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

Awesome. Great. Thanks, everyone, for joining us. Happy to be hosting this discussion with Sage Therapeutics' Chief Medical Officer, Steve Kanes; and the Chief Operating Officer, Mike Cloonan. Thank you guys for joining us virtually. Everyone's audio and video working well?

Yes, Paul, we can hear you good. You can hear us?

Absolutely.

Great. I appreciate you joining us outside by a lake. That's a beautiful scenery there.

Exactly.

It's awesome. So maybe we can start. I think just for everyone on here, for the flow of this discussion, I'm going to kick it over maybe to Mike just to give sort of a quick update and set the stage for Sage. And then we're going to actually initially talk about tremor and 718. And then we will go from there on the key components of the 217 discussion. So Mike, please take it away. Thank you.

Thanks, Paul. Yes. I'll be brief so we can get into the questions. But as many of you tuned into our earnings call just a couple of weeks ago, we're really pleased with our execution. And what is a year of execution for Sage. We feel really encouraged where our clinical program is at this point despite what's happening with COVID. We are on track with all of our clinical programs. We're excited for the future, right? We've set ourselves up for a very catalyst-rich next 12 to 18 months through our execution that we are currently performing well. And we really have this opportunity now that we're well positioned to advance this multi-strategy franchise that we've talked a lot about. And as you just said, Paul, we have our depression franchise. We have our neurology franchise, which is headlined by SAGE-324. And then we have our neuropsychiatry franchise, led by SAGE-718 and our NMDA platform. So a really exciting time for Sage, right? We set out this year very focused to execute against the clinical trials. We're doing just that. We think we've positioned ourselves very well to take advantage of this next 12 to 18 months, which could be very transformational for Sage. So we're all very excited. We want to thank the Sage team. They've been doing fabulous work despite everything that's going on, like every company, navigating through COVID. And the last thing I'll say is, given COVID, right, you probably have read this, Paul, and many folks have read this, is that the pandemic obviously has affected many people from the disease itself. But when you think about what the spillover could be, right, the next wave, you probably have read a lot of this around, what could happen to mental health, right? And what is happening to mental health? They're on the rise. The depression is on the rise. Suicide is on the rise. So we believe our mission now more than ever is supremely important and very committed to patients and families to make a difference with the molecules that we have. And so looking forward to the conversation today, and thanks for giving us the time.

Yes. No, thank you so much, Mike. That's a great intro.

So yes, I think it will be interesting to talk about tremor since that's the next key big event for you guys, and it's also something, I think, on the investor side, there's been some discussion of, but maybe we can get a little bit more into the weeds. But maybe we can start, Steve, with just, on SAGE-324, can you talk about the unique properties of this molecule and, I guess, why it's better suited to a tremor population than 217, understanding that, mechanistically, there's probably more similarities and differences?

Sure. I mean, just to remind everyone, what we've been doing in essential tremor is looking in a very systematic way across our portfolio for many years now, starting with SAGE-547, which is ZULRESSO. We've done prototype work with SAGE-217, as Paul has alluded to. And now we've landed on 324 as the molecule we're going to bring forth for the indication. And what we've seen across all of those trials is a very predictable response where when the drug is on board and when the patients are having high rates of exposure, tremor has improved. And as the drug leaves the system, the symptoms potentially recur. So what we've identified in SAGE-324 is a drug that we'll be using chronically, which has a broader therapeutic index than 217, but one that we can give chronically. And so the KINETIC study, the ongoing study, it's a 1-month study, placebo controlled, and it really has all the features that we believe is necessary for chronic therapy for essential tremor. So there's a lot there, and look forward to sort of getting into some of the details, but really identifying it based on its pharmacokinetic properties as well as its overall profile.

And Paul, maybe just be good to remind folks. I mean, essential tremor is the biggest movement disorder that exists. Over 6 million people, right, in the U.S. suffer from essential tremor. And as you know, there hasn't been any innovation in over 50 years, right, in this space. So that's why we think another really compelling case to be made for being disruptive and innovative with essential tremor.

Yes, right. Okay. That's great, Mike. Thank you. Yes. So I think kind of on this topic of probability of success and the mechanistic rationale and 324 itself. I guess when we've done some work on what seems to be effective in tremor, right, I think there's some efficacy out there for benzodiazepines are evidence, right? You can see alcohol as kind of an effective short-acting benzo-like drug in tremor and then also gabapentin. So almost I think a GABA approach should work. But on the flip side, when we've talked to physicians, they've said that it's really all about therapeutic index and avoiding significant sedation in an elderly population. How do you think about that? And maybe we could talk a little bit more again about why the properties of this molecule itself you feel like is well suited.

Well, I think it goes back to a few things. And I agree with you. The profile that you're describing is really important. Among those is the ability to avoid tachyphylaxis, right? Meaning the drug can't wear off if it's going to be given chronically. And that is the fundamental challenge with some of the medications and probably why they're not approved for the treatment of essential tremor. So going back to the nonclinical data that we have with 324, we see that we're able to have efficacy in animal models of tremor without attendant semblance. So much less sedation. So you can separate those 2 out, probably more difficult with drugs that are not specifically identified. Moreover, given all of our data, we have not seen tachyphylaxis. We haven't seen efficacy wear off. But those are the kind of questions you ask yourself in -- or we ask ourselves when we start a Phase II program, right? So we've dosed for up to a week. We've shown a continued increased efficacy, this is in 217, over the course of that time. So rather than seeing effects wear off, at least within multiple doses, we've seen an overall improvement. That's likely due to the activity at the extrasynaptic receptors as opposed to some of the other receptors that are more reactive, but those are the kind of questions we'll be looking to answer during the Phase II program.

Yes, that's great. Can you talk a little bit about clinical endpoints in tremor? It's a space -- again, as you said, there hasn't been a lot of innovation. So it would be helpful, I think, to understand, one, the key endpoints and a meaningful effect size; and two, how you just sort of think about the potential regulatory path here.

Yes. So I mean, I can tell you what we looked at. As I said, we've been looking at essential tremor now for over 6 years, and we've been doing it repeatedly, dialing in how best to assess patients, how to understand what endpoints are the most robust, which ones that give us the best signal to noise and how also to take unique sort of patient-focused looks at what data might be meaningful. So what have we looked at? We've looked at more standard rating scales that are done by physicians. We've also looked at the effects of our medication on wearables, and it allows us to see improvements of tremor in real time. And what's been nice to be able to demonstrate is very high correlations between all of those endpoints. If you see them on one, you see them on another. We're also dialing in those endpoints that we think are going to translate to the most clinically meaningful benefit. So we've looked at things like TETRAS. We've looked at things like risk-based accelerometers and other kinds of things, all of which show very clear efficacy. We include them all, and the question is which endpoint might be a regulatory one. Those are for discussions for another day after we've had FDA interactions. The one thing I would say is we're very confident in our ability to identify endpoints and be able to see them robustly across a variety of methods. So I don't think there'll be a challenge there in identifying some of the [indiscernible] with regulatory pathway.

Okay. Yes. I think some investors are maybe prime to think that this is a tough area for adjudicating efficacy. Some of that is due to the lack of innovation and maybe some physician comments. But the other thing interestingly that maybe has some parallels to what Neurocrine went through in tardive dyskinesia is this data set from Cavion, the company that Jazz bought, where they saw a discordance between clinician raters, video raters. You guys always are pretty thoughtful in how you look at clinical endpoints and really try to limit variability. What's your thought process here on what you need to kind of get right to have a tremor study that doesn't, say, fail a drug that works?

Well, so I think there's a few things. And I won't go into all of the details, obviously, as this is an area where interest is heating up. But what I would point to is how that we've done -- how we've done it across the study so far. We've done our clinical rating scales the way they've been validated in person. We've also done remote ratings with things like wearables, where you're able to really, as I said, measure tremor in real time. And those are highly, highly correlated. And what's nice about that is, in contrast to perhaps some other areas within neuroscience, tremor is an objective endpoint. You actually can measure its amplitude, its frequency, when it comes, when it doesn't. So there's something really hard about the endpoint that allows you to really assess the impact both in -- with the drug as well as on patients' overall benefit.

Got it. Sorry. Unmuted myself. Okay. Great.

Don't do that.

Yes. Well, you deserve this. If we were in person, I'd just interrupt you the whole time. Maybe just lastly to finish this discussion off, going back to kind of what's a clinically meaningful effect. And look, I think within that, this study is 60 patients. So it's not tiny, but it's not huge. Are you powered statistically for a clinically meaningful effect?

Yes. Well, I -- well, obviously, we're looking for clinically meaningful effects across the board. But one thing I would just take the opportunity to share with everybody, Phase II is the initial exploration period where you're looking to see the benefit, and you use that to then lean in and identify whether or not you've seen maximal effects or how best to optimize that effect. So we're looking to replicate in a placebo-controlled trial with a month-long dosing the effects that we've seen, identifying whether those effects are maintained with repeated treatment, does it match what we've seen previously, which is improvement over time. Those are the things that we're looking for in this first study. As part of the overall program, then you optimize those effects, use that for decision-making into later phase development. So we really are at the first steps of that journey. But what we've seen so far, and this is based on data that we've been collecting over many years, we're really starting to see what would be clinically meaningful effect, this 30% to 50% improvements in large percentage of the patients. And I think that gives us confidence that the KINETIC study is on the right track.

Yes. Okay. Very good. So before we switch to 217, I want to make sure we have a good amount of time to go through some of the 217 questions. But just on SAGE-718, I don't want to brush over it and not give it its due. But I think maybe we can skip over the mechanistic rationale stuff. Boosting NMDA activity is something that's talked about a lot in cognition. I want to talk a little bit about your development strategy. I know in MDD and PPD, right, you ran these small studies teasing out looking for kind of a really big and obvious effect. In the cognition space, it goes without saying, right, that effect sizes are usually really, really small. So how do you think about what would be kind of clear-cut derisking data from this open-label basket study you're running?

Well, so thanks for that. One of the things that we've seen so far with 718 is some very unique benefits yet. So as you said, a lot of people have been interested in NMDA-positive modulation for cognition for a long time. What we're finally doing is getting those data and understanding those benefits in real time. But the first step in doing that, and it's very similar to the way we moved into the mood disorder space, is you start by interrogating the system in very scientifically verifiable ways. So we started first in healthy volunteers in a model system and showed benefits in executive function, a very unique finding. Then we extended that into a population of patients with Huntington's disease and showed very similar findings. Now this is with very short dosing administrations, but the fact that we're seeing those signals means that we're able to administer the drug, we're seeing predictable effects in the systems that we're affecting, and we're starting to get some understanding about how best to measure them. What we're looking to do right now in the program is extend that to understand whether this is something specific about these model systems or in Huntington's? Or is it something that can be extended out to Parkinson's, to Alzheimer's, to other areas? And we'll use that all to then design how best to optimize it and move forward into subsequent phases of development. So new science, really exciting area, and the data that we've seen is quite compelling. I'd say executive function, for people who are not familiar with what executive function means, this is the ability to plan and then carry out complex activities. It's exactly the kind of challenges the patients with these early phases of dementia face. Improvements in executive function translate into independence, the ability to do things. And that's what people are looking for, particularly in these horrible neurodegenerative diseases. So a lot more to come there, but really compelling initial data.

And one thing I'd add, Paul, too, and this may be a good segue into zuranolone, is that what Steve is articulating is we're taking this portfolio approach to 718, right? Because we've got these multiple indications that we're looking at. We've got to follow the science and then determine, okay, what is the right path to pick which indications do we think makes the most sense. Then we have to take it up like another level and look across the 3 franchises, right? As you've heard us say many times, we really believe in the 3-franchise approach, having that depth and breadth across each of the franchises. So we have to look at not only the decisions Steve just articulated with 718, but how do you start to think about life cycle management across zuranolone and 718 and 324. So that's really what we're doing, trying to take this really disciplined approach with our resources so that we can make sure we place the bets in the right way for investors, for patients, et cetera.

Yes. Yes. Okay. Very good. So maybe let's talk about SHORELINE. So you recently announced some data, right, showing durability, efficacy, safety for 50 mgs. But I guess I'd be kind of interested in how do you think the SHORELINE data are going to be perceived by the clinician community in this space? And do you think -- given that sites are so used to chronic treatment, do you think these data without a true placebo arm will be kind of seen as believable and informative enough to be useful in clinical practice?

Yes. So this is a really important question. The way I think about SHORELINE, so I'm a psychiatrist as well, as in practiced for many years, is these are the kind of data that you need in order to understand how an episodic treatment might be used in clinical practice. Just to back it up. I mean what we're doing with zuranolone program is reconceptualizing potentially the treatment for depression, wherein a 2-week course of therapy is efficacious. Patients are able to be maintained without additional therapy, if appropriate, and only be treated as needed. And that's a medical approach. It's a very patient-focused approach. It's one that we think is really going to move the needle for folks where there has only been a very standard paradigm for many, many years, chronic therapy and so forth. But in order to get there, you have to show the drug works, right? So if they have the efficacy trials, then you also have to establish the -- for want of a better word, real-world evidence or naturalistic data that shows that you can, yes, treat patients repeatedly and have a predictable response each time, that there aren't additional safety issues that pop up by doing intermittent treatment, that patients are able to monitor their own symptoms. Of course, we know this as clinicians. But if it demonstrated, that they're able to monitor their symptoms and present themselves when they need it. And those are the -- and because these are real people, the only real way to do such a study is to do it in an open-label design. So the kind of data we've created are actually the ones that physicians, payers, the -- payers, regulators really want to look at. Like what does this look like in real life is the kind of thing that we can use for communicating. Now it's not a substitute for placebo-controlled trials. It's the kind of thing that adds to the overall program to illustrate how an entirely new paradigm will be used. So it's something we're really, really proud of. Of course, the study is ongoing. You mentioned 50 milligrams is ongoing. But what we've seen so far is quite compelling. Patients -- 50% of the patients only needed 1 treatment in a year. Greater than 70% only needed 2. And we think that's fundamental. It's a great proof point that the idea that patients don't need to be on medicines chronically, and it's something that we believe will be very impactful as we move forward with the program.

Yes, yes, yes. Okay. Great, Steve. On the regulatory side, I think you probably heard me ask this question to you 10 times over 4 [indiscernible]. But one of the most interesting things about 217 on its strengths, right, is maybe you don't need to redose all that often. On the flip side, it creates a much more complicated question, at least in my mind, surrounding what constitutes a 1-year exposure on ICH guidelines. And to that point, when you had the REDWOOD study before it was paused, right, it seemed like, and you can maybe clarify if I'm right, that the idea was to describe safety when a patient gets the max amount of course of drug per year, which is 6 courses. In the SHORELINE update, no one needed 6 courses. So what do you see -- assuming that one of these acute studies is positive, what do you see as kind of the regulatory need to generate long-term safety data, repeat exposure data? And where is kind of your head at now that you actually have this huge sort of real-world dosing database?

Yes. So just to raise it up one notch. First, we've had more than 2,500 patient exposures with the drug. We have a very clear understanding of the overall benefit risk profile, and it's a very positive one. So as we move forward with the program, that's something that's fundamental to where we are. We're -- what we've said, and this is still true, is that we require long-term data, and those data are intended to come from both SHORELINE and REDWOOD. But we believe, as you said, there's a possibility that we can talk to the FDA in the context of positive data from WATERFALL, in the context of SHORELINE data to understand what is the most efficient way to move this medicine for patients. I'll let Mike speak to the urgency. But right now, we know that the rates of depression are going through the roof. And so if there is a possibility, and we've identified 3 distinct pathways to get this medicine approved, we'd like to understand that opportunity. So Mike, maybe you want to think about the 3 -- you also mentioned them. Can you mention how...

Yes. [ Very interesting ]. That's where I was going to go, Paul. You know this well, but maybe for the rest of the folks listening, right? We intentionally set up the 3 distinct paths that you know. So PPD has its own path, right? That one additional study, which is SKYLARK, also readout next year, that could create a path in PPD. And then we have the 2 paths within MDD, which we are sort of alluding to. We have the RRT indication, the rapid response therapy indication, which is co-initiation with standard antidepressants. And then we have, which is what we're talking a lot about today, is what we think is still the endgame, is get to this episodic or treat as needed paradigm shift in depression, right? And so as Steve was just articulating, the SHORELINE data, very compelling across the overall profile of the product, right, the stats that Steve mentioned. So we really believe that, that's the right approach that we're going to continue to get there. But we've given ourselves these 2 distinct paths, right? We've got the single study of CORAL, right, which will be starting this year to get to the RRT indication. And with the treat as needed, we still believe we'll need some longer-term data, but what will that look like is a continuum. As Steve said, we will continue to bring new data that the FDA hasn't seen yet. They haven't seen SHORELINE. They haven't seen the 6-month MOUNTAIN's follow-up study. This is all data that we will share with the FDA and continue those conversations. But what we feel confident is giving ourselves those multiple shots, those multiple opportunities to get this to the market, as Steve said, to drive what is really a high unmet need and growing, and we're fully committed.

Yes. Yes, Mike, that's -- I'm glad you commented on that because I think the one question I wanted to kind of ask you that I get a lot from investors is really what is the NDA filing time line look in different scenarios. And I think, look, there's the regulatory side, right, which needs to be worked out. But then there's just the strategic side, right? You only get one chance to launch a drug. So let's say that, hypothetically, WATERFALL misses its primary endpoint, but the other 2 studies succeed. And now you've derisked a drug for acute treatment. Would you just go ahead and launch it as an acute drug? Or would you still say, you know what, we still have -- like the path to episodic treatment is still kind of just a safety-driven path. Why not still go kind of the same route you were always going to go, right? Like how do you think about those permutations from a strategic perspective?

Yes. I think you hit on the head. I think that's exactly -- the 3 paths gives you that opportunity, right? So independent of what happens in any one path, you move forward on the ones that are successful. And I think that is the strategy, Paul, that we laid out. Again, PPD can be its own indication if that was the only thing that hit. If we get RRT, we believe that's an indication. And obviously, as I said, the continued goal is to get to episodic and treat as needed because we think that's best for patients, right? We actually think that is the best outcome. And so when you think about -- if you take the RRT, right -- let me play out a scenario because we don't know yet in terms of filing time lines and success of the trial, et cetera. It is a hypothetical, right? But let's just play out the scenario that say RRT came first, say that it was positive, we had an opportunity to get there. I actually think that it's very synergistic, the RRT indication with treatment as needed. And the reason we say that is because what we hear from physicians today, they're really intrigued by treat as needed, but we know it's been a chronically treated disease in the past. So that's changed, right? You're requiring change in a marketplace which takes time. The rapid response therapy indication, the co-initiation with a current standard antidepressant, it's much more in line with how physicians treat today, right? So if you give them that option initially, and they can treat in a way that they're comfortable with, now they're going to gain real-world experience with the approved drug of zuranolone in a setting they're comfortable with, while, at the same time, we'll be able to publish data on SHORELINE and 6-month durability of MOUNTAIN, et cetera. So that by the time the treat as needed indication gets there, you've already started to move physicians and patients and payers down that path of recognizing what the potential benefit of zuranolone is. So we see them as very synergistic. And they're just 2 different approaches, same patients, but just 2 different approaches as to how you want to treat MDD.

No, that makes sense. And I know that each study can have different sort of regulatory implications. But ultimately, I think the point that you're making is that for either one of these MDD trials, if one -- if you go one out of 2 in either scenario, I guess, maybe I don't mean to put words in your mouth, Mike, but 5 years from now, do you still see the drug being used in the same manner?

In the same manner as?

Episodically, I guess.

No. I think -- listen, again, I think that's what -- we want to get to that point. And we do believe that, again, based on what we've seen in SHORELINE, based on what the data is telling us and what the market is telling us, right, in terms of what the unmet need exists, we absolutely believe that treatment as needed is fundamental to that approach, right? And the data is going to drive it. The data is going to tell us and the trial is going to tell us. But again, we -- every step of the way, we've shown that zuranolone is rapid-acting and has a strong durability, a strong tolerability profile. And as we get -- but that remaining question has continued to be the durability and how often do I retreat, but we're getting that data now, right? And as we get that data in, it's validating what we think is the right approach. The potential to have that treatment as needed is something we absolutely are committed to.

Yes, yes. Okay. Very good. We only have a few minutes left, but I want to kind of just finish up on a couple of WATERFALL questions related to learnings from MOUNTAIN. So maybe -- I know some of the stuff is proprietary. But I guess, one, what can you say about WATERFALL that has been optimized relative to MOUNTAIN learnings? And two, I think on ClinicalTrials.gov, you recently upsized the study, which happened with MOUNTAIN after, I think, looking at the data and maybe on a blinded basis, seeing certain HAM-D scores that were lower. I kind of forget the exact nuance back history and what you've disclosed. But maybe you can just sort of clarify what's going on with MOUNTAIN and if there's anything that you're seeing in enrollment that wasn't as planned.

So, well, I'll start with that question. The thing that we're seeing in enrollment is a lot of interest and a lot of velocity in the enrollment of the trial. And likely for the reasons that Mike pointed out, the rates of depression are going up, and there's a lot of interest in this drug and this mechanism. And what that's given us is the opportunity to, as you say, increase the size of the trial to allow us to have more power against secondary endpoints. We've done a few things. If you recall, when we started the study, we initially gave guidance that it would be -- we'd have the data readout in 2021. We've tightened that up to the first half of the year. And even within that, there's plenty of opportunity to enroll and give us additional power against some of the subgroups and other kinds of things that people are interested in. So that's the opportunity that we took. And we'll keep looking for opportunities like that, quite frankly, because as the study moves forward, you really look for the way to maximize the use of studies of this sort. They don't come by every day. So we want to make sure that we have -- we can glean as much information -- as much specific information as soon as we can.

Yes. Okay. And then just on MOUNTAIN, right? You did a lot of [ talking ] and looking at that data. And I think the things that you talked about were patient mix of enrollment, maybe things like timing of dosing. What have you [ up ] maybe getting rid of certain sites? What can you say?

Yes. So eliminated the sites where there wasn't great performance. We do that typically. We increased the HAM-D cutoff for enrollment. I think that's a big part of it. And perhaps the biggest driver is something that we've been talking about a lot, which is also increase the dose and ensure that all patients are above what we believe is going to be a lower threshold for efficacy. So all of those, plus a lot more were included based on the learnings from MOUNTAIN and are in line with what we've done in other trials, including the positive 201 study as well as the ROBIN study. So we're building a really foundation on our demonstrated profile that we've seen. And SHORELINE data just adds to that, the MOUNTAIN follow-up data just adds to that, and we look forward to having the MOUNTAIN -- the WATERFALL data in the first half of the year.

Awesome. Okay. Last question. Just one follow-up from an investor that was on my list, but I might have not gotten to it otherwise. On the CORAL study, I think that study is on top of background Zoloft exclusively. Is that a study that could be -- if it works, could be generalized to being combined with all antidepressants?

Yes. So here's what I'm going to tell you right now. Zoloft is in the mix. It is the most commonly used antidepressant. We're looking in it to see whether or not we want to extend that. We believe that it is a registration study. It's being designed in close contact of the breakthrough with the FDA. But Zoloft across all of our studies was the most commonly used antidepressant. And so that's a definite piece of that trial.

Okay. Okay. Great. Anything else to close with, Steve or Mike? This was great.

I'd just say thanks again, Paul. I think -- as I said in the beginning, we're really excited about the future, right? We think we put ourselves in a really good position, right? The execution to date has been really strong for the company. These milestones, this catalyst-rich time that's out ahead of us is really exciting for the company, and we're looking forward to future updates and keeping you guys posted as we make progress.

Awesome, awesome. Well, thank you both very much for joining, and thanks, everyone, for listening. And we'll see you again soon. Hopefully, you have a good productive rest of your day. Appreciate it, guys.

Thanks, Paul.

All right. Thanks, Paul.

Take care, everyone.

Bye.