Hi everyone, and welcome to the atai Life Sciences Q3 Earnings Call. I'm Stephen Bardin, CFO of atai, and I'm very pleased to have you with us. I'm joined by our CEO and co-founder, Florian Brand, as well as our CSO and co-founder, Srinivas Rao. Today, we'll be speaking with two of our long-term analysts, Andrew Tsai from Jefferies and Ritu Baral from Cowen. Over the next 30 minutes, Andrew and Ritu will be asking us about recent progress across the pipeline and upcoming catalysts. Now for the standard reminder that our discussion today may include forward-looking statements about atai's future results and performance, subject to risks and uncertainties that may cause actual results to differ. Atai does not undertake any obligation to update such statements, which speak only as of today.
Before kicking off the interview, I'll comment on the key financial takeaways included in today's earnings release. First, our Q3 operating use of cash was $28 million and was in line with expectations, so no surprises there. Next, I want to remind everyone of our previously announced debt facility with Hercules, which provides access to up to $175 million. Back in mid-August, when we closed the facility, we drew down an initial $15 million. We ended Q3 with $304 million in cash. This cash balance, plus the Hercules facility, gives us a strong capital position. We reiterate our cash runway extending into 2025. Let's start our interview section. I'll hand the mic over to Ritu from Cowen to get us started. Ritu.
Thanks, Stephen. Florian, there's been a lot of news flow around your company as of late. Before we get to it, could you please provide a brief high-level summary of your most recent progress and upcoming milestones for the ones who might not be too familiar with it, with atai?
Absolutely, Ritu, and great to have you back, both of you, Andrew and Ritu, on this earnings call interview format. Quickly as a recap, our ultimate goal at atai is to achieve clinically meaningful and sustained behavioral change in mental health patients. As of today, we have eight drug development programs in the clinics, and we also have four enabling technologies that we deem to be very complementary with those eight programs. All of those eight programs target several major indications, neuropsychiatric indications. In addition, all of these programs have either already achieved proof of concept in the case of COMP360 or have a path to efficacy data over the next two years. We expect that these upcoming efficacy data sets are really going to be value inflection points for atai.
Looking at last quarter very briefly, in addition to multiple clinical trial initiations, we have announced two very encouraging positive initial Phase I results with GRX-917, our deuterated etifoxine compound, and with KUR-101, our deuterated mitragynine compound. Both of them showed very, very intriguing PK/PD results and encouraged us to drive the development of those programs further. Last but not least, of course, we have PCN-101. We announced that we completed our enrollment in this study. That's the Phase IIA proof of concept study in TRD with R-ketamine, and we are very much looking forward to the data coming out of this trial around the end of this year.
Florian, speaking of PCN-101, can you share more details about the planned target clinical profile of that drug?
Sure. Our plan with PCN-101 R-ketamine is really to be differentiated from other depression treatments on the following dimensions or elements. We're targeting unsupervised at-home administration, rapid onset of effect within 24- hours, and we are looking or anticipating an intermittent dosing regimen. Looking at approximately two times dosing per week and are of course, very much focused on developing this compound in treatment-resistant depression. Not MDD, but TRD.
If the profile forces you to choose, if the PK/PD forces you to choose, would you aim for a therapy that has stronger efficacy or one that is far more tolerable than current treatment options? How do you see the biggest unmet need and the biggest opportunity?
Yeah. The key for this program is really, demonstrating data that allows at-home use. That's why we also like to demonstrate dissociation or sedation profile of PCN-101 that allows that, to happen. Basically to administer this compound at home without supervision, while, in terms of efficacy, we're looking, in terms of our internal product goal, to be comparable to other approved depression treatments such as Esketamine, as kind of a prominent other enantiomer of ketamine.
Got it. If you do reach that, the previously disclosed target safety risk ratio of 2, I think, on sedation and dissociation, exactly what does that risk ratio of 2 mean practically? What's the math behind that analysis?
Yeah. Let me jump in on that point. Basically, the risk ratio is the way we're calculating risk ratio is very similar to what was used with SPRAVATO and their summary basis of approval. Essentially, you can. You know, there are normal ranges for both the MOAA/S and the CADSS, and you can just, you know, basically break the patients into those that were abnormal and normal in each arm, right? Now you have a percentage. And of course, that gives you. You can then take those percentages and compare them across arms. And that's essentially how you generate a risk ratio, right? If it's 20% abnormal, or sedated, for example, in a drug arm, it's 10% in a placebo arm, then obviously that's a risk ratio of 2.
Rather than on an individual basis, it's more of a population analysis on a binary basis.
That's exactly right. The agency looked at normal versus abnormal, and they also looked at more severe incidents. Severe sedation, for example, was something that they also looked at. This is the general tack that we'll be taking. Of course, the idea, as Florian mentioned, is that we are really targeting at home use. The reason that we defined risk ratio as less than 2 is, you know, it's a criterion to say that it's comparable in many ways to placebo. That's essentially the argument. We know that for sedation, there are plenty of drugs out there that are quite sedating and in fact, you know, sedating many fold higher than placebo. There's less data around dissociation per se, but there's certainly data around hallucinations that can be quite prominent. Eszopiclone is a fantastic example. That's Lunesta.
It's a sedative hypnotic. It's pretty widely used. 0% hallucinations for placebo, 3% hallucinations on the drug. I mean, you get a sense of how prominent some of these, at least from a risk ratio perspective, side effects can be. We're saying, "Okay, fine. Let's get a number, let's get a ratio that's comparable to placebo, markedly better, markedly better than SPRAVATO," and you know, put our best foot forward in that regard.
What's the practical profile of a drug like that that will best address that unmet need, that at home need? You know, how far can you push that, and what's a good comp, maybe other than Lunesta? Are there other at home profiles that would represent a good comp for what's accepted?
Yeah, that's an interesting question. The FDA, in the approval of SPRAVATO, did something that I think is a little unusual. When you talk about dissociation normally or just talk about any adverse events, you include sedation, et cetera, you normally are thinking about adverse events, right? These are the spontaneous adverse event rates. In the case of SPRAVATO, there was a focus on two instruments that were specifically designed to measure those particular outcomes, sedation and dissociation. Sedation, they used MOAA/S, and dissociation, they used CADSS, and that was their point. That was the basis of everything. There's a certain logic to that. Adverse event rates can change, as the doctors, you know, as the doctors get trained on something. Dissociation is kind of a tough one to pick up on. I mean, what does it mean?
What does a patient report to a doctor when?
Right. Right.
You know, "I'm dissociated," right? You have to map certain terms that the patient is using onto a term, a metric term, a preferred term, and that can get a little bit complicated. Using CADSS, using MOAA/S sidesteps that entire issue. The reason I bring that up is we don't have an exact analog in that regard. It turns out that no drug has either the CADSS or the MOAA/S in the label in the United States, to the best of our knowledge, and we did a fair amount of exploration in that regard. We went back to, you know, again, hallucination rates and, you know, things of that nature. Certainly, Belviq is an interesting one. Belviq or lorcaserin is an interesting analog. It's more around multiples. We know that at multiples, that drug is hallucinating, creates quite pronounced hallucinations.
That's an interesting kind of product because that's kind of what we're trying to get to here, right? Dissociation. We're trying to get to a therapeutic index. We're trying to get a multiple of the therapeutic dose that results in dissociation. Very analogous in many ways.
Thanks. Andrew?
Sure. Thanks, Ritu. Like, I think you guys, we touched on safety enough, but I do want to stay on PCN-101 because it could be a big data readout for you guys. On that framework, maybe let's shift to efficacy then. How did you decide, five points versus placebo at hour 24 to be, quote-unquote, the bar? What went into your calculus? You know, why not four? Why not three? Why five?
Well, as Florian mentioned, what we're looking for ultimately in our label based on the pivotal studies is on the order of four points MADRS change at the final endpoints, which could be, you know, four weeks, six weeks, et cetera. We clearly have not got that, you know, we haven't guided on that, and we don't have a good sense of that just yet. That's what we're looking for ultimately. We kind of backtrack from there in essence, right? What should it look like considering this is a Phase II trial? This is a single dose study. It's a different formulation. It's our first experiment in this regard. That's how we kind of backtrack, and that's how we got to five. Also, you can triangulate on this number.
If you look at other trials of ketamine and S-ketamine that have similar size, you know, on the order of 30 patients per arm, your numbers are in the five kind of range, five to seven, somewhere in that range. Again, these are sort of the two different data points that we use to triangulate on the number.
I see. You know, naturally, Wall Street tends to compare cross-trials, you know, maybe when you report the data, they might be tempted to compare what Spravato shows at hour 24. I guess why would that be fair? Or if not, why is it not fair to compare directly, I guess?
It is potentially depending on the study, right? The larger studies, the Canuso 2018, those types of studies are not unreasonable in terms of the size. I think where you're headed with that is focusing on that very first Phase IIA trial that J&J conducted with the IV form of Esketamine.
Right.
There are some features of that trial that may have exaggerated the magnitude of efficacy there. First of all, it's a small study, right? It's only 10 per group. In fact, the 0.2 mg/kg group was only nine subjects. So that's certainly one issue. It's, you know, they had many fewer sites than we do. We have over 22 sites. They had nine. We don't know how many of those are actually actively recruited, of course. So that may be another factor. The most important factor, however, is that Esketamine was quite functionally unblinding in that trial, right? So if you look at the CADSS numbers for placebo were very small.
The CADSS number for either 0.2 MPK or 0.4 MPK were really quite high. By definition, there was a very pronounced functional unblinding that was occurring in that trial that could very well have swayed the results. Obviously, that is not something that we're anticipating with the doses that we chose in our Phase IIA trial.
Makes sense. You know, this is a single-dose study, but you know, we will be curious what durability could show. It seems like patients are followed up for a week or two in the study. Correct me if I'm wrong. What should we expect on durability of efficacy?
Yes. You're absolutely right. We're looking at the primary, of course, is MADRS change at 24- hours, but we are looking at seven days post-dosing as well as 14 days post-dosing. That's what those are obviously secondary endpoints. We do, you know, the clinical data here is the previously published third-party study of Arketamine. They showed a very pronounced effect at 24- hours or roughly 20-point change, placebo-corrected, of course. They did show a durable response that went out to seven days, which is their maximum time point. That was the last time point that they actually assessed. They did show a robust efficacy at that point. That would be wonderful to see, to replicate that result, of course, and then, of course, follow that out to 14 days.
Makes sense. Like, bigger picture, you know, what would cause you to reassess moving forward? What is the scenario in which you say, "Hmm, maybe we shouldn't go into a Phase IIB," or something like that? What would you see on efficacy or safety?
On efficacy, I think, you mentioned it before, so this is really a single-dose study, right? Of PCN-101, while the TPP actually foresees multiple doses per week, as we discussed earlier. So it's important to remember that this very first study is really all about testing whether, as Srini mentioned, we have a greater therapeutic index compared to other approved treatment options. Regarding potential thresholds, I won't go into any specifics, but, what I can tell you is that, of course, efficacy of a single dose below what is considered to be clinically meaningful would certainly lead to us reassessing the current clinical development plan in treatment-resistant depression.
Let me emphasize also that based on all the existing preclinical and clinical data that we have from third parties and internally studying Arketamine, that we remain very confident to achieve the objectives that we guided on, disclosed, and discussed in our R&D Day in terms of trial objectives for this specific Phase IIA.
Okay. All right. Thanks. Ritu, I'll hand it back to you.
Thanks, Andrew. How did you guys select the two doses for this trial? Would you consider bringing two doses with two different profiles, again, going back to that safety versus efficacy question forward into Phase III?
Yeah. The dose selection was based on our Phase I, full stop. We conducted that trial two years ago now, and we looked at a range of doses up to 150 milligrams, so 30, 60, 90 in between. We picked doses that were consistent with our hypothesis around therapeutic index. What I mean by that is we picked one dose that was sub-dissociative, essentially, and then the other one, and that's 30 milligrams, of course, and the other one that is essentially threshold dissociative, 60 milligrams. This allows us to properly test our hypothesis that we will have a therapeutic index, pretty much by definition, based on the results from our Phase I healthy volunteer study. Now, to your second question about whether we go forward with additional doses, I mean, yes, potentially, of course.
I mean, it depends on the results of this study as well as subsequent trials. There's a rich history of bringing forward multiple doses. Doctors love that, in general, to have that flexibility. Exactly right. I mean, there may be sort of different balance of efficacy and tolerability at the two different doses that may be very important, again, from a physician's perspective, vis-à-vis a particular patient that's sitting in front of them.
Thank you. As you think about your subcu bridging study, what are you trying to replicate from a PK/PD perspective from the IV? Is it the peak? Is it the AUC total exposure? I mean, what feature of that PD do you think is most important to confer an antidepressant effect through this mechanism?
It's a fascinating question in the sense that there's not a lot of discussion on this. I mean, what is driving the efficacy of ketamine and S-ketamine? Is it the AUC? Is it the Cmax? There seems to be some coalescence around the concept that it's a Cmax driven effect with S-ketamine. I don't know how strong that data really is, but that seems to be where people have come to. However, if you think about Auvelity, which is another glutamatergic compound that has an NMDA antagonism, it seems to be more AUC. Insofar as the half-life of dextromethorphan, and when buprenorphine is around, is really quite long. So it's just kind of sitting around there. You're building up a level over the course of the first five days or so, four or five days.
In that case, it's probably AUC. It's probably not being driven by Cmax. We have these two data points that are a bit discordant. In the end, the most important thing that we really need to get a handle on is the true bioavailability. One anticipates the bioavailability from a subcu perspective should be 100%, right? I mean, there's not likely to be any kind or close to 100%. There's not likely to be any local metabolism, for example. Nonetheless, that's the most important thing that we really want to assess first. Then, of course, we'll look more closely at Cmax. Overall duration of effect is another one, you know, the total time. All these factors, I mean, all these things will be looked into. Then you basically do the lineup against the effective doses.
You may do, depending on what those numbers look like, you may do some bracketing, right? You may pick a dose. For example, let's just say 30 is effective. You may pick a dose that actually replicates the AUC, and then another one that might be a little different on AUC, but actually replicates the Cmax. This is pretty much what J&J did in their development. They went, as we've discussed, from that IV trial to an intranasal, but they did do dose ranging around the intranasal once again.
Got it. Andrew?
Great. At the end of the day, I think it's just, you know, placebo-adjusted efficacy of five is ideally what you want to see. On safety, something comparable to placebo on sedation, dissociation, hallucination, something significantly less than what SPRAVATO shows. Is that kind of a fair takeaway?
Perfect summary. Yep.
All right. Very good. Maybe we can shift to COMPASS, actually. They recently announced their Phase III plans. In my notes, you know, it's actually pretty different. One's a placebo-controlled, another one is a repeat dose, I believe. They're obviously different, two different studies after meeting with the FDA. In your view, is the FDA flexible with them doing either option, or is it your view that the FDA wants sponsors as they move to Phase III to do both types of studies?
I mean, it's an interesting question. The short answer is we don't know for sure, right? We have the information that COMPASS provided to us. What I can say more broadly, however, is that the FDA is comfortable with different sorts of controls, depending on the drug and depending on the indication. There's the ICH E10 guidance that says, you know, dose control is okay, and placebo is kind of a dose control. The dose is zero, right? They're fundamentally okay with that. MAPS is obviously doing placebo-controlled only. Why COMPASS settled on one that's dose-controlled and the other is placebo-controlled, really not 100% certain. On the other hand, it is interesting. It's helpful as we move forward through our own programs.
You know, obviously something that we'll take under advisement as we design those next trials.
You know, maybe one more on COMPASS is, you know, they are even using integration or as well as adding digital support now in Phase III. You know, just how important to you, in your view, is the use of these tools in terms of maximizing the outcome, I guess? Is there a way to quantify or is this? Is it your view that these tools serve to augment the psychedelic? Or go ahead and talk about what you think, I guess.
Yeah. I mean, certainly this is an area of significant focus for us, and you guys have been with us for quite some time. This is something that we've talked about for a long time. The original concept behind IntroSpect really came together in late 2019, and IntroSpect itself was formed in early 2020, as I recall. This is something that's been really near and dear to our hearts. The reason for that is that digital therapeutics offer a potential means of allowing for scalability but also reproducibility, if you will, or standardization of therapy. There is a strong suggestion that, well, certainly prep work is important, right? That's something that COMPASS has also indicated. I mean, it is a very unique experience. A psychedelic experience is very unique.
If you're uninitiated, that can be, you know, that can be very disconcerting. You know, telling the patient what to expect, setting intentions, setting expectations, all of these things are really critical prior to the psychedelic effect. There is this concept, of course, of behavioral plasticity that occurs after the administration of these compounds. It's interesting to think about really harnessing that behavioral plasticity to affect behavioral changes. That is also a central motif in, you know, our strategic vision, right? This concept of behavioral change. It's one thing to just change a symptom, and an SSRI, for example, can do that. It's another thing to actually be able to change behavior to allow for a more durable response, right?
Whether that behavior is improved socialization or improved exercise or something, if you have this period where there may be more flexibility, there may be more malleability to change some habits, that's fantastic, and one should really kind of take advantage of that. That's what we want to do with our digital therapeutics. Yeah, it's really important to us. It was fascinating that there was no factorial design, and we highlighted that in our R&D Day. The caveat, of course, is, how is COMPASS viewing these digital therapies or these therapies, right? Do they? You know, we have been very specific in indicating that we would love to have our products develop as combination therapies, right? Combination therapy being a term of art, it's a drug and a device put together, you know, inseparable in their label.
That's obviously our vision, and that's something that we are working towards. You know, again, COMPASS has not provided guidance on how they're approaching it.
Thanks. Makes sense. Ritu, I'll turn it back to you on other questions about the pipeline.
Thanks, Andrew. I want to move to 007. You know, as you look to the next trial, the Phase IIB trial for RL-007 in CIAS, how did you choose that primary endpoint, the MCCB scale, efficacy endpoint for this trial? And would it be an approvable pivotal endpoint? And then that secondary endpoint, the VRFCAT, what does that assess?
Yeah. The MCCB is in fact or has been in fact supported in the past by the FDA as an approvable endpoint for CIAS. Actually, they have been involved or has been developed with the FDA's input. While the MCCB broadly assesses cognition, the VRFCAT assesses the impact of changes of cognitive function by basically simulating activities of daily living. That basically went into then also the decision, of course, around endpoints for this specific trial.
How did you choose the 20 mg and the 40 mg doses for this study?
The doses were essentially selected based on our biomarker study in patients where we read out the data or where we published the data earlier this year. The 20 milligram and 40 milligram doses both showed clinically meaningful pro-cognitive effects, both of them. That basically then was also the basis to go forward with those doses in the larger Phase IIB study that we anticipate to initiate still this year.
Great. Andrew?
Thanks. Maybe one more on that topic. I believe Biogen actually failed with their CIAS asset this year. Does that kind of shake your confidence at all? Do you know if they had any biomarker changes in their initial data sets? Presumably you're following that asset, I guess.
We are following that asset. We're certainly following aspects of the path that they've placed. You know, this is obviously a challenging indication. There's no doubt about that. That, of course, is part of the appeal of this indication. Obviously, you know, there's a huge unmet medical need here. The cost of these patients to society and, of course, the personal cost is very, very high. You know, even small changes in cognition and as reflected in the VRFCAT, for example, and of course, ultimately in actual activities of daily living, that's going to be pretty key. In terms of the biomarker data, I don't recall seeing anything published on that personally.
Right.
It will be interesting to get a better sense of that or even get a more granular read on the results that they had.
Great. We're already hitting the top of the 30 minutes, unfortunately. Thank you so much, Andrew and Ritu, for being with us today. Really grateful that you joined again.
Thank you.
We covered a lot of programs in this session, and that, I believe, personally speaks to the breadth of innovation that we, or that is happening at atai. Looking at the last quarter, we've made meaningful progress, in my view, across our entire pipeline. We have delivered results that have met the high standards that we set, ourselves for our own programs. Looking ahead, we cannot wait to share, the eagerly anticipated top-line results of Phase IIA of R-ketamine, of course. We anticipate that for around the end of this year. This specific data set can be potentially, as Andrew and also Srini mentioned, be a significant event for atai as, we're targeting an at-home therapy for treatment-resistant depression.
With our strong balance sheet and runway to 2025, we're in our perspective very well-positioned to achieve additional milestones beyond R-ketamine across all of our other programs. That makes us very excited, and we anticipate a lot more to come over the next quarters and years. With that, let me thank you for listening in today, everyone, and have a great week.