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

Thanks, everybody, for joining us today. My name is Laura Chico. I am one of the senior biotech analysts at Wedbush. It is my great pleasure to introduce our next presenting company, Sage therapeutics. And joining us today, we have CEO, Jeff Jonas. Jeff, I am thrilled to have you here today. Thank you so much for participating. Before I pass it over to Jeff, let me just set the agenda real brief. First, we'll get started with Jeff delivering some introductory remarks, a brief overview to Sage and giving us a quick update on recent developments. And then we will transition to Q&A with the remaining time. [Operator Instructions] So Jeff, with that, I will turn the floor over to you.

Thanks. Firstly, thank you, everybody, for joining us today. Hope everyone is healthy and your family and friends are all healthy. I'm going to be really brief because I know people don't need to hear a can presentation from me rather hear the Q&A. We've just released our quarterly earnings. And I guess the top line is we've made really good progress across our pipeline. Things are moving along really well at Sage. The teams have really done a great job executing along our clinical programs and -- as well as with our execution elsewhere. So we're pleased with how the quarter came out. We have had some new data with -- I think I'm going to probably get questions about this from MOUNTAIN. And our early discovery group has been very productive, which led to our recent announcement for a Phase III program for COVID ARDS with ZULRESSO. So overall, things are moving on very well and -- despite the way the world is. And I'm going to stop there and then just turn it over to Laura for Q&A.

Sure. Thanks, Jeff. [Operator Instructions] But Jeff, yes, let's start off. It seems most appropriate with MOUNTAIN given the update. So earlier this week, you did report some 6-month follow-up data with a proportion of patients at 6 months showing some good durability of response. Help us understand how to put this in context, just for the audience, remember that MOUNTAIN enrolled a few different zuranolone doses versus the placebo. Study did not meet significance, but a few different ways to cut the results here. So how does that 6-month time point compare to your previous expectations? And perhaps, how should we think about it in relation to existing antidepressants?

Well, first, Laura, when we started this program, we sort of had a vision, which was how can we change the treatment paradigm in depression. And our vision, based on the science that we had at that point, was thinking about a way to avoid chronic pharmacotherapy. The current psychiatric treatment model, and especially for major depressive disorder, is a chronic therapy model. And so as we conducted our studies, we have to follow the science, with first was ZULRESSO with this mechanism. And then with zuranolone, we always had follow-up out to about 42 days. And the sort of -- and the observation we'd always made was, well, out of 42 days, we didn't see much deterioration of response after they give initial endpoints, regardless of other patients who're on other prophylactic SSRIs. So when we did MOUNTAIN, we made a decision at that point, let's ask the real question which clinicians and the label requires, which is what happens to patients over the long term. Now the conventional wisdom -- and I think many of you know, for depression is, it requires chronic therapy. And so we often get asked, well, what happens to the patients after two weeks? So in the first group of studies, we've -- in all of our studies, we've always seen rapid response. That's been an ineluctable finding within a few hours to days with 217. And that rapid response has been pretty consistent. We've always seen good tolerability, and we've never seen evidence over the 42 day follow-up out to 1.5 months or a month -- 1.5 months, I guess, a rebound withdrawal relapse. But we needed more convincing data. So when we did MOUNTAIN, we amended, in the middle of the trial, to allow patients to continue -- this before we do the results, out to 6 months in a blinded fashion. And about 50% of the patients agreed. So that was really the antecedents for why we did the study. We also wanted to do the study to really take a look at the natural course of depression. But many of us believe that the older data may no longer be relevant in today's world. And so we thought a longer-term follow-up would be necessary. So we did that with MOUNTAIN, and we got a very good response where about half the patients agreed to do the follow-up. And surprisingly, if you follow psychiatry studies, about 80% remained in the follow-up about to 6 months. So we had a very -- this was a very robust population with the usual caveats that technically randomized. So we did that. We -- first thing we wanted to know was what happens to patients who get better. And it turns out for the original population -- and if you think about this as a medical disease, they were 2 major findings. With the overall population out to 15 days, we just missed significance, as most of you know, about 0.1. And -- but we were significant at all the rest of the time points. So we did have early separation. And so for psychiatry, that's one thing. But if you think about this as any other medical disease, we demonstrated rapid onset of activity and with patients getting better faster, which in most other areas of medicine has value. But we recognize that in psychiatry that people want to know about follow-up. So we then said, what happens to these patients over 6 months. So in the original study -- in the original cohort of acceptance, that's the ITT population who agreed to the follow-up, what we found was, as you might expect, there was no difference at the endpoint. It was 0.1. And that was pretty much maintained out to day 182, I guess. And that -- and overall, about anyone who responded -- 75% of the people responded, maintained their response. If you think about this from a medical standpoint, it's not dissimilar to what you might see in a migraine study. You can get people faster with a migraine drug. But at the end of the day, if you respond as a placebo responder, your response curve will look the same. But it did tell us something interesting about the natural history of depression. And that was 2/3 of the patients overall who got better -- and this is an area of misunderstanding, remained better, whether or not they were treated. And it turns out 2/3 had no ancillary therapies. So that runs counter to the basic thesis that if you're depressed, you are a chronic patient. So that was in the population where there's no difference at endpoint. So we became interested in the more pertinent question, which is what about the group of patients where we knew we had significance and where we knew the drug we demonstrated clear activity. And that's about 2/3 of the overall population, and that's 24 and above. And there, we had a very interesting finding, which gave us more confidence about the durability of the drug, right? Because when you compare the population when there was no difference at day 15, there's not much you can say after that. But in the population where we saw a difference, that's a date -- that's a 24 and above at day 15, that affect the difference between placebo, and this includes comparison to placebo. The delta between placebo was maintained consistently out to the 6-month period. And that was done, that maintenance of effect was accomplished where -- again, 2/3 of the patients were not on any other background therapy. So that was, I thought, a very counter intuitive finding. Certainly one that goes against the current conventional wisdom. And there was no difference if you took an SSRI or not. And even though we have -- that's really only about 1/4 of the initial population and so it's really not powered for significance. By day -- by the 6 months, that almost hit significance. The difference between placebo, who are also filed at 6 months, that delta was maintained with a P 0.06. So we almost made significance at that point. So from that, we take -- we really take apart 2 points. One is, our early data from the science suggests that this drug works differently than antidepressants. It's not tweaking receptor's system that cross-talk with each other. It's doing something different. So it's really -- I wouldn't even -- we don't like to even call it an antidepressant because the mechanism looks different. And we think -- our early data has always suggested there might be a fundamental reset in neural circuitry. And these data are yet other another step, and we think confirming the hypothesis that with zuranolone therapy, if it's approved, could offer a distinct and sort of disruptive opportunity for patients to be on a drug that doesn't require chronic pharmacotherapy. So that's how we looked at these data.

Interesting. And yes, I think the relationship you're mentioning with regards to whether or not antidepressants were on board is also very interesting with regards to that. Okay. We have talked about this before, Jeff, and also done our own evaluation of the PK data. The 50-milligram level for zuranolone would certainly seem to offer more extensive exposure basis. But curious, what gives you the most confidence that this is the dose level. And I think there's a lot of early data with zuranolone looking at the tolerability profile. And really an interesting dose response relationship. But -- yes, how -- what's your comfort level with regards to the 50 milligram dose?

So there are a couple of points. Let me just back off for one moment, I'm going to be annoying. I was on a call earlier in the day with another psychiatrists who worked ERs. And we were talking about people trying to extrapolate liquid formulations to solid. And I was making this point, which is, if you've worked in ER, you've given someone 50 or 100 of benadryl intramuscular, it will knock them out. You give your kid 50 milligrams and if you have that I have, it barely does anything. So it's really -- you really can't extrapolate from [ bolusing ] and -- or from rapid uptake. And so with what we've done, one of the things we have is I have to say, right now, we probably have the largest clinical database with a postsynaptic GABA agonist, a positive allosteric module, I guess, say, in the world. And one of the things that MOUNTAIN -- we don't want to learn from failure too much, but you do learn from failure. And one of the things we learned from MOUNTAIN is that we were very fortunate in the sense of having the 20-milligram arm that was enacted. Because that gave us extensive dose response information, which we couldn't have gotten otherwise. You can't get those response on an open-label study. Of course, you can't. That's sort of silly. And so you have to have placebo and you need an arm that doesn't work. And -- so I don't want to learn this way again. But what we learned in MOUNTAIN gave us extensive population PK data. And that gave us population dose response. And we learned and we could see where the variability occurred with 30. And we were able to calculate based on dose responses, this is a well-behaved molecule. What the dose is that ought to attenuate or mitigate the kind of variability we saw in MOUNTAIN. So that's -- so we really use some basic since that. The other point is that a lot of people conflate this kind of drug with a benzo. And that's, of course, inaccurate. You can't titrate to sedation, which is foolish and, of course, makes no sense. And you also -- there are biomarkers, things like beta power, which work in some levels, but are really not adequate to give you dose response. And -- so we as a company have developed some proprietary information, which we're not sharing really for competitive reasons. And that really led us to the 50 milligram. The one thing that I'd emphasize is in any -- 50-milligram was not the top dose limited by adverse events. That has never been an issue and as -- with respect to the oral formulation. And as we made clear in our release, we've had doses up to 90 milligrams that are well tolerated. So we've treated more -- thousands of patients now, more than 2,500 with these doses up to 90 milligrams. I've never seen -- we didn't expect to and when we haven't seen lots of consciousness or anything like that. So we chose the 50 based on an extensive internal database of dose response population PK, and some things that are frankly proprietary. Because we know there are companies that are following along with us, which -- -- imitation is, of course, flattery, but we're really not interested in providing more of a road map.

Sure. Understood. And if you could see me, I'm nodding in agreement.

I just see your glamor photo, as I said earlier.

That's -- I think that's really helpful color. And just can you remind us when we might get to see some of the detailed PK data? Because I think that is really the key question right now in terms of the dose response. And to your point, you have kind of accumulated a really robust database there. So when might we get a little peek under the hood?

Well, I think, not for a while. I think the agency has it. We'll see the clinical data with WATERFALL and hopefully, this 50-milligram with SHORELINE. But with all due respect, we are we know there are people who are fast following. We're pleased that the GABA mechanism we've pioneered is now being mimicked. But we also -- I think MOUNTAIN shows clearly, dose response in this type of molecule with this particular mechanism, is not straightforward. It's not beta power and a sleep scale. It's much more elaborate. And so there's no safety concern. So we're going to keep as much proprietary as we can. Basically public disclosure because it's material from a safety standpoint.

Okay. Understood. Maybe not necessarily a MOUNTAIN question, but one, different data question. Can you remind us on the abuse potential for zuranolone and kind of where you're at in formally studying those aspects of the drug?

So we're sort of -- we're a bunch of people who have done a lot of drug development. So those studies are completed. We made a decision early on that it didn't make strategic sense to do things like abuse liability, driving studies, drug-drug interaction while you were still doing your Phase III. Because you really want to be able to design your Phase III appropriately, and you really can't unless you have those kinds of information. So we have completed those studies, and we haven't made them public yet, but there are no troublesome signals in any of those studies. So we're very comfortable with what we're seeing. And of course, as you probably well know, those -- many of those studies involve super therapeutic doses, and -- which is why we can say we've treated up to 90 milligrams. Because we have. So all I can say there's no troublesome signal, nothing that is stubbornly rate-limiting and nothing that would have impacted how we selected our Phase III dosing.

Okay. Very helpful. Maybe let's transition a little bit to SHORELINE. It might be appropriate to say what pandemic the enrollment here appears to be moving along pretty well. I'm just curious if you could kind of elaborate a little bit on steps you've taken to maybe increase recruitment rates. And I'm not sure if that's actually what's happening here, but is this a function of site expansion? Or are there just other drivers at play?

Yes, there are actually multiple drivers. One is the team. We've been preparing for this for a long time, and we have a bunch of -- well, most of the team are veterans. And every trial has certain obstacles and COVID was just another, and they did a great job preparing for it. So let me just give them those kudos. That being said, there are a couple of things that we've been picking up from the sites. And as we've said, both SHORELINE and WATERFALL have been rolling ahead of expectations. And so there is some feedback we've got. One is, there's more depression right now and people are depressed and with actual diagnosis of major depression, not like bad mood states. The other interesting feedback we've gotten is that people find this dose regimen very user friendly. And if you think about this, running a study that require 6 to 8 weeks of antidepressant, plus 6 months of follow-up on the drug, monitoring safety, doing blood labs it's very cumbersome. And it's cumbersome for trial sites and it's cumbersome for patients. And this really reflects directly what we hope will be the value proposition for zuranolone. We know this is different in the way psychiatrists and other people think. But this is a drug that can be administered over 2 weeks and then you're done. And that is very user friendly. And patients have been really very willing to undergo our studies. And I think it's a direct reflection of the product profile and the potential benefit that it offers. So -- and plus work by the team. So we've been familiar with telemedicine. We have some other interventions we've done that probably are more proprietary. But it's really a combination of what I think is the target product profile of the drug, the user friendliness of this approach to treat depression and, basically, the work of the team to prepare in advance for this type of obstacle.

Okay. And just for our audience, just recall, SHORELINE is evaluating zuranolone on an as-needed basis over a year. The zuranolone program is pretty comprehensive if you look at the Landscape Program. If -- you're on track to get the 30 mg dose data later this year. I think you might have also mentioned on the earnings call, there might be a possibility of getting 50-milligram data. Could you just talk about the drivers a little bit there in terms of what could pull that forward? Is that entirely dependent on recruitment rates? And then just given the discussion we had so far on dosing, maybe help us understand, is one of these data sets more important versus the other? And by that, I mean, the 30 mg versus the 50 mg, is that -- is there one that we should be paying more attention to?

Well, I think the 30 mg will be more complete. And remember, this is -- if you go back and remember what people -- the typical complaint about clinical trials is that people have said, this doesn't reflect real-world usage. And so what SHORELINE does is that it's a real world study. And we believe it's the largest of its kind that has been conducted. So what you'll learn from SHORELINE, remember, so this is a real world study, which asks a more fundamental question for clinicians and the label, which is if you get better, what do you do with the patient? What happens? And that's what SHORELINE will answer. So we'll know -- we'll get a sense of retreatment of adverse events, relapse rebound over a very long period of a year and we'll get a sense of how the drug performs. We are allowing patients to be retreated with 50. So we might get some anecdotal incremental ideas about benefit from that. And we are going to do -- and we're also initiating a cohort and we have initiated a cohort with 50 milligrams to get experience directly with the 50s. So hopefully, we'll see all of that by the end of the year. We'd like to tie it up in one announcement. And really the rate-limiting steps, frankly, are data cleaning in this kind of study. Enrollment has gone very well. A lot of data cleaning often has to be done at the site, and that's often the rate-limiting effect. And a lot of people think you get last patient, last visit, and the next day you have your data. It doesn't really work that way. So that's really what will be the time limiting effect for our announcement. But we are encouraged by our progress, and we hope -- we're optimistic but we can't guarantee that we'll have the 50. I think it's fair to say the 50 -- although we're not seeing any signals at all among 50 and above, clearly, I think people want to see more data, and hopefully, we'll have it by that. But right now, we're comfortable, especially with the safety profile.

Okay. And just because you mentioned you want to put those 2 together, should we be assuming that if the 50 mg data is not in hand that you kind of hold these, could that drift a little bit towards, I guess, the first quarter of '21. Just to give...

We're going to put out the 30 because we said we would. So we will.

Okay. All right.

Yes. No, the 50 we'll do -- I'm hopeful we'll have it done by the end of the year. Remember the 50 won't be a year, though. So the 50 is going to give you more of a clue about initial response and tolerability and maybe some retreatment. So really the 50, it's something we want to do. We want to take a look at the dose. But we're pretty comfortable already with it. But we've already committed to it, so we'll try to do it. But it's enrolling very quickly, as you mentioned earlier. That and WATERFALL both really, enrolling quickly at this point.

Okay. Maybe last one on zuranolone. I do want to leave time for the pipeline because I feel like that's not getting enough fair time lately. But last one here. Your -- actually, your recent deck is really helpful because you have the grid of different studies going on for zuranolone. And you've highlighted 3 potential indications for zuranolone. So basically 3 ways to win, if you will. How should we think about the path with the most rapid trajectory to market, just to put that in context?

Well, remember, when we started this, we had the approach that we did one more study to get 2. And so the postpartum and the MDD was dependent on that. When we made a decision, we didn't want to have that interdependence anymore. So each of these studies individually will allow a filing without having to wait for another study. So that's another acute study. So we basically do have 3 shots on goal. The sequencing is probably too early to project right now. I think we need to -- with WATERFALL, we'll want to see SHORELINE, see what kind of design we might need for REDWOOD. And then the other 2 depends when they come in 2021. So this is really just going to be a tactical decision. If they come in within a month or 2, we might do one thing. If they're going to be separated by 5 months, we may do another -- and kind of we roll in or not. So there are a number of regulatory strategies. I think it's -- right now, it's probably too soon to speculate on sequencing. And I guess all I can say is right now that everything is on track or ahead on schedule.

Okay. That makes sense. I did want to leave time for the pipeline here. So I'm going to go with 718 first. So I think I like to sloth around and uncover little nuggets of Trivia. I actually struggle to find something about NMDA. So the first one I would say is just that pubmed citations for NMDA modulators are actually on the decline since 2015. So maybe stepping back for a moment. Can you just tell us about NMDA as a target? Yes, just kind of what we should be thinking about how this fits into the overall Sage strategy, but also the potential for candidates to target NMDA?

So as you know, we have an NMDA positive allosteric modulator, which we think is maybe one of the leading activators right now in development. They've been very hard to develop. And that's because NMDA, in general, is an activating receptor system in the brain. It's long been known to be involved in things like cognition -- on the positive side, cognition as well as long-term processing, knowledge acquisition and things of that nature. When we started looking at these, we made a decision that you can go high, with PAM or go low in the NAM and block. And our belief at the time and remains this is that on the negative side, there's a little bit more cross-talk between receptor systems that makes best efficacy harder to ascertain. They tend to be more pleomorphic, for want of a better word, or maybe promiscuous in terms of receptor activation. So for example, last year at ACNP, the Stanford group presented data that suggested the effect of esketamine could be blocked by naltrexone -- naloxone. We -- but people say, well, NAMs PAMS have no opioid activity. So we decided not to pursue that, although we do have some interesting NAM programs that we may bring forward, but for different types of indications. So we focused on the positive allosteric modulators. In order to be fair, there's not a lot of literature on them because people have had a hard time making them. And our discovery group and our translational group has done a really good job in producing this library of chemical equity. So we decided to focus on the PAMs with the idea that it was a less competitive space, there is more unmet medical need and really knowing -- and we just thought that, that offer us order was a better opportunity as a company. So that was our initial thinking. And then we did have some surprising data. We expected the NMDA PAMs to be active, primarily in disease states, where there was low levels of NMDA activity. And that was, of course, Huntington's. And there, we saw this effect on executive function, which was distinct from what you see, say, with modafinil or with amphetamines, where you really have a performance benefit, but not a cognitive benefit. So that was unique, but then we found the same effect in normals. So it really offered us a really quite a unique opportunity with a novel mechanism that was very, very active. The trick with this mechanism, and the issue has always been that at higher levels, you get excessive activation, potentially seizures. So finding the right molecule at the right margins was probably the tricky part. So one of the gratifying things we've discovered so far in our clinical trials is that the adverse event profiles have been extremely benign even in the face of these clinical effects. So we think this is a unique opportunity. We are looking at some other indications before we commit to a double-blind study. We're looking at Parkinson's dementia. We're likely to look at Alzheimer's, potentially other maybe schizophrenia as we move forward, because we really don't -- we really don't have a handle yet on how this impact may be manifested clinically and how it is best measured.

Okay. And I actually went back and found a press release dated April 27, 2017. That was the start of first stage 708 -- 718 dosing study. So it does feel like there's been a lot of work going on in the background to kind of tease out some of these dynamics that you're talking about. And you just mentioned Alzheimer's. With that in mind, I was kind of thinking big picture here, too. You've looked at Huntington's, we're getting the Parkinson's disease data soon. Is the end goal here to look at a broader application like Alzheimer's or schizophrenia? And then to that end, we've got the Parkinson's disease data coming. What is the goal here? And what does good "data" look like?

Well, good data for us is going to be -- remember -- so there's one thing about open-label data is you have to be cautious about using it. But when no one's ever tried to do mechanism, I think it's typically more reliable. And for what -- for us, we want to see what the effect sizes are. One of the things we've done we've run small trials, and they've been small because we've used these open-label trials to guide our designs of our Phase IIs and Phase IIIs. So we'd like to see what the effect looks like in an Alzheimer's, in a potential schizophrenia as well as Parkinson's dementia, before we commit to where the science takes us. This is going to be driven by where the drug operates best and where the science gives us the best data. So -- and that's how we'll pursue it. We think the Huntington's data looks good. But I was in Florence when we filed namenda. I don't think there's anyone who works in medicine who wouldn't love to see something that can attenuate symptomatology in Alzheimer's. And certainly, executive function. Failure of executive function is one of the principal causes of disability in Alzheimer's and other diseases where it's impaired. So it's something where we may be able to make a big difference. So we'll wait to see what the data tell us and then we'll commit. It's not -- it won't just be driven by market. It will be driven by the science and where we think we have the best opportunity to create patient benefit.

Okay. Makes sense. And I see we only have a couple of minutes left, so I'm going to throw out one more on 718 and then a FutureCast question. So 718, is there any chance we see a combination therapy regimen? So maybe an NMDA with GABA? And then second, for FutureCast, that's coming up in September, any chance you're going to be willing to give us a little teaser in terms of what we should be looking for or expecting at the update?

Well, it's interesting. We have a lot of chemical equity. And so one of the things we'll see at FutureCast is some of the stuff that's gone on in the background, not necessarily things that people think of material. But by example, we've done work in inflammation, which is how we got to the COVID program. We've been looking at peripheral GABAs. We've done work in other central anti-inflammatory pathways. We may not have made targets, but we have a lot of work there. We actually have molecules that look to be combination NMDA GABAs in one molecule. So we haven't decided what we're going to show in terms of our early pipeline because it's more Skunkworks. But a lot of it has been very productive. We'll talk a lot about, obviously, 324, because we're very excited about the opportunity in essential tremor. We'll do some -- we'll do talk about 718, talk about some new stuff on 217, more licensed pharmacoeconomics. And then we're going to talk a lot about what our discovery and our earlier teams have done in order to enhance the company's future and productivity over time.

Well, definitely looking forward to that one. All right. Well, I think we're kind of coming up on the end of our time with you, Jeff. But thank you so much for joining us. Always a pleasure to connect. Everybody in the audience, thank you as well for joining us. And with that, we will conclude the Sage Therapeutics presentation. Have a great day, everybody.

Thanks, everybody. Appreciate the time.