Good afternoon, everyone. Thank you for joining us for the Alto Neuroscience Fireside Chat here at the TD Cowen New Mechanisms in Neuropsychiatry and Epilepsy virtual conference. I'm covering analyst Ritu Biral, and joining me is my associate and colleague, Athena Chin, who works with me on my neuropsych names. Joining us from Alto, we have Amit Etkin, Founder and CEO, as well as CFO and Chief Business Officer, Nick Smith. Nick, Amit, thank you so much for joining us today. Since we only have 30 minutes, I want to kick straight off and give a few minutes to the platform. There's once again increasing focus on precision psychiatry and biomarkers in the field of neuropsychiatry, which has, of course, been one of the most significant challenges, as well as being the holy grail of the field.
Can you talk about, maybe start off giving us an overview of your platform and how you select biomarkers for each of your programs?
Yeah, absolutely. Thank you for having us on. I think the simplest way to boil down what we're looking for from a biomarker frame is actually much like the rest of medicine, knowing what a drug does biologically and therefore who to give it to. When you look at it from that perspective, there are two kind of important pillars for us when we go about doing any biomarker work. One is the deep understanding of the biology of these disorders and how to measure them in humans with tools like EEG, cognitive tests of various sorts, wearables, and the like, things that tell us about brain function in a very hypothesis-driven manner. The other is a data-driven pillar where we use machine learning to try to understand, for example, out of EEG data, what might predict outcome with a drug.
In all cases, we also tremendously prize prospective replication of any one of our findings so that we really know what we're building on. The overall goal then is to get past what is really purely clinical and phenomenology-based groupings to things that are more meaningful. We'll talk about the individual programs and where those come in. At each of our programs, we embed a biomarker-based go/no-go, which is really to say we want to know what we're doing and have the outcome be directly reflective of that expectation so that it actually meaningfully de-risk that next step, which could be advancing a drug or could be not advancing a drug. Either way, we want to do that confidently.
Is there a correlation threshold? What are the sorts of things that are possible biomarkers? You had cognition in one program. There is certainly EEG. What is the gamut you can choose from?
Yeah, so I think principally for us right now, it's EEG, which takes various forms and measures various aspects of brain function. What we would call cognition really is a broader category, so not just traditional cognitive measures like memory or processing speed, but also more modern cognitive neuroscience measures like how you respond to reward and make decisions. That's already a pretty broad range. We also do look at things like wearables. We do gather genetic data, genomic data. I think from a regulatory acceptance perspective, EEG and cognition are fairly far ahead for a range of specifics.
When you say wearables, do you mean like actimetry, like activity as a mental state?
Yeah, exactly. Mental state, yeah. That's right. Actigraphy tells you about your circadian rhythm, tells you about sleep and activity patterns, all of which are incredibly meaningful and robust. There are challenges around how you use a device that might be a third-party device in that case, which is why we tend to, at this point, favor EEG and cognition also from a regulatory acceptance perspective. All of it falls under the same data analytic, data science orientation.
Got it. FDA's perspective on biomarker enrichment, how do they approach it, and has it evolved recently?
The nice part is that we're not the first to do this. There are many other areas of medicine that have laid the groundwork for us. There is an enrichment guideline that the FDA published in 2019 that really lays that out. Essentially, think about it this way. We don't, for our efficacy measures, touch the endpoint. That's an agreed-upon, for example, in depression, moderacy, or HMD, agreed-upon measure that the field and the FDA have long functioned under. What we're doing is, at the front end, changing the inclusion-exclusion criteria to also include a biomarker. You have to show various things about that biomarker: its reliability, its meaningfulness, and so forth, and its predictive value. In a sense, what you're doing is tweaking the inclusion-exclusion criteria, so your primary outcome is in those people who have the biomarker. You then get data across the spectrum to understand risk-benefit.
Framed in that way, it's not all that different from existing frameworks that people use. It's just a heck of a lot more specific for the person and follows exactly the guidelines that FDA has set out. Additionally, our Chief Medical Officer came from J&J, where he led the Cell Torexant program into phase three, which used the enrichment framework. We really benefit from a lot of expertise and experience in this and then have gone in front of the FDA, for example, in a type C meeting for ALTO-100, laid out our understanding of that framework, which seemed to match well with what they had previously put out publicly and then what FDA told us more directly for our program. Everything obviously relates at the end to individual programs, but the concepts will bear well across them.
Understood. If you could take just a couple of minutes and walk us through the pipeline as it stands, you have a number of assets. As you do that, if you could highlight the programs of corporate focus right now for Alto.
Let's start with the first program that is going to read out in the coming months, which is ALTO-101. This is a PDE4 inhibitor, a well-validated mechanism used in the immune system, but now used in the brain where it's been long of interest in terms of its procognitive capacity. We're developing that for cognitive impairment associated with schizophrenia, or CIS. CIS is the basis of schizophrenia in many ways. It is the thing that comes first. It's most impactful long-term for these patients, and there are no treatments available for it. Ours is the latest stage product, in fact, in development right now. There, we're using biomarkers as outcomes, biomarkers of the pathophysiology itself to tell us whether we've moved the needle in a meaningful way on the biology, as well as then looking at cognition. We can dive more as we go through this discussion into individual programs.
That's an earlier stage, very biomarker-driven process. At the very end of our series of four major readouts over the next two years, we have ALTO-207, which is a fixed-dose combination of pramipexole, which is a D3-preferring dopamine agonist, and ondansetron, an antiemetic. That really harnesses the ability that has now been pretty clearly shown in a number of very impressive studies of pramipexole to drive potentially outsized efficacy in depression. Here, we're targeting treatment-resistant depression with a co-formulation supporting both rapid treatment, rapid escalation in dose, and reaching a higher dose, which is needed for effective treatment, but limited currently by the nausea and vomiting induced by pramipexole when used alone. We also have ALTO-300, which is an adjunctive treatment in depression. Here, think about what the options are for these patients around basically just antipsychotics, maybe ketamine for the later stage folks.
Antipsychotics are tremendously burdensome in terms of metabolic effects, obesity, movement disorders, and the like. ALTO-300, also known as agomelatine, has been very well tolerated and been shown to be effective in depression, approved as a monotherapy in Europe and in Australia. This is therefore a U.S.-focused strategy. There, we have an EEG biomarker that we've linked to the drug's mechanism of action. Finally, ALTO-100, which is an entirely novel mechanism that enhances neuroplasticity in the brain, in particular within the hippocampus, a brain region important for both mood and memory. We select patients, in this case, with bipolar depression based on an impairment in neuroplasticity as reflected by an impairment in memory. That program, in fact, ALTO-207, ALTO-100, and ALTO-300 are all phase IIB studies, very late stage, all driven by patient classification using some sort of biomarker-informed approach and all targeting populations that are in tremendous need.
In that bipolar depression example, all they have are antipsychotics, which neither gets you tremendous efficacy plus all the tolerability challenges. There is a lot to be excited about, frankly, with every one of these programs. You're muted, Ritu.
Sorry, I know how to use Zoom in 2025. Let's start with 207. This has been the lead inbound, I would say, question on Alto that we've gotten here. You recently acquired this asset from Chase Therapeutics. Can you talk about the mechanism as a D3 dopamine receptor-preferring agonist and 5-HT3 receptor antagonist, that sort of balanced signaling, how that supports potential clinical activity in treatment-resistant depression?
Yeah. Let's start with dopamine and that really being the target here. Dopamine, we have as a field understood, is important in depression, low dopamine related to low motivation, anhedonia, and the like. We've done a lot of work around dopamine with our biomarkers, in particular with EEG biomarkers. It's pointed not only to low dopamine in depression, but particularly low dopamine in treatment-resistant depression. Treatment-resistant depression, as clinically defined, is actually already an enriched population from a biomarker perspective. The question is, how do you best address that? Both our own EEG biomarker work and that of others has pointed to the fact that you need to hit dopamine receptors directly. That is a direct agonist stimulating dopamine receptors as opposed to simply increasing dopamine levels, which other drugs have done and are not as effective.
Is there a subtype of dopamine receptor that's more important than another?
Yeah. The observation has been that the D3 subtype of a class called the D2-D3 class is probably what's most important, especially within the reward system, that being kind of an obvious link to anhedonia and motivation. It goes along with a lot of clinical data for pramipexole itself. Pramipexole, approved for Parkinson's and restless leg syndrome, when given to patients with depression, both in older studies and much more strikingly in a recent Lancet Psychiatry 150 patient TRD study out of Oxford, showed outsized effectiveness signals. In that case, a Cohen's D of 0.9, three times what you expect from a normal drug placebo difference in a very treatment-resistant population. Studies of other dopamine agonists have not supported that kind of efficacy. However, with efficacy in this case comes a main drawback of pramipexole, which is its tolerability challenges.
That really keeps you from dosing it quickly to the right level, even which level you can reach in the first place, and is mainly in the form of nausea and vomiting. That, as you said, is where the 5-HT3 antagonist comes in, which works on a synergistic pathway in the brainstem that then blocks the emetic effects of pramipexole. Now we can dose it faster, five times faster, and to a higher target, which is really what's needed to both get real efficacy in the clinic in practical terms, but also by dosing it faster, you can have even faster onset than you would otherwise.
Got it. The EEG biomarker that you mentioned, is that associated with, you mentioned anhedonia and one other feature of the depression. I'm forgetting what.
Yeah, I mean anhedonia, motivational deficits.
Motivational deficits, right. Is the EEG biomarker related to those aspects of the depressive spectrum in these patients?
Yeah. It's related not only to the phenotype clinically, but it's also related to essentially behavioral measures of that process. If you look at how people learn based on reward, getting that feedback, which is blunted in depression, that's particularly blunted in those depressed patients with low dopamine as measured by the EEG biomarker. This EEG biomarker also goes up and down when you increase or decrease dopamine. It's a really nice measure of the underlying process, which helps us both to understand what the right population is and the right intervention and can be a complementary biomarker for the development itself. It won't be required as a companion diagnostic in this case because TRD is already an enriched population. It might be helpful to further enhance the.
Understood. Pramipexole was previously studied in depression, but it was discontinued because of the side effect that you were just discussing. Can you review the prior data and effect sizes from its prior antidepressant activity beyond the Cohen's D that you mentioned? I think that was the 0.9. What other aspects and scales were generated in those clinical trials?
The early trials, which the first one was really 25 years ago, focused mainly on depressive symptoms. I should note that even in those early trials, which were really just figuring out how to dose it, they still saw effect sizes.
Was this MADRS and HAMD still?
Exactly.
Okay.
Exactly. They saw effect sizes in the 0.5 to 0.6 range, which is double a typical antidepressant effect. This is all placebo controlled. What they did in this Lancet Psychiatry Oxford study, and these were patients from the NHS, 10 years of median duration for failures on average, so very treatment-resistant population, is they assess a number of symptoms and everything got better by that same tune. Especially when you look at it, not just depressive symptoms, but anhedonia symptoms. Anhedonia really is not something that moves well with existing treatments. It's been a motivator for development of new treatments. It predicts resistance overall. That moved just as well as depressive symptoms, as did anxiety, as did suicidality. Functioning improved all with very similar scales. That really points to a very meaningful clinical improvement overall.
In the phase IIA data from Chase Therapeutics, from whom we acquired this asset, they saw an even larger effect size on the MADRS, so the 1.1 Cohen's D at eight weeks of treatment. Importantly, saw that the CGI tracked just as well, which is usually a measure, I mean, it's a measure of overall disease, right? That's usually not terribly sensitive to drug versus placebo differences. Yet, not only did it show that same effect, all of these symptoms and functioning scales showed very early separation from placebo. Really encouraging for the overall differentiated clinical profile here.
Can you go into remission from TRD on the MADRS and have no change in anhedonia or next to no functional change in anhedonia? Is that possible?
You can.
Go ahead. Sorry.
Partly, it's a nature of the measures themselves. The measures are not ideal, yet that's the only tools we have. There are certainly situations where anhedonia is just one of the most sluggish symptoms to respond, and you may still have significant anhedonia at that point.
If you have residual anhedonia, will you see continuing impairment on the Sheehan?
Yes, you see impairment in functioning. You also see that there is greater likelihood of relapse. There might be just greater chronicity measured in different ways. It's one of those areas that, perhaps not surprisingly, if you think of the tools that we have, if it's a dopamine-type phenomenon, and most of the tools that we have are serotonergic and noradrenergic, our treatments are just not well designed.
Got it. Moving back to safety and tolerability, with the combination, ostensibly, and in prior data, there should be lower nausea, lower vomiting. What would that leave the most common side effect of the combination to be? What should our expectations from that data be?
Yeah. The things that would not be targeted would be things like somnolence or insomnia, which can happen. That's certainly going to guide how we think about the dosing of the drug. There's a potential side effect that I actually don't think is an issue with depression, but is important to recognize here, which is that pramipexole, when given to patients with Parkinson's, can lead to multiple control disorders.
Right. Psychosis or impulse control? Is there?
Impulse control. In Parkinson's, things get a little wonky when you look at psychosis. Dopamine and psychosis is, obviously, it's a complex mix for those folks who have very low dopamine. Pramipexole itself is not really a driver of psychosis in a meaningful rate as much as impulse control disorders. What they did in this Oxford study is very carefully monitor and quantify impulse control symptoms throughout the study. Keep in mind, by the way, their primary outcome was at 12 weeks, but the study followed people randomized for 48 weeks. Inefficacy was consistent throughout. Equally consistent throughout was really no increase in impulsivity. That's important because it seems like it's not an issue of the drug so much as a disorder. Patients with Parkinson's who have very, very little dopamine may be more sensitive. If you think of low dopamine, you increase responsivity postsynaptically in those receptors. They may be oversensitive.
It's an outsized reaction to.
That does not seem to be an issue for depression.
Got it. When do you plan on initiating the phase IIB with ALTO-207? What's the study design that you have in mind right now?
Phase IIB will start first half next year. We're going through consultations with the FDA. We'll say more about the design in specific terms when that's sorted. Expect it to look like a well-powered one-to-one randomized phase IIB because what we're hoping here is that it may have an opportunity of being part of that registration package.
The pivotal package, yeah.
Yeah. Our plan would be one phase III afterwards, and the design will reflect.
Robust nature. Briefly, before we move on to 101, could you review the IP around that?
Yeah, that IP is a series of methods of treatment IP, much like what.
On the combination.
On the combination, exactly.
Are there specific doses or ratios or receptor activity ratios mentioned within?
Various things around, for example, granted IP around the core formulation itself and use cases and special populations and so forth. I think to get an idea of what that looks like, you can really just reference what Karuna's IP portfolio is, which is entirely a method of treatment, or Axsome's Auvelity, which again is entirely a method of treatment. Both have held up very, very well. Frankly, when you have a novel combination approach that achieves something unique clinically, I think it brings a lot of value and is protected by a lot of unique IP.
Understood. Let's move now to ALTO-101 in CIAS, cognitive impairment associated with schizophrenia. ALTO-101 is a small molecule PDE4 inhibitor. Top line from your phase IIB are anticipated second half. Can you review how the PDE4 mechanism ties into cognition, specifically cognition and schizophrenia, as well as the phase II design?
Yeah. PDE4 is what people have known really primarily from the immune system as an anti-inflammatory, as well as the like. It's actually been a mechanism of study for a long, long time in the brain. It's one, in fact, that I worked on as an MD PhD student. It's been around for a long time. The idea is cyclic AMP as a second messenger that is downstream of receptors in the synapse, transduces signals to drive neuroplasticity, to drive changes in excitability, coming off of a range of different receptors. It's kind of a nice common target. PDE4 is our enzyme that breaks down cyclic AMP. Inhibiting PDE4 will increase cyclic AMP, improve cognition and neuroplasticity because of that. What we've understood about schizophrenia and CIAS is that there are abnormalities in patients in cyclic AMP signaling, but there's also abnormalities in synaptic signaling writ large.
Glutamate receptors, acetylcholine receptors, and so forth have been the focus of previous development efforts. Those previous development efforts have suffered from, in large part, not really knowing what the right target has been physiologically, not necessarily bringing that through development. That's something that we're addressing here. What we're doing is a couple fold in the design. One is use EEG as the primary outcome here, but specifically try to find the right EEG measure that is the most sensitive to disease pathophysiology. We looked at multiple large data sets and prospectively replicated an EEG biomarker called the theta response that best differentiates between cases and controls. We've done this in multiple data sets, including very recently.
We reported out on results on that front and correlates with cognition and in phase one data is most sensitive to the effects of the drug and preclinically is most sensitive to the effects of the drug. It's really a setup for knowing it's the right signal to move that's clinically then meaningful in terms of the biology of the disease and responds nicely to the drug. That's the primary outcome in a crossover design study, everybody's control. We look at cognition as secondary outcomes, not necessarily expecting much movement in the 10 days of treatment that everybody gets, but if we do, that's going to be a phenomenal outcome.
Sure.
is really a careful de-risking in an area where a lot has failed in the past, but boy, the need is massive. Right now, there's just no treatments at all. It is the driver of disease writ large. The opportunity here for moving the needle meaningfully clinically is just fantastic.
Got it. Athena, I will leave the 300 questions to you.
Yeah. We do want to touch on 300 before we wrap. Your agomelatine compound in phase IIB adjunctive MDD study, recently you announced your pre-planned interim analysis outcome back in February. Based on the outcome, you're upsizing the study. How is that additional enrollment going so far? When should we expect completion of the study enrollment?
Yeah. We continue to guide to mid next year, and enrollment is continuing. I think it's important to understand why we did that interim analysis and what it means. The interim analysis was really motivated by findings we had around the ALTO-100 program in MDD and particular sites reflecting the risk that I think is now prevalent throughout the neuropsychiatry space in bringing in a fairly large number of professional patients who we assessed as people who were saying that they were taking the drug but repeatedly detected none in their blood. What we did is a blinded case review of the ALTO-300 data, took out data from sites where those sites showed those risk factors, did an interim analysis. It showed drug-like signal that supported continuing slight upsizing in the sample, most importantly replacing the people we took out.
That's really a first piece of what's now something we're doing across all of our programs, which is bringing everything to a much, much higher level of rigor than has been done in neuropsychiatry in recent years. Things like requiring medical and pharmacy records. You have to show that you're getting refills prior to coming in, blood measures or urine measures that you're taking the drug that is the underlying antidepressant, antipsychotic, or mood stabilizer as is relevant. Bringing a lot of AI tools in to understand in the patient interviews what can we learn about those patients.
How frequently are you requiring those urine and blood samples from the patients?
Beginning and end of the double blind, as well as at the end of open label. That's a lot of touch points. We're learning a lot of things about populations that normally aren't even screened for that and oftentimes are negative for those factors. That really is important, not just in the ALTO-300 story, but across the board, that the level of site risk is substantial. We had to execute differently to de-risk the studies purely from an execution perspective on top of the science, in this case with agomelatine or ALTO-300, where we know it's an effective antidepressant, well tolerated. We saw a positive signal in the interim analysis. Now we understand the biology of the biomarker as well as reflecting, in a sense, the inverse of drug action on the brain.
What is your trial powering after that upsizing?
The Cohen's d of about 0.4, 80% power.
What do you hope to see in those phase IIB results that would support advancement to phase III studies?
I think that one is straightforward. Statistical significance on the primary outcome. It's a population in need of a new augmentation agent. Antipsychotics are really the only option, so a lot of clinical opportunity there if we can hit the primary.
What should we expect from the placebo group in terms of the range of placebo response?
All of our placebo, for example, for ALTO-100 has been pretty reasonable and well within historical norms. I would expect that to continue to be the case.
Got it. With that, we are at time. Thank you both for joining us today and those who tuned in. We will be hosting our Epilepsy KOL panel next. We hope to see you there. Thank you.
Thank you for having us on.
Thanks.
Thank you.