Good afternoon, everyone. Thank you for joining us for the final session of the 6th Annual Novel Mechanisms in Neuropsychiatry Summit, by TD Cowen, and the Fireside Chat with Alto Neuroscience. I'm Covering Analyst, Ritu Baral. I am joined this afternoon by my Associate and VP, Athena Chin. And with us from Alto, we have Amit Etkin, who is CEO and Founder, I believe, of the company from the very beginning, from your Stanford days, I believe, Amit.
We want to start with ALTO-207. This is the program with the highest degree of investor interest for Alto. It's a fixed combination of pramipexole and ondansetron, currently in phase II-B development, as an adjunct treatment for TRD. For those less familiar with the program, could you review pramipexole's development history in depression and the challenges that limited its use such that this combination makes sense?
Yeah. First of all, thank you for having us, and as they say, save the best for last. Hopefully we'll come up to that billing. The question's a great one because pramipexole has actually been around and looked at in depression in different ways for over a quarter century, and the results have been very clear, both on the efficacy, which is outstanding, and the main issue, which is nausea and vomiting, preventing dosing high enough and fast enough, and that's really the motivation for this combination. Let me take a quick step back and just explain the context here. Of course, we broadly understand dopamine is important for depression, especially things like anhedonia. There's not been a direct dopamine stimulation or dopamine agonist strategy apart from pramipexole, which is a D3 preferring D3, D2 direct and full agonist.
Antipsychotics increase dopamine levels a little bit, but that's really not the same as directly stimulating dopamine. Pramipexole is approved for Parkinson's and restless leg syndrome, and usually it's dosed very slowly along with, in Parkinson's in particular, along with the disease. But trials in depression have consistently pointed out large effect sizes. Most recently, the trial PAX-D from the U.K., which is a nine-site, really well-done industry-like study in treatment-resistant depression, showed a Cohen's d of 0.87, so nearly 3x the average drug placebo difference in depression, and this is in TRD, in 150 patients. The efficacy signal is very clear. Most of those studies have been limited in terms of the dose that they've used. Most are around 1, 1.5 .
Some of the more recent studies, a little bit higher in terms of milligrams of pramipexole, because the more you give of any dopamine agonist, pramipexole amongst those equally, the more nausea and vomiting you get, which precludes people from getting high enough, fast enough. You have clinical trial data, and you have case report data that all say that you need to actually go higher in treatment-resistant depression in order to get the efficacy signal really coming out. What motivated then the combination, originally developed by Chase Therapeutics, was this idea that if we can mitigate that effect, very much like the COBENFY concept of mitigating adverse events so you can get your full agonist effect clinically. If we can mitigate nausea and vomiting, then we can get to a much faster titration schedule.
They used a titration schedule 5x faster than the label. Most prescribers go far, far slower than the label when they try to use pramipexole by itself, and we can get to a higher target. In their phase II trial, they got to 4.1 mg, which is the highest of any trial out there. With that, they were able to show both good tolerability and excellent efficacy, an 8-point delta on the MADRS, a Cohen's d of 1.1. That's really the motivation here, and then we can get into more of the specifics of the program itself, the way we look at TRD and the way we look at adjunctive and monotherapy use. But in terms of the groundwork, I think it's quite clear for the efficacy of pramipexole.
I want to dig in further on what you mentioned about how doctors use it now. It is used only as a monotherapy because it's pretty much only available as a monotherapy right now. But it was actually right about a year ago, and we had our KOL dinner after this conference where we asked about it as part of our conversation, and we were so surprised when our KOL turned to us and said, "Oh, yeah, I use this a lot. It works. It's very useful." It was the first time I'd heard this KOL who'd used it for years talk about it, because we had never asked, apparently. But there does seem to be a fair amount of real-world use, which we've come to understand in this past year.
Can you talk to, again, that titration curve that's used in the real world, how far up they can manage to go before they do hit that ceiling with nausea, and then what sort of patient this real-world monotherapy suggests the commercial opportunity is for 207?
Yeah. When people use it in the real world, and we've actually looked at this data directly through the NIH All of Us dataset . The vast majority, so about 75%, go at least 1.5x slower than the label. The average time to get to a milligram, which is the point at which you start to get efficacy in clinical trials, is over a year and a half on average in this dataset. So people go low and slow. The label is low and slow, but they go lower and slower. But often what you even see is at that next iteration when you're increasing the dose, and it's seen in that All of Us dataset equally, people immediately down titrate because they're not able to tolerate it. As a clinician, it's very hands-on.
Some people won't touch the stuff because you can't even get to 0.125. They'll just have a reaction in terms of not tolerating it. Then you go up a little bit, you wait, maybe the person has side effects, you go back down, you re-challenge. It's a very hands-on process. But once you get people to a good dose, clinicians tend to really like it because it works, because the efficacy has been very clear on a number of fronts, right? So one is it's being given right now because of what's involved in titrating it to very resistant patients, and yet they're responding well. You have an anhedonia effect that clinical trials point to as like 0.6 Cohen's d to 1.0 massive effects.
Nothing else has these anhedonic effects. It's clear clinically that that has an anti-anhedonic effect in just clinical practice. That target to be solved is how to get it to be tolerable, how to get titration to go faster. Remember, Chase used it 5x faster than the label titration, which is probably an order of magnitude faster than a clinician would use in regular practice, and hit a higher target. And the important part here, and this is important in understanding what TRD really means to us and to this drug, is the very broad range of depression patients, if you now have a well-tolerated, unique, really first-in-class type mechanism, the broad range of the depression patient population should benefit from it. When we talk about TRD, which is two-five treatment failures, it's really different populations at two than it is at five.
In all cases, you need at least two failures to get insurance reimbursement. You're really talking about at two failures, somebody who's failed like an SSRI and an SNRI, or two SSRIs, primary care populations. And that becomes a really important target where you can differentiate. General psychiatry, where they're comfortable using antipsychotics and various other kinds of drugs, but you're mainly still in all cases prescribing standard take-home drugs. That's another great population. And the much, much more severe end of the spectrum of TRD, where they're thinking about interventional psychiatry, and psychedelics, and esketamine, and what have you. Well, why do all those things with all the REMS program and needing to come in and so forth versus just taking a take-home[crosstalk].
So much easier. Right.
So much easier, right? It is about that differentiated clinical profile, the clear evidence of efficacy in prior pramipexole studies, and the tolerability that comes with the combination.
So your II-B is ongoing. Could you review the study design, very specifically the titration protocol that you are using in comparison to PAX-D, as well as the target dose, which I believe is 3.2 mg of pramipexole, 15 of ondansetron, and how that 3.2 compares to the final PAX-D dose as well?
Yeah. The PAX-D study, as I mentioned, is the seminal trial, really well done. Done like an industry-sponsored trial and powered like 150 patients. What they did was adjunctive treatment. That is, you come in on your antidepressant that you failed, you do not change that antidepressant, you add pramipexole on top, and had a huge effect size. So that is the basis for essentially the design itself that we are using. It is an adjunctive TRD population, two-five treatment failures come in and stay on their stable antidepressants, and then we add ALTO-207 on top. It is a titration period of two weeks, total treatment period, including titration of eight weeks, 178 people randomized one-to-one drug versus placebo.
That titration period is a custom titration schedule that to the patient just looks like you take a pill twice a day. They get a pack, a titration pack, that come commercial launch will just be a starter pack. Super straightforward for them. The content of that pill changes and that titrates them to that right target, which is, as you said, 3.2 mg of pramipexole.
They just have to identify the appropriate pill for the appropriate day in the titration.
Yeah. Which is not too hard, right? If it's on Monday.[crosstalk].
Right. Especially if it's in a titration pack, yeah.
Exactly. That just gets them there comfortably, and it's 15 mg for all days of ondansetron. Then they stay at their maintenance dose for the rest of that eight weeks. The powering of that trial is also done deliberately and conservatively. All of those prior pramipexole studies at an average dose of about 2 mg gave an effect size of 0.64. We're using 3.2 mg, and we've shown a dose-response relationship in those prior data, so higher dose. PAX-D reached only 2.3 mg, so considerably higher than prior data. But that 0.64 effect size at that roughly 2 mg dose, we're powering conservative to that. Our study is powered at a Cohen's d of 0.45.
At 80% power, and statistical significance, that is 50% power, begins at around 0.3 Cohen's d. In other words, anything from standard of care on up would be well powered, and importantly, every single effect size reported at a milligram or more in any pramipexole trial in the past would be statistically significant given our powering. We tried to position this trial conservatively and in the right way for success, where seeing frankly a 0.45 effect size would be an absolute home run, especially with this clinical profile and differentiating on things like anhedonia.
How does the protocol handle patients who can't tolerate that titration to 3.2? Because, as you just said, even if trial patients as a whole didn't get near the three, you could still have a successful study.
That's right.
You could still use their data, right? How is that handled in the analysis?
The base expectation, first of all, is that the clear majority of patients get to that top dose.
50%+ will get to 3.2.
Well over 50%, just judging by the Chase data, which was in without the formulation that we've developed, which is a modified release formulation that PK matches these two drugs. Pramipexole is a longer half-life than ondansetron, so you don't want it sticking around creating problems when ondansetron drops off. We PK match them with a custom modified release formulation. The titration schedule itself is optimized and is a little bit slower than what Chase did as well. All of that's geared to maximize tolerability. Your question though was what if somebody doesn't reach that?
Let's say they reach 2.4 instead of 3.2 as they're titrating up, and the protocol just has them basically exiting into maintenance at that tolerated dose. We optimize for each person the max that they're able to tolerate as they go into then the following six weeks o f maintenance dosing.
Their efficacy is the efficacy that comes from that patient.
Correct.
That's that data point. Got it. What do you anticipate the overall dropout rate to be? Would it be any different than your standard 10%-20% rate for most clinical trials?
Yeah. Our powering assumes 15% dropout, which is that standard rate. All of our trials have come in within that. That is sort of industry standard. In terms of the dropout rate that really matters, though, which is the AE-driven dropout rate. Our target is basically CAPLYTA, which I think is sort of a nice, very clearly commercial target, which is a 9% AE-driven discontinuation rate. Anything around there would be absolutely fantastic and exactly what we're setting the program up to achieve, which is a well-tolerated drug that, in this case, would deliver differentiated efficacy in a form where now you have everywhere from primary care on up able to take it with no issue.
What are the safety events that you are monitoring most closely? We talked about nausea, but one very well-documented potential risk for pramipexole is impulse control disorders. How are you mitigating that risk through trial design, study monitoring, and how will that flow through to potential real-world use, do you think?
Yeah. Let me just kind of define terms, I think, for people who may not understand necessarily what that bit means with respect to the mechanism of pramipexole. Pramipexole is stimulating dopamine. You are stimulating the reward system. It stands to reason that is a good thing for the treatment of depression, right? But for some people, that stimulation of the reward system might be a little too much, and they might do things like gambling or playing online games more. A lot of data has shown that that is really an issue that you see in Parkinson's, which is a very different setup, dopamine system-wise. They have fewer dopamine neurons and a hypersensitized post-synaptic system.
If anything, in depression, it is actually a hyposensitized post-synaptic system, and the rates of these impulse disorder symptoms is much, much lower in depression. It is really more of a disorder than a drug thing. Actually, data suggests it is the same in pramipexole as it is in aripiprazole, where it is, by the way, also on the label for aripiprazole, and somewhere around 0.1% based on a recent Swedish national registry study. So very low rates. PAX-D also pointed out additionally that there were two people who had increased online gaming behavior. Both of them were 2.5 mg. Those symptoms actually went away when they were down titrated to 1.5 mg, and they continued on the drug. These are things that are actually quite manageable. We are monitoring them. We are expecting them to occur at a very low rate.
Probably something you see more in an open label long-term exposure than in a short-term exposure. Usually, it would take a while to come out. All in all, we are certainly doing what PAX-D did, which is excluding folks with ICDs at baseline to limit that rate. Those are just a handful of people in the PAX-D study who were screened out for that reason. This will be something I think ultimately clinicians will monitor for. It is already in the pramipexole label because of the Parkinson's that will be inherited in the combination in all likelihood. It is in there, pramipexole label, and it is just not something that has been a huge barrier. The main adverse events we track are the ones that keep people from being able to tolerate the drug, which is really nausea and vomiting.
There are other things that occur either at lower rates that you do not worry about as much, or frankly, things like somnolence, which can happen with chronic dosing that we will be looking at. Frankly, the only side effect to which ondansetron might itself add is things like constipation, which obviously can be readily managed. We also monitor, because of the combination, EKGs to be able to rule out any QTc effects on the ondansetron end of things. The ondansetron dose is really below where you see QTc prolongation, and our modified release reduces the peak levels that also are what drive QTc. All of those things are being monitored, and everything is proceeding well according to plan.
Enrollment is progressing. What is the latest guidance on top line data?
Latest guidance is still the original guidance, which is second half next year. Everything is following plan very nicely on that front. We had delayed this spring our 100 and 300 readouts. As you know, by putting in quality measures to avoid things like professional patients, it slows down recruitment because you really want to make sure you are getting the high-quality compliant patients. That led to those trials being pushed out a bit. All of that was already baked into the assumptions for ALTO-207's recruitment, and so we remain on path for second half next year.
You are going to start the phase III before that. The plan is to start the phase III 207 study in early 2027. What drove that decision to start it before the readout, and how much flexibility do you have to modify that phase III design if the II-B results differ from your expectations?
Yeah. There are two things you learn from unblinding and end of phase II, that kind of outcome at a phase II-B stage, right? One is efficacy, the other is tolerability. Well, efficacy, we have 25 years of clinical trial data, including some very definitive seminal studies that clearly point to efficacy, then we have powered conservatively to those. So that feels less like a reason to wait for the end of phase II. Tolerability, we have just talked about, that can be assessed out of blinded data. FDA in our Q4 interaction, planned interaction here to align on the design of that phase III will get blinded data, so they can assess tolerability.
If you have both of those two things understood from different quarters, then actually the opportunity for accelerating your phase III program and therefore NDA filing is then quite exciting, that is what we did, and we will be seeking alignment with FDA on that design later this year. The flexibility in terms of the design, within reason, there are various things that are flexible, certainly we will be open to feedback from FDA.
Potentially, because of the conservative powering, we will want to revisit the sample size once we read out the phase II-B, but those are good problems to be solving at that point, and whether you do or don't, that's really a decision strategically more than anything else. Having everything on route allows you to bring your NDA filing timeline in, and it's the same thing that we are also considering now for the monotherapy trial, which is the second phase III. The package here would be the phase II-B plus the first phase III as adjunctive treatment.
The monotherapy TRD still, but monotherapy trial that would kick off second half next year also assuming alignment with FDA would also come before that phase II-B readout. That gives us a really full package of both monotherapy and adjunctive use of the drug being on label, which no standard treatment in depression that is not like an interventional thing like esketamine has. That is a quite unique profile.
As you think about these phase III's, are there aspects of the trial design that could better highlight potential anhedonia signals, and what could that mean for indication, labeling language, et cetera?
Yeah. I think everybody understands anhedonia is important, and it will certainly be part of the publication plan in very clear ways. We are discussing right now with FDA what it would take to validate a scale that would be used in our trials to assess anhedonia as a secondary outcome. I think as we all know, this has been something that has been a frustration for the field for years, that there is no validated scale. You can pull out items from your MADRS, and that's sort of a default that you could do, and other programs have argued for that. But there are some more modern anhedonia scales, and those are discussions we're having with FDA right now. It would be really great if there is a path there for validation to more formally indicate utility for anhedonia.
Regardless, it'll be in very clearly, I think, in the publication plan for physicians to know and see.
Understood. Let me turn it over to Athena.
Amit, let's close out with your broader pipeline, including ALTO-300, your agomelatine compound in phase II-B study for adjunctive treatment of MDD, as well as ALTO-100 in phase II-B for bipolar disorder. Could you just quickly recap the design as well as the modifications that you made to patient eligibility and site screening, and when we should be expecting top-line data?
Yeah. ALTO-300 is a drug agomelatine that's approved in Europe and Australia as a monotherapy in depression. We're bringing it here as an adjunctive treatment, the goal being to displace antipsychotics, which have huge side effect burden. In ALTO-300, agomelatine has been historically very well-tolerated. We're developing the 25 mg dose here, so it's being studied adjunctively. We also have a biomarker to further enrich response, which is an EEG biomarker that we established and replicated based on a machine learning signal. That's a first half next year readout. As I alluded to earlier, that was pushed back a bit in part because of these quality measures. These quality measures are things like medical and pharmacy checks, looking at making sure you can measure in their urine their underlying antidepressant.
We have reported both for ALTO-100 and 300 that this has resulted in at or close to 100% compliance as measured in different ways for those different studies. That is a really positive signal about making sure you have the right patients. So excited for that outcome. It is certainly a drug that I think should be on the U.S. market because it offers a very different kind of profile versus everything else, and certainly versus antipsychotics. ALTO-100, in a similar vein, in bipolar depression, all they have is antipsychotics. There would certainly be a welcome opportunity for something that is not an antipsychotic, but especially so is a novel mechanism of action that ALTO-100 is.
It is a pro-neuroplasticity treatment that increases BDNF signaling and release, and it is being given to patients with bipolar depression enriched for a phenotype that we found in major depression predicts better response, which is impairments in learning and memory as an index of poor neuroplasticity, matching the effect of the drug on enhancing neuroplasticity. That is a mid 2027 readout, and very excited for both of those programs, frankly. A lot of investor focus, obviously, on 207, but that positions us across our pipeline as now having multiple phase II-B readouts next year, all with drugs that could have tremendous value for patients.
Just quickly, what do you hope to see from both of these studies that would support advancement into phase III?
I think it is very clear for both its statistical significance on MADRS change. Given the setup there of well-tolerated drugs looking to displace antipsychotics or first in class, that's really the bar to cross.
Got it. With that, we are over time, so I'll leave it to Ritu to close out.
Great. Thank you, Athena. Thank you everyone for joining us for this 6th Annual Neuropsychiatry Summit. We covered a lot of new mechanisms, approaching data sets, approaching PDUFAs, and paradigm shifts in neuropsychiatry. We hope that many of you will join us for our adjoining KOL dinner this evening. Looking forward to seeing you there. Please contact your TD Cowen rep if you have not received the invite and would still like to attend. I think we have some spots open to discuss everything that we covered today. We'll be there with Athena and my colleague, Joe Thome. With that, thank you everyone for joining us.