Sage Therapeutics, Inc. (SAGE) Earnings Call Transcript & Summary

Thank you for joining our live fireside chat during our Piper Sandler Conference. My name is Yas Rahimi. I'm a senior biotech analyst here at Piper Sandler. I am so thrilled to be featuring Sage Therapeutics. Joining me on Zoom is Barry Greene, CEO. We have quite a bit to cover in the next 30 minutes. And first, I'm going to start off on congratulating you on 2 great announcements this morning. One is the data set -- the 12 months data set from SHORELINE across the 50-milligram dose group. And the second one is the unveiling of the design of the SAGE-718 Phase II DIMENSION study. For investors who may have missed that new information, it is available on the website -- of Sage's website, and we'll talk about that.

But let's kick off the discussion with the new data that's just off of the press, the 12-month SHORELINE data across the 50-milligram dose scope. So Barry, maybe we'll start off to maybe help us understand like, what did we learn new this morning? That's one. And then the second one is put it into contrast efficacy and safety data for 50-milligram versus 30-milligram dose group with the SHORELINE data set.

Right, Yas. So look, thanks for having us, and thanks for the organizer, Piper, for having us here in this live event. We appreciate it. We were excited to have a couple of announcements this morning. But if I take a step back, as you'll remember, we started the year talking about this being a big, data-rich year for Sage and the strategy of expanding and accelerating the pipeline. And it's -- I find it quite remarkable that we sit here toward the end of the year, and all 3 programs that we started the year moving forward with -- zuranolone, SAGE-324, SAGE-718 -- are moving forward with really rich data in hand. So I'm very excited by the execution we've had this year and what the future of our pipeline looks like. You asked specifically about the new SHORELINE data, and we're very excited to announce the top line data. I'll remind you that SHORELINE is a naturalistic study where patients assessed with MDD, HAMD-17 and MADRS -- HAMD-17 over 20, and MADRS over 28 -- were included in the trial. The 50-milligram group, 199 patients with MDD, we enrolled for a dose of 50 milligrams. The drug was taken nightly for 14 days. And quite remarkably, about 75% of the patients achieved a response that's about at least a 50% reduction from baseline, and over 40% achieved remission. That's HAMD of less -- greater or less than 7. We did have a very low withdrawal rate. Only 3 withdrew in the prior to 28 days, so very, very low dropout rate, in fact, very high retention rate. What's really incredible is that majority of patients, about 55%, received one treatment course and stayed well for the entire year, and about 80% only needed 1 or 2, 2-week courses in the entire treatment year. So as we compare that to the 30-milligram, we're seeing better efficacy with 70%, with 30% that needed 1 or 2 2-week courses [indiscernible] without more safety liability. We're not seeing more severity of adverse events. So we're really pleased with the SHORELINE. And again, given that this is a naturalistic study, our take is that this is sort of how the drug would respond -- or patients would respond in the real world. Very exciting results.

Yes. Thank you, Barry. I completely agree with you, really remarkable data. A question for you. So when you reported out the 30-milligram dose group, you guys noted the [ TE ] frequency between the first dose to the fifth dose. So here, we didn't see that in the 50-milligram dose group. So if you could just comment on -- do you see the frequency of [ TEs ] also go down somewhere to the 30-milligram between first and fifth treatment. And then how does the sedation and somnolence compare in the 50-milligram versus 30-milligram dose group, especially during retreatment?

Right. So given that these are top line results, there's only so much we report at this point. We did comment that the 50-milligram safety was very much like the 30-milligram safety. We have slightly more incidence of adverse events but no more severity. So there's a little bit more incidence. The thing that I think is important here, and I'm actually happy to see it is -- you know this -- most people living with MDD have problems sleeping and often take a sleep aid. So the fact that while they're on this drug, they sleep better is actually a very good thing. We also -- we don't have the data yet. But in previous data, we've noted that the HAM-A also is greatly reduced, which is very different than we see in standard of care. So again, very rich data set with SHORELINE and the totality of the zuranolone data.

And then, Barry, can you remind our listeners on what is the requirement in SHORELINE to get the second dose? And what's the requirement to get the third and fourth dose?

Right. So repeat assessments are done every 14 days virtually and then, I think, every 3 months live. And both HAMD and PHQ-9 are assessed. And for a patient to be retreated, they have to have a PHQ-9 of over 10, or equal to 10, and HAMD of greater or equal to 20. To put that in sort of more simple terms and the way I think this translates to the real world is, if you've taken zuranolone, you're assessed at the end of treatment and asked, how is your depressive mood? Are you still dark or not? Are you sleeping better? Or how is your anxiety? And if your depressive mood is lower, you're sleeping better, you're lowering anxieties there, you're well. If your darkened mood comes back, your anxiety gets aggravated, you're sleeping less, well, that's a sign of -- come back to your doc and ask if the retreatment is necessary. But again, the fact that the majority of patients who knew they weren't on drug only required one 2-week course is actually quite remarkable.

I think a question that has come up quite a bit from clients is that how do you manage that in the real world, right? What would the label guide doctors in terms of [ ranges ] think if it's necessary? Like how do you think about that?

Yes. Yas, it's sort of what I just talked about. So you can imagine that someone gets diagnosed with depression in the physician's office. They're given a prescription of zuranolone, maybe even a refill in case you need it. Not unlike maybe your Z-Pak if you've got a lower respiratory tract infection. It might work in 14 days. You might need to recourse. And then with digital assistants, physician extenders, you check back into that patient. What's remarkable here, again, is that the drug works so fast. So people will know as early as day 3 after 2 evening doses that they're starting to feel better. The office can check in, and they -- or they can check in. And then in the out times, 14 days, 28 days, it really -- we have to codify this, but it really isn't much more complicated than, are you well? Is your depressive mood back? Or is your anxiety back? Are you sleeping less? If the answer to that is, yes, I'm getting worse, then a retreatment may be necessary. The data we have to date suggests that for most people, they only require one. Some will require a second in the course of the year. So the fact that we see 80% only require 2 or 4 weeks of drug in the course of the year, again, is quite telling of what we believe will happen in the real world.

And then we also noticed this morning the press release that 40% of the patients in the 50-milligram dose group were on antidepressants versus the 60% that were not. So when you look at the HAMD reductions of almost 16 points that goes down at day 15, is there a difference between these groups, patients that are already on antidepressant versus not?

Right. In this study, we're not seeing much of a difference between those on monotherapy or those on a stable antidepressant. I'll remind you that in the 201 study that's been published in New England Journal, those on antidepressants and zuranolone did slightly better. In the WATERFALL study, they did slightly less better. So it's really hard to say that there's a big difference between monotherapy and those on a stable antidepressant. We'll need to continue to analyze data to understand. For me, the good news is that if you're depressed and you're on something that's not working as well, you can add zuranolone, and these data suggest you get better. Or if you don't want an antidepressant because you're worried about some of the unpleasant side effects -- weight gain, sexual dysfunction, euphoria, inability to sleep -- then zuranolone as a monotherapy can work. And of course, the CORAL study is up and running, where we're testing giving concomitant antidepressants with zuranolone. So we'll have those data as well.

So that leads me to the next topic, which is CORAL. Thank you for the transitioning. Question to you first is enrollment is complete. Publicly, you've said the data is expected early 2022. Especially at the beginning of the year, the first 2 weeks of [indiscernible], there's a lot of news announcements that come out. Should we be expecting CORAL data ready to be at that time point? If you could provide us, some granular details on timing would be helpful.

Yes. So we have completed the enrollment of CORAL. Just to remind everybody, until we had the pre-NDA meeting, it was not wise of us to close the CORAL study. When we confirmed with the FDA that the data we had in hand, not CORAL, was sufficient for filing, we rapidly closed CORAL. But of course, those patients that were queued up for screening, we did not want to block from the trial. So that puts us out to early 2022. Early for us is first or second quarter, and we do not anticipate the data -- this is very important -- we do not anticipate the data in the first 2 weeks of the year. It will be after that.

Okay. Thanks, Barry, for the clarity. So the next question for you here is we learned on the last earnings call that the primary endpoint is changed for HAMD-17 to day 3 from day 15. I guess a question here that has come out basically, so with [ HIT stats' stake ], will it be counted part of the efficacy package? If it misses for some reason, then it will be counted for the safety analysis. So like what are the different scenarios based on the outcome of this data?

Right. So if I take a step back, Yas, I'll remind you, at the beginning of the year, we were quite excited by the LANDSCAPE and NEST program of zuranolone. And we communicated that in meeting with the agency, we defined 3 different Phase IIIs, 2 in MDD and one on PPD. Any one of which was positive, we believe, we had the filing package to go forward and file the NDA. With the positive WATERFALL, we confirmed with the FDA in our pre-NDA meeting that WATERFALL plus the rest of the data was sufficient for filing, so we're moving forward. But that then did it -- it enabled us to look at CORAL for the design that it actually was originally intended, which is a quick start when someone's also given an antidepressant. As you know, standard of care antidepressants take 4 to 8 weeks to work, if ever. They -- patients often change every -- on average, every 7 weeks because of some of the unpleasant side effects they face. We talked about that, things like weight gain, sexual dysfunction, euphoria, inability to sleep, heightened anxiety. And we're not seeing that, which is quite good. So what we want to know for CORAL is when given concomitantly with an antidepressant, do we see that rapid effect at day 3? It'd be shocking if we didn't, given the fact that we've seen that in 3,000 patients thus far. And then, of course, what we don't want to see is at the end of the treatment course or anywhere in between that there's a rebound that, for some reason, zuranolone on top of a brand-new concomitantly prescribed antidepressant works fundamentally different, and patients might get better at day 3 but then get worse. Again, we have not seen that in 3,000 patients who expect it. And these data, if positive at day 3, and we don't see that kind of rebound, will allow us to educate the health care community on being able to give zuranolone as monotherapy on top of a stable antidepressant working concomitantly with a new antidepressant. So all kind of 3 approaches that a patient or a physician might want.

Great. And you know how investors get when they get close to data, they're trying to predict their results. I guess the question for you here is, what is the probability that CORAL is going to achieve a 3-point difference over placebo at day 3 or greater, in your view? I think it's important to discuss data expectation just given that CORAL is a very different design from the rest of the LANDSCAPE study, so maybe investors need to be really mindful of comparing, contrasting them to the rest of the LANDSCAPE. So if you could just provide some commentary on expectations, that would be helpful.

Yes. I mean what we're looking for is a statistically significant difference between drug and placebo. Whether it's 1.7 or 3 points, that range is less important because, as you know, there's very significant expectation bias on depression trials. And when people know that zuranolone works so well, we're seeing very robust placebo. The more important point, though, is we're seeing better results on drug. And study after study after study, including the SHORELINE data we just talked about, patients respond clinically and stay better longer. So we're counting on that out of CORAL as well.

So Barry, the next question. So we recognize the FDA will be looking at the totality of the MDD data package, right, for approval. I guess the question that constantly comes up is, yes, they will look at across all the LANDSCAPE studies, but how will the FDA focus on -- how much will the FDA focus on response rates, remission rates and durability at day 42 from WATERFALL versus maybe the rest of the LANDSCAPE trials?

Yes. So we know from discussions with the agency and as evidenced actually by the ZULRESSO AdCom, that what they're looking for is a rapid response and a sustained response without rebound. So we do have that in all of our trials out at day 42. We believe the WATERFALL trial for those patients measured at day 42, we saw about an 86% retention of the HAMD reduction that we saw at 15 days. Very importantly, and we presented this at ECMP, when we look at patient-reported outcomes, the SF-36, patients themselves who were blinded reported when you look at the integrated data, a statistically significant difference across all eight domains of SF-36 at day 42 on drug versus placebo. So the both the physician assessment but, even more importantly, the patient-blinded assessment suggest that those on drug stay better and are statistically significantly better at day 42. And I count a lot on how the patients are feeling.

Yes. That's helpful. Another question that always comes up from investors is they recognize the SHORELINE data and, being a realistic study, the importance of almost 80% of the patients requiring 1 or 2 courses of therapy, especially in the 50-milligram dose group. I guess the part that the question comes up, it's an open-label study, so it doesn't have a placebo arm. And so the question always posed to me is, well, how will the FDA get comfortable with the significance of the SHORELINE data set in the absence of a placebo arm?

Right. So the SHORELINE -- the importance of SHORELINE is not the efficacy package per se. We already have, with positive WATERFALL, a significant number of well-controlled positive studies that will [ go into fund ], including the Japanese study, which is a monotherapy study at both 20- and 30-milligram. So those data constitute the efficacy package. What we're counting on SHORELINE for is more safety as well as the data on retreatment. And these data will help us guide health care providers in the real world about what to look for and when retreatment might be warranted.

I know recently you shared the combined data set from the LANDSCAPE studies, and this broke down the safety analysis across all trials. I guess the one thing that investors always keep asking me is, what percentage of the patients fall within severe versus moderate versus mild sedation? So if you look at the totality of the data, like what's the breakdown? And then the same question as for somnolence. So if you could just remind us [ something ] when we look at the totality of 3,000 patients, what is the frequency within the buckets between sedation and somnolence?

Yes, I guess probably the best way to look at that, Yas, rather than spat-out statistics and numbers is the fact that the adverse events are mild to moderate, and there -- no interventions are required. They sort of -- they wear off naturally, which is great. And the discontinuation rates across all of our trials are remarkably low. They're almost placebo-like level. So what's important when we talk about the real world are the patients staying on drug and the fact that it's 2 weeks, we're seeing a remarkable number of patients or 95% of patients stay on the drug for their 2 weeks. And then obviously, we've already talked about the benefits that those patients are receiving after 2 weeks. So in the commercialization end of this, what counts is compliance and adherence. It's very, very high.

And then, Barry, the next question for you always comes up is remind us like what is the shape of the curve of sedation and somnolence during the 2 weeks of treatment course, right? When does it show up? When does it get worse? And when does it improve? Or does it just pop up early on and stays consistent?

Yes. It's interesting to look at. So I'll remind you that zuranolone is instructed to be taken at night with a meal, and patients are reporting, after they take their drug, some point later that they're sleepy; they're tired. The adverse events can be reported any time. So if they are assessed the next morning and asked, did you take your drug, were you sleepy, and the answer is yes, that gets marked as sedation. As we said before, most people living with MDD are also prescribed a sleep aid because they do have problems sleeping. So the benefit of sleep is actually while it's reported an adverse event, and when you talk to your KOLs, they will tell you this, it's actually a benefit in this patient population. What patients report discontinuations on more readily with standard antidepressants, as we talked about, are sexual dysfunction, weight gain, and that's -- or unpleasant feelings, and that's what gets them to drop standard of care antidepressants, which is why, on average, patients are changing antidepressants, if they do change, every 7 weeks.

No, that's helpful. And then, Barry, I know you guys are completing the driving studies and additional PK studies with the 50-milligram dose group. So based on the data that you have seen, do you think the label is likely to require not to operate any machinery during the entire treatment course or just over the first few days? So how should we be thinking about that?

Yes. I think that's not a fight worth fighting with a drug that gets into the brain or any drug that will be Class IV. So it's likely that we'll have sort of standard language: Don't operate heavy machinery or driving until you understand the effect of the drug. Not too dissimilar from what you see with over-the-counter Benadryl, frankly.

And then I think the question that I think a lot of investors also struggle with is how it will fit into the current treatment landscape, right? And how will pricing play into it, given that this will be not a chronic treatment? So how do we think about it? What market research work has -- Sage has done? And what glimpse can you give us right now? And what should we expect at R&D Day? Will we get a little bit more on that?

Okay. Lots of good questions in there. So let me break down the questions. So first of all, look, we're -- the MDD and PPD market -- well, the MDD market is very fragmented, and -- as you know. We are excited by the potential of zuranolone to help a wide variety of people with both MDD and PPD with their associated symptoms. I'll point back to we know with the patient-reported outcomes that patients also, in addition to the HAMD and other assessments, they're reporting feeling well. They're reporting lack of anxiety, reporting that they're sleeping better, as evidenced by all eight domains of the SF-36. So we've seen this consistent data. There are no oral meds available that get people better as early as day 3 after 2 evening doses. And we believe that with education, both on the physician side and the patient side, that people will reach for zuranolone. And if it was you or someone in your family, who wouldn't want to get better fast? And I think Anita Clayton, when she did the ECNP Encore, basically said that. If it was someone in her family, she'd reach for zuranolone. So with education and with patients requesting, we think that there'll be really significant use in the real world. And as we talked about, we will have data, we believe, at launch, if successful and approved, that we'll have monotherapy data, zuranolone on top of standard depressive -- antidepressives that are stable and concomitant with new antidepressants. So sort of all 3 scenarios for a patient or health care provider that might make them comfortable, we will have data to educate. In terms of pricing or that side, it's too early to talk about specifics of price, but we intend on doing, and we've talked about this, is kind of leaning in on the payer side and being very proactive with value-based agreements. And the essence of that is that if you, payer, let the script go through, which we're asking for, the drug works well, that's great. We've made the patient better; we deserve to get paid. If it works less well, then we'll get paid less, and they're not at risk of a more expensive nongeneric type of price. So we're hopeful that the payer strategy, the data are sufficient and really there to make sure that if patients want something to get better, they can get zuranolone.

So now let's transition to 718. Maybe we'll start off -- like you unveiled the DIMENSION study. This was the Phase II Huntington's disease. So just kind of give a glimpse of the design that was shared this morning.

Yes. Look, we're encouraged by the data we've seen to date with SAGE-718. I'll remind you that it's a first-in-class NMDA PAM. It was specifically designed to upregulate NMDA in a very specific way. And the data we've seen from preclinical to early clinical now through Huntington's and Parkinson's has been consistent. So as we articulated, we plan to now open what we're calling the DIMENSION study. The study will treat patients for 3 months with an additional month of follow-up. The primary endpoint of the DIMENSION study will be changed from baseline in composite score of Huntington's disease cognition assessment battery or HD-CAB. This is a cognitive battery developed for use in assessing cognition in HD trials, really important and kind of an important assessment in Huntington's disease. The key secondary endpoint will be the standard independent scale, the UHDRS scale. Patients will be randomized 1:1 with 718 versus placebo. The 718 group, they'll receive 1.2 mg of 718 daily for the first 27 days, which will get blood concentrations at the efficacious dose as evidenced by the PK/PD studies we've previously done and then move to 0.9 mg for day 28 through 84. The long half-life of SAGE-718 suggests that that's the right dosing range. And as you noted, Yas, we did post the study slide on our website, so folks can get more detail on the website.

And then, Barry, what's the size of the study? How many patients?

Sorry. A little -- clear my throat.

Yes, no worries. The size of the study, how many patients?

We -- sorry, hold on one second. Got to be able to talk. We haven't talked about the size of the study yet. That's something we'll come forward with in a bit.

Okay, good. Something to look forward on R&D Day.

Sorry about that.

Yes, no worries. All good. Okay. So the next question I have for you is, I think the challenge -- so Street recognizes that potential of 718, right, and Parkinson's and Huntington's and potentially in the Alzheimer's data set that we'll be seeing also. The challenge for them always exists is that these studies have been in small open-label [ variants ]. And so how can they get comfortable about the totality of the data derisking, I guess, future development? So I think that's -- I think they recognize the opportunity but maybe fail to figure out the translation of the open-label small studies to larger studies across the 3 indications.

Right. Well, I understand, and actually, we agree that we're sort of a placebo-controlled study away from definitively declaring that what we're seeing in these tests are drug versus placebo. What I would point to, and we'll be presenting the Alzheimer's data in the near future, is the consistency of data and how rapidly improvements are seen across learning and memory and executive function. It's unlikely that those are learned tests or placebo effects, given the data we've seen to date, where the drug works when it's on, it stops working a little bit when it's off, for most of those assessments. But we'll have across the Huntington's trial and soon the Parkinson's trial, which we said we're going initiate early next year, at least 2 shots at placebo-controlled trials. And those data will demonstrate definitively that the benefits that we're seeing across learning and memory and executive function are, in fact, drug benefit.

So now let's move to 324. So we're eagerly awaiting the design of the Phase IIb, but I guess the question that I have for you -- so you've publicly said that you're doing a dose-ranging study. You're going to be changing the time of the dose administration at night versus the morning. So I guess, given these changes, like what magnitude change do you expect to see with lower doses in efficacy and safety? So if you could just provide some commentary because the Phase II data looked very, very encouraging. So the question here is, if you change some of the variables, how will that impact the data?

Right. So we are running a -- it will be a placebo-controlled study. We talked about it being of longer duration. But we'll have more, at FutureCast, on the actual study design. And we are looking at various dose ranges. What I -- what we're looking for is a dosing frequency that allows for continuous drug. It will be a chronic medication, and at sort of that 30% to 50% change in essential tremor as well as changes or alignment of changes with activities of daily living. And what we saw with the Phase II, while it was dose at the high end of the dosing scale, we saw that 30% to 45% reduction in essential tremor. And very importantly, we also saw a statistically significant alignment between those changes in essential tremor and the changes in activities of daily living. And we'll hear from patients, but things as simple to you and me as feeding ourselves with the spoon, buttoning our buttons or texting friends, going from not being able to do that to doing that is a life changer. And that's what we'll be looking for.

And then maybe the last question for you, so what should we be expecting to learn at the R&D Day?

Yes. So in FutureCast, we'll have a bit more on zuranolone, less on the data readout -- because we've put the data out there -- and more on the experience of what treating physicians might be thinking about in the real world. We'll go through SAGE-324, the Phase II design. And then we'll also talk about SAGE-718, the totality of data. And if available, we'll talk about the Alzheimer's data as well.

Great. Thank you so much, Barry. Thank you. We're coming to the end of the fireside chat. The 30 minutes really flew by. So thank you so much for making yourself available. And thank you for all the great work and dedication that you've brought to Sage, and so we're really excited to work with you and seeing 2022 will be a great year. Thanks again on behalf of all of us here at Piper Sandler.

Yas, Thanks for having us. It's been a great year. I'm looking forward to next year. And let's remind everybody that while zuranolone is incredibly exciting, we have an entire pipeline of Sage-invented products that we'll be bringing forward and talking more about in the months and years to come. So really great time for the company.

Great. Thank you. Thank you, everyone.