify for priority review vouchers. Our second franchise is leveraging the TV platform for common neurodegenerative diseases, where each indication has a commercial potential of more than $5 billion. With our 2 Alzheimer's programs targeting Abeta and tau currently being tested in the clinic, we have the most advanced wholly owned set of BBB-enabled product candidates targeting the 2 key pathologies of Alzheimer's. We are pursuing these 2 franchises in parallel with the enzyme business providing the near-term commercial growth and neurodegeneration creating substantial upside. Slide 22. Zooming into the next 12 to 15 months, this slide highlights our 3 main priorities: first, building on AVLAYAH's strong launch, we are on track to become the U.S. market leader in Hunter syndrome; second, advancing zafinofusp alfa toward a planned BLA filing in 2027; and third, generating initial clinical data with our 2 Alzheimer's programs in 2027. In short, building commercial momentum while positioning the pipeline for important milestones next year. Turning now to the financial results, Slide 23. AVLAYAH generated $17.8 million in net product revenue in Q3, and $21.4 million in 2026 year-to-date. Operating expenses in Q3 and year-to-date were essentially flat compared to the same period last year despite the commercial build for the launch of AVLAYAH and the progress on the pipeline. This reflects our effort for increased capital efficiency. In Q3, we recognized net income of $70.1 million, which is primarily due to the onetime gross proceeds of $195 million from the sale of the PRV. We ended September with approximately $1.01 billion in cash, cash equivalents and marketable securities. With this strong balance sheet, we have the capital to execute on our 3-year goals, which Ryan will speak to next. Ryan, back to you.

Slide 25. Thank you, Alex. The progress we discussed today gives us a foundation for our D3x3 goals through 2028. We aim to build 2 growing commercial brands, establish 5 clinical proof of concept and bring 4 to 6 additional programs into the clinic. AVLAYAH marks an important step in delivering on the promise of our TransportVehicle platform. The work ahead is to build on that first launch, pursue clinical evidence across the pipeline and bring more medicines to patients. This goal gives us a clear way to measure our progress through 2028. We are deeply grateful to the patients, families and clinical trial investigators who make our work possible, and we are inspired by the dedication of the team that carries it forward. Thank you again for joining us today. We are now ready to take your questions. Operator?

[Operator Instructions] Our first question comes from Jessica Fye from JPMorgan.

This is Adam on for Jess. I just wanted to know, with your guidance in 3Q, so now we have 4Q guidance, can you help us think of the trajectory and where the S-curve inflection point is being pulled forward on top of that with another quarter of data? Are you in a position to share maybe patients on drugs or cumulative start forms?

Great. Thanks for that question. This is Katie. Yes, we recognize that it's quite a jump from last quarter to this quarter. And of course, the rapidly expanding coverage really drove a big portion of that. So definitely, many more patients were able to get on treatment. And we were able to move in the Phase I/II patients on to commercial therapy much sooner than we expected, and they're already at the maintenance dose. So they contributed to the revenues. In terms of going forward and what we're looking at, of course, we're looking at a compilation of different pieces of data. And one of the reasons that we guide with revenues is because we feel like that most represents how the business is evolving, whereas other metrics are still, this early in launch, moving. So we feel -- with the base of patients that we see today and the new patients that we hope to get on treatment over the next quarter, we feel like our guidance from a revenue perspective is the most accurate.

Our next question is from Salveen Richter of Goldman Sachs.

This is Tommie on for Salveen. And congratulations on the progress here. Just a follow-up on that. When we're thinking about the current J-code, can you dissect a little bit more how much that factors into the 4Q revenue step-up versus new patient adds? And just a quick one on 126. Maybe you could provide some color on your expectations for the additional Phase I/II data in first quarter next year and how you're thinking about differentiation versus the therapy now in the market.

Great. Thanks, Salveen (sic) [ Tommie ], for your question on the J-code. So the J-code is effective October 1. And certainly, with rare disease, having a J-code in place helps with the reimbursement processes. So it will help the access piece go much more smoothly than if the accounts were using the miscellaneous J. So of course, that's been incorporated in our thinking on the guidance. And then your second piece on the question -- sorry, there was another piece.

Just on the additional 126 data and the differentiation here that you're expecting versus the therapy now in the market.

Yes. Thanks for the question. This is Peter. As you know, we presented preliminary data from dose-finding cohorts earlier in this year at WORLDSymposium. And for the final data cut, we'll have the full 20-patient cohort from the study with additional follow-up as well as patients treated with the final dose regimen. We're excited about the data. Even with the preliminary dose-finding cohorts, as we noted, we saw an 80% reduction -- mean reduction in CSF heparan sulfate with the ability to normalize as well as peripheral effects, as we covered in the presentation, on urine GAGs as well as normalizing liver volume. And so we believe these are the strongest biomarker data that have ever been presented in MPS IIIA and that portends the potential for clinical benefit. We continue on the accelerated approval path with heparan sulfate as a reasonably likely surrogate endpoint.

Our next question is from Andrew Tsai of Jefferies.

Great execution. So I also wanted to ask more about the pipeline since you have a lot reading out in 2027. On the FTD progranulin program, when that data comes in first half, it sounds like you'll share a series of proximal biomarkers, but you'll also share NfL out to year 1. So on NfL, what kind of percentage reduction would you want to see for you to seek an accelerated approval? And then secondly, what would you want to see on the functional outcomes? Is that connected to the accelerated approval thinking as well?

Yes, Andrew, this is Ryan. Thank you for the question. Just a little bit of history here. As you know, we now have full control and ownership of the PTV progranulin program. And essentially, we made the decision in the middle of this year to extend the endpoint to 1 year to focus on NfL. And the reason for that is the proximal biomarkers, which are biomarkers of lysosomal dysfunction are very modestly elevated. That was what we had seen previously, but actually, I think importantly, this is probably what we expect to see in this ultimate cohort, specifically Cohort B3. And as a result, we shifted our focus primarily to NfL as the endpoint. In terms of percent reduction in FTD granulin, you're looking at about fourfold to fivefold elevated NfL, not too dissimilar from what we're seeing actually in the lysosomal storage diseases like Hunter and Sanfilippo. And if you look at, historically, our data, it takes some time for NfL to reduce. But then when it does reduce, it's robust. We haven't set a point of what type of reduction, but believable reduction is going to be important for us to move forward. I think one other sort of really relevant side note is that FTD is on a spectrum with ALS genetically and even within families that share the same mutation. And ALS has also used NfL as a surrogate endpoint of reasonably likely to predict clinical benefit. So obviously, we would be taking the first medicine forward using NfL as a potential surrogate. But fortunately, there's some history with ALS. So that would be the goal. But obviously, the key point now is getting that data in the first half of the year before we decide what the regulatory path would be.

Maybe just to add, Andrew, on the regulatory side of things, I'll just remind everyone that there are no approved therapies for FTD. And so this would be paving a path with positive data and trying to leverage precedent. I'll also note that -- as a reminder, that the DNL593 program received orphan drug designation earlier this year from the FDA, again, recognizing the unmet need.

Our next question is from Tazeen Ahmad of Bank of America.

Just a follow-up on Sanfilippo. I'm sorry if you already mentioned this, but what do you need to complete before your BLA submission happens next year? And are you still planning for an accelerated approval in calendar '27? And then one quick question on Hunter. When do you think you'd be in a position to feel comfortable providing sales guidance?

Okay. Tazeen, I'll take the first one on what we're going to see in the BLA package. Peter can add to that. And then in terms of sales guidance, obviously, we've provided guidance quarter-over-quarter. So we'll get -- Katie can add a little bit more detail there. So basically, the data package that will be in the BLA filing will be the 20 patients. I think as already articulated by Peter, the first 8 patients we used to identify the right dose. So we actually changed the dose regimen and then we dose escalated. We went from every-other-week dosing to weekly dosing until we really identified the dose. And even in those patients where we were dose finding, we achieved greater than 80% reduction of CSF heparan sulfate with the ability to normalize on CSF heparan sulfate as well as some additional lysosomal biomarkers. And so our plan is to take that data. Now all 20 patients have completed the 49-week portion. That data will be presented at WORLD, but also that will be the data that drives our BLA filing. The timing of filing and approval and, frankly, really the timing of filing is driven by Module 3, which is the manufacturing piece. And what we've guided to is essentially that we'll file in 2027. We're moving as quickly as possible. Just a reminder that we are manufacturing that molecule ourselves. So we built our own facility. Ultimately, this will be the best in terms of COGS to be able to manufacture ourselves. We've transferred that already within the last several years. But this will be the first product that we're launching ourselves. So that's what's driving the timing of the filing. And then subsequent to that, obviously, we would expect a priority review and look for approval. So we have less clarity on when we think we get approval as we experienced with AVLAYAH. But again, filing in '27. The data will be in hand and it will just come down to Module 3. It will all be based on manufacturing. In terms of the second question, which is around sales guidance or revenue guidance, can you just add a little bit more clarity on what you're asking exactly because, of course, we're guiding...

Annual guidance.

Oh, okay. Yes, I missed that part. Do you want to comment?

Yes. At the moment, we haven't decided whether or not we're going to guide annually. So as the launch continues to evolve, we'll definitely look at all our metrics and have more information for you all.

And our next question comes from Paul Matteis from Stifel.

This is Matthew on for Paul. Congrats on all the progress and performance with AVLAYAH. So my questions are more on the pipeline. So for the Abeta program, maybe once we see the Phase I data, what are your thoughts on continuing to advance this yourself versus partnering this with another company? And then just another question on the Pompe. What biomarkers should we be looking for in the initial data that might help establish differentiation from standard ERTs with your delivery platform?

This is Alex here. Thanks for the question. I will address the partnering question. So at this point, what we guide towards is that we have the capabilities and the capital to advance the entire portfolio forward. We put it in the context of our D3x3 goals. And with $1 billion in cash, we're well financed through that. Now in the long term, of course, the large neurodegenerative diseases are ideal partnering opportunities. We also have a track record of doing that in the past where we have earlier on signed collaboration agreements for Parkinson's and Alzheimer's disease. So this is something that we would consider in the future. But for now, the focus is very much on executing and generating the clinical proof-of-concept data.

And then on the question on biomarkers for Pompe, we're looking at urine Glc4 as well as serum creatine kinase. The idea here is looking at glucose-related metabolism as has been with other Pompe medicines and then, of course, muscle-related biomarkers. Obviously, our goal initially is working -- we're enrolling LOPD patients that are switching to ETV:GAA. And essentially, those are the biomarkers that we'll be looking at early on to show that we can maintain a robust biomarker response.

And our next question is from Sean Laaman of Morgan Stanley.

This is Mike on for Sean. Congrats on all the progress. Thinking about those 375 pediatric patients, where does, like, U.S. penetration top out before the COMPASS adult label and assuming, like, treatment of neurological manifestations matters to all those patients? But, like, could you just give us an understanding of the target profile for the new patients you're hoping to get next quarter? Any color there?

Yes. Thanks for that question. So in terms of market share, our belief is that all eligible patients should be on AVLAYAH. And we've shared historically that almost all patients have neurologic manifestations. So as we commented on earlier, our expectation is that we will become the market leader, and that's both in treating incident patients as well as prevalent patients that AVLAYAH would be the physician's choice for treating Hunter.

And our next question is from Michael Yee of UBS.

This is [ Madeleine ] on for Michael. Congrats on the quarter. I have a couple of questions maybe on Alzheimer's. So the first one is Biogen sharing some updated data from their tau Alzheimer's trial through some long-term extension data, subgroup and biomarker data, at CTAD. I guess what would you want to see from this data to gain more confidence on tau as a target? Or how should we -- things that you're looking for in the subgroup analyses? And the second, on the Abeta program, just a question on how you position that asset in what's becoming an increasingly competitive space. You're obviously looking at the biomarker-positive patients, not necessarily with cognitive symptoms. So maybe how you think about the Lilly TRAILBLAZER-ALZ 3 trial, which is coming sometime between now and 2027.

Okay. Great question. I'll take this initially, this is Ryan, and then if Peter has anything to add. So in terms of the Biogen data, I cannot really predict what else they'll show. But what I can tell you would be helpful is just to understand a little bit more about the biomarkers. As you may recall, although there was a doubling and ultimately, quadrupling of dose, there wasn't a robust difference in the tau reduction across those doses. However, there was a difference, obviously, in the safety profile. So a biomarker that would be quite interesting to look at would be NfL, and the question is, is the higher dose, is there some type of either inflammatory or some type of sort of biomarker that the higher doses are not as tolerated as the lower dose? And I think the confusion right now in the field is what is the target level of tau reduction that someone needs to achieve in order to drive maximal efficacy. And if you take the Biogen data at face value, you're not striving for super robust tau reduction. But I think that, that could be complicated by the fact that it's intrathecally delivered. And in our case, of course, we're delivering systemically and knocking down tau. So I think that's point number one on tau. Do you want to add anything else on tau, Peter? Okay. So let me talk about the differentiation of the ATV:Abeta program. As you may recall, we published data in Science of last year around the unique properties of DNL921. So this is our ATV:Abeta program. And what's unique about that molecule is essentially the effector function on/ off switch. So when this antibody is bound to transferrin receptor, we don't engage Fc gamma receptors. And as a result, we would see little to no hematological findings with DNL921. However, when it's bound to amyloid plaque, you can engage the Fc gamma receptor and have robust plaque reduction. And so the data that we've generated, we've compared against the competitors that use a traditional shuttle architecture. And what we see is more sustained brain exposure, higher brain exposure, better plaque reduction and better stability. Remember, the transferrin receptor binding is integrated into the Fc itself, which is separate from what you see with all the shuttle programs. So ultimately, we should expect a robust plaque reduction with a better tolerability and safety profile. Just one other note that there also is infusion-related reactions and some immunogenicity with those shuttle architectures that we expect hopefully not to see with DNL921. So there are multiple areas in which we expect differentiation. The initial healthy volunteer data will actually be important to understand the hematological components of things. And then from there, obviously looking at plaque reduction. Ultimately, I think we're going to need larger patient numbers to look at the rates of ARIA.

Maybe a couple of things to add to what Ryan has just said about the 921 program. I think relating to the immunogenicity, there is a potential to differentiate on infusion-related reactions. Now we'll have to gather the data to determine that. But based on the architecture, that might be a possibility. The other thing is we are exploring the subcu administration, which could be a great convenience in the crowded marketplace that you've highlighted.

And our next question comes from Mayank Mamtani of B. Riley Securities.

Congrats on a strong commercial quarter. Sorry if I missed this, are you able to break down the de novo new patient adds versus the rollover from the Phase I/II? And just broadly thinking about the peak sales, how you're thinking about that for AVLAYAH? Is that really just moving -- the curves moving to the left or your broader view on penetration share is also evolving until the COMPASS data becomes available and folds in the adult opportunity and, of course, ex U.S.? And maybe just point to any similarities with zafinofusp alfa commercialization to AVLAYAH as you think about that 2 years out? And lastly, maybe a follow-up on tau MAPT. Just remind us why you are able to fast track into Alzheimer's patients, that program. I ask that in context of also the siRNA peer data we recently saw in healthy volunteers.

Thanks, Mayank. I'll start with the commercial questions, and then I'll turn it over to Ryan and others. In terms of patient numbers, we've always said that we share what we think is most reliable and helpful for you all to understand how the launch is progressing. So we feel like guidance on the actual revenues is the easiest way to look at to see how our launch is progressing. In terms of the base of patients, what we've shared is that, of course, every quarter, we continue to get new patient starts and the speed in which we can get a patient from interest to infusion is really driven by reimbursement and coverage. And we've shared this quarter that we've seen another positive move in terms of that momentum getting faster. And in terms of the overall opportunity, we estimate 375 pediatric patients in the U.S. And our goal definitely is to capture all of those patients. We believe AVLAYAH is the best treatment for patients, which is why our expectation is that we would have leading market share over time. In terms of the similarities between zafi and AVLAYAH, we're very pleased with the biomarker data because you could see, with the MPS II market, there's really great belief in the normalization of the disease biomarker. And this has translated to the tremendous conviction we've seen from physicians on getting patients on AVLAYAH. And we're very hopeful that these are the same physicians that will be -- or similar physicians that will be treating MPS IIIA is that they will also see the normalization of biomarkers that's been very meaningful in MPS IIIA as well.

And then on the question around going directly to Alzheimer's patients, I mean, Peter can add any additional details, but I'll just say that, that was our decision. And I think mainly because it's most important to understand the biomarker in the context of the disease.

I think the only thing I'd add is it just reflects the mindset that we take to development, recognizing the importance of generating benefit-risk data as quickly as possible, given the unmet need.

Our next question comes from Ananda Ghosh of H.C. Wainwright.

Congrats on the quarter. I have 2 questions. The first one with respect to Sanfilippo. As Sanfilippo moves towards BLA, what are the learnings from the AVLAYAH launch challenges? And how do you plan to integrate or streamline 2 products going forward? The second question is, Takeda guided to 9% decline at a constant currency for ELAPRASE. Does it reflect the switching dynamics you are seeing with AVLAYAH?

Okay. I'll start first on the synergies between AVLAYAH and zafi. I think we've shared before that the infrastructure -- commercial infrastructure that we've built today for AVLAYAH will actually be the same infrastructure that will launch zafi. And the reason is because the customer base is similar. So what we're building out today should help expedite how we think about launching our second product. And with the credibility that we have on launching our first brand, the engagement that we've had with the community and the trust that we've built with the stakeholders, we feel like this will accelerate our zafi launch as well.

And then in terms of the sales reported by Takeda, the decline in ELAPRASE, we have no comment on that.

Our next question comes from Joseph Thome of TD Cowen.

This is Jacob on the line for Joe. So this past quarter, we saw some data from a competitor using a more brain-penetrant approach at tau reduction and they reported some CSF tau lowering data. I was wondering if you had, like, any thoughts on that data and whether, like, we might expect similar sort of reductions using more brain-penetrant methods? And you had also said it's maybe difficult to compare some of these other methods versus IP injections and how those affect CSF. I'm just curious if you have any additional thoughts on that, now that we've seen these data.

Yes. I try not to make it a habit of commenting on others' data. I know we've already discussed briefly the intrathecal versus our approach using the TransportVehicle. But I think it's a very exciting time in the blood-brain barrier field. And it's great to see, of course, in any area where you're making progress, many others will use similar approaches. And in this case, I think the data you're referring to uses transferrin receptor to get an siRNA across the blood-brain barrier in healthy volunteers. And so I think a question earlier in the call asked why do we go directly to patients. And I think our focus on patients just to understand the biomarker nature in patients is key. And I think what we've seen in our own labs is that any type of oligonucleotide, depending on if you engineer it, either an ASO or an [ SI ], we can successfully get it across the blood-brain barrier. So we're really excited to be in a highly competitive space. And I think the fundamental question is what level of reduction do you need to drive clinical benefit, and that's just going to take more data. And obviously, crossing the blood-brain barrier is very different than intrathecal delivery in terms of biodistribution. And I think that's -- other TfR approaches will have more similarity, but good to see others getting data, seeing safety, seeing pharmacodynamic response and what's becoming a competitive field.

Our next question is from Marc Goodman of Leerink.

Yes. Can you remind us of the pricing that you expected for AVLAYAH and whether 3Q is tracking to that level? I mean, gross to net or anything different from expectations? And have any of the patients discontinued? And if so, do you know why?

Thanks for the question. I think we've shared before that the current price is $5,200 per vial. We've not shared gross to net numbers. We said an average 30-kilogram patient is -- the cost would be about $811,000. And your second part of your question, I'm sorry...

That was on discontinuation. I can take that. There's a very high level of patients remaining on drug. Remember, this is a lifelong disease with continuous weekly infusions. And so historically, you look, the numbers are really high in terms of patients staying on drug.

Yes. So when we look at ELAPRASE's persistence, it's very, very high. So it's a chronic therapy, and it's too early to tell now, but our expectations is that persistence would be high.

And just back on that pricing that you were talking about, I mean, if we just cut that into 1/4 of that number, I mean, is that generally how it was playing out in the third quarter as far as the 30 kilogram on average?

Yes. No, because there's dose escalation. So the patients start, for the first 8 weeks, they're on 3 milligrams per kilogram and then they go up to 7.5 milligrams. So there's dose escalation. And the number that I shared, the $811,000, that's based on the 30-kilogram patient, and we're going to have -- we're going to see a lot of variability in terms of weight. Given this is a pediatric indication, patients start very young all the way up to 16, 17 and older. So it's hard to say what is an average.

And I think, Marc, that's why we guide to revenue because there are so many variables that go into dose escalation, weight. And obviously, as these patients continue on drug, they will need more vials over time as well.

Our next question is from Myles Minter of William Blair.

Congrats on the quarter. One on safety related in the Sanfilippo program. I think at WORLDSymposium earlier this year, you presented some data, I think it was from 8 patients, and there was 1 patient that had an infusion-related reaction, mild to moderate, but it forced them on to a second infusion bag. Has that happened again at all to your knowledge? And then the second question is just with the gene therapy being approved earlier this year. I think they had some anti-SGSH immune responses that actually required some B-cell depletion in a couple of patients to get them back on track. Have you seen any sort of antidrug antibody reactions with your product? Just curious there.

Yes. I'll provide a top line and then Peter can provide a little bit more detail. Actually, infusion-related reactions are very common across all of these ERT patients, including in Sanfilippo, and immune tolerance protocols are very often used even for ERTs. It's seen often, like for example, in IOPD. And so these are the type of things that we would -- depending on how a patient reacts, we would consider these type of approaches. A little bit different than gene therapy, which has a secondary sort of risk and a lifelong risk around the immunogenicity related to the AAV. But for enzyme replacement therapies, it's very common to see ITI. So I don't know exactly what you're referencing on the patient that needed another infusion bag, but we see the same rates of IRRs in Sanfilippo that we see in -- essentially in Hunter. Do you want to add anything?

Yes. I'll just add that we don't have the details of additional patients at this time, but the team is actively analyzing the data comprehensively across the 20 patients and the full duration of exposure. And so we'll have more information at the time that we present the data.

And I'll just add one other point to answer your question specifically. ADAs are very common in ERTs as well. So I'd say, like, the vast majority of patients in all of these enzyme replacement therapies develop some type of antidrug antibody. And so that's why we've actually always pushed dose and we build tolerance at higher doses. And then as long as patients don't miss doses, they actually can tolerate it well. And that's what we saw and was published in New England Journal. And I think that's why a lot of ERTs are ultimately underdosed historically because they're not driving towards biomarker normalization and they have to develop basically tolerance over time with dosing.

Our next question is from Laura Chico of Wedbush.

I wanted to circle back on your comments on DNL921. And I guess the underlying hypothesis here, DNL921 generates less immunogenicity than perhaps other brain shuttles. What would be an acceptable degree of antidrug antibody incidence and neutralizing activity that you'd expect for 921? I guess, is there a maximum level that you want to avoid? And I guess I'm trying to better understand your thinking here on how antibodies might impact exposure and plaque clearance over time.

Thanks, Laura, for that question. This is Ryan again. It's not really a quantitative question. It's qualitative in terms of driving efficacy. So antidrug antibodies and infusion-related reactions are always in the context of efficacy. Obviously, it's a very different context in risk-benefit when it comes to Alzheimer's disease. And so what you'll see in the early days of trontinemab in particular was it's really about tolerability and how many patients basically dropped out of the study. And so they then, of course, shifted to pre-dosing with steroids, which is not necessarily ideal in the Alzheimer's population. So our goal here is a drug that is tolerable to the point where we have a lot of persistence on drug and drive maximal efficacy. And I think the other side note around differentiation is just the hematological findings, which, of course, we hold a high bar on what we want to see in humans, especially in elderly that are taking this drug. So hopefully, that answers your question.

Our next question is from Michael DiFiore of Evercore.

Congrats on the impressive progress. Two for me. At this stage of the AVLAYAH launch, what are the average times to authorization and first infusion? And does this differ between commercial and Medicaid? And my follow-up question has to do with COMPASS. In a scenario where there's strong heparan sulfate normalization, but Vineland misses, could you provide any color in terms of what has been agreed upon with the agency regarding a favorable totality of evidence benefit?

Great. Thanks for your question on the time from start form to infusion. For ultra-rare disease, unfortunately, there's not a standard time measurement. It really depends on the patient's insurance plan as well as the logistics of the hospital institution that they're working out of. Of course, we shared that we got our J-code, which will speed up the reimbursement process, but there's not an average time per patient. And you highlighted commercial versus Medicaid, even that varies. Generally, commercial is thought to be faster, but sometimes Medicaid can be fast as well. So it is very individualized, unfortunately.

And on the COMPASS question, the first thing I'd say is we're confident in the design. We leverage the data from the Phase I/II study to inform the design of the COMPASS study. And it covers a broad population with both CSF HS and Vineland, as you pointed out. Now we also, as you pointed out, have a totality of evidence that includes 47 patients from Phase I/II with an additional time of follow-up from the original accelerated approval. We'll have the COMPASS study itself, and we have open-label data following the COMPASS study. And to your point, as is common in rare disease, we will be submitting a totality of evidence. And so I think, across the expanded clinical development program, we're confident that we'll be able to show both CSF heparan sulfate reductions as demonstrated in the Phase I/II, and we should be able to demonstrate clinical improvements as well.

Our next question is from David Hoang of Deutsche Bank.

Congrats on the quarter. I'd love to ask about the upcoming POC data for 921 and 628 programs in Alzheimer's. How should we think about the bar that you would use to justify moving these programs forward based on that data? And then once the data is in hand, could you talk about how you think about next steps for these programs? How quickly, for example, could you move into registrational-type programs?

Yes. As we're getting close on time, I'll just give a quick answer and then we can discuss more, David. But I mean, for us, believable and robust biomarker response is key. We avoid giving a percentage because it's all contextualized within dose, for example, and then moving forward. But our goal in the Phase I study is to identify a dose with a robust and safe profile -- robust biomarker profile. And then in terms of the design of either a Phase II or a Phase III registrational trial, we have a lot of internal discussion and thoughts about that. We have not shared publicly what that design will look like. But speed is going to be key. And for both tau and for Abeta, especially for Abeta, there's precedent, right? So the key now is designing the differentiated study going forward. So I'll leave it at that, and say stay tuned, but a robust biomarker response with a safe and ideal dosing regimen is what we're looking for.

Our next question is from Charles Moore of Baird.

Congrats on the quarter. Just a couple on tau. I was just curious how you're handling background Abeta treatment and more generally, how you're thinking about combinability there. If you have any thoughts on what you expect the dosing frequency to be, and if you could just remind us if you're working on a subcu formulation for later-stage trials or the commercial setting.

Yes. So I'll take it in reverse. Subcu will be the ultimate goal with both tau and Abeta programs, and those will evolve as we see biomarker data. Our initial study is not designed to combine with amyloids, but that is certainly where the field is going to go. And I think the idea of targeting both Abeta and tau is going to be ideal. But in these early trials, we're looking for biomarkers, not in patients that are also on Abeta program -- on an anti-amyloid.

And our final question comes from Rudy Li of Wolfe Research.

Congrats on the strong quarter. Maybe just another quick follow-up question for DNL628 data in first half. Can you determine the dosing regimen based on CSF tau data? Or do you need to wait for additional biomarker imaging data to mature? Maybe just talk about the next step, what additional work is needed before moving to that pivotal trial?

Yes. Just the short answer is tau will be what drives our dose selection. So the short answer is we can determine dose frequency and dose level based on CSF tau. That is the primary proximal biomarker.

This concludes the question-and-answer session. I would like to turn it back to Ryan Watts for closing remarks.

Thank you for joining us today on the call. We appreciate the interest and obviously, very excited about this quarter and looking forward to the future.

Thank you for your participation in today's conference. This does conclude the program. You may now disconnect.