Ultragenyx Pharmaceutical Inc. (RARE) Earnings Call Transcript & Summary

Okay. Good afternoon, everybody, and thank you once again for joining us for the 41st Annual Cowen Healthcare Conference. I'm Yaron Werber, biotech analyst here at Cowen, with my colleague, Brendan Smith, and my team. And it's a great pleasure for both of us to moderate the next session with Emil Kakkis, CEO and President of Ultragenyx. If you have any questions, by all means, you can either throw them into the Wall Street Webcast, and I have it open here, and we'll read them anonymously on your behalf. Or send them either to Brendan or myself, and we'll be happy to take them as they come in. Emil, great to see you. Thanks for joining us.

Thanks for having me, Yaron. Happy to be here.

So let's start with -- we want to run through GSDIa, then touch on OTC, then touch on Angelman and then probably close it up with osteogenesis imperfecta. For GSDIa, the updated Phase I/II, I think there was an update in January, showed a 77% reduction in cornstarch across the first 3 cohorts at 1 year and more than 90% reduction in the 3 patients in cohort 1 who are 2 years. And then you also showed a nice impact in continuous glucose monitoring and how much time people are spending in euglycemia and also a reduction -- about 12% reduction in body weight. As you now are moving into Phase III in the first half, it looks like cornstarch intake and maintenance of glucose control are going to be the co-primaries. Can you give us a sense how did you design the study? How do you power it? Which ones are secondary end points? And the biggest question is using that proof-of-concept Phase I/II data, how do you actually power Phase III.

Sure. Well, we -- one of the things we -- in rare disease drugs, Yaron, you need to really be ready to learn from your first study and not get fixated. While there had been emphasis on looking at time to hypoglycemia, time to hypoglycemia has some complexities in interpretation sometimes because of the hormonal issues going on. But what was very powerful in the Phase I/II study is how fast their starches need to be reduced, especially if we're using continuous glucose monitoring, because we saw within days after getting the gene therapy that glucose levels were soaring as their own body started releasing glucose, and so we would have to reduce their starches. So in some ways, it became clear to us that the thing that's the best replication of what the liver is doing for glucose release is figuring out how much starch you have to keep giving. It's just a more direct measure of how much you're relying on your liver for glucose versus oral starch. And so in some ways, that is more at the core biochemically what's going on. The second thing is patients don't really measure time to hypoglycemia, but they do know and they measure every day how much cornstarch they're eating. And if you talk to patients, the burden of the cornstarch, how often they have to take it, how much they have to take, how they feel after they take it, and the fact they get -- they gain weight because of their taking it, they almost become like diabetics in the sense that they're taking so much sugar all the time. I don't know if you have eaten a lot of birthday cake. But occasionally, if I eat birthday cake, I feel bad afterwards. I don't know about you. I'm at the age where you start feeling it. The truth is they feel like that all the time. They're on -- sugar overloaded. So we learned also that cornstarch sugar is not only a very powerful changing parameter, but it is actually one that's clinically meaningful direct to patients. That's what actually affects them every day. And we've had great feedback on that. And so we sort of learned and flipped and realized this is at the core both biochemically and clinically what's going on in these patients. It's what they have to do to survive. It's the drug that keeps the gun from -- that's at their head every day from going off. And if they don't have to do it, then it changes things for them. So what became implicit though in all of this is that if you're going to change that parameter, you ought to show that you're doing it while maintaining glucose control, right? So the second end point is saying, look, you can't reduce starch and be hypoglycemic all the time. That is not okay. So it's about creating a context for cornstarch reduction. It's cornstarch reduction while maintaining good control. And we believe, over time, as their metabolism gets adapted to the new state, that the highs and the lows will get better. A lot of the patients are actually running high all the time. In fact, they're not having lows, but they're running high all the time. And so -- which isn't a healthy way to be. So we think we need to put that together. That is a new end point. Both things are new. The agencies never dealt with this end point before, and it's different. And obviously, you and I looking at, well, cornstarch sounds kind of low tech, right? But sometimes your own -- the best end points are low-tech obvious ones, right in front of you, right? And I believe it is the right one. So we have the data to support that, both from patient interviews and evaluations in our EMA meeting. In fact, the patient representatives, they had told EMA that -- in fact the CHMP that that was the right end point, that they agreed with the company on how important that was. So the FDA is kind of doing something similar. We'll get that settled, we believe, here and we're -- we believe we'll get the support for it. With regard to power, the treatment effect size is very large and even with any kind of modest effect on the placebo group, we believe we're very well powered because the changes, at 57% by 12 weeks, is so large that we feel very, in fact, overpowered in the sense that we can detect probably smaller declines. So even with conservative, we're more than 90% powered to see a difference. The CGM part, we're looking at more like maintaining control rather than being superior to control. We will look to superiority as a secondary end point, right? The thing with the CGM is that it takes time for that to kind of settle in as your body adapts to the low-starch environment. But we think the CGM just provides the right context for interpreting the starch data.

Okay. So the secondary end points will be actually looking at superiority on those primary end points?

Well, we'll look at CGM particularly, superiority, that you have better control than you did before. So you spend a higher percent of your time within the 60-120 range. And we've shown -- that's the data we showed. You mentioned at the beginning of the segment, Yaron, was that the time spent in the range tightens, a lot of the tightening has to be because they are hyperglycemic, in fact, a lot of the time, right? But they're doing that to keep themselves from being hypoglycemic, right? But whatever you're doing, it's still out of control, not in a good control range. So we think that will tighten. And if you look at the patients further out who've gotten monitors now, they have really good, steady control now, and their insulin levels have fallen finally. And their cortisol levels have come back into the normal range. So you almost see, at the second, third year out, their metabolism is starting to adapt to the new world. And your body doesn't adapt -- they've been -- some of these patients were on starch for 20, 30 years, right? I mean you can imagine that the physiology of your pancreas and beta cells and everything else doesn't exactly flip back after 30 years of starch into normal for a little bit longer. So if you talk about durability of the gene therapy, when you look out 2 and 3 years, the patients, they get better and better as time goes on. So we actually think it's not only durable, it actually looks better when you look at the 3-year-out patients.

And from a reimbursement perspective and outcomes, what are the important outcomes? I mean cornstarch intake is, that's one end point. But I imagine the maintenance of glucose control and continuous glucose monitoring is probably going to be more important for reimbursement with some longer-term outcomes.

Right. Well, we're also having some functional scales in the secondary end points with some other supportive PROs in the other. So there will be supportive information. We'll also look at frequency of dosing as a way of measuring inconvenience or impairment. But starch reduction itself, starch is cheap. So you could argue, well, that's not -- losing starch doesn't really matter. But you have to understand that it's the quality of life of feeling that you have a gun to your head. And if you talk to patients who've been in the trial and we did the surveys, the fact that they don't have to take the starch all the time -- in fact, some -- most of them are off daytime starch completely. Just get a little at breakfast, nothing all day and they have a little bit at bedtime. It's not even clear to me that they need that, frankly, right? If you're going all day long, then clearly your liver is making enough glucose. So I think it's more of a habit than anything. But for them, they don't have to bring starch with them all day. So the quality of life stuff is very important to them. Plus they tell us, when they exercise, they used to have to be worried about whether they're going to go low or not. But now they don't. They don't have that problem. They don't have problems where they have feeling good, feeling bad, feeling good kind of during the day as their sugar goes up and down. So we think that those things will play out both in the functional score, which is the secondary end point, as well as some of the quality of life stuff, which will help us in reimbursement. So we feel comfortable there should be enough power in those end points to give us a meaningful result for reimbursers.

Okay. Great. And so let's move to OTC. That program is also on track to start later on this year into Phase III. So the prior data showed 6 out of 9 were responders, including 3 complete responders who normalized their diet and discontinued alternative pathway meds. They were still maintaining their ammonia control. And then 2 of 3 patients in the current steroid prophy cohort have been treated. I don't think we've seen the final data on those. So your thoughts on -- it sounds like you're moving toward using prophy steroids, right? Is that you're thinking on that study?

Yes. For OTC, we'll use prophy steroids that start at the time of drug administration. We won't do the third. We did -- we were planning 3 patients, but then the third patient got hung up with -- wherever they were, the COVID, it was in Europe. But we decided we'll just go with the 2 and see how they do. We'll put out some data later. So far, everything has been fine. So our main point was to make it convenient and easy to apply. And it could help improve potency. But at this point, it was more about making it practical. And if it helps potency, then that would be great. But we'll get a little more information on those 2. But so far, it's gone fine.

And with GSDIa, you'll wait for your prophy data and then potentially put -- introduce that into the Phase III later on?

Yes. We have 3 patients that did get prophy steroids before. We're trying to figure out -- and GSDI is a little more complicated because of one reason, that the transgene is actually steroid-induced. So if we start at the beginning, it's also inducing the transgene immediately upon the gene therapy. So we might do something where we just do a delayed prophy where we do the dosing, and a couple of weeks in, just start them on steroids. What we did in cohort 3, which worked pretty well, was we waited for their baseline LFTs to come up, and that comes up at about 3 weeks, before we started the steroids. But we could just start the steroids right before that point would have normally happened. The reason is that we don't want to induce the transgene too early, too fast. There already is a surge in starch, and so it has a little complexity there to the GSDI story. But the 3 patients have done well, and they produce their starch and are doing well. We'll put out more information on them. But it is a little bit more complicated. By the way, it's a very cool and special thing, Yaron, that we actually have an inducible transgene. And I don't think anyone else has one. Like we can actually give our patients a little bit of prednisone to actually induce the transgene expression, right? I think that's a nice feature, actually, and it really does seem to work. Like we have the data already showing it really does induce the transgene, so -- which gives me actually more confidence that the transgene is where it needs to be, right? So it actually responds the way we expect.

And so with OTC, I know you're still in discussions, but as you think about what would a Phase III -- what could it look like? And what would be the primary? I think -- Emil, I think you might be on mute.

It flipped on mute by itself. We're planning to 50 patients just like the -- but 1:1 randomization, instead of 2:1. The 1 -- people do 2:1 because they want to be nice to the patients, to have more chance of being on drug. But the truth is, it's a fundamental truth, which is that you lose power. And so we're going to go 50 patients 1:1. The truth is that even though we added only 11 more patients, Yaron, that actually doubles the power, so -- by having it 1:1 randomization. So we're going to do 1:1, 50 patients. Based on the FDA's requests and our feedback from scientific advice, ammonia level will be a primary end point. How we manage that will be different, obviously, for some patients who have normal ammonias versus high ammonia at baseline, right? There are some differences in how you manage it. But the idea was to show good or improved ammonia control, and we would hope to show substantially improved ammonia control, especially if they have high ammonia. At the same time, we'll be looking as well at reduction in the need for the scavenger medications as well as diet control as other end points. The ureagenesis assay will include the secondary. It is a biomarker end point. I do believe it's useful. It's still an experimental assessment, and it hasn't been proven in the regulators' eyes at this point. But we do think it's helpful in trying to assess whether the biology is right, the drug is working. But we'll make it secondary at this point and include certain other clinical end points, including a kind of functional score for OTC, which relates to how well they're thinking and performing and as well as some other end points around crises and such. We don't expect these patients to have any crises. So we can't really look at hyperammonemia crisis as a parameter. It's just not enough events to be able to conduct that evaluation effectively. But we'll include it because it's important.

And how long do you think that -- and the primary is going to be at 52 weeks on that study?

48 weeks, yes.

48 weeks. And is that the same with GSDIa as well?

Yes.

Okay. And so...

Yes. The natural question is, will the regulators want more than 48 weeks, is that what your question is?

Yes. And also, how long do you think enrollment will take?

The regulators are concerned about durability. I think the data we have from the first set of patients will help support durability. So I do think that we'll be -- we'll find a way -- it will be acceptable. But if there's any loss of effect in the Phase I/II patients, then it would put the Phase III patients then in the line of fire for proving durability. So far with -- we've had what we thought was excellent durability in the first couple of patients that had -- that were successfully treated. They've done well. They're living well. They're not on diet or drugs at this point. So the question of enrollment. Most of the time in enrollment is getting the sites set up, frankly, right? It takes -- get through the contracting, all that process. So normal assumption is that it takes a year to get enrolled. We would want us to go faster, and we're trying to press the team to get it done sooner than that. But it's mainly because it takes months to get sites up and then put patients in play. And because of translations and regional regulatory requirements, which have only gotten worse -- times worse because we do have potential COVID lingering impacts going on in the year. So it's not going to be all straight smooth sailing. And a lot of countries are having to deal with a lot of other issues that are not trials, right, first and foremost. So I wouldn't press for less than a year at this point. Although we're going to try for it, but I'm just saying to you from a timing, planning standpoint, I would assume that.

Great. Brendan, you want to move to Angelman?

Yes. Great. Thanks again, Emil, for joining us. So as Yaron mentioned, we want to touch a little bit on Angelman. And then if we have some time, get to OI a little bit towards the end here. So obviously, in Q4 last year and in December, again, at the hunting -- excuse me, at the Angelman meeting, you guys released some data and then updated it. Really encouraging stuff. You saw, basically an average of 2.4 points across all the treated patients on at least 2 to 3 -- or excuse me, 2 out of the 5 dimensions of the CGI-I-AS scale. But as you also noted that there were some extremity weakness that you pointed out, rightfully so. And is there going to be -- or I suppose you have already submitted some updates to the trial design and protocol. So can you kind of just remind us where you're at in that process? What are kind of the main updates that you have submitted or plan to take and implement and kind of what steps are left to getting alignment with the FDA on that? I think you're on mute again. Sorry, Emil.

It seems to be going on mute all by itself. That's not me doing it, so. In any case, it -- last December, we submitted a major amendment, which included a lot of information that the FDA had not seen. We've received some questions. We've been dealing with questions, requests, concerns. I think their most focused concern is around what is the nature of this event, really. And it's hard to do because they're not doing pathology here, right? We're looking at clinical findings. And the clinical findings, the MRI findings, in our mind, look like a very localized inflammation of the cauda equina, the tip -- sort of the end -- the nerves at the end and the meninges there. We're still working with the FDA to -- for them to understand that and the steps forward. What we proposed is starting at the low doses and going a very slow titration with more monitoring and with sentinels and so forth. We're still having that discussion with the agency at this point. We think we have a decent plan. We're also talking to authorities outside the U.S. as well. And we've worked with a number of them in the past that I think are very competent and effective and would allow us to open up the direction for the program. We're highly committed to getting the program going rapidly as possible and dosing. But the FDA has its own requirements, and we'll be looking at those. But we feel very confident that, at the lower doses and managing the way we would plan to manage, that we could reduce or prevent the incident of this -- of the safety event while still getting the efficacy we require. There's very few times in medicine where the safety event is so distinctly anatomically separated from where the efficacy is, right? You just don't get it that obvious, right? And usually, it's all mixed up. But to have something very localized and have the CNS effects be completely separate, it allows me to believe that you can -- we must be able to find a way to get through that. So we're confident. We'll continue our discussions with the agency. We're opening discussions with others, and our plan is to get the study started and treat some patients. And hopefully, later in the year, we'd get a handle on can we repeat dose at the lower range of what we tested before and achieve the kind of efficacy we saw, which we think, from everything we've done, including the animal models and the historical data, should be achievable. So we're excited about this. The community is excited about it. And we'll work internationally to make sure it gets going as promptly as possible.

Great. Right. So I think one thing about these patients that is pretty notable kind of across the board, regardless of who you speak with, is really the heterogeneous nature, kind of, of their symptoms and of the disorder as it manifests kind of full scale there. So when we look at something like the CGI-I scale, are there certain dimensions within the scale that you think are a little bit more amenable to treatment with something like GTX-102 or maybe would have more significant impact on physician responsiveness, marketability for the drug when we're kind of looking at functional data like that?

You're getting at a really important basic question is how do you prove efficacy in a heterogeneous disease? And the one thing I would say is that Ultragenyx, among all companies, is probably the best positioned to be able to deal with this question because we've been developing in a method, which we use in the Mepsevii program, called the Multi-domain Responder Index, which is another way of doing it. The CGI, by the way, is a way of letting the investigator integrate information, right? But the problem with it is it's kind of a nonspecific way. There's not -- in fact, it would still be better to use individual domain scores that are actual measurements of function, right, to score whether you're having a clinically important improvement, and then adding up scores in the communication domain, the behavior domain, gross motor, fine motor and sleep and seizures. So we believe the Multi-domain Responder Index approach is really the best approach. CGIs, I think, were accepted by the FDA and could be accepted. But often they lack the underpinning basis for the decision. And the instrument is, to some degree, subjective on what the doctor is thinking. I think it's better to establish an end point that's based on actual functional testing of what the patient can do quantitatively and to come up with a scoring system. The Bayleys has scoring system for expressive communication, which we showed the data on at December, which showed very nice improvements, both expressive and receptive communication, which both are important, like what words you say is one thing, but also listening to your name or following instructions, by the way, which patients had not done, were starting to do. Even the ones that hadn't spoken words were listening -- were responding to their name, which they had not responded to their name before. For most parents, that was actually a miracle first step that the kid responds to the name and actually turning to look at them, right? So I think expressive communication will be a core one and the Bayleys has a test. There's also the ORCA, which is a different test for communication, which is relayed from the caregivers kind of method. And that was validated for Angelman. Both fine motor and gross motor has elements in the Bayleys that you could use and there are potentially other tests. The [indiscernible] is another thing that looks at movement activity, which showed a nice improvement, but it's a little less well established. And sleeping is kind of easy in terms of scoring, how much sleep is going on. So -- and seizures or EEG could be things that you could look at quantitatively as well. We did show some data on delta power showing improvement in the patients, and you could use that kind of an endpoint. Our view would be that a Multi-domain Responder Index would be based on 4 or 5, potentially 6, clinical end points, right? And that the biomarker-based end points would be supportive information of the -- verifying or validating that the clinical improvements are based on biology, right? So that's the way we'd look at it. Now the MDRI has not been used in a major study primary end point. We've had multiple discussions with the FDA outside of this program, and they're very interested in it. P.K. Tandon, my Head of Biometrics, and I have written a paper on the topic, which will come out soon, and we're actively pushing the FDA toward understanding the value of it. There's a beauty to it and it's a beauty that is particularly important for the Angelman community, which is that we could enroll a broad array of patients in the study, right? You don't have to all have the sleep problem or the seizure problem or the language. You -- although almost everyone has the language problem -- or everyone does. The truth is you can enroll a heterogeneous group of patients. That's the value, that's what FDA was interested in. Instead of having a highly selected subpopulation that has the exact primary end point problem that you want to study, which all of us would say, well, what about all the other patients that don't have that problem, right? What about them? So we think this is what -- is the power of this is that we've broadened the enrollment criteria and improved the understanding. And that's why I think the MDRI is a paradigm-shifting approach to clinical development. And I think would be -- I think it will be the next-generation approach to establishing efficacy in heterogeneous populations.

Great. Great. Okay. I would love to get just one question on OI, unless you're hearing some other really pressing questions here that we should address. So just kind of let me know. So just really quickly for OI. Obviously, you announced a partnership with Mereo BioPharma in December. On the Q4 call, you kind of confirmed plans to start this Phase II/III dose-finding study later in the year really focused on bone production near term, and then maybe later term for a pivotal really on fracture reduction, bone mineral density. So just, I guess, really quickly here, can you kind of just help us understand the commercial opportunity in OI for the drug? And if there are some types that you're really planning to target? And maybe where you see this drug falling in kind of the treatment paradigm for [ these ]?

Sure. First, let me just say that the study will have a dose testing stage where we compare 10 to 20 to 40 mg per kilo, and we'll go seamlessly into the Phase III randomized set straight through with a decision made on dosing with a single dose arm that continues, and we will then expand. So that's the design. So it would end up becoming a pivotal study. Now the question about types, I think from the data we have already from Mereo's collected in the 90 patient study, they had both type OI type 1s, type 3s and 4s. Type 2s are lethal ones. So type 1s are due to deficiency in collagen. They have a -- many patients grow up, are not deformed but have a frequency of fractures. And so among high-frequency fracture patients, there would be potential benefit. There are some patients whose fracture frequency may not be enough. Type 3s and 4s, I would say, are the more physically deformed and really higher unmet need. But we would expect to have type 1s and 3s and 4s in the study. Type 1s that have a lot of fractures, Type 3s and 4s all have a lot of fractures. So that would be the types that we'd focus on. There are many other types, but I think those are the key types that are -- have been studied, and we think basically cover the majority of OI patients that are out there. So it's a pretty broad population. I think talking to PIs who are treating using Crysvita for XLH today, I've talked with half a dozen in the last few weeks, they usually have 2x as many OI patients like that than XLH patients, right? So our sense is that the true population existing in major clinics of key opinion leaders is double probably with what you see with XLH. So we think it's a larger opportunity in terms of both the types as the total population size than XLH.

All right. Kick it back to Yaron.

Well, Emil, thank you so much. I think we're just about out of time. So thanks so much for joining us. As always, really appreciate it. And we'll continue to...

Thank you, Yaron. Thank you, Brendan.

Thank you very much.

Stay safe. Thank you.