Good morning, everyone. Thanks for joining us today. I am Athena Chin, an associate at TD Cowen. We have here today Amit Etkin, who is the CEO of Alto Neuroscience. He will be giving a corporate presentation update on their pipeline. Following that, we will open up the floor for Q&A. Amit, please take it away.
Thank you. Thanks for having us, and thank you for joining us today. This is the disclosure regarding forward-looking statements. Let me start, though, before updating on specific programs, to kind of give you a bird's-eye view. We've been developing a precision approach to understanding psychiatry and developing psychiatric disorders, the idea of tapping ways to understand biology scalably and with regard to things that you can do both in R&D and clinical practice. We've had a number of trials, and I'll talk about several of them read out, both as a private and public company. It's important to understand just how much activity has been going on here. We've dosed, let alone screened, hundreds and hundreds of patients in these trials. There's active patient selection happening across programs, across indications, across multiple trials at the same time.
I'll share some of that progress with you, with a lot more to go for ongoing trials with cash runway into 2028. What I'll tell you about here are two aspects of what we do. One is using a biomarker, basically an objective measure of brain function, things like EEG, behavioral measures, and so forth, early on to understand how to dose a drug and select the right population. Later on, in phase II, phase III, and beyond, for patient selection in a way that would ultimately drive the clinical use of that drug. Both of those are things not really standard within the field, but we think are critical to changing the odds of success and the impact of these drugs clinically. We've already seen quite a bit of success in applying this kind of approach in different ways.
We have found biomarkers that we've prospectively replicated for multiple different mechanisms, both across investigational drugs and drugs that are already out there and approved, like SSRIs and ketamine. We've demonstrated target engagement in allowing us to understand how to de-risk later-stage trials through biomarkers, and now engaged in a series of large placebo-controlled trials where we select patients based on these biomarkers, and then, if positive, advance it to pivotal trials. Across our program, we have four different novel agents in the clinic right now. Two of them are in phase II-Bs, and that's Alto 100 and Alto 300. I'll dive more into these programs in a second. Alto 300, we just had an interim analysis, and that is on track to read out mid-2026. Alto 100, also in 2026.
Alto 203 and Alto 101 are earlier-stage programs, really, that are using biomarkers to understand how the drug works and therefore to target it. Those are on track to read out this year. Let me start with Alto 300, give you a little bit of an understanding of the drug itself, and then what we've done with it. Alto 300 is also known as agomelatine. It's a drug that's approved in Europe and Australia for the past 15 years and has been in active use. It's a different mechanism as an antidepressant from all other antidepressants out there. It stimulates melatonin receptors, which has circadian effects, and blocks 5-HT2C receptors, which leads to disinhibition and increase in dopamine and norepinephrine.
These two functions of the molecule work synergistically across a number of different functions in the brain, and that you can see in both clinical and behavioral measures. You can see that listed on the right. It is an interesting de-risk molecule in the sense that it is used clinically, and practitioners like to use it because it is also a very well-tolerated drug, even better tolerated, in fact, than SSRIs and SNRIs. Here, we are using it, however, as an adjunctive treatment in depression because of that tolerability and because of the clinical and commercial landscape with respect to adjunctive treatments, which is really those patients who have failed a traditional treatment who really need something different. The approach we took to this particular drug is relying on an EEG and machine learning-based way to find biomarkers.
We've done this when I was a professor at Stanford in my lab there, developed and validated our ability to find de novo biomarkers that are data-driven to identify those patients, prospectively replicate them in independent populations. That's then what we did in this case. The biomarker that we found turned out to be a fairly simple signal related to the neural variability, a measure we call entropy. You can see in the top trace there that the more variable, the better the response to the drug; the less variable, the lower the response. That was then prospectively replicated in the bottom left in predicting an independent group of depressed patients in a phase II-A study that we ran, also adjunctive in depression, that if you're biomarker positive, you have a better response.
That's then what led to the development program that I'll talk about in a minute. The other side of these biomarkers is trying to understand the why. It's not strictly speaking necessary for developing the drug. I mean, we don't really truly understand the why of an SSRI, but clearly they're very important drugs clinically. That goes across the board. It's important for us in building further and being able to ultimately inform clinicians and the FDA. We've started to do that with experiments like that on the right. Remember, the drug is a 5-HT2C antagonist as well as a melatonin agonist. We gave animals the opposite of that, a 5-HT2C agonist that makes animals more anhedonic, more depressed-like. What we saw is a dose-dependent increase in that biomarker.
In other words, we directly linked that EEG neural variability biomarker to the activity at a receptor that is targeted by Alto 300. That's actually pretty cool when you think about just where this came from, a data-driven discovery of a biomarker that's replicated clinically. Now we're understanding that directly links, even though it was data-driven, turns out to directly link to the drug mechanism. That work is ongoing with more to come. On the clinical trial itself, the way we do this is we screen patients for the biomarker based on EEG, but we capture all of our other biomarkers so we can better understand these patients. We include both patients with and without the biomarker, but the primary population here, the biomarker positive, is what we've powered on.
We randomize people to a standard six-week double-blind placebo-controlled period here with 25 milligrams of Alto 300 given at bedtime. Everybody then goes into an open-label extension, and we use a lot of kind of common approaches to characterize these patients, things like the SAFER interview from MGH. I'll talk in a minute about some of the further strengthening that we've done in the context of our learnings from Alto 100, which I'll come back to later, but also a case review that we just conducted and read out as part of an interim analysis on this drug. As you know, we had a setback this fall with the Alto 100 program, which failed to meet significance. I'll dive into that. There were learnings that directly came out of that program.
A lot of those learnings, in this case, related not to the biomarkers themselves, which in that case showed the kinds of signal that we expected, but rather the conduct of the trial, especially at the site level, which leads to a greater likelihood of professional patients, a greater likelihood of data that you have a harder time trusting. Having seen that this fall, we immediately turned our attention to the question of how much of it is a risk on Alto 300, and then how do we mitigate that risk. The way we did that is then with a blinded prospective case review of all of the enrolled participants to understand at the patient and site level what is that risk, remove in a blinded way those patients, and then conduct an interim analysis to understand, are we on the right track?
How do we then go forward with this study based on the information seen? 171 patients had been enrolled. 52 were then removed in this blinded process. Looking for things that are really systematic issues at a site, we found evidence of data manipulation at a site. Another site, we found evidence of coaching that led to data that, frankly, we just couldn't trust and removed from the trial. Those people are not in any of the analyses going forward. In the biomarker positive patients, conducted an interim analysis with three different outcomes: either stop the trial early for futility, stop the trial early for success, and there was a fairly conservative threshold there so that we had minimal alpha hit on the final p-value when the trial reads out, or continue the study.
Under the option of continuing the study, we allowed for sample size re-estimation. The results came back clearly that there is signal there that's supportive of the powering and the approach that we took in setting up the trial, and a recommendation to slightly upsize the trial to improve our odds and powering of the study, frankly, to the same size study as we had for Alto 100. This is 200 patients who are biomarker positive, as well as, of course, a small group of people who are biomarker negative. That's exactly what we did. We upsized the study and now on track with a study to read out in mid-2026 rather than the first half of this year.
A lot more confident that what we've seen from our learnings from Alto 100 implemented in our conduct for the Alto 300 study and looking forward in terms of the Alto 300 readout has taken a lot of those learnings in and put us in a better position to see positive outcomes here. Transitioning to Alto 100, where we have a bipolar phase II-B program ongoing, I want to go through some of those specific findings and learnings and then mention the bipolar program. Alto 100 is a novel agent that was developed to enhance neuroplasticity in the hippocampus, which is a brain region important for both cognition and mood, and where things like neuroplasticity underlie the ability of the hippocampus to encode memory.
The drug enhances that function, and we identify that people with impairments in memory, so poor hippocampal neuroplasticity, have a better clinical response in terms of their depression symptoms to that drug. The drug's mechanism acts in part by increasing the release of BDNF, brain-derived neurotrophic factor, which is really at the core of neuroplasticity mechanisms that we've understood for a long time. What we did over the last number of years with that program is conduct a large phase II-A study where we, just like with Alto 300, honed down on the biomarker and then prospectively replicated its ability to predict better response, and then conducted a 300-person phase II-B study of which 200 or so patients were positive for this biomarker. As I mentioned, that failed to hit statistical significance, but a few really critical things came out of it.
We included both monotherapy and adjunctive use of the drug, and that was all pre-specified in the nature of the design. We expected and saw about 70% of the patients being monotherapy and the rest being adjunctive therapy. Within those two pre-specified populations, there was clear evidence of signal for the adjunctive group that was not powered and did not hit statistical significance, but it was consistent with our expectations, not in the monotherapy group. I'll dive into the numbers and show you in a second, but it came down to an issue of medication compliance being much lower in the monotherapy than the adjunctive group. That obviously drove learnings that we applied for 300.
We've looked at the data in any number of ways, and it all continues to point to, and what then drove our 300 case review process is sites with issues in their conduct that you can identify in this way really drove the null effects. Those analyses, some of which we're now putting figures to, but we've put words to before, are what gave us confidence to continue the Alto 100 bipolar program, which I'll talk about momentarily. On the left is the change in MADRS scores from baseline in the adjunctive population where we saw week 6 a Cohen's d of 0.47 effect size, a p-value of 0.09. It was near significance, even though it was not powered. It was a pre-specified analysis. The question really became, why? Why do we see an effect there and not in the monotherapy?
As I alluded to, it's a story about compliance. It could have been about something we didn't understand, like biology that we'd not anticipated or clinical populations we hadn't studied. Coming into this phase II-B, we had studied both monotherapy and adjunctive. We saw enrichment in both that was replicated, but we also saw very high compliance in both. These were in a set of sites that actually had also high compliance in the phase II-B and a nice effect size. When we enlarged the study as we went and added sites in the second half of the study, we saw as the non-compliance rate, which was over 40% here in the monotherapy arm, was clustered. That is not just every site had the same non-compliance rate, but rather some sites brought in a fair number of patients with a high non-compliance rate.
It is really about site issues as opposed to tolerability, for example, of the drug, which was excellent and not a driver of non-compliance. As you can see, all of the adjunctive patients were compliant, and that certainly supports the interpretation of this adjunctive pre-specified signal. This is all the pre-specified. If you then, with all the appropriate caution in looking at those patients who are positive for drug and also positive for the biomarker, you can see again evidence of a suggestive drug signal here. That is on the left, comparing those biomarker patients, positive patients who had drug in their blood to the placebo patients across the whole study. On the right is comparing those patients with drug in their blood when you contrast those with a biomarker or biomarker positive to those or biomarker negative.
All of that supports that there is a signal there consistent with what we'd seen before, but that it's a trial conduct at certain sites issue, and that's what we really focused our execution to solve. All of that at the time also encouraged us to continue the Alto 100 bipolar phase II-B. This is a study that's designed exactly like the Alto 100 MDD study, except for obviously the difference in diagnosis, and it's all adjunctive. This is adjunctive to a mood stabilizer. Already it strengthens the argument here in terms of being able to execute on this, but we further strengthen that since looking at PK, for example, in small batches as we go so we can continue to monitor for non-compliance.
For both Alto 300 and 100, institute of centralized sponsor eligibility review, we measure the underlying antidepressant in Alto 300 and the underlying mood stabilizer for Alto 100 so we can know as much as is possible about these patients and their level of compliance. The mechanistic rationale here for bipolar depression is actually very similar to major depression. If you look at postmortem data, gene expression data, and so forth regarding impairments in neuroplasticity in these patients with poor memory. Mechanistically, that's the core thesis. In terms of the data that we've seen and the execution, strengthening the execution, it's all very much encouraging us to proceed with this study, as well as funding that we've gotten from Wellcome Trust to help offset the cost here. I'm going to transition to two earlier stage programs, Alto 101 and 203, before rounding out here.
Alto 101 is a study in schizophrenia. It's a phase II POC study on track to read out second half of this year. It's a PDE4 inhibitor. We know a lot about PDE4 inhibitors, obviously, in other parts of the body. It's been heavily used in the immune system, but it's been a mechanism that's long been of interest as a pro-cognitive intervention. The questions have been twofold. One is, how do you actually target it? How much do you give? What are the effects you expect? The second is a class-wide issue around tolerability. As you give more of a PDE4, you get nausea, vomiting, diarrhea, which are effects of the drug no matter whether you're looking at the immune system or the brain that have to be overcome. We solved those two problems in two different ways.
Using biomarkers to understand first what the right biomarker using EEG is of cognitive impairment in schizophrenia, which is our target population here. We defined a measure that's called a theta frequency response or a low frequency response to auditory stimuli. It tells you about excitation and inhibition balance in the brain and basic sensory processing that drives cognitive impairment. We found that that marker is the best measure of this impairment in schizophrenia that is our target. In turn, found that in a phase II, 40-person crossover single-dose study with this drug, that on the left here, you see a dose-dependent increase in that theta response, which is more intense red blobs there. There is also a dose-dependent improvement in cognition in the form of processing speed, which is a core element for cognition and cognitive impairment. This is all placebo adjusted on the right.
Moreover, the degree of EEG change correlated with a degree of cognitive change in these people. Mechanistically, it makes sense. We understand the dosing and the targeting for patients with schizophrenia with a biomarker that directly maps on in an objective way to their phenotype. That is the dosing part. On the right is how it is delivered to be able to get around these adverse events. It turns out the speed of absorption is the driver of these adverse events. We reformulated the drug from an oral immediate release, which would be a BID dosing, to a transdermal patch, which would be a QD patch. Going from the blue to the pink, slowing down that absorption not only dramatically reduced these adverse events, but actually tripled the amount you get from an AUC perspective. Despite that higher exposure, you get much lower adverse events.
I am pretty excited about where that lands us. That's a crossover design with 10 days of dosing with drug and placebo. Everybody's their own control on track to read out second half of this year. The final program is Alto 203. It's a little bit earlier stage. This is a trial of an H3 inverse agonist, histamine inverse agonist. The idea there is to drive dopamine release by inhibiting the H3 receptor and especially to drive dopamine release in the reward system for many obvious reasons.
What got us excited about this program is that the originator had found also in a 40-person placebo-controlled crossover study that single doses led to an immediate positive subjective response, a decrease in alertness and mood symptoms, which you see on the right here that are comparable to or greater even than modafinil, which releases dopamine in the reward system, as well as positive effects on cognition and other objective outcomes, things related to sleep and so forth. That is exactly what we're looking at, is in that same kind of design, single-dose crossover study in patients with depression and elevated anhedonia, symptoms looking at both the subjective and the objective outcomes. That is on track to read out first half of this year. Where that puts us is a lot of activity here.
Earlier stage programs this year, later stage programs next year, four different readouts with cash runway well beyond that into 2028. A lot of opportunities to see the promise of precision psychiatry. A lot of exciting programs in areas of really, really high unmet need. Areas like cognitive impairment, schizophrenia have no approved treatments with none on the horizon. We are really excited for where that puts us and look forward to your questions.
I'll kick off with Q&A and open up to the floor. First, I would like to ask about Alto 300, specifically on the EEG biomarker. Can you speak to how long kind of that EEG testing takes in the clinic and how that might translate to practice if it was indeed predictive in the phase II-B? Thank you.
Yeah. Great question. The recording only takes eight minutes, and it's minimal time to set it up. The biomarker itself is captured from a single electrode. As you think of scalability, that's pretty well built in. We've been doing a lot of work on the hardware and software side to get us ready for that and understand even for phase II, how do we do this in the most scalable way possible? Our view is either in a clinic or in a patient home by themselves, you should be able to put on an EEG system with no expertise and then automatically score the biomarker. There's no need to know how to read EEGs because psychiatrists don't know how to do that.
This is a system already built by the nature of how we define the biomarker and what we know about the stability of EEG signals like these to be clinically scalable to the size population that we need to serve.
Understood. In our Neuroscience panel this morning, we were discussing about how recent MDD trials have had really high placebo responses. I was wondering what you're doing in all your different studies to kind of mitigate that response and whether you're using centralized raters, what else you're doing to try and obtain that expectation from patients.
Yeah. Placebo has always been a challenge. It's actually one of the things that has gone really well in our trials is keeping placebo under control. The placebo response in that adjunctive arm, for example, that I showed you for Alto 100 was seven points. People have seen double that in trials. We do a number of different things. We try to manage expectations, for example, enrolling patients with and without the biomarker and patients are blind to that biomarker. Everybody gets open label drug afterwards, so there's not a need to kind of be concerned about whether you're getting drug or placebo initially and potentially kind of inflate expectations around that. We're also doing a lot of very careful patient selection. It comes to a lot of other smaller factors.
Our inclusion is determined by a MADRS, for example, done by an MGH SAFER interview rather than by sites. There's no opportunity for score inflation by sites, which often leads to placebo response and other little things we do along the way to ensure that the patient is clinically stable so they're not kind of on the way to improving even in the course of the screening.
Understood. Audience, if you can raise your hand if you have any questions, I will pass the mic to you. Otherwise, I'm going to keep going. I do have another question on the bipolar study. You study Alto 100 in MDD patients. What is different about the biology between MDD and bipolar depression that you think will kind of support success in this indication?
It's actually less the difference, but more the similarity in the sense of the biology being similar. What we understand about the role of the hippocampus in mood and cognition is shared, the role of impairments in neuroplasticity, whether looked at at the molecular or cellular level, reductions in volume in the hippocampus, impairments in memory, the relationship of cognitive impairment to poor either functioning or poor clinical outcome with standard of care treatments, all that's actually shared between them. The outcome measure, the MADRS, is shared. A lot of this actually acts to de-risk programs in that way. The issue that became as I've kind of gone through here, right, the problem in the 100 MDD study is a site and execution issue.
That part is the difference that we are now very proactively solving in our assessment of patients, our assessment of sites, measuring the underlying mood stabilizer, for example, to know that people are what they say they are. I mean, they are very clear clinical cases. So we're feeling good about where we are in that program.
Got it. On Alto 300 again, as part of the top line readout next year, what can we expect in the top line as well as kind of any post hoc analyses on any subgroup breakdowns for biomarker versus biomarker negative, et cetera, et cetera?
Yeah. It will be much like for the Alto 100 program. What we discussed, a top line would allow is understanding the people with the biomarker and those without the biomarker. Of course, the powering and the statistics are in the people with the biomarker. Those without the biomarker, it's really a question of is there numerical evidence of enrichment, but not really an expectation because it's not powered of any statistical comparison between the two. There's not subgroups within this like there was the monotherapy and adjunctive. These are all adjunctive patients. In a sense, it's also a simpler top line readout. Just like with Alto 100, advancing to phase II really requires evidence of efficacy of the drug.
Lastly, on your cognitive impairment associated with schizophrenia program, so Atai is also studying RL-007 for the same indication. What are the most distinct features of Alto 101 that kind of makes it stand apart from other compounds in development?
I obviously can't speak to their programs. Not a lot has been disclosed there. Alto 101 taps into a well-understood mechanism, PDE4 inhibition, that's been of interest for decades in terms of improving cognition with already evidence in humans in small studies that both in schizophrenia and in healthy people that a PDE4 inhibitor does improve cognition, work done with a drug, roflumilast, including in similar designs, actually, as our phase II POC study. We also have a very direct link, as I showed you, to biomarkers of change and how mechanistic EEG biomarkers tie to the cognition programs. We're trying to set this up obviously as well as we can. There's patient selection that's happening here, which to my knowledge has not happened in any other CIAS trials.
That's a core part of our DNA is finding the right patients objectively using biological tools, and that's happening here. We hope will increase the probability of success.
Got it. Thank you guys so much. We are at time. Have a good rest of your day here.
Great. Thank you.