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

Hello, and welcome to the Denali Therapeutics webinar on interim data from our Phase I/II study of DNL310 in patients with Hunter syndrome or MPS II. I am Laura Hansen, Vice President of Investor Relations, and I'd like to thank you for joining our webcast today. Please note that the press release we issued earlier today and the slide deck for this webcast are available in the Investors section of our website, denalitherapeutics.com. Before we get started, I'd like to note that the presentations given today and the responses to questions will contain forward-looking statements regarding Denali's future plans, business strategy, product candidates, planned preclinical studies and clinical trials, among other things. Such statements are subject to numerous important risks, uncertainties and assumptions. Should any of these risks or uncertainties materialize or should our assumptions prove to be incorrect, our actual results could differ materially from these forward-looking statements. These risks, uncertainties and assumptions are more fully described in our filings with the SEC, including our latest quarterly report on Form 10-Q and our latest annual report on Form 10-K. Any forward-looking statements are based on information available to us as of today, and we disclaim any obligation to update any forward-looking statements except as required by law. On the webcast today, I am joined by members of Denali's management team, Ryan Watts, Chief Executive Officer; Carole Ho, Chief Medical Officer; Alex Schuth, Chief Operating Officer; and Steve Krognes, Chief Financial Officer. I would like to take a moment to review the agenda and Q&A logistics for today. We have scheduled approximately 1 hour for the webcast, including presentations and the Q&A session. Ryan will begin with introductory remarks and Carole will present the interim data from the Phase I/II study of DNL310. Then Ryan will provide concluding remarks and begin the Q&A session. [Operator Instructions] And now I would like to turn the program over to our Chief Executive Officer, Ryan Watts.

Thank you, Laura. Very excited to be here today. Also very excited to share some of our data in our DNL310 program, actually extensive data, biomarker data and clinical data. I'd like to start with our purpose. Denali is -- has set out to defeat degeneration. And as we've highlighted in the past, we have a number of ongoing studies or completed studies across all of these therapeutic areas, including rare neurodegenerative diseases, ALS, Parkinson's and Alzheimer's disease. Today, we'll focus on the rare neurodegenerative diseases, specifically lysosomal storage diseases. We follow a set of principles as we develop each of our medicines. These are our discovery and development principles. First is what we call the degenogenes, genes when mutated that cause neurodegeneration. Today is a highlight of that as we focus on IDS in Hunter syndrome. Second is engineering brain delivery, so inventing medicines that readily cross the blood-brain barrier. And again, today, we'll focus on our Transport Vehicle technology, which is designed to cross the blood-brain barrier for large molecules, such as antibodies, enzymes and antisense oligos. And our third principle is biomarker-driven development, and I think that's a major highlight of today's presentation. We'll show extensive biomarkers as we approach Hunter syndrome. And then finally, our goal is to have a patient impact. To increase the likelihood of success, we follow these 3 principles in each of our programs. So just a reminder that our portfolio is split roughly between small molecules and biotherapeutics, about 50-50. We're focusing on the biotherapeutic portfolio today in our most advanced program in Hunter syndrome, DNL310. I would like to highlight, however, that we've now advanced DNL126 for Sanfilippo into the IND-enabling stage. Let's talk about biotherapeutics in the blood-brain barrier, and I'll begin with the blood-brain barrier challenge. The blood-brain barrier is a major obstacle for brain delivery of biotherapeutics. And the way that we've approached it is to use natural transport mechanisms in blood vessels in the brain, and specifically the transferrin receptor as shown here. We have engineered molecules to latch on to the transferrin receptor, which is constitutively endocytosis, bringing iron into the brain, with a goal of bringing our biotherapeutics across the blood-brain barrier and into the various cell types in the brain. To highlight the technology that we're using, the Transport Vehicle technology, we published 2 papers last year back-to-back in Science Translational Medicine. The first paper was -- outlined the invention of the Transport Vehicle technology and its application to antibodies, with proof-of-concept in both small animals as well as nonhuman primates. The second paper utilizes the Transport Vehicle technology to get an enzyme across the blood-brain barrier for Hunter syndrome. Here, we use the IDS Hunter mouse model, and we can show robust reduction in substrate in brain, which correlates with reduction in cerebral spinal fluid, one of the key fluids that we use in our human studies. In summary, the Transport Vehicle achieves high concentrations and broad biodistribution of biotherapeutics in brain. It also can achieve a dose-dependent reduction in brain substrate, and we're very excited to show data even at lower doses today for our DNL310 program. So important for us is the broader potential of the Transport Vehicle technology, both in various modalities but also in therapeutic areas. Our initial focus is in neurodegenerative diseases as well as lysosomal storage diseases. However, we see potential in broader neurology, oncology and infectious disease. And importantly, we're applying the Transport Vehicle technology to antibodies, enzymes, proteins and even antisense oligos and advancing a number of programs in this space. Before I hand it over to Carole, I'd like to summarize the interim data that we'll present today. We're very enthusiastic about the ability to rapidly and durably reduce heparan sulfate in cerebral spinal fluid across all patients tested and across all doses. In addition, we'll be presenting data on exploratory biomarkers of lysosomal function, where we see reduction consistent with improved lysosomal function. We see a high variability in neurofilament, observed both pre and post treatment. We'll also show for the first time improvement over 24 weeks based on clinician and patient reported global impression of change at clinical outcome. In addition to this, we see enhanced peripheral activity. All patients are switched from either sulfates to DNL310. And here, looking at both the urine and serum, we see improvement in activity with DNL310. And importantly, the safety profile is consistent with standard of care replacement therapy. The totality of this data has led us to make the decision to accelerate our program into a Phase II/III study. And with this, I'm going to hand it over to Carole to go into great detail on this data.

Great. Thank you, Ryan, and I'm pleased to be able to present the DNL310 interim Phase I/II data. So DNL310, as Ryan described, is an IV therapy that uses Denali's Transport Vehicle technology to deliver IDS, the enzyme that's missing or defective in MPS II, to the brain and the body. The technology of a fusion protein of IDS to ETV has a differentiated binding profile to the transferrin receptor, which in preclinical models demonstrates broad distribution through the brain to neurons, astrocytes and microglia. DNL310 is currently in an ongoing Phase I/II study. It's delivered systemically IV once per week, which is the same dosing frequency that standard of care Elaprase uses. In this Phase I/II study, patients on Elaprase switched to DNL310 in order to evaluate the effects of DNL310 on both the body as well as the brain. The development of 310 is intended to replace current standard of care by delivering efficacy in the brain, a current unmet medical need, while also delivering differentiated efficacy to the body. This next slide outlines the study schema for the Phase I/II study. This is an open-label, 6-month study followed by an 18-month safety extension in approximately 30 MPS II patients aged 2 to 18 years old. Eligible patients are either treatment-naive or on approved IDS therapy for longer than 4 months. In the data presented today, all patients were on IDS therapy and were switched to DNL310 without a washout period. There are 3 cohorts in this study designed to evaluate the safety profile of DNL310 and enable dose selection for our subsequently planned Phase II/III study. The primary endpoints are safety, which includes adverse events, infusion-related reactions and total urine GAGs, which will demonstrate the effect in the periphery because patients switch off of IDS. Key secondary endpoints and very relevant to the blood-brain barrier crossing capability of this Transport Vehicle include CSF heparan sulfate and then also urine heparan sulfate. Exploratory endpoints include other GAG measures of dermatan sulfate, heparan sulfate and keratan sulfate, which we'll present today, including lysosomal CSF lipid biomarkers and CSF and serum neurofilament. Clinical outcome results in cognition, behavior and global impression are also assessed in this study. And today, we will present data on the Global Impression scales. Now I'll go through the detail of the cohorts. We'll present data today for Cohort A and B. Cohort A is a within-patient dose escalation cohort and includes neuronopathic patients 5 to 10 years old. Today, we are sharing 6-month biomarker data and safety data up to 43 weeks. All patients in Cohort A remain in the safety extension study of 30 mg per kg. Cohort B now is a dose-finding cohort to evaluate lower dose regimen of DNL310 in both neuronopathic and non-neuronopathic patients aged 2 to 18. Dosing modifications, as you can see, can be made in Cohort B1 after 12 doses have been administered. Cohort B2 and B3 study stable doses of 7.5 mg per kg and 15 mg per kg. All patients in Cohort B will roll over to the safety extension at the end of the study at 15 mg per kg. Cohort C is a cohort that will enroll neuronopathic patients younger than 4 years of age and sibling pairs to further evaluate exploratory clinical endpoints, such as behavior and cognition, in an age before normal development is markedly impacted by this disease. Dose selection in Cohort C will be determined based on ongoing and emerging data from this study. So now I'll show the demographics of the patients that will be reviewed data in this interim data cut. The demographics are shown here. The median age of both cohorts is 6 years of age, and all 17 patients as noted were on IDS treatment prior to enrollment. Except for one patient in Cohort B, these are all neuronopathic patients. There are 5 patients in Cohort A, all of whom completed the 6-month study. Cohort B is ongoing with 12 patients included in this interim analysis. These 12 patients have completed 12 weekly doses. The largest racial group in this study, so far, is white and the majority of patients are non-Hispanic or Latino. This provides a further detail of the data that will be presented today. As noted, safety up to week 43 and 25 in Cohort B. Heparan sulfate, which is our key secondary endpoint in Cohort A up to week 24 and in Cohort B up to week 13. We'll be also presenting lysosomal biomarker data up to week 24 in Cohort A, specifically GM3, BMP and GlcCer, and we'll be presenting Cohort B GM3 data up to week 13. In addition, we are presenting exploratory neurofilament data on Cohort A up to week 24, and exploratory clinical Global Impression of Change from both an expert clinician as well those as a parent/caregiver up to week 24. So I'll start with the interim safety summary, which is in the safety population of 17 patients. Very importantly, all patients remain in the study with no discontinuations. All 5 of the Cohort A patients continue in the safety extension at 30 mg per kg. And of the 12 patients presented today in Cohort B, 10 are continuing in the 6-month study period and 2 have advanced to the safety extension. All of the data presented today was reviewed by an independent data monitoring committee on July 9, 2021, and importantly, recommended continuing the study without any modifications to the study protocol. All treatment-emergent adverse events were mild or moderate, except for 2 severe treatment-emergent AEs, which were IRRs in one patient, which I'll describe below. Infusion-related reactions were not unexpectedly the most common treatment-emergent adverse event incurring in approximately 71% of the patients. The majority had mild or moderate IRRs, but one patient had a severe IRR, which was an SAE, which will be detailed below. Of the 12 patients with IRRs, 8 required standard interventions to prevent subsequent IRRs, and 2 required additional dose and infusion rate reductions. Most IRRs to date have occurred between weeks 3 and 6. And in the patients -- 3 of the patients who experienced IRRs and advanced to the safety extension, most of the pre-infusion medications have been discontinued. There have been 2 patients that have had SAEs related to infusion-related reactions in this study. One previously reported was a patient who had an SAE at week 4 that was hospitalized for overnight hospitalization, but then was discharged without any further sequelae. The second patient had 2 SAEs, severe IRRs that met Sampson criteria of anaphylaxis at week 3 and 4. These were managed with pre-infusion medications, dose and infusion rate reductions. Importantly, this patient remains in the study and has now tolerated subsequent weekly doses, including dose increases. Regarding safety laboratories, there were no notable abnormalities or trends in safety laboratory evaluations except for anemia, which was also described previously. There are 4 patients that have had treatment-emergent events of anemia, all were considered not related to drug. These were graded mild in 3 and moderate in 1, and all 4 are improving or have resolved despite continuing dosing at either 15 or 30 mg per kg. In summary, the weekly IV infusions of DNL310 for Cohort A and B were generally well tolerated at doses between 30 (sic) [ 3 ] to 30 mg per kg with a safety profile that is consistent with standard of care therapy. I'll now move to the safety biomarker of urine GAG. So as a reminder, this was collected in the study to monitor clinical response to DNL310, particularly given patients were washed off of IDS standard of care therapy. Total urine GAGs were measured by a color metric, CLIA-certified assay and were used in the study to monitor peripheral response to 310. Across Cohorts A and B, total year-end GAG levels overall decreased, including in the lowest dose regimen of Cohort B1 at 3 mg per kg. After switching from IDS to DNL310 without a washout period, total urine GAG levels declined almost in all patients, suggesting improved peripheral activity with DNL310. Consistent with the improvement in reduction of urine total GAGs, we also assessed whether DNL310 could further reduce elevated levels of heparan sulfate and other peripheral GAGs, including dermatan and keratan sulfate. We have previously published an increase ranging from 2 to 3.7-fold in all of these 3 glycosaminoglycans in the serum of MPS patients, which are depicted in orange and red. Notably, these patients are on standard of care enzyme replacement therapy or have received hematopoietic stem cell transplant. You can see that these are elevated compared to the dots in gray, which are non-MPS pediatric controls. We have observed now with treatment of DNL310 in Cohort A across these 3 different serum glycosaminoglycans, there is a reduction at week 24. Notably, keratan sulfate is a dominant GAG present in articular and growth plate cartilage where bone grows. In certain MPS diseases such as MPS IV, where keratan sulfate is the dominant GAG that accumulates, the disease course is characterized by skeletal dysplasia. This reduction in these peripheral GAGs, including keratan sulfate, suggests that DNL310 could have increased potential long-term effects on peripheral clinical endpoints, including bone remodeling. Now I'll turn to the CNS-penetrating capability of this Transport Vehicle technology and DNL310. So importantly, I want to highlight again a slide that we've shared previously that shows the importance of heparan sulfate in monitoring the effects of blood-brain barrier crossing activity of DNL310. So as you can see in the family of MPS disorders, those that have CNS involvement highlighted in light orange, you can see all have heparan sulfate accumulation. For those diseases that do not have central nervous system disorders or cognitive symptoms, you can see that heparan sulfate is not represented here. So similar to the previous slide, I want to remind everybody about the increase in CSF heparan sulfate that is seen in MPS II patients, even though they are currently on standard of care therapy as shown in the upper left. There's roughly a 10.7-fold increase in CSF heparan sulfate in MPS II patients compared to controls. As you can see here, across cohorts A, B1, B2 and B3, all 15 patients had normalization of heparan sulfate at the end of the dosing period up until this time, with rapid response in 12 of these patients by week 7. This percent reduction is 90% in Cohort A and ranges between 86% to 92% at lower dose regimen demonstrating the efficiency of DNL310 for crossing the blood-brain barrier and reducing the most important glycosaminoglycan in the CSF that is correlated with clinical symptoms across the MPS disorders. Dermatan sulfate is also a key substrate of IDS and so we share the data here again. Again, you see a 31-fold increase in dermatan sulfate in the CSF and MPS patients compared to non-MPS controls. And what you can see is that there's a rapid response in 12 patients by week 7 in the reduction of dermatan sulfate in the CSF. I think notably, data that we're now presenting in Cohort B also shows that at these lower dose regimen, we continue to have very efficient reduction in CSF dermatan sulfate. So to extend the understanding of the effects of DNL310 in the CNS, as you know, we have profiled also lysosomal lipids which accumulate in the cells of both neurons as well as glia in the central nervous system. And we use this as a biomarker to get further evidence of effects in the brain. As you can see here, GM3, which is a key nervous system ganglioside, this has elevated 3.7-fold in MPS patients compared to non-MPS controls. As you can see here, across Cohorts A and B, 10 of 15 patients achieved normal CSF GM3, including a dose -- lower dose regimen, as you can see in Cohort B. This demonstrates that DNL310 is getting into the brain and is reducing these lysosomal lipids that accumulate. We also looked at 2 additional lysosomal lipids that have been associated with lysosomal dysfunction, including BMP and glucosylceramide. As you can see here in natural history, cross-sectional data, we do have an increase that is modest in BMP and glucosylceramide. And you can see in both of these biomarkers, there is a mean reduction in these biomarkers at week 24 in Cohort A. So next, we explored neurofilament, a biomarker of neuronal structure, which has not previously been characterized, to our knowledge, in MPS II aside from publications from Denali. In 2020, Denali was the first to publish cross-sectional data demonstrating an observed increase in CSF and serum neurofilament. We have extended this analysis now to look at natural history data to further characterize the longitudinal trajectory of neurofilament. As you can see in the middle panel, these are 3 patients from our natural history study that is studying patients with MPS II. And these patients have the opportunity to roll into Cohort A. Three of these patients in the natural history study are Cohort A patients that then rolled into Cohort A. As you can see here, there is significant patient variability that is best evidenced by the collection of neurofilament in short time frames within a short span. You can also see that there is a mean increase, a market mean increase, of 94% in the 4.5 to 6 months pre-dose. After treatment, there is a modest increase of 15% and 36% increase in neurofilament in the CRM and the CSF. This within-patient variability in neurofilament, as well as the increase in neurofilament that is seen in the natural history study, leads us to conclude that, at this time, the utility of neurofilament in MPS II as a treatment biomarker requires further investigation across the field, specifically in generating additional natural history data. Notably, we also looked at another biomarker of neuronal structure Tau, which shows no change and near normal levels over the 24-week dosing period. I'd now like to turn to our data on the cognitive effects based on clinical global impression change scales. Before doing this, I'd like to provide an overview of cognitive milestones in Hunter's patients. As you can see in this graph, this graph shows the correlation between developmental age on the y-axis and calendar age on the x-axis. As you can see, in normal development, these should track very much along this 45-degree line. However, you can see that there are 2 patient populations that are evident in Hunter syndrome, those that have neuroparenchymal involvement that show regression in their developmental age over time, noted in orange, compared to those that may have milder neurocognitive effects or no neurocognitive effects that track along a normal trajectory. As you can see that this change in the trajectory is very evident by age 5. And all of the patients that we have studied in Cohort A are above the age of 5. Therefore, we would expect very little change or regression in their cognitive development over time. So here's the data on the clinical and parent-caregiver Global Impression of Change questionnaires, which were used to rank overall MPS II symptoms, cognitive abilities, behavior and physical abilities on a 7-point scale from very much worse to very much improved. Worsening is depicted in shades of red, and blue is depicting improvement. In Cohort A, all patients were neuronopathic and range between 5 to 8 years of age. Interim assessment of clinical outcomes by expert clinicians and caregivers suggest improvement in overall MPS II symptoms, cognitive abilities, behavior and physical abilities in 5- to 8-year-old patients. In addition to the scales shown below, we have also extracted excerpts from the comment field from the CGI-C, the clinician -- expert Clinician Global Impression of Change that was entered by the investigators into the database. These include comments such as language much improved, more complex sentences, now able to hold a pencil, more conscious of conversations surrounding him, physical features that are much improved, building much more complex LEGO towers, better stick-figure drawing, and use of new words and completing 3- to 4-word sentences, less aggressive following directions. In summary, this data suggests that in Cohort A patients who are greater than 5 years of age, assessment of clinical outcomes suggest improvement in overall MPS II symptoms, cognitive abilities, behavior and physical abilities as assessed by an expert clinician and parent-caregiver. So in summary, the data from the ongoing study of DNL310 in Phase I/II supports a well-tolerated safety profile of DNL310 at doses ranging from 3 to 30 mg per kg weekly. DNL310 demonstrates efficient blood-brain barrier crossing and durable reduction in CSF heparan sulfate and dermatan sulfate over 6 months with sustained normal levels in all 15 patients in this data cut. DNL310 effects on GAGs and lysosomal lipids was -- were also observed at all dose levels, including 3 mg per kg weekly, demonstrating activity on lysosomal function. In addition, global impressions suggest improvement in symptoms as assessed by an experienced clinician and caregiver. The utility of neurofilament in MPS II requires continued investigation and additional natural history study data given the variability observed and the marked increases in neurofilament observed in patients rolling over from the natural history study on to the Phase I/II study. Our safety profile is consistent with standard of care enzyme replacement therapy in MPS II, with safety data now up to 43 weeks of dosing. Based on this data, we look forward to enrolling in Cohort C, designed to further explore clinical endpoints including behavior and cognition and an age range for which treatment effects on development milestones may have the highest likelihood of impact to be observed. In addition, we are accelerating our activities towards a registrational study to start in the first half of 2022 to demonstrate patient benefit in both neuronopathic and non-neuronopathic MPS II. And with that, I'll turn it back over to Ryan for conclusions and to start the Q&A.

Excellent. Thank you, Carole. So I'd like to summarize and we'll dive into the Q&A here. So in conclusion, I think, importantly, we showed that one of the hallmark biomarkers of Hunter syndrome, specifically heparan sulfate, we saw a robust and sustained reduction, including normalization. I think, importantly, even at low doses such as 3 mg per kg, we're seeing normalization. I think this highlights the power of the Transport Vehicle platform and further validates this platform as we use it across other modalities. We also show, for the first time, global impression scales, improvement in cognition, behavioral and physical function, including some exploratory biomarkers such as lipids, which are downstream of heparan sulfate. The safety is consistent with standard of care. And then what does this mean in totality? So for us, we're now accelerating the Phase II/III study to begin next year. And we're further building out our Transport Vehicle franchise with a focus on the enzyme Transport Vehicle in addition to other Transport Vehicle-enabled programs, highlighting that we'll be bringing at least 2 more Transport Vehicle-enabled proteins to the clinic in the next 6 months. I think with that, we'll take questions. I see that there's a number of questions. Great. Excellent.

So maybe we'll start here. Let's see. Let's start with questions related to the infusion-related reaction. So I think there's about 3 questions related to this. So how do the rates of IRR compare to Elaprase? And Carole, I'll hand it to you.

Yes. So as noted, we believe that our current profile is very consistent with standard of care, where, per the Elaprase label, there is a 57% to 69% rate of hypersensitivity reactions that are observed that are in line with what we've observed. In addition, there are severe hypersensitivity reactions involving more than 2 to 3 body systems that have been reported in Elaprase in approximately 15% of patients. I think just in terms of our IRRs, these were managed by standard infusion-related medications for IRRs. And so again, we feel that this is very consistent with standard of care.

I think we'll continue on this theme and maybe just ask 1 or 2 more questions related to IRRs. So how do we gain confidence that these IRRs are not related to the Transport Vehicle but rather related to Elaprase or to IDS?

Yes, it's a great question. And I think this is something that we've observed and looked at our data very carefully. And I think the best data really that gives us confidence is that the patients in Cohort A remain in the study and at 30 mg per kg. So these are the patients that have the longest dosing up to 43 weeks. And as I've noted, we actually see a decrease in these infusion-related events over time. Similarly, in Cohort B, those individuals that have had infusion-related reactions continue in the study. And overall, we see that these are very manageable, very much like what you see with standard of care therapy and management of ongoing infusion-related reactions. I think maybe the last thing that I'll say is that just in terms of, very importantly, looking at the durability of response of reduction in CSF HS, if infusion-related reactions were related to antidrug antibodies that, for example, decreased the efficacy or the -- of the enzyme, we would expect to see changes, and we don't. And I think very importantly, also, with peripheral GAGs, urine total GAGs are constantly monitored in the clinic by physicians. And then as well, we have our mass spec very quantitative assay that has shown no increases in these GAGs over time.

There are several questions around ADAs. Any new understanding of ADAs to 310 or other related effects that we can speak to? And maybe I'll just ask 3 questions in a row because I think they'll be related. So for patients previously on IDS treatment before entering the study, were any positives for antidrug antibody -- neutralizing antibodies? If so, how did this impact biomarker reductions from the treatment? I think you've sort of highlighted this already, Carole, that there was no effect on the biomarkers. And could you characterize the presence of any neutralizing antidrug antibodies? How does this compare to rates observed with pediatric patients treated with Elaprase? So those are 3 questions around ADAs.

Yes. So ADAs are quite common in this patient population. So yes, there were patients with pre-existing antidrug antibodies. And we've mentioned this previously. So for example, in Cohort A, there was one patient that, as you may recall, had a slower reduction in their decline of CSF HS. But very clearly, that patient has normalized at this time. And so I think the ability to dose higher certainly allows us to dose through these ADAs, and that's something that we're evaluating further in our Cohort B where we have more stable dosing across the 3 doses of 3, 7.5 and 15 mg per kg.

And I think I'll just add to this is that this is heavily influencing our ultimate dose selection as we go forward to the Phase II/III. And the ability for us to be able to dose higher is critically important because, obviously, ADAs are par for the course with enzyme replacement therapies. And we've definitely seen a correlation between preexisting ADAs and the ability to reduce heparan sulfate. So for example, if you look at patient in the B1 cohort, the one that didn't respond immediately, had higher preexisting ADAs. Very similar to Cohort A. So that ability to have a higher dose is critical. So let's turn our attention to anemia. So a question on do you still believe that anemia events are blood draw-related.

Yes. We absolutely do. So the patients in Cohort A that initially had reductions in hematocrit, we follow those patients very closely and also actually made protocol modifications to reduce the blood draws that we had over time. And as you may recall, also in the early part of the study, there were multiple dose escalations. So Cohort A went from 3 to 7.5 to 15 in a short period of time. And every time we did that, we needed to actually draw blood to monitor for PK, ADAs and other endpoints. And so what we've seen is that these patients over time, despite dosing up to 30 mg per kg, have essentially either resolved or near resolved the anemia back to baseline levels. And so we're quite confident that this is not related to the Transport Vehicle technology.

Okay. Some questions now related to neurofilament and the high variability. And the following question, what do you know about variability of neurofilament in different diseases and the time course of change is dependent on the pathology of the disease and, say, Hunter syndrome versus MS where you have relapses? And what does this teach us about the value in Hunter?

Yes. So that's a great question, and we're certainly learning as we go along the way. And I think we're a bit surprised to see the variability and then also the increase, the marked increase, that we saw in the natural history data. And as noted, we're the first to demonstrate effects or the increase in neurofilament even in this disease. So we're really characterizing this and sharing this data as we go along. In terms of variability, there is variability, I would say, that is seen across multiple indications, ALS, AD, that we have profiled neurofilament. But I think that -- and there's also published data, for example, in COVID that they've looked at neurofilament and MS. Let's say that over short periods of time, there can be quite a bit of variability in neurofilament, which certainly makes it challenging to look at the effect of neurofilament as a treatment response biomarker. I don't know, Ryan, if you want to add to that.

Yes, I'd love to add to that. It's interesting as the data that we published on neurofilament showed a very widespread neurofilament in both serum and CSF. However, that was just a cross-sectional look at neurofilament. What is, I think, most interesting to us is that within a single patient you're seeing high amounts of variability in short periods of time, including this increase at least in the 3 patients that subsequently enrolled in the treatment study. Now I don't know how to compare that. People have generally thought that neurofilament was stable in some of these other diseases, such as ALS. I mean stable within reason. But obviously, it makes it difficult to interpret, especially with such few patients, these data, with that type of variability. So let's stay on the neurofilament questions. Is there evidence that heparan sulfate kills neurons slowly? So removing heparan sulfate might have a long tail of neurofilament fall off, whereas the modest increase you see more consistent with partial correction likely implicating PK in deeper regions of the brain? So maybe I'll take this one, Carole, and you can add to that. We're just really careful about not over-interpreting the data from this marked increase in patients to the modest on treatment. That being said, the other really interesting observation is that Tau doesn't really change, and it's near normal levels. And a lot of times, you're seeing this relationship between Tau and neurofilament, which are not seen in MPS II. So I think just with this variability, we're going to need to get more data to imply something, as asked in the question, that this is actually evidence of deeper brain penetration. I think our strongest evidence of brain penetration is obviously the robust effect on heparan sulfate but also the downstream lysosomal biomarkers, which are fully corrected. It's worth noting that we actually never achieve that level of correction in the animal models that we're seeing in humans. So when we looked at brain levels or CSF levels on animal models, probably the maximum reduction we had was about 70%, and it didn't return to normal levels. So in this case, I think that my conclusion is that there is a pretty big capacity for transport in the human brain. There's 400 miles' worth of blood vessels, and using transferrin receptor as a very efficient way to get across the blood-brain barrier. And just recall that for every capillary, there's an associate -- for every neuron, there's an associated capillary. There's only 1 or 2 cell body distances you have to travel once you've crossed that capillary. I think related to this -- and here we'll ask this question around the preclinical data which, again, I think is very challenging to compare, but we definitely have some insights. So how do you believe the preclinical Hunter mouse data is translating for exploratory biomarkers such as neurofilament and clinical outcomes? The patients in the study, with the exception of 1-, 2-year old are a bit older. If you were to start treatment in younger-age patients, can you speculate what neurofilament levels would look like over time, perhaps according to what preclinical Hunter models data show that's, of course, our own data? Was there any notable stabilization or reduction in neurofilament for a 2-year-old patient? It's a great question. Carole, I'll hand that to you.

Yes. So these are great questions. And I think going back to also the other question about variability in neurofilament and over time, I would say that in pediatrics, as compared to MS, we have much less data on what the normal trajectory is of neurofilament. I would say that probably the data, probably the best pediatric data comes from SMA1 and SMA2, 3, but their -- the effect or the correlation between treatment effects are really strongest with SMA1 and not so much with SMA2 and 3. So I think just as we look at these biomarkers as exploratory biomarkers, I think we really have to understand more about the trajectory. So now going back to the question around starting early, we do certainly think that starting earlier will have a greater impact on the disease, simply because at that earlier stage you have lost fewer development milestones and so there may be more of an opportunity to have a greater impact. As far as neurofilament, I think it's still just early days because even the natural course of neurofilament may be very different in 2-year-olds. We know that neurofilament actually starts out much higher, then there is a period of active pruning and remodeling in the brain that then the neurofilament starts to go down, and then it increases slowly over age in normal individuals. But I think there's this dynamic period during development that we don't really understand what's happening with neurofilament, and when there are changes in the neurocognitive status with treatment interventions, how this affects the remodeling of the brain and, therefore, effects on neurofilament. So it's maybe a long answer, but I don't think we know right now how treating a 2-year-old from a perspective of neurofilament changes we could anticipate what that would look like. I would say and the reason that we're enrolling in Cohort C is that from a clinical perspective we're even more likely to see effects in the younger age range. And given the exploratory data in Cohort A with regard to the global impression scales, we're very enthusiastic to explore further in younger patients.

Yes. I think that answers another question based on some of the other presentations at MPS 2021, obviously, ongoing today. There appears to be an emerging trend of better responses with earlier treatment. While it may be early with Cohort A -- or it may be early with Cohort A subjects, any thoughts on how this may apply to DNL310 treatment? What can we do to facilitate earlier patient identification? So I think, Carole, in some ways you've answered that. But maybe just to reemphasize, it would be worthwhile. Yes.

Yes. Well, I think this is something that we're really enthusiastic around engaging the community. And in terms of newborn screening, it certainly would help identify patients earlier. In some states, that is now being added to the newborn screening panel. But for the most part, it is not currently the standard of practice to have newborn screening. I think just as we and others in the field are developing therapies for neurocognitive symptoms, there is certainly awareness. And we're hearing every day from the patient community of individuals that are being diagnosed earlier, and particularly siblings where one sibling already has the disease and the parents are then seeking diagnosis and treatment for the younger sibling, this is another opportunity to enroll earlier patients.

Yes. I definitely think the mantra earlier is better. I will say that we were very excited about the data that we saw, the clinical data, treating these more severe patients in Cohort A, albeit a small number of patients. With the understanding of the hazards associated with inter-study comparisons, can you comment on how the 6-month DNL310 data may differ from JCR's IDS TfR molecule, JR-141, 6-month data? So I'll be happy to address that. And Carole, you can add to this. So we actually -- at the mechanistic level, we compared a JR-141-like molecule to DNL310. This paper is actually available on bioRxiv. We presented this data at WORLD. It's really a question around architecture and brain delivery. But I think it's really worth noting in that paper, looking at both brain and CSF reduction at different dose levels, the high differentiation of the Transport Vehicle-enabled IDS or ETV:IDS relative to the JR-141-like molecule in terms of architecture. We believe this is related to the affinity and when you're bound tightly the transferrin receptor not readily crossing the blood-brain barrier. In addition to that, although the assays are different, you can look at percent reductions, right? And we, of course, have internal control data. Now thinking about the clinical data comparing it, you ask what are the normal levels of a healthy individual versus MPS and then how does the treatment effect. And when we're seeing 90% reduction and normalization, that's in contrast to maybe 30% reduction early on and maybe 60% max reduction. And so there's a very clear difference, I think, both how rapid the response is. And I think to add to that, even at doses as low as 3 mg per kg, we're seeing a normalization in at least 3 out of those 4 patients. Again, the 4 patients having higher ADAs, which we can dose at a higher dose. So I don't know, Carole, if you want to add to that. But...

Yes. And I think that summarizes it very well. I guess I would really just want to emphasize, again, the clinical data where we see a larger magnitude of reduction and this normalization to the normal levels as opposed to, say, non-neuronopathic levels. And then I think just in terms of clinical data, we also are very pleased to see that CSF dermatan sulfate is also substantially reduced. And then just in terms of the overall activity of the enzyme and penetration even to peripheral organ systems, we shared the CRM data across heparan sulfate, dermatan sulfate and keratan sulfate, which demonstrates substantial reductions as well. So we think that this suggests that the binding profile, as Ryan outlined, of our transferrin receptor binding and the Transport Vehicle is clearly differentiated.

Let's turn to the Phase II/III. So how do you think about powering a Phase II/III study now to show a neurocognitive benefit? How long does follow-up need to be and how large a trial? And what is the basis for the assumptions that lead you to that design? And related to that, Phase II/III, have you met with the FDA yet? What is the gating for starting the Phase II/III at this point?

Yes. So we've certainly been thinking a lot about the Phase II/III study. We have met with global regulatory authorities to discuss this. I think as many know, the regulators have spoken at national meetings, including WORLD, and they've indicated that the FDA would like to see a randomized study compared to standard of care, Elaprase. And so that is factoring into our thinking around the size of the study and the duration of the study that would be required to see a treatment effect. In terms of looking at progression of neurocognitive symptoms, we would anticipate needing to study these patients in the randomized study for the 1- to 2-year time frame. But that being said, we're certainly continuing to push forward to explore in Cohort C with open-label data to better understand these cognitive and behavioral end points and start to understand whether we can see clear evidence of clinical impact even before that when you complete the Phase II/III study.

Let's turn our attention back to neurofilament and a series of questions. Hopefully, we can address these pretty quickly. Do you have a view on how neurofilament may be impacted by age? Could there be a different outcome in under 5?

Yes. So I think we addressed that question earlier in that we don't really understand the natural history of pediatric neurofilament and how -- not only do we not have a very clear picture of a normal development, how that changes. We do know again that it starts high, then goes down. And then as people age into their decades, it increases again. But we don't know how it would be impacted by treatment paradigm. But again, we do think that in this younger age range, we're more likely to see clinical changes.

There also could be a difference in the nature of the IDS mutation. I think it's really fascinating when you look at our data, the 5 patients that there is a real spread between the different levels at baseline of neurofilament, including one patient that's relatively low and stays low but has deemed neuronopathic and saw cognitive benefit. So there seems to be a disconnect there. So it may be as much the nature of the mutation as it is the age and timing of intervention. Okay. Can you comment on the clearance for neurofilament? Do you have enough data to suggest you are slowing neurofilament growth versus pretreatment? And I think maybe I'll answer this. We're not claiming anything actually, that we're slowing the reduction. We're showing the data as is. We did not have enough data to be able to claim that. That is one interpretation of our data. But it's a very good point that there's a balance between production and clearance, and neurofilament has a relatively long half-life. And what we can tell you is that it's variable between patients and within patient, and that's the data that we have. Why does neurofilament increase in the serum in the natural history of MPS II patients but is variable in the CSF? What might explain that seeming disconnect? So I should make this really clear. And then, Carole, you can -- if you want to add to this, we only have natural history data for serum. As you might expect, yes, it requires anesthesia to take CSF. And in these observational studies, we're not taking CSF. So the data that you see is just in serum, pre and post, and then just in CSF on treatment. And I think that answers that. So there's really not a disconnect. It's just that you don't have the comparison.

Ryan, maybe I'll just add to that they may have been referring to cross-sectional data, which we do have cross-sectional CSF neurofilament data that we showed. But these are the samples that were taken opportunistically before we enrolled our natural history study. And obviously, the limitations of cross-sectional data is exactly why we started enrolling this natural history biomarker study? And so this was really -- for us, it was a learning to see that in this longitudinal data there was this marked increase over time which was not apparent that we would see that from cross-sectional data.

Another question. Did -- has there ever been data showing CSF neurofilament or other biomarkers in CSF with standard enzyme replacement therapy? How does DNL310 compare to those? I think the answer is no. There's no existing data that we're aware of besides the data that we've generated. So obviously, the data we generated in the -- in our manuscript is on standard of care enzyme replacement therapy as well as stem cell therapy. So that's the data. It's cross-sectional but not longitudinal. Okay. So okay, another question here. It seems like the plan for Cohort B may be changed a bit. It used to involve dose escalation versus 3 parts. Why the change?

Yes. So we're continuing to learn as we go along in this study, and the changes in Cohort B are to better understand the difference between 3 mg per kg, 7.5 and 15 mg per kg.

I would say that...

You can see we're seeing clearer activity at the lowest dose level.

Yes. And that's the reason for the change is that, because we have activity at all 3 doses, we decided to stay on dose longer, essentially stay on the lower doses longer. Related to lysosomal biomarkers, so GM3 data suggests lysosomal correction is incomplete. Do you believe that result might arise from the intermittent dosing? Will you explore other dose approaches with your improved enzyme? Explain. Examples might include gene therapy or cell therapy. So maybe I'll comment on this. I actually don't -- I mean, you can look at the data, it's either normalized or near normalized. So I guess I wouldn't interpret it as incomplete. And interestingly, even at the 3 mg per kg, we're seeing normalization in GM3. The reason that you may not see it completely all the time is, again, probably some variability in GM3, is the way I would interpret that.

I would agree with that interpretation and note that I think where there may have been a question of incomplete, we're also looking at Cohort B data with time points as short as 7 weeks. We already know from our Cohort A data that the effects on GM3 appear to follow the reduction in CSF GAGs, which happens much earlier, which would be logical that you would see the reduction first in the substrate then the improvement in lysosomal function.

Okay. So here's a question. Any comment on time course for removal of standard-of-care peripheral treatment -- mean, median and by dose? And maybe this just requires a little bit of clarification. So for DNL310, there's a switch from standard of care to DNL310. And so basically, they're no longer on standard of care. Basically, DNL310 treats both peripheral and central. And I think it's an important point that we're seeing enhanced activity in peripheral biomarkers in addition to normalization in CSF, heparan sulfate and robust effects on lysosomal biomarkers. So the idea would be that DNL310, we're basically developing it to replace standard of care. Okay. So again, on Phase II/III plans in neurofilament, how do you think about setting inclusion criteria with respect to baseline neurofilament levels for the Phase II/III trial?

We would not use baseline neurofilament levels for inclusion. I think there's really no data at this time that clearly demonstrates any sort of prognostic or predictive utility of neurofilament as a treatment response or a prognostic biomarker.

Okay. So a question here, is it possible that for CSF neurofilament, there is a very variable but dramatic higher progression and increase over time in the natural history for patients on either sulfates? If so, might a much more modest increase in CSF neurofilament, and I would add serum, seen on DNL310 represent improvement? And so I think as we commented, that could be one interpretation. I think we don't have enough data to conclude that certainly. And then I think related to this, could we elaborate more on sort of our thoughts or our hypotheses on the neurofilament data? I don't know, Carole, if you want to summarize your thoughts or hypothesis on the data.

Yes. I think I've mentioned most of these points previously, but I think the variability that we see was surprising in the natural history data and I think, again, just speaks to the importance of having longitudinal data because these are pediatric patients, again, that are enrolled at different ages. And so baseline values are very different. We knew that actually already from cross-sectional data. But you layer on the variability on top of that, it's really -- I think it becomes very challenging to make any sort of conclusions at this point. I think this is something that is going to take additional data and quite a bit more data, including natural history data, which can be very challenging to get. I think there are other programs that have samples that would be great to understand more about these neurofilament levels in well-characterized data sets. But I think until that time, we really can't make any conclusions about the utility of neurofilament.

So in the next 4 minutes, we're going to do some rapid fire Q&A here. So where do you think DNL310 could fit into the treatment paradigm for MPS II relative to a onetime gene therapy? Maybe I'll answer this, and Carole, you can add to that. So I mean, at this point, what we're seeing with gene therapy is a highly variable response when you look at, for example, heparan sulfate in the CNS. And also gene therapy targeting the CNS is different than gene therapy targeting the periphery. So I mean we view that there's certainly a major role for enzyme replacement therapy until you can have sort of absolute rescue. Our goal is exactly that, to absolutely rescue both in the periphery and in the central nervous system.

Yes. I think I'll just add to that, that someday, the vision would be, for patients, that there's a gene therapy that's a onetime administration and it addresses both the central and peripheral manifestations of disease. At this point, I think until that happens, this really allows patients, if this continues to demonstrate the profile that we've seen, to get weekly therapy that addresses both. So in certain gene therapy approaches, there may not be complete efficacy in the periphery. And therefore, patients may have to end up getting 2 therapies, one for the periphery and then the gene therapy, which would have addressed the central nervous system, where this would address both.

Given the degree in variability in neurofilament you observed in Cohort A patients, how are you thinking about the pooled Cohort A/B data? Do you still think there is a potential to observe meaningful reduction with additional time, longer number -- larger numbers of patients?

Yes. So I can answer that. I think 6-month data, even with more patients based on what we've seen in the natural history data variability and increase over time, is not going to be informative. I think as we've seen in other therapeutic areas, specifically, maybe I'll mention in lysosomal storage disease, for example, CLN2, it may take a long time before you can see these changes in neurofilament. And I think based on our data and what we've characterized so far, again, very limited. I think our focus is going to be more on our lysosomal biomarkers, heparan sulfate reduction and clinical outcome measures.

I think I'll just add that the CLN2 data, the levels are actually substantially higher relative -- in terms of full difference than what we're seeing in MPS, this is related to that. So based on what we know about the pathophysiology of MPS II, do you think GAG levels are more important than neurofilament? Is there a significant neuronal loss in this disease that drives the pathology? So I think the last question, it's a great question. We actually don't know if there's significant neuronal loss, and that's obviously why we looked at Tau. If there was significant neuronal loss, you would expect that Tau would be elevated as it is in other diseases. And so it may be possible that what you're seeing in elevated neurofilament is some type of neuronal dysfunction or that it's within error in terms of variability. We most certainly think that GAG levels are the most critical biomarker. It's what's correlated in the periphery with benefit. It's what we believe will correlate in the central nervous system as well. And although sort of perplexing neurofilament data, the totality of the data very strongly supports that DNL310 is crossing the blood-brain barrier, rescuing lysosomal function and is translating or should be able to translate into clinical benefit. With [ hindsight ] that's certainly the case. So to summarize, we're very enthusiastic about where we are with DNL310. We're expanding the Phase I/II study, obviously with Cohort C. We're able to keep patients on the lower dose for longer because we have normalization even at lower doses, again, sort of illustrating the robustness of the TV platform. And so we look forward to sharing more data as we go forward. So I think with that, we'll thank everyone, and we look forward to connecting more on this data. Take care.

Great. Thank you.