Thanks very much. It's my pleasure to be moderating this chat with Amit Etkin, CEO and founder of Alto Neuroscience. I'm going to ask Amit to give maybe a three to five-minute overview and update on the company, and then we'll do Q&A. Thank you. Appreciate it.
Awesome. Alto is a psychiatric drug developer. We focus on areas of tremendous unmet need in things like treatment-resistant depression, bipolar depression, and cognitive impairment, and schizophrenia. We take a biology-forward precision approach, so really try to understand the biology of the disease, but especially of specific individuals. What characterizes the disease in those people? What are the operative mechanisms? What are the right drugs to develop for subsets of people defined in various ways? Really trying to take a page out of the precision medicine playbook that's been very successful in other areas and applying it here to psychiatry where it's never really been applied. All of our programs are in phase 2, mostly late phase 2, one earlier stage phase 2 proof of concept trial. All in patients, all meaningful data sets to come in the next two years.
One of those programs, which was actually our most recent acquisition, is launching a phase 2 B as well as a phase 3 in the coming year. A lot of activity really directed at meeting needs in the clinic and getting into the clinic commercially before too long, and all with a lens of trying to de-risk as much as you can, either on the drug or the biomarker, ideally both. We can talk about that in these various programs, but looking for anchors in mechanism of action, in the characteristics of the people, and what do we know about different subgroups of people, and then what the effects of the drugs are on the brain itself so that as you develop a drug from the earliest stages, you can effectively de-risk with biology, again, not something that has been done in the field in the past.
Ok, great. You want to talk about the 300 program and just kind of the history and the data generated to date and sort of where we are right now?
Yeah. Alto 300, agomelatine, is a drug that's actually approved as an antidepressant in Europe and in Australia. It has a unique mechanism of action. It's a melatonin agonist and a 5-HT2C antagonist. That latter part leads to a disinhibition and increase in dopamine levels. It's very well tolerated. Here we're developing it as an adjunctive treatment in depression, targeting the U.S. market. The motivation for an adjunctive treatment is because right now, basically, that's the realm of antipsychotics, which are generally not all that well tolerated. They have modest effects clinically but carry risk of things like metabolic syndrome, weight gain, tardive dyskinesia, movement disorders, and so forth. The opportunity to displace it with a well-tolerated novel mechanism within the U.S. is exciting.
The further ability to identify those patients through a biomarker, in this case based on EEGs, is even more exciting, so the ability to really target it. What happened here is that we found in a large phase 2 A study in the same adjunctive use in depression, a biomarker based on EEG defined through machine learning and then prospectively replicated, so people with a biomarker respond better. Though it was a machine learning data-driven approach, we then were actually able to translate that EEG signal, which is just basic neurophysiology, excuse me, back to animals, as well as actually to humans, to understand mechanistically how it relates to the drug. It turns out to map directly to the drug's mechanism of action. Essentially, what the biomarker reflects is brain states that are the opposite, especially of the 5-HT2C action of the drug.
There is a logical reason. There is clinical data on the drug. It had historically, in terms of its approval path, had positive phase threes, had some negative phase 3. These are typical things that you see in psychiatry, especially in Alzheimer populations. For those reasons, it never ultimately made it to market here in the United States. We are now in a phase 2 B, guiding to a mid-2026 readout. That is going to be in 200 patients with a biomarker, a small number of people without the biomarker. It is progressing. Pretty exciting to have that outcome coming up.
Yeah. You ran this interim analysis on a blinded basis of this study, leveraging learnings from a prior trial of a different drug that didn't work. Can you just talk about that interim? Does that interim analysis affect at all our ability to extrapolate potentially positive data from the study onto a future trial?
Yeah. The reason we did an interim analysis here is actually not because we want to do interim analyses or actually try to avoid them. It is because what happened this past fall, a year ago, with Alto 100 pointed to systemic risks in terms of how psychiatry trials are run. What happened specifically was in Alto 100, which did not meet its primary outcome, we had two different subpopulations, those taking the drug as an adjunct and those taking the drug as a monotherapy. These were pre-specified, stratified populations. We had a nice effect in the adjunctive group, no effect in the monotherapy group. What it boiled down to was that the adjunctive group was fully compliant with a drug based on blood draws, and the monotherapy group had only a 55% compliance, nothing to do with side effects or anything.
That was really a site-level risk, with certain sites bringing in a disproportionate number of non-compliant patients that basically had a professional patient profile. Seeing that, we then completely changed over our requirements for all of our trials in terms of the level of documentation and control over the patient phenotypes clinically, requiring medical records, pharmacy records, and blood measures of the underlying or urine measures of the underlying drugs, whether it be in depression, bipolar disorder, or schizophrenia.
For 300, which was already in process at the time, we conducted a blinded case review using just the baseline clinical data to identify those risk factors, took out data from four sites, ran an interim analysis that pointed to a drug signal and a signal to continue the study, upsized slightly from 150- 200 biomarker positive, and then continued the study from there, obviously replacing sites as we went. To the point of drawing conclusions, that study should be very much what a phase 3 program would ultimately look like in terms of design and population and so forth. It should reflect directly on what we think the ultimate effect size is. We've also seen that now all the changes that we've made and been very public about have been followed by others.
Big pharma, midstream in trials, has required what we are now we've been requiring and been very loud about. Biotech companies have done the same. Frankly, the question should be, why aren't people doing it at this point when this risk is just so apparent? We've been very pleased, for example, with our Alto 100 bipolar program to now see 96% PK positivity with these changes implemented. It makes a big difference, these execution risk factors.
Yeah, that makes sense. How do you think of the bar for efficacy in the enriched population? Is the bar a better effect size than average because of this element of enrichment? Conversely, what is the bar for showing that enrichment matters between the biomarker positive and biomarker negative patients?
That's a kind of a bit of a complicated set of questions, right? Because one is really, what is the bar for the drug being useful and moving forward?
Yes.
In my mind, that's just statistical significance.
Even though people have to step through the EEG and it's no guarantee, why do you think that?
Because their option right now is an antipsychotic. And an antipsychotic, as a psychiatrist, I would rather prescribe anything other than an antipsychotic because of its long-term side effects. Its modest efficacy, but really, its long-term side effects. Replacing an antipsychotic, and that's the case also for how we think about 207, right? That has tremendous value clinically. We do not want to set the bar too high for something that clinicians will want to use just inherently. The EEG side, we've done a lot of work on developing an EEG system that's very easy to use that patients put on themselves.
Takes about 15 minutes to do the recording, very low cost, and has the benefit of now giving you objective information as a patient or provider about the disease, which itself, I think, enhances people's desire to take the drug, to stay on the drug, understand their disease in a way that has not been done before. We do not think it'll be a barrier. In fact, if anything, work in the opposite way. In terms of what a label would look like and kind of how the EEG comes in and the biomarker comes in, the feedback we'd gotten from FDA on the Alto 100 program, but can be generalized across the board, driven by their 2019 enrichment strategy guidance, is that if the drug works in the general population, it gets labeled for the general population. Alto 300 agomelatine has worked before in the general population.
If it also works in the population of the biomarker, that is considered an enrichment marker that's basically a complement to that AlcoMer label. If it only works in the biomarker positive population and doesn't work in the general population, then that's a companion diagnostic, and you'd be required then to do the biomarker. It really depends on how the phase three ultimately lays out.
There's no actual downside if it turns out that the drug just works everywhere?
Correct.
It doesn't complicate the development path, you don't think?
No. I mean, you'd still power, in this particular case, you're still powering for a primary population in your biomarker positive.
Right.
At some point in the phase 3, you'll have to run the whole population because you have to demonstrate the risk-benefit profile anyway.
I understand. If it can be used more broadly than just the biomarker population per label, does that undermine the IP strategy at all?
Not at all. The IP, because the primary outcome is in the patients with the biomarker, that's in the clinical trial section by definition. Therefore. That's the issue IP anywhere in the label covers that.
Yeah, ok . And then just on 300, what's your comfort that at the dose you're exploring, you're going to avoid a liver safety issue?
Yeah. Great question. As background, agomelatine is dosed at 25 mg or 50 mg. Historically, we're only developing the 25. Servier's initial development never measured LFTs at baseline. They were then somewhat surprised to see an LFT elevation really just at the 50 mg dose that was transient. There's never been any liver failure issues. It's basically an adaptive liver response, but they'd not really been prepared for it. Novartis, when they were developing the drug afterwards, screened for LFTs at baseline, and they saw a placebo level of LFT elevation with 25 mg, but the expected higher rate of LFT elevation with 50 mg. A huge amount of clinical practice pharmacovigilance data has continued to back up the real-world safety that this is basically just an adaptive liver response that you see in some people, a couple percent within the first few weeks, and it goes away.
That really has pointed to safety consistently. We've also in our own studies, and this is why we chose 25 milligrams to not have that LFT monitoring requirement, we've not seen elevations in LFTs and in fact have a stopping rule that nobody has triggered in the phase two B on LFT elevation. All of that makes us feel pretty confident that there's really no meaningful liver signal there.
Ok . Have you seen data on a blinded basis?
Yeah. So I mean, we have a stopping rule, right? If you go above three times upper limit of normal, then the drug is stopped just as a monitoring thing, not necessarily for risk per se. That's not been triggered.
Ok . That's all encouraging. Do you want to switch gears and talk about 207? Give a little bit of a backstory there.
Yeah. 207 is a fixed-dose combination of pramipexole and ondansetron. Pramipexole is a dopamine agonist, so it directly stimulates dopamine receptors and specifically within a class called the D2, D3 class. It prefers the D3 target. Dopamine, we know, generally is important in depression. Our own biomarker work with EEG, where we've developed a biomarker of dopamine levels in humans, has pointed to treatment-resistant depression as being particularly enriched for a hypodopaminergic phenotype. You then take a direct dopamine agonist strategy for people with low dopamine, makes logical sense, right? Unfortunately, with pramipexole, which has been approved since 1997 for Parkinson's, as with every dopamine agonist, the more you give, and especially the faster you give it, you get nausea and vomiting as dose-limiting effects. That is fundamentally a barrier for use of this drug in the clinic.
That's why it's paired with ondansetron as an antiemetic that works on a synergistic pathway. In the backdrop here, there is a lot of clinical data that's been shown on pramipexole showing its efficacy in depression by virtue of that D3 selectivity and really showing outsized effect sizes. Whether you're talking about the initial older studies, which showed a Cohen's D of around 0.5, standard of care is around 0.25-0.3, so that's already large, or a trial that published in Lancet Psychiatry this summer, which was 150 TRD patients drawn out of nine sites in the U.K., which showed an almost 0.9 Cohen's D effect size. That was after 12 weeks of treatment. They kept people randomized for 48 weeks, and they saw consistent separation throughout the entire time, really striking effects.
As you would expect, because of the nausea and vomiting issues, that has limited people's doses and led to a 20% dropout in that study with drug compared to 5% with placebo. The opportunity then for a co-formulated product is really exciting. You can dose higher and faster. Chase Therapeutics, from whom we acquired the compound in a 32-person phase two A trial, so randomized trial in depression, showed an over 8-point delta on the moderates over Cohen's D of 1.1 at week eight, with the combination here allowing you to hit a dose that's higher and to titrate much, much faster, five times faster than what the pramipexole label allows.
I think that becomes, across all those studies, a potentially really exciting tool to bring into the clinic in a population TRD or adjunctive use for that matter, where you're looking at antipsychotics or esketamine as the only treatments, modest efficacy, lots of tolerability issues. We're excited now to push this program forward, which is launching a phase 2 B. We actually just pulled the timing of that up a little bit with our Q3 release this morning. We had guided to mid-2026, but now first half 2026 to launch a phase 2 B, which we think has the potential to be part of that pivotal registration package. Then we're able to, with additional recent funding, accelerate the launch of a phase 3, which will be by early 2027, and puts us on a good path with that drug and bringing it to the clinic soon.
Ok. That's great. I know you've talked about the combo mitigating some of the prior tolerability limitations. Any safety considerations here that we need to be mindful of with just kind of boosting dopamine activity in the brain?
Yeah. So beyond obviously the benefit of doing that, right, with engaging people, behavioral activation, anhedonia, by the way, the drug has had consistent large effects on anhedonia.
Yeah. It's an obvious mechanism of regulation.
An obvious one, right?
Yeah.
Much larger than comparison drugs, either in development or approved. The main things you want to avoid beyond obviously nausea and vomiting, which is what the combination is geared to do, is at high doses, you can get somnolence or insomnia. Somnolence probably more frequently. We have thought about our dose range from that perspective. In a Parkinson's population, which is a kind of a hypersensitized dopamine population, they have so few dopamine neurons, their dopamine system is just hypersensitized. There have been reports of a fairly substantial rate of impulse control disorders emerging when you treat with any dopamine agonists, including pramipexole. That does not appear to be the case in depression. If you look at the PAX-D data, which is this Lancet Psychiatry paper I mentioned, 48 weeks of exposure, really no impulse signal. That has been the case consistently throughout both randomized and case reports.
How much is this used in the real world today off-label?
That's a really interesting thing. The data, like the first data for the drug in depression, is from 2000.
Right.
Especially TRD specialists really like using it. If you talk to these folks, they think it is a go-to. The problem is you can't really dose it effectively outside of the handful of people who can like draw your hand holding.
There's only certain types of docs too that I think would be willing to kind of make their own sort of thing like this that's not approved. Like it's a special type of person.
It's a special type of person.
Yeah.
To get somebody to a decent dose, you escalate, you get side effects, you back down, then you re-challenge. That is just a pain in the ass to dose, right? From a practical perspective, it has been out there, and people who are in that kind of subgroup like to use it, but it has just not been a practical tool that people can broadly use. That is exactly why we are doing the development path that we are.
Yep. Ok. Can you talk a little bit more about some of the other RCTs that are out there for this? I mean, the effect size you cited is just like otherworldly, but we know in small depression studies or academically run studies, right, that you can get this kind of overstatement or a lower placebo response. How comfortable are you that this historical evidence here really stands up to scrutiny?
Yeah. So very comfortable. If you look at some of the older trials, one was run by industry by Upjohn, who was the original co-developer along with BI of the drug, actually trying to encourage its development in depression. There's sort of a long history there of Upjohn as a psychiatric developer. They showed a clear effect, Cohen's D of 0.5, 0.6 in their trial, and that was at a lower dose. Another older drug, similar effect size, again at a lower dose. There is the PAX-D study was nine sites across the U.K., so yes, academically run, but in size, 150 people and scope, much like an industry-sponsored registrational trial. There was a study from the University of Lund, single site trial that just came out as a preprint showing an outsized effect on anhedonia, a selected anhedonia population. There are real-world case series, large case series.
For example, one out of Italy that looked at pramipexole versus aripiprazole, followed people for 24 weeks. These are really long exposures and showed much larger efficacy for pramipexole versus aripiprazole in similar populations as a case series, so not randomized. That is a lot of very consistent data, but on top of that, you have randomized trials for depression in Parkinson's. Obviously, it is used for motor symptoms in Parkinson's, but you can also look at it for depression. That has been positive. It has been very, very consistent. Whether we get the 1.1 Cohen's D 8-point delta that Chase saw, or we get a 5-point delta, as some of these other smaller studies have shown, those are phenomenal outcomes.
Yep. OK. So starting the phase 2b in the first half of next year, and then starting the phase 3 early 2027?
Yeah, by early 2027.
What's the thought process of starting that phase three, I guess, before you get the phase 2 B data?
We know enough about the drug, about its dosing, about tolerability, to know how to design that. That's the main question, right?
You don't need more info for this study.
You don't need more info. That's right. The opening was that we had a conversation with FDA, and they outlined essentially what we need to do in order to start it, which was bridging talks to kind of chronic talks from acute data, and then a dose ranging on the phase one on the ondansetron side to figure out how little ondansetron you can get away with, essentially, and still block the nausea with pramipexole. Both of those things are now in process and are eminently doable. That allowed us to then accelerate that phase three initiation. That pulls in the NDA filing target, which is ultimately our goal.
Yep. Ok. Good. Anything else to add there?
I think it covers that one pretty well.
Ok. Great. Last few minutes, you want to talk about the rest of your pipeline?
Yeah. Alto 101 is actually the next readout. This is a PDE4 inhibitor, but for cognition. PDE4 inhibitors, as you all well know, are well-validated compounds in the immune system. They've been of long interest in the brain. In fact, a lot of the work on learning and memory and plasticity pathways and so forth is focused on cyclic AMP, and PDE4 has therefore become a logical modulator of that. The problem has been that there are two things going on with PDE4s that made them challenging. One is we haven't known what the biological sort of phenotypic target is with respect to cognition and de-risking that and measuring that in the right way. We've taken an EEG biomarker approach to get there in schizophrenia. The other is tolerability. All PDE4s carry actually nausea and vomiting as well as an effect.
We have hypothesized that that is really due to a pharmacokinetic action. The faster the absorption and onset of the drug, the more you get those side effects. We solve both problems in different ways. Now in a proof of concept trial in schizophrenia, focus on cognitive impairment in schizophrenia, for which there are no treatments at all. It is a really, really important area. It is the primary determinant of long-term disability in these patients. We showed that we engage EEG biomarkers with the drug in a large healthy population, as well as actually preclinically, that we have then linked that biomarker. We had linked it to the phenotype of cognitive impairment in schizophrenia repeatedly. It is the best biomarker that we can identify for cognitive impairment in schizophrenia. That proof of concept trial then making use of that as the primary outcome.
If that's positive, that moves the drug forward. We think that that de-risked the program. We also reformulated the drug from an immediate release oral to a transdermal patch. That slowed down the PK and solved the nausea and vomiting adverse events. That was the other part of that package, as it were. We will also in that trial be measuring cognition, but because this is a crossover design, 10 days of dosing with drug or placebo, everybody's their own control. We really have positioned it for the EEG outcome, and seeing something on cognition would be above and beyond. That said, there are additional data, for example, with a drug called Roflumilast, which is a drug approved for COPD, in a small 18-person crossover design with eight days of treatment, showed both EEG and cognitive benefits in patients with schizophrenia. Really exciting opportunities there.
If that's positive, that goes into a longer treatment phase 2 B trial. The readout here is in Q1, so coming up. The final program is Alto 100, which is a novel neuroplasticity-enhancing compound for patients with bipolar depression characterized by neuroplasticity deficits. The way we identify them is based on a test of learning and memory, learning and memory requiring neuroplasticity for its basis. That's a phase 22 B study. 150 out of the 200 patients will have the biomarker and looking for efficacy. The same kinds of things that you'd look for in MDD, moderate change in patients who are either on a mood stabilizer or an antipsychotic, so they are non-responsive or partial responder patients with bipolar depression. The opportunity in that population is all they have is antipsychotics as the only treatment for bipolar depression approved by FDA, and so potentially even become first-line treatment from there. That's a second half 2026 readout.
Great. And cash runway?
Cash runway is into 2028, which covers us through all of that. Just to give you kind of the calendar, right, you have a Q1 2026 101 readout, mid 2026 Alto 300 readout, second half of 2026 Alto 100 readout. The Alto 207 program initiates the phase 2 B. The first half of 2026 will read out in 2027. We'll guide more specifically when in 2027 as we launch. Its phase three launches by early 2027. That's a lot of activities all covered by current cash and supported. That acceleration of the Alto 207 program was supported by a PIPE we recently completed led by Perceptive that brought $50 million into the company and allowed us to bring that phase 3 forward. Doing all of the kind of prep and then launching the trial, not funding the full trial, we really tried to keep the raise pretty focused, but a lot of opportunities here for capital formation on the back of positive data, really guided by which drugs have clear value and move forward, and how do we position those in the right way.
Great. All right.