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

Great. Good morning, everybody, and thanks for joining our Third Annual Neuroscience Conference. Definitely great to be starting it off with Barry. It's the first discussion. I think we're doing on camera since you joined Sage, maybe we did one at a conference in the fall, but more -- so no online questions today on Sage. But thank you, man. And maybe you want to kick it off and just sort of set the stage with the update in zuranolone trials, tremor and NMDA, and then we can get into specifics, Barry. That sounds good?

Yes. Yes. Thanks, Paul. And I appreciate Stifel for having us. And it's wonderful to kick off your third annual and be the lead company. So yes, things are going incredibly well. I'm excited by the progress we're making at Sage. As I've said before, we're positioned to be a top-tier biopharmaceutical company and the leader in brain health. You asked specifically about kind of zuranolone progress, the overall Landscape Program is going incredibly well. And if I step back, I'll remind you that we have 3 distinct paths for opportunities with zuranolone, 2 in MDD and 1 in PPD. Obviously, the one people are focused on is WATERFALL, which is zuranolone as a monotherapy, but we also have the CORAL study, which is zuranolone co-administered with SSRIs, which reads out late this year. And the SKYLARK study, the second study in PPD following an already positive study. So this presents 3 unique opportunity sets and 3 unique pathways for approval. Obviously, we hope that they're all positive, but there's some risk in drug development. So it's good to have multiple shots. We're feeling really good about what we've done with WATERFALL. The kind of accrual rate went incredibly well. We were able to upsize the trial a couple of different times, which provides a lot more flexibility in kind of subgroup analysis. In addition to that, overall progress, we started the year saying there'll be 10 different data readouts. We've already had 1 SHORELINE, which showed the full 30-milligram data set confirming the interim 30. And then very importantly, the 50-milligram showing improved efficacy without any additional safety liability. So it's, again, 1 out of 10 so far, but it's setting up to be a data-rich, catalyst-rich year.

Yes. Awesome. So maybe let's talk about WATERFALL. On the upsize of the study, I guess how do you guys think about that internally, right? Like was there any kind of cost benefit calculus to whether or not there's a certain sample size or maybe the benefits of more power kind of decrease and you sacrifice some patient quality? Like what was the thought process there? And obviously, you and I both know, right, the context surrounding MOUNTAIN being bigger than expected and then not working. So any color would be helpful, Barry.

Yes. The first thing I'll say is that the upsizing had nothing to do with looking at any data, even blinded data. So it wasn't see something and make -- and take action. It really had to do with the opportunity set. So not only at Sage, but all across the industry, when we design Phase III studies, we're looking for the most efficient effective study we can design in terms of capital and time frame. So we look at effect size, we look at variability, we look at -- and you're trying to design a trial to be as capital-efficient as you can. What happened here, and unfortunately, the world depression rate has gone up three to fourfold driven by the pandemic. And therefore, again, unfortunate for the world, but it gave an opportunity for this trial set to accrue very rapidly. So we set time frames to have the data in the first half of 2021. So without really changing capital structure, without changing time frame, without changing number of sites, we're able twice to upsize the trial in greater than 525 patients. Obviously, what that does is give better powering to the primary endpoint of 15 days, the secondary endpoint. But very importantly, this allows kind of a richer analysis of some of the subgroups, men versus women, black versus white, age group differences to understand the effect size across all those potential subgroups. So having rich data set, assuming it's positive can only help.

Yes. Yes. Okay. Well, so it's interesting, like I think Sage talked about this, too, and some of your peer companies definitely talk about the art of psychiatry trials and selecting the right patients and homogenizing the population. And one question I've gotten about the WATERFALL study is whether or not there's any sort of risk in what the population may look like because we're in a pandemic, and it's this kind of like oscillating environmental driver that affects everybody's mood. You guys have talked about how the pandemic has created a bigger issue with depression and maybe even allowed you to recruit faster. So how do you think about that in terms of, I guess, I don't want to say this in an insensitive way, but like patient quality or really having the right population in your study and knowing that there's no added variability with it?

Yes. And look, it's a great point. And for some companies and some drugs, it can present a risk. At Sage, this study, first of all, was designed within the context of the fact that we're independent. So site selection, remote monitoring were all designed into understanding that we're, in fact, in a pandemic. So it was designed kind of in the old world with, "Oh my God, now what we're going to do?" It was designed with the pandemic in mind. So that's really an important point. The second point is that Sage has hypothesized for quite a while by really deep in understanding kind of brain chemistry and brain architecture that while depression is there and is an underlying -- is underlying us all the time, it's actually not a chronic condition per se. It's triggered by certain events. And these trigger events cause depressive episodes. So whether that trigger is you've had a baby, whether the trigger is that you've had a spousal issue, whether the trigger is you lost your job or whether the trigger is I'm trapped because of COVID, it's a trigger that drives you into a depressive event. And the hypothesis that Sage had, and I think well proven by the SHORELINE data, is that when someone hits this depressive event, his HAM-D is over 70 -- over 24, that by giving them 2 weeks of zuranolone, your rewiring brain architecture, helping to get them back to a normal state. And the data that we've demonstrated, now over 3,000 patients in subjects, is very consistent. zuranolone gets people better, faster and keeps them better. We see positive effects as fast as 3 days with durable effects. The SHORELINE data continue to show in the 30-milligram group that 70% of patients only need 1 to 2 2-week dose in the course of year. So just think about being off drug for 48 weeks of the year, it's an incredibly different profile and the safety liability is incredibly consistent as well.

Yes. Yes, makes sense. So you mentioned powering secondary endpoints. How do you think about the regulatory and commercial importance of hitting statistical significance or not at day 42 versus placebo? And if the p value is not the kind of hurdle, how do you interpret those data?

Yes. So what's really important, and these are slides that we're -- that are decked that our health economics group has generated. What we see in zuranolone, looking even at the previous MOUNTAIN data and the previous SHORELINE data, is that this is the first drug of its kind in 35 years that's a fundamentally different benefit risk profile. And we've just talked about this. If you can get people better faster and keep them better with literally 2 or 4 weeks' worth of drug versus a chronic drug that has significant levels of side effects causing people to drop off their drug, a p value is kind of what we need here. That's the thing that will help understand that this, in fact, the drug works. It works well. We've demonstrated statistically. Whether some of those later endpoints hit or not is less important than the 15-day endpoint hitting us having that period. And really, it's because of that differential benefit risk.

Makes sense. Okay. I guess, how do you think the FDA thinks about those data? Like one thing that was talked about at the brexanolone panel is interpreting the drug arm and thinking about how placebo effects are variable later. But I guess, like should we be nitpicking the behavior of the drug at day 42? Like if it's up a couple of points, is that bad? Like what -- I don't know, I'm sure you don't want to set it like an actual quantity of hurdle. But like qualitatively, what is demonstration of durability? And what isn't demonstration of variability? Understanding that like in depression, you're kind of playing for a couple of point benefit over placebo in the first place.

Yes. Again, I think the evidence that we have so far -- let's take the 50-milligram data just in SHORELINE, we have more than 80% of patients report positive response at day 15. That's just remarkable. Let's compare and contrast that to other choices they have. Other choices are drugs that they have to take chronically that take 6 to 8 weeks to work and only work in about 1/3 of people. So the fact that you've got 80% of patients reporting a positive benefit at day 15, but then they say is durable, is quite remarkable. And even those patients that have to take -- and this is not a lot, because 70% will need 1 or 2 2-week doses, but in those patients that might have need a third or fourth dose, they continue to respond. And in fact, the adverse event profile lessened. So if you can treat people episodically when needed, even if it's 3 or 4 different times a year, that's quite remarkable and quite different from any other opportunity they have today from a medical perspective.

Yes. Yes. Okay. How do you -- maybe put the permutation game with me for a bit, like if the -- so depression drugs tend to be approved, right, on acute efficacy data and some demonstration of long-term safety. If WATERFALL fails and then CORAL works, what then kind of happens, right? Because like CORAL is framed as this sort of -- it's a study for an acute label, but so is WATERFALL, right? And they're both acute trials. Like are the regulatory outcomes for the drug really different if it's CORAL as your positive study and WATERFALL as a negative study? Or are they more the same? Like how do you think about that?

Yes. Well, I'm quite enthusiastic about the potential for positivity in all of these studies. So that's...

What is depression, so who knows, man? Yes.

Look, Paul, as you said, we've had Prozac. Prozac, I think, had 3 or 4 failed Phase IIIs before the successful Phase III. So the question is a variable and risky business. Again, I look at just how fundamentally different the understanding now is of depression, how fundamentally different zuranolone works, getting people better, faster in such a rapid form. So again, we said from the beginning, there's 3 unique shots with zuranolone of approval. Those 3 different studies that I already talked about, WATERLINE, CORAL and SKYLARK, any of which, if positive, provides a potential regulatory path for approval. And what CORAL's asking the question is if we co-administer zuranolone with an SSRI and it's known that the patient's on SSRI, how old the curve differentiate? I think with any positive Phase III, given the Phase IIIs we've run, given some of the post hoc analysis on MOUNTAIN and given the long-term durable data we have on SHORELINE, that's a pretty data-rich package. And of course, we'll have to sit down with the agency with breakthrough status. So open dialogue with the agency to find out just what the right package for an NDA submission is. But these -- again, these 3 trials do us 3 unique shots of getting this 1 medicine on the market.

Yes, yes. Okay. So we talked about this a bit. And I guess, going into these acute readouts, to what depth have you had discussions with the FDA yet on what you'll need for long-term safety and redosing? And what sort of counts as like a 1-year exposure? Like I said, 1 retreatment, is it 4? Like is -- what is duration of exposure? Is that the right way to think about it? Or is it really thought about as number of courses? Like how much have you waited for good data to really firm this up versus because you have breakthrough, how much is this an ongoing discussion?

Yes. I mean, I guess it's a little bit of both, Paul. So we -- because of breakthrough, it's an ongoing discussion with the agency. And what I can tell you in the agency discussions that I've listened to and in terms of the agency documentation, they understand that challenge with depression. They actually understand even with the availability of 30-plus agents on the market, many of which are generic, that this is an unmet need, that there isn't a medicine on the market that does with zuranolone. So there's acute understanding in the division about what the unmet need looks like. We've agreed with the agency on these 3 unique paths. Then, of course, data matter. It matters what the data say in terms of what we can sit down and how quickly we can drive to approval. What we have said, and this is something we'll have to confirm hopefully after a positive WATERFALL, is there's a Phase III that we paused called REDWOOD that was asking a question about treating on a very regular basis, and that was designed before we had any of the data like the SHORELINE data. So that was an important question without data. Now that we have data, that's no longer a question that we think is important to study. There may be other kind of post-approval studies, but that's not one of them. So again, assuming WATERFALL is positive, we'll sit down with the agency with that full robust discussion about what the filing package looks like, and hopefully what the post-approval commitments look like.

Right. Okay. Maybe premature, but like do you think there's a path to approval where there's no -- like a path to kind of near-term approval where there's no restriction on the number of retreatments?

Again, the -- we have to rely on the data. And the SHORELINE data suggests that even after -- and we're talking about dwindling in, so -- which is good. After the third, fourth or a couple of handful of patients on a fifth trial, they continue to respond. So think about that. Why would you reach for a different drug that may not work in 2/3 of you when you continue to respond on this drug? And I'll go to kind of the worst-case extreme by data we have. And again, it's a very few set of people. But if I said every time you have a depressive episode, take this for 2 weeks, you'll respond. And you have to do that for 10 weeks out of the 52 weeks a year. That's still your best choice.

Yes. Right, right, right. Okay. All right. Great. Let's talk about tremor a little bit. Maybe just -- can you just quick update the status of that program? I think some folks, myself included, thought the data might have come by now. So where exactly are you with that IIa? And you've kind of already announced you're doing a IIb later this year. So you must have some confidence in activity. What led you to kind of already start planning the next study even a few months ago?

Well, just to be clear, what -- we've designed SAGE-324 for chronic treatment. And we're looking at diseases where unlike depressant, which is sparked by specific episodes, depressive events, and we thought and now are showing SHORELINE that we can kind of rewire brain architecture with a couple of weeks. We're studying diseases like essential tremor or maybe epilepsies where the brain constantly kind of is misfiring, and an episodic treatment probably isn't required, more of a chronic treatment that constantly keeps drug level in the brain at a consistent rate. So 324 was designed for bioavailability and long half life. So that's why the drug was designed. The initial studies in essential tremor of single-dose studies showed that if you can get blood concentration to a certain level, that you dramatically decrease the amount of essential tremor. So what we're asking in the KINETIC study is can we continue to keep blood concentration levels at a certain part and maintain that 30% to 50% change in essential tremor that's durable over the 28-day study. So specifically, that's what we're asking for IIa. What I'm looking for is that 30% to 50% improvement in essential tremor that's sustainable over 28 days. So no tachyphylaxis, nothing like it works for 5 days and stop working thereafter. And then we're looking for an adverse event profile, Paul. It's no surprises. We think we know what these drugs do in terms of somnolence and sleepiness and dizziness. So we expect all that, and we've given the maximum -- kind of maximum dose of 60 milligrams. So we're really looking for durable efficacy over a point in time. Now we said, just to be clear, that, that should -- that we see that, then we anticipate running a Phase IIb. So we didn't commit to IIb no matter what. If we don't see a drug effect, we're done, then we move on. If we see the drug effect, then what we're asking in IIb is, okay, now that we've seen that, what dose and frequency maintains blood concentration levels that maintains that durable effect where we now dial back the side effects? Because we want to make sure that the drug works and that people stay on the drug. So that's what we'll be asking in IIb.

Right. Okay. No, that makes a lot of sense. So you mentioned it here, and I think Steve mentioned it on the earnings call, just kind of alluding to maybe 60 migs is at the top of the dose response curve. What gives you that thought? Like what gives you the thought that maybe we can dose lower and maybe we should? Is it because of blinded adverse event rates or something like that?

No, it's not based upon the current study. The KINETIC study is a blinded study, and we committed that it was early 2021, which means first to second quarter. So we'll hit that early '21. People may have expected earlier, but it's an experimental study. The team wants to make sure that they collect all the data, they lock the database, they clean it. And we understand all the data before sort of going out with the data. So that's ongoing at the team level, and we'll hit our first half. So the view of 60-milligram is not based upon the KINETIC study, it's based on all the previous work. And what the team has done is dose escalation. And we know that the efficacy is driven by AUC, area under the curve, so constant blood concentration. And that most of the side effects are driven by Cmax, so -- like what you get right after you take the drug. So we know that 60 milligrams drive a large Cmax but also creates the best area under the curve when taken daily. It's possible, Paul. You've seen this that you take drugs for small -- you take lower amount of drug for longer periods of time, and it might take 4, 5 or 6 days for the efficacy to kick in. But -- and those are the kind of profiles we'll be playing with in that Phase IIb.

Yes. Yes. Okay. All right. Good. How do you -- last tremor question before we switch to NMDA and cognition. But how do you think about the regulatory path in this space? There's some drugs out there, but all the studies here are so old. Like is that -- have you guys done any kind of regulatory refresh in your thinking?

Well, look, as you said, there hasn't been a drug approved for essential tremor, the large movement disorder we have out there in over 50 years. So the agency -- the agencies all understand the big unmet need here and the fact that there really aren't very effective pharmacologic intervention. So we've got some time to work with the agency on exactly what those endpoints look like, but I'm pretty confident that we'll be able to figure that. There's enough ways of measuring tremor and measuring the other effects of benefit effects that we will have a path forward.

Yes. Yes. Okay. All right. Great. Last 5 minutes, do you want to talk briefly about SAGE-718, the ongoing studies and then we can do a few follow-ups?

Yes, for sure. So SAGE-718 now is our first-in-class NMDA receptor PAM. It's in development for oral therapy, for cognitive disorders associated with MDA receptor dysfunction. The first place we're studying it is in kind of neurodegenerative diseases. So without -- we don't have enough time, but we're not going to into too much detail. The biologic hypothesis came from certain deficits in Huntington's patients, this 24S-hydroxycholesterol. So the first studies we did was in Huntington's, and we saw a beneficial impact on cognition, executive function, which is critical for independents. So we're following that up with the PARADIGM study in Parkinson's, which will read out early this year as well as the LUMINARY study in Alzheimer's, which will read out late this year. And pretty excited by -- I've mentioned this before, I'm really excited by 718. It's a wholly owned program of Sage, and it really follows really sound biologic science.

Yes, yes, for sure. These initial data, I know Sage has this history, and I guess it's not fair for me to nitpick because it's worked out. This history of doing these small open-label studies and then, "Hey, look, they're predictive of actual success in a placebo-controlled setting." Like that said, I don't know what the equivalent of this is in cognition, but it seems unlikely that, in cognition, you're going to show the equivalent of a 17 point HAM-D drop in a couple of days or something like that. So I guess, how do you think about the clinical hurdle for what is good data in an open-label study in PD, dementia or Alzheimer's that makes you confident in efficacy?

Yes. I mean, look, we -- there's a number of -- there's a battery of tests, things like work in memory, things like problem solving that even if you -- were a placebo effect, you can't work yourself through a placebo effect. You're either able to solve these problems, you're not able to solve this problems, even if you think you're on a drug or not. So it's early. But when you think about the biologic hypothesis, the preclinical model, the challenge, the ketamine challenge models with imaging that we ran in the early days of Huntington, the biologic scientific story continues to line up. So obviously, we're going to have to continue to run the larger trials to agree with the agencies on what the endpoints look like. There's a way to go. And it's early. It's early, and it's a smart bit of data, but it's an exciting opportunity.

Yes. Yes. Okay. All right. Great. What do you think ultimately this looks like in terms of which indications you select forward? Like obviously, you've got this broad mechanistic rationale, but it seems unlikely, right, you do multi-hundred patient studies in every dementia. So how are you going to prioritize that?

Well, look, as I've mentioned, we've got a couple of other studies going that will give us data in Parkinson's and Alzheimer's. We have the Huntington's data. So what we plan on doing is sitting down with the agency and sorting out. Again, since the study of cognition executive function has not been done, we have to sit down with the agencies to understand what the potential path to approvals look like. What's very clear, Paul, is in all of these neurodegenerative diseases, if you have a medicine that gives people more time with independent living, that's a huge advantage. So how do we answer that question? I'll give you the bookends. One bookend is we have our Huntington's data, we run hard at Huntington's. And 718 is initially Huntington's drug-orphan disease, good leeway, something that we could commercialize not in the United States, but multiple countries. We're -- we see data across indications that give us a basket study. And those are the kind of the bookends, and there's lots of different ways to go in between. We really have to map that out with regulators before I can give you more definitive thoughts.

Yes. Okay. Fair enough. So we should get some data soon, right?

Yes.

Some initial data soon from that, some initial data from tremor soon. And...

Absolutely.

Have you guys given any more color on the timing of WATERFALL, like I guess based on when you finish enrollment? It feels like late May, early June, something like that.

Yes. So what we've said -- so what we did at the beginning of 2021 is we moved our guidance to early, mid and late with the exception of WATERFALL. We didn't want to make anybody scared. So we kept that first half of 2021. And I'll say that's until June 30. So it will come out in the first half of 2021.

Okay. Okay. Awesome. Great, Barry. I have one question from someone listening in, just real quick, and that is, when you guys did the Biogen deal, you talked about accelerating work on other indications like bipolar, anxiety. What's the status of that?

Yes. So the value of the Biogen deal is kind of -- is multifactorial. First of all, strengthening our balance sheet allows us to say definitively we're expanding and accelerating the pipeline. We just talked about 718. That's a perfect example of something we're able to tell you in January. We're going to do all 3 of these studies and run hard at them without pause. We've also been able to bring other aspects of our pipeline forward like 904, 689 that we're kind of on the shelf and waiting. So these are wholly owned programs of Sage that we have global rights to that we're now accelerating. Major benefit. The other benefit of Biogen is their commercial power and their global power. So we can run studies for zuranolone 324 on a global scale with them rather than just uniquely U.S. focused. And then it gives us tremendous commercial optionality. By leveraging their commercial infrastructure, we can build where we need to, but not take risk of building too large or too great until we see levels of success. And then obviously, if 718 continues, we'll have commercial infrastructure that gives us ability to launch kind of 718 on our own. And then yes. And then finally, it does allow us to expand into other things like GAD and other anxiety and other areas with both zuranolone and 324. So we're planning on doing that. The first job #1 with us in Biogen, though, is to win in PPD and MDD and really secure the data sets that allows the world to understand that zuranolone deserves to be given early in the treatment paradigm because it is a different benefit risk than anything we've seen in 35 years.

Yes. Yes. All right. Very good. Well, looking forward to seeing some of the data, Barry. It's always a pleasure. So thank you for joining us.

Thanks, Paul. Appreciate it.

Have a good one.