Everyone, thank you for joining us at the 44th Annual TD Cowen Healthcare Conference. I am Joe Thome, one of the senior biotech analysts here on the team at TD Cowen, and I am joined by my colleague, Athena Chin. Today we have three excellent panelists for our orphan epilepsies corporate panel. We have Alex Aimetti from Marinus Pharmaceuticals, Jeremy Levin from Ovid Therapeutics, and Marcio Souza, CEO of Praxis Precision Medicines. Thank you all very much for joining us today. Thanks for having us. Maybe just to kick things off, we're going to start with a few thematic questions, and then we'll dive into some round-robin on the specific company programs. But maybe just what are sort of some of the important regulatory and clinical considerations when developing drugs, specifically for orphan epilepsies, relative to maybe broader neurological indications? Maybe how cooperative is the FDA in factoring in these designs? If you have your smiling for you on a side. Oh, sure. Very well. I'll kick it off then. We always give Jeremy the start to. By the way, these guys are going to give the presentation in my company sometime soon. But to be fair, your question is very pertinent because the FDA is highly facilitative. Let's put that on the table. The considerations, if you haven't read it, is that they understand walking into them with a patient group, understand the end groups that map to them, that map to the patient, is clinically meaningful sometimes, and therefore it allows you to work with them to define neuropsychiatry. We've done that a couple of times, and I'm sure that my colleagues have done the same. But that is a critical component. Yeah. I think there's a big improvement as well on what Jeremy just mentioned, where the agency, I would say in general, is cooperating and collaborating and understanding a lot more where genetic epilepsies are a new thing. We tend to talk about this as being around the concept's been around for a long time, but genotyping patients or understanding whether or not there's one or multiple reasons they're having, I would say, decomposing the LGS in specific buckets into actually real buckets is kind of new. So they are trying to understand from us, from patient groups, what is the history, if any, natural history here, and sometimes there's none. What are the most important things? Seizures don't manifest equally. As we know, there are different types of seizures. They're clustered for some conditions. They're more generalized for other conditions. But I think, in general, the fatality of these specifically has not been brought to the table so much with the agents. So, I can tell, recent interaction we had showing them a little bit more of that increased their urgency to actually get to treatments with patients, which I think is an incredibly important point to discuss. Yeah. And I think, again, just to differentiate between broader neurological indications and orphan epilepsies from the clinical perspective, I just want to emphasize what Jeremy said is really partnering early and often with the patient advocacy groups. And we've seen huge benefits in doing so, really helping us identify where the right institutions are to conduct these clinical development programs and which providers we should be talking to, and really helping to raise trial awareness within the patient community. I think the other thing you mentioned, Joe, was just around kind of the regulatory front. I mean, again, the FDA has been extremely supportive in thinking about creative ways to bring new, safe, and effective medicines to patients with rare diseases. And that's certainly been our experience as well. Then just quickly to touch on some of the reimbursement side, since our drug currently is approved for the treatment of seizures associated with CDKL5, which is a rare developmental and epileptic encephalopathy, I think some of the payers that we've interacted with really understand the significant unmet clinical need in this patient population with such limited therapeutic options that today we've essentially had 100% coverage and approval on all patients with CDKL5, which we think is just a testament to the understanding of the significant unmet need here. Maybe can you elaborate that a little bit, I guess, with the payer landscape? I guess, as you were launching, is it all me? Was there any pushback to payers that maybe you want to have to try a broader ASM before sort of now that you do have an approach for genetically defined seizures? Yeah. So a lot of the coverages that we're seeing are very similar to label in our clinical trial, which is the trial of two prior therapies. Now, our indication is approval for seizures associated with CDD in patients two years and above. And these are patients that, unfortunately, seizures manifest very early in life. So getting through the first few rounds, and they're highly refractory to existing treatment options, so getting through the first two therapies has not been a barrier to access for Ztalmy in these patient populations. I think that the common feature for this area, it's a great advantage, is that you definitely will have to try a generic. There are several generics which are often used. Once you've gone through that, they don't push back. Yeah. I don't think anyone of us, at least knowing my colleagues here, would be developing these drugs if generics worked for those patients. So it makes the entire conversation with payers, to be honest, with clinicians as well, a lot simpler. They fail. They are refractory. They are at high risk. So there is no pushback there. Definitely from other areas there, we haven't properly tried it. And that is different. That is not the case on those epilepsy at all. Maybe just broadly, how do you pick the indication that you're going after? Obviously, we've seen a lot of development in Dravet, LGS, TSC. Maybe some of the other orphan encephalopathies maybe have trailed behind a little bit. So how do you make that decision between going for it? I think it's really important. Some people, and we've heard recently about the idea of a DEE label, would be thrilled if that happened. But in our case, we're now in phase III with soticlestat. Data will come before September. And if you look at the history of that, that was a voyage of understanding exactly which indications to go after. So unlike others, we took this novel mechanism, and we did a 1b/2a, 18 patients, seven different DEEs, then spoke to the FDA and said, "Can you look at a basket?" And you couldn't at that time. Maybe it's changed. This is five years ago. The FDA may be responsive to it. Once we'd looked at those seven different disorders, we picked the ones that we felt would be the most responsive. There were conditions of that, obviously, a significant market, obviously, other competitor drugs. But the reality was, could you make a difference in those? We picked four: CDKL5, DUP15Q, LGS, and also Dravet. We then ran, not powered for efficacy, joint trials in LGS and Dravet. And we ran a similar trial in CDKL5 and DUP15Q. And now those patients have been that was the right decision because now our phase three is DUP15q with Takeda, is Dravet, and Lennox-Gastaut. That's 140 patients, 240. That is powered for efficacy. We're very confident about that. But we have something else. And this is the real advantage. We have three years of data, open-label data, of which we reported on two years, where these patients in the rare disorders have gone through these patients, never seen any kind of interaction with all the multiplicity of their different drugs, never seen a side effect of any consequence. I think we are feeling really good about the status of having picked the right disease, having follow-on disorders, and then being able to evaluate how your drug works in multiple diseases over multiple years. With a bit of luck coming up this year, that data will be positive, and it will show that that decision process was right. Yeah. I think that's a very thoughtful way to look into this, the way Jeremy and Ovid laid out. I think a complementary way to look into this as well, right? So in our case for PRAX-562 and for other molecules, if you're looking preclinically, this could be used in the case of PRAX-562 across many, many DEEs. Then you have to rank the ones that are more likely mechanistically on this case as linked to genetics to use. Then you go on the other end of this spectrum. These patients have used different drugs. So if you inquire the U.S. claims database or EMR systems and you realize that a lot of those drugs are not doing anything for these patients on the other end, so the basket approach is being attempted. I think the Europeans, specifically, they incentivize companies before to look into that. Are a great theoretical idea, but those patients don't behave the same way. I think that's a very good comment. By not behaving the same way, right, not getting into all the nitty-gritty details of statistics, but it complicates significantly to show the benefit for the ones that would benefit. I think at the end of the day, we want drugs that work. These are fatal diseases, as I said a couple of minutes ago. We don't want to simply explore or play as chance as the main factor. Chance should never be the main factor in drug development. It should be the thing you're trying to control for. We touched a little bit at the beginning around the FDA and sort of the quality of your interactions. Obviously, the past couple of years, the CNS divisions of the FDA have gotten a lot of attention with Aduhelm and the Leqembi approvals. Really? Yeah. I guess, is that broad across CNS and going down to epilepsy indications, or is this left for the U.S. news? No, no, no. Don't mistake U.S. news for FDA. If you like that approach, then you should love epilepsy because whatever their boldness was in Aduhelm and in neurodegenerative disorders, there was lack of mechanism. There was real need. Here you have an area which is broken open. You have mechanism. You know exactly how to do the trial. You know that, in fact, all the old stuff is pretty useless and it's not working. And oh, we know now that you can effectively reduce seizures. And by the way, in some cases, you have the hint, just the hint, you're going to have disease modification. We certainly are hopeful we'll see that with soticlestat. However, again, I'll reiterate it. If you love their boldness in Alzheimer's, boy, that's the area you should really be focused on because epilepsy then opens up not just the idea of seizure suppression, helping 3 million patients in the United States. It also opens up the possibility - and this is really important - the concept of hyperexcitability in the brain because hyperexcitability is not just epilepsy. It's psychosis. It's depression. It's pain. And one of the hints that we have from the past is that old epilepsy drugs had some effect on these other disorders. So I'm very comfortable now saying that I think we're at the cusp of a complete change in certain areas, scientifically, of hyperexcitability, epilepsy being the opening door and facilitated by the boldness that we're seeing. Yeah. Indeed. Indeed. Completely agree. On the subject of placebo response, we've seen a lot of variability in these epilepsy studies. What are you doing to kind of mitigate these responses, and what are some other strategies that you are also pursuing? Maybe start with you, Alex. Yeah, certainly. There's a few different ways that we think about placebo responses in some of our epilepsy trials and ways in which we can use education and training in protocol design to help control for it. I think the first really important thing is in these ultra-rare diseases, sometimes these are the first placebo-controlled trials that they've been exposed to. So I think that requires a lot of education on the parents and caregivers around the importance of unbiased reporting of seizures, consistent counting of seizures as a way to potentially minimize placebo response. I think, second, it also requires a lot of education and consistency of recording specific seizure types over time. And so we and many others in the space work very closely with the Epilepsy Study Consortium to really help drive that consistency within our clinical trials as well. I think, lastly, through protocol design, I think more of a statistical consideration is we really try to avoid what's known as kind of regression to the mean and really make sure that we're enrolling the right types of patients that really have consistent background baseline medications and consistent, unfortunately, uncontrolled seizures prior to coming into the study, which I think give the trial and the drug the best chance to demonstrate efficacy if there were to be some. Jeremy? Yeah. No, this is a great consideration. And as in the epilepsy field, people always point to Fintepla and Epidiolex and say, "Gee, check out. They're doing well. And by the way, they're doing commercially very well, and they are growing." However, if you look at their phase two, they had an observational phase two. We decided not to do that because of the reasons that you described. Great phase two results, less good, just as you point out, Alex, phase three. So what we did was to do a randomized placebo-controlled double-blind trial in our phase two with the express intent to tease out what is placebo. To Alex's point, how do you select these patients? What's exactly the right kind of patient for the trial? I'm confident, based on what you said, it's the Epilepsy Consortium's role finding the best possible training that you have for your sites. You must have that. We carry that through also to our 329 program. All our programs carry this through. You pick real specialist areas. What you do then is, in fact, your entry criteria becomes super tight. For example, in the LGS trial, we sharpened up for our phase III based on our phase II data that your entry criteria, we noticed that these patients were being defined as having drop seizures. They were having 5,000 of them in the phase II. That's not possible. It means that, in fact, somebody is not experienced in measuring what is a drop seizure. Drop seizure, you fall to the floor. And so that criteria was changed to be very clearly major motor drop seizure, which, by the way, is a part of CDKL5, as you know. We then also tightened how you actually bring patients in by looking at those who are breaking through the highest number of their therapeutic regimes. I think by doing that, you can really control for it. I'm confident that we have. Yeah. And I think even the concept of placebo amplification or response in epilepsy is a little bit misguided, to be honest. Because when you look into everything Alex and Jeremy just mentioned, it's really controlling training and entry, right, of the data. I have a very hard time believing that patients can just like, let's say, pediatric patients, we were talking about DEEs here. They're not even the ones reporting, for the most part, right, in certain of those conditions, specifically the pediatric ones. So how are they actually having a placebo effect when they are unaware they're on that drug? But there is a reporting bias that you have to control. Those are all diaries for most of the seizures. They have to be counted properly. They have to be trained properly. And I think different companies took different approaches there. I think we all take very similar approaches, right, education, training, making sure zeros are entered as much as numbers that are positive entered, which being a problem in the fields. And then, obviously, whatever response we see on the placebo arm, we call placebo response. But we might as well call that variability of this patient population, I think, would be equally appropriate. And I do think you should focus on this because in all rare diseases, the families are hopeful, the doctor is hopeful, the patient is hopeful that they are able to be conscious of it, and they misreport results. And we've seen this in certain areas like Angelman's as a real challenge. That's why we're no longer in Angelman's, because unless you run a strict placebo-controlled trial, and I do mean double-blind placebo control, you end up with a situation that everybody's going to be positive. Everybody's going to be positive because you take that patient from their home, you take them to the site, at the site, they get stimulation, and they start feeling better, and they record better. And so you have to have very objective endpoints. Now, in epilepsy, we have the advantage of all the previous trials. We know how to control for this. And so I'm thrilled that we've been able to do that. But in the absence of a gold standard, this double-blind randomized placebo-controlled trial, I fear that you often get anomalous things creeping in, particularly if you haven't picked your site, trained your people, and made jolly sure that they know what they're measuring. Just one more thing to add on in terms of picking your site. One of the things that we've observed and has been reported in the epilepsy literature is to be very thoughtful about the geographies in which you conduct your clinical trials. We've had some experience where there's been some higher placebo responses in specific regions and things to consider as we think about future studies. I think that that's very important to be very thoughtful when considering what clinical trial sites and investigators to work with. Thanks, great color. And we kind of touched upon this a little bit ago about genotyping. So with the increased use of genetic testing for diagnosis or classification, how has this changed your thinking with drug development, clinical trial design, or even reimbursement negotiations? Maybe you said it was the most recent? Yeah. We started there, right? So for us, the inspiration for all these programs was through genetics. So to some extent, we rely upon increasing the number to actually pick the targets, understand the population. I think what is quite interesting, maybe I'll share a little bit about our 222 program because we look into the epidemiology over time, and you look into like 10 years ago, the gene itself association with epilepsy is not much older than that. So no one was genotyping any of these patients other than large research hospitals in the world. And of course, there was a very small number of patients, right? They would go to another bucket. you look into right now, a patient with SCN2A gain of function is always identified in the first three months of life because they seize in the first three months of life, and they get a genotyping. But those numbers of incident patients, they don't make sense because they're much higher than the expected. So it made us ask the question, what's happening to these patients? Why is the prevalence still around 1,500 or so patients in the U.S., right? And I think what we realized, they're all dying in their teenage years because there's completely inadequate care right now. I think the more we understand about the genetics, the more we're going to be able to decouple patient types, as we were discussing before, have better therapies for them, better associated care because it's not only the pharmacological treatments, as we know, and develop even better drugs and better combination of drugs, right? I think all of us here, I know we shared the intent is not only hope, hope's never a strategy to modify the disease. It starts with they living better, less seizure, more days with seizure-free, but ultimately, development for these patients and independence for their parents and so on. Genetics play a huge role. In the United States right now, it tends not to be definitive on everything, but it's literally not a problem for new patients. It's still a small percentage. Don't get it. But for new identified patients, it's not an issue. But for the 10-year-olds, 15-year-olds out there, it's still an issue that we need to fix to get them the definitive diagnosis. More treatment, more diagnosis. That's normally how it goes. I think the role of genotyping is different across the spectrum of when the patient is defined through to when you pay for your therapy. On the payer side, we've not seen an entry of this yet. Okay? It may be in the future. As you go down, you find, certainly in clinical trial before approved, FDA is interested, but it hasn't yet made definition about it. As you go into the clinical trial, it's helpful. Let me give you an example. It's helpful for we're going into phase II this year in an area called cerebral cavernous malformations. 1.7 million patients have such things in their brain, of which approximately one-third are genetic. There are three different genetic kinds. It's important for us to know what is the genotyping of these patients. Actually, physiologically, you can pick them out. But as you go further where you have pure clinical syndromes like LGS, I think this is where what you're describing is so important. What will happen over time there, right at the beginning of this clinical trial process, we will be able to gradually tease out the syndromes within the clinical syndrome not, sorry, the genotypes within the clinical syndrome that match with a specific type of behavior. And that's important. I don't think it's yet had a major impact, but it's likely to in the future. I think we're at such an important inflection point for genetic testing in the field of epilepsy. There's just a ton of momentum. It's the number one topic at the major medical meetings. And I certainly am a huge proponent of genetic testing to really help understand the basic mechanisms of disease and really help these patients, parents, and caregivers find a community and also enable future disease-modifying and precision medicine. I think, though, as we continue to identify more and more and more pathogenic genes that lead to epilepsy, the genetic pediatric epilepsy space may become more and more and more fragmented, which, again, I think is critically important to know. I think that it really does improve patient care and all of that. I think as we move forward into the future, I think similar to as you're talking about LGS and whatnot, I'm hopeful that patients that do get a genetic test and do get that genetic diagnosis, if in fact they have one, but also carry a clinical syndromic diagnosis as well to help enable access to broader clinical trials, potentially, or access to potentially approved medications as well. I think, again, there's an opportunity for that. But again, I think that we're at a really exciting point for genetic testing in terms of lower barriers to access of these tests and the basic science understanding of the diseases. Perfect. Maybe we'll do a couple of rounds of company-specific questions now. Alex, we'll stay with you alphabetically. The company had an update last night, what we've all been waiting for, the enrollment of the interim for the RAISE study. So maybe if you want to just highlight the background of the RAISE study and maybe what was the update last night. Yeah, certainly. So just really quickly, At Marinus, we're developing ganaxolone, which is a neuroactive steroid that acts differently via a GABAergic mechanism, distinct from other compounds that exist out there. We're developing it in an oral formulation used for these rare epilepsies as well as an IV formulation, studying it in refractory status epilepticus. And as Joe mentioned, we did announce last night that as of January, we did reach the predefined enrollment target for us to conduct a prespecified interim analysis. We anticipate that interim analysis will be conducted in the first half of the second quarter of this year. We're extremely confident in some of the powering assumptions and statistics in the interim analysis. And if we do achieve the prespecified stopping criteria, we then would announce top-line data shortly thereafter. Can you go into a little bit of the powering assumptions? What are both of the primary endpoints? Maybe how did you set those powering assumptions based on the phase II? Yeah. Before I go there, I guess just to provide a little bit of background on the study design. So in refractory status, these are patients that have failed first-line benzodiazepines and second-line non-sedating IV anti-seizure medications. There have been previous placebo-controlled clinical trials in super-refractory status epilepticus, which is further in the line of therapy. The treatment objective of that study was trying to get patients off of third-line IV anesthetics. I think what's critically important to understand in our study design is that we're really looking to avoid the initiation of third-line IV anesthetics, which has been shown through multiple publications to be associated with significant clinical complications and increased healthcare resource utilization. So in the phase III RAISE study in refractory status, we're looking to enroll patients essentially on the verge of IV anesthesia and randomize them 1-to-1 to ganaxolone or placebo. We do have co-primary endpoints. These co-primary endpoints were developed in collaboration with our clinical thought leaders and FDA to be clinically meaningful attributes of a new therapy in RSE. And that is rapid onset of action. So the first co-primary is looking at the proportion of patients that have status cessation within 30 minutes. And then the second co-primary is really aimed to demonstrate more of a durability of effect. And again, looking at that avoidance of IV anesthesia. So that's the proportion of patients that do not go to IV anesthesia at 36 hours. Again, I think when we look at our open-label phase II data and we looked at them, and that was in 17 patients, we saw 100% of the patients did not go to IV anesthesia at 24 hours. Now, when we now transition to a phase III placebo-controlled trial, we certainly don't believe that to be 100%. But again, I think we have a lot of confidence in the therapeutic effect of ganaxolone in a very similar patient population. Now, one of the things that we often get asked quite frequently is now that we have a placebo arm, what are some of the expectations around a placebo response, particularly on that second co-primary of IV anesthesia? I think what's really important to note here is how thoughtful and intentional we were in the protocol about how we're trying to identify the right patient population. And again, as I mentioned previously, these are patients on the brink of IV anesthesia. A clinical decision to randomize a patient in the study is essentially a commitment to IV anesthesia if they do not improve post-randomization. We think that that's an important way to control for a potential placebo response in this study. Conservatively, from a statistical perspective, we've thought potentially up to a 25% placebo response. I will say that looking at the baseline demographics and our intense education of sites, I personally would be surprised to see a placebo response north of 15% on that second co-primary endpoint. Due to some of the clinical enrollment delays, there have been some questions around sort of the overall commercial opportunity. Obviously, you've talked a little bit about the unmet need. Can you maybe talk a little bit about why the pace of enrollment in the interim doesn't match what the commercial settings look like? Yeah. Conducting clinical trials in an ICU with a lot of hospital turnover, starting it during a pandemic, has certainly been challenging. In refractory status epilepticus, we believe there to be about 35,000 episodes of RSE per year. I certainly believe that enrollment in a clinical trial here, and there is to be a potential disconnect between market opportunity. Conducting a clinical trial in an ICU with multiple stakeholders with different backgrounds requires seamless coordination. So it's taking a lot of it's a very high-touch study that we've gotten many folks within Marinus on the medical affairs side and on the clinician side to really help form those deep and trusting relationships at the site and really help on the education around the study. I think that a lot of that work has started to bear fruit in what we've seen in terms of an increase in the enrollment rate in the last six months. Jeremy, let's start with OV329, your GABA-aminotransferase inhibitor in phase I study. Could you just quickly review the study design? I don't believe you've disclosed doses tested and when we should expect first data. Yep. But for those in the audience, we actually have four different programs. One is in phase three, so we'll get the results before September. The second, which is 888, is in cerebral cavernous malformations. That's going into phase two. Then what you're talking about now is our SAD and our MAD that's being conducted for 329, which is a GABA-aminotransferase inhibitor designed to subvert and change the area of vigabatrin $600 million a year in sales, but a putative background problem. And then behind that is a complete platform of breaking open the area of KCC2, which is going to be an area that all of you as investors, if you're in the interest of hyperexcitability, should be paying attention to. It's going to be a very big change in the area. Now, going back to 329, Alex, I compliment you. The guys at Marinus cracked open what is a huge market: IV and acute refractory seizures that bring a patient to the hospital. We have taken a very clear view that known mechanisms like yours and like ours are mechanisms that can make a difference if you get them to the patient at the right time in the right way. In our phase we've conducted right now a phase one MAD and SAD. We'll report the data out on the oral chronic data by the end of this year. That uses TMS. In addition to that, we have added really stringent criteria for how we want to look at dosing. You can get a hint of dosing. We're not reporting that out until we have the full data. But the hint of dosing, this compound is in animals. The dose is roughly 3 mg/kg versus vigabatrin 300 mg/kg. In our case, we know that we are going to be able to detect we're looking to detect using these TMS and a multiplicity of different measures as to whether or not we're seeing a good result from the oral. If we do that, we'll be able to advance it, hopefully, into a phase II and then into a phase III. At the same time, we've taken learnings from Marinus and others of the scale of the unmet need in the hospital setting of those who go into anesthesia. We've advanced an IV formulation of this. So we have an IV formulation, and we expect to be able to put that into patients next year, not this year, next year. So data coming end of the year, and it's all looking exactly on target. Awesome. So 329 also shares the same mechanism as vigabatrin, as you just mentioned. There are some syncope signals that were observed with the use of vigabatrin, such as retinal damage. How common is that? And what have you seen so far with 329? So the side effects or the putative side effects are the opportunity. The very fact is the mechanisms are known. If you inhibit GABA-aminotransferase, you increase the amount of GABA in the extraneuronal space, and you should see an effect on both phasic and tonic currents in the cell. And so the first question you have to ask is, are these side effects real? Well, the fact of the matter is, 20 years ago, when they looked at these patients who were coming in, they didn't really study the ophthalmology of the children at the time. That said, let's not kid ourselves. If you take a look at the rat and you give them a dose, there's no therapeutic window for vigabatrin. You give them a 300 mg. You see an effect on the retina. Now, we don't yet know exactly what the mechanism behind that is. I do know that if you give the same dose, the same dose equivalent, three milligrams, you don't see that. In addition to which, we designed a different molecule that binds completely differently to GABA-aminotransferase. It has a completely different pharmacology. So, for example, repeated low doses of this drug, of our drug 329, have a long-term effect, very long-term effect. And unlike vigabatrin, which is fairly rapid, very quick. So we're very confident that we've understood the pharmacology of this sufficiently for us to say that our information from the SAD and the MAD will allow us to say, can you dose this chronically, which we hope to be able to do? We know we can dose it acutely because in an acute drug, those types of toxicity issues are not of any consequence whatsoever. Awesome. And you also mentioned that you hope to see good data from your oral OV329. What would you constitute as good data from this? Well, any kind of data is good because, to be perfectly honest, we've had to reconstitute the entire body of information around vigabatrin. It's really remarkable that people have not done publications of any rigor on this. They haven't done any true science around it. So for us, that's going to be good data. That'll tell us, should we be investing long-term in a drug, what is the exact dosage you want to use? And good doesn't mean it works. It doesn't work. At this stage, we will know how much GABA is generated because the way we've designed the trial and the detection capabilities we have is really fabulous. You can actually measure roughly how much GABA is being produced. And the advantage of having vigabatrin is that you know how much GABA you have to have. It has to be about 150%-200% increase to say, I'm going to hit the seizures. So lots of interesting stuff from vigabatrin. But good data will say to the field, this is what a GABA-aminotransferase should look like. Inhibitor can be used chronically, or it should be used acutely. And we're pretty confident we're going to get the data we need. Awesome. Thank you. Marcio, you're top on the order from your phase II POC study of 628 and photosensitive patients is due hopefully shortly. There's only three more weeks in the quarter. Yeah. Do you expect partial or complete suppression of the PPR response? And how does this translate to broader therapeutic drug efficacy? The approach we took of 628, and I think it's not very dissimilar to what the discussion we have here, right? We know how those drugs work. Can we widen the therapeutic index here? If you look across the board for 628 preclinically, the phase I, the translational work for the phase I of the EEG, because using TMS, we decided to use quantitative EEG, po-tay-to po-tah-to, on how we actually measure the effect there. But it's there. It's in the brain. It's doing what we expected. It kept surprising us, I would say, from a therapeutic index perspective. We know what it's binding. We know it's a different site that has not been explored before. But what it's generating from a very good set of animal models for translation is unprecedented efficacy for focal epilepsy and for others as well, not only focal, but we'll be no pun intended focusing on focal epilepsy right now. So the question we had at the point that we designed the PPR study was whether or not the therapeutic concentrations we were expecting were in the range that we could move comfortably to a very efficacious drug, as we expect this to be, in a larger placebo-controlled focal epilepsy study. And there are not a lot of ways to determine that. One way that if you look into the literature that's being explored thoughtfully is you've got a patient with epilepsy that has, upon stimulation by stroboscopic lights, an epileptic discharge in the brain that starts focally but generalizes very quickly, whether or not that range is reduced or eliminated. Not a very pleasant study for the patient, right? So several hours being stimulated, being measured with a hooked-up EEG, but very good if you want to know whether or not a drug is doing what it's intended to do in a patient population. Again, if you measure our internal criteria for success, or the literature it says, 50% partial response, that should be a yes. Go forwards. Arguably, one could argue 25%, but I think we set a minimum of 50 for ourselves. What we said in January was, with our first cohorts, that is 15 milligrams, which gives a general exposure above 1 multiple of the MES-EC50 equivalents in humans, we exceeded our expectations. So I know we didn't say much more than that. People keep asking, what does that mean? But I'm going to repeat it. If we say 2 out of 4 with partial, that is not a lot of other permutations to exceed that. So it's actually pretty straightforward. But what we meant to say there is, that's it. We could move forward with this compound. We should move forward. That is indication that all the boxes have been checked. At the point in time, the decision we had was to dose the 45 cohorts because the patients were scheduled. So we have, as we know, it's very hard to reschedule all of those patients, especially the ones in PPR studies. So why not explore the safety, the exposures, the response, despite the fact that most of the studies before did not show an exposure response, but rather not a binary response? And there has not been, specifically with sodium channel modulation, a false positive. And what I mean by that is, should be obvious, but what I mean by that is, every single sodium channel modulator that was tested on PPR, and there are many, ended up being very successful on focal and only limited by safety on the use, not very different than what I was discussing before, vigabatrin or other medications. So when you look into the 628 profile with safety, and knock on wood, continue to be exactly how it is right now, really not an issue at all. Look into the translational for efficacy, very straightforward from the MES. Safety is really only humans. And what we are seeing there is exposures that are, on average, 10-20-fold the MES-EC50. We cannot even dream of exposures like this for any other compounds, right? You hit a wall with everything else. Around three have been safe so far. So if you can combine both of them, this should be what the field of focal needs. That is a drug that doesn't require titration. Neurologists are very comfortable on prescribing, not these specific specialized epilepsy centers, and brings efficacy, driving resolution of seizures in these patients. So we're incredibly excited. Keep an eye for our president in the next three weeks. Thank you. Perfect. Moving back to Alex, one more question on the RSE opportunity, then we'll get into some of the oral programs. But the company has indicated sort of the importance of some of the secondary endpoints that you're monitoring in the study. Kind of what are these? And will these also be coming with the top-line update if the trial is successful, or how are you going to lead those out? Yeah. No, I think that's a critically important question. So we were very thoughtful in the design of the phase III RAISE trial in refractory status epilepticus. And we recognize the challenges associated with launching a new product in a hospital system and how critical it is to demonstrate both clinical as well as the economic value of a new therapeutic. So there are multiple secondary endpoints in the protocol that really look at healthcare resource utilization that we think will be critically important to those formulary decision processes that will occur in the future. Examples such as length of stay in the hospital, days in the ICU, time on mechanical ventilation, the percentage of patients that progress to super-refractory status epilepticus. So there's a multitude of endpoints that we were very thoughtful that we included in the protocol and will be captured. I think what's really important to note, though, is that by conducting this interim analysis at two-thirds information, which will turn out to be about 83 patients, and we reach that enrollment towards the end of January, we now are north of 90 patients enrolled in a double-blind fashion. By the time that the anticipated DMC review and interim analysis is conducted, it's very possible we'll have north of 100 patients enrolled in a double-blind fashion. So we will take then at that point, and probably the rest of the summer, to collect and clean those data. Some of those key—let me rephrase—very important secondary endpoints will then be released towards the end of the summer or at some major medical meetings here in the fall. We'll include that total 100 patients, which I think will really help with some of the sample size and the powering of those secondary endpoints will only be tested more descriptively. Perfect. Maybe we'll jump over to the TSC opportunity. Can you remind us what we saw in the phase II? And maybe to lead a little bit, you did alter the titration protocol for the phase III that's ongoing right now. Why and what do you see? Yeah, certainly. So we do have an ongoing phase III study in Tuberous Sclerosis Complex. We think TSC is about a 4-5x market opportunity compared to our approved CDD indication. As Joe mentioned, we did run an open-label phase II study in 23 patients. We reported a 16.6% median reduction in seizure frequency. And we at Marinus were surprised with that number, to be completely honest with you. And we thought it didn't truly represent the potential clinical benefit of the molecule in this patient population. However, with all phase II studies, I think that they are very important and help inform phase III designs and development. And this one certainly did that as well. What we noticed when we looked at our phase II study is that we saw a higher discontinuation rate and a higher report of somnolence in that patient population. Now, I certainly don't believe that there's fundamentally a different safety profile in this patient population, but perhaps because they're less cognitively impaired and able to report somnolence, that potentially could have an impact. There's some evidence in the literature that suggests that both somnolence could potentially lead to seizure exacerbation. And also what we saw is that when somnolence was reported, we saw multiple dose reductions, potentially not able to reach therapeutic levels. As a result, and when we moved into phase three, one of the things that we observed is that when somnolence was typically reported, it was during the titration phase. So we took more of a scientific approach and based upon population PK modeling on how we could more effectively titrate patients up, better tolerability to potentially effective therapeutic levels. As Joe mentioned, in our blinded phase III data, we're really encouraged to see less than a 7% discontinuation rate, which, again, is a significant improvement relative to our phase II study. We're not hearing, again, anecdotally, reports of somnolence nearly at the same rate that we did. That piece of information, albeit in the blinded data set, at least gives us some hope and confidence that the change that was made to the titration schedule may potentially be having the intended effect that we were going for. Maybe last question, just in terms of when we're going to see data from this program, and what are you looking for? What's the clinical meaning from the signal of a reduction in seizure burden? Yeah. So we're looking to report our top-line data in the first half of the fourth quarter this year. The study is 90% powered to show a 25% delta. I think the powering assumptions were similar to what we saw in our CDKL5 phase III study with a 27% treatment delta. These are patients that are highly refractory. When we look at some of the baseline demographics, these are very similar to CDKL5 patients. They have 50-60 baseline seizures, and they've tried multiple therapies, including a lot of the branded drugs that are out there, such as Epidiolex. And so I think that seeing an effect that way in this highly refractory patient population will be clinically meaningful. Jeremy, before we move to your KCC2 transport program, let's talk about Soticlestat. It's in two phase three studies. With regulatory filings expected around the next half of this year, maybe 2024, could you just review the timeline, the data, and frame maybe expectations for those? Yeah. So it might tie a little bit into Joe's question as well, if I might, because secondary indications are very important. So this is a molecule which originated in Takeda's labs. We took it over, and we did the following: phase I, then we did a 1b/2a, 7 different DEEs. And FDA said, "You can't have a DEE label. Maybe that'll change in the future. Longboard, we're crossing our fingers." Then Takeda most surely will get it. Then the next thing was a phase II, IV different disorders, and then the phase III, when they took it over completely. And they will owe us $660 million and double-digit royalties up to 20% on what should be a multi-billion dollar franchise. Now, when's the data coming? So what we know is the following. Number one, Takeda is guided that it will have regulatory filings no later than the end of their fiscal 2024. That's March of next year. So we do know that they've completed enrollment in both the LGS and Dravet. We also know that they say that the data will be out before September. So we're in a very good moment here. But if I can, just for one minute, just turn to CDKL5. Very instructive discussion. One of the early phases that we did was CDKL5. CDKL5, essentially a small cohort, open label, but very, very instructive. We did both seizure reduction and secondary endpoints. And basically, when we handed the program over, we asked Takeda to continue this, which they did. Their open label data now is three years plus. And they reported this at the Lulu Foundation, which is their data, not ours. And we're not. We didn't actually present that data. It's difficult to find online. What they showed is a 50%-80% reduction in seizure, consistent for many years, not just a little bit. And that's all seizures, okay? Very substantial reduction. So it should work very well with Marinus's drug. And I think we're on the cusp of a change there. Now, the secondary endpoints, which is really important, what they also reported was the changes in cognition, behavior, and the communication. These are very interesting. So I would increasingly expect other trials to ask the question, "Are you seeing a change in these secondary endpoints?" Because they represent profound change in the way that we're thinking about epilepsy. Epilepsy is not just seizure reduction. It is a change in the way that the new medicines are working. And here, you have three different mechanisms, completely specific. Each one can play with the other, by the way. Although we haven't yet cured epilepsy, there's a whole this is why I was so excited about telling you about Alzheimer's. If Alzheimer's is great, you should be looking at epilepsy. The reason is sitting in front of you here, three different mechanisms, each one of whom can play with the next. Oh, by the way, until you get a cure, 40% of all epilepsy breaks through. Then they have to add another drug. That's why we're excited, not just about the mechanism, but also the secondary mechanism. I hope I've answered your questions. You should stop me talking now. No, I have another follow-up. What's the status on your OV350 KCC2 transporter? Well, that is a great program. For those who don't know, KCC2 is the controller for hyperexcitability in the brain, okay? And we have animal data that demonstrates that the lead molecule, which is 350, interdicts and can reduce completely in status epilepticus, which is caused in animal models, which is resistant to benzodiazepines, can make you re-resistant, can reduce it, can stop the seizures. Now, we repeated that with the U.S. Army. They use rats. They give the rats a neurotoxin. The rats go into epilepsy, and they then try and reverse that with the benzodiazepine. It doesn't work. The rats die. But if you give them our drug, it does. They live. So we know it works independently. Where the status is right now, we've really begun to unpack hyperexcitability. So we also have data not just on epilepsy, but on acute psychosis. So that acute psychosis is really interesting because now, as we unpack this mechanism, we not just have 1 molecule, which is 350. We have 100 molecules. There is the 4000 series and the 5000 series. So we have now the potential to drive into the clinic an IND for many years to come from this array. We're taking a very careful study as to whether it should be in epilepsy, whether it should be in psychosis. We'll make that decision and make that announcement when it's ready to go. But we are going to have an IND before the end of the year. And it will be; we're going to look at both oral and IV, same kind of strategy here. Whatever the disorder is, if you have an oral, that's great. You can use that for long-term chronic treatment. IV is acute treatment. IM is acute treatment. But the point here is we're very confident about our data in psychosis. Acute psychosis, this was an unexpected but predictable finding. So that's it. But you also should ask the question about my favorite disorder, which is CCM. But that's a question for another. That's in phase two or will be by the end of this year. Understood. Thanks, Jeremy. Marcio, next on Praxis T-Type or ASO. So your interim EMBOLD study showed an interim 43% median seizure reduction for patients. Can you just walk us through current plans for future study designs? What additional time points will you be investigating, and what can we expect from the rest of this year? Yeah. So PRAX-222 is a quite interesting molecule in a sense that is by far the most potent ASO for any of the CNS conditions out there, right? So if you look at the preclinical data, we're able to not only reduce the seizures, actually, preclinically completely eliminate the seizures, but independently of the point of entry, we're able to keep the animals alive. And these are disease animals harboring the human mutation. So it was incredible from the get-go. And thanks to Steve's lab and all the funding from actually parents to start that program. When we took it all and we look into all the human primates data and the different animal models, we could see efficacy at very low levels of exposure. But we took, I would say, the more traditional approach to, let's start with adults, that is not going to harm patients, number one, would show some efficacy, but maybe there is some space to continue to play with it. So for patients, a lot of what my colleagues just said, good sites, very good intervention, well, all the diary data entry and so on. So first on the safety, right? We dose 4 monthly doses, intrathecally delivered, 1 milligram each for these patients, no treatment-related adverse effects. And that is in itself a pretty big deal for an ASO. I know there is an urban legend about the distribution of ASOs in the brain. ASOs distribute incredibly well in the brain. That's not an issue here whatsoever. For as long as you have the right mechanism, the right dose, and really measuring it properly, what wasn't expected was the degree of improvement. We reported both mean and median were incredibly similar. We know the distribution there, like over 40%, a very significant increase on the number of days without seizures. If you think about just the quality of care, quantity of care, healthcare utilization, not seizing in a given day for patients that seize every single day, it's a really big deal. Incredibly happy, very consistent results across these patients. Have a few patients outside of the U.S. as well for emergency access, very consistent with the U.S. cohorts. And then begs the questions like, "What do we do with a drug on a disease that has nothing, nothing controls seizures, nothing changes the course of the disease here?" And by the way, there was a number of improvements that we didn't report that are what I would consider disease-modifying improvements. That's not what we are looking for here. We're incredibly happy to hear from the investigators, from the parents, on what was happening on the daily life of these patients. Maybe that adds to the enthusiasm that we didn't talk much about. But it was a change in course. So we wanted to make sure major regulatory agencies, including the FDA, the EMA, and others, were appreciating, number one, the gains. The second, were sharing with us the idea on how to develop this drug. As you probably know, we have a PRIME designation in Europe. It's a very straightforward dialogue with European regulators. They want to help us develop this. As Jeremy mentioned before, the FDA is pretty open for these conversations as well. But there are three major geographies, and two of them I just mentioned, that are fundamental for patient recruitments and for the size of the markets. We want alignment with all three of them before we start the global program. We are two out of three on the scorecards for now. We want to get that done by mid-next quarter and then start the pivotal study before the end of the year. We expect to be a single pivotal study here, which would put us no promise at this point in time. But I think we can enroll this incredibly fast. That's in our specialty and get that done, hopefully, within 12 months of initiation and then an NDA submission for that drug. Last but not least, Praxis' sodium channel blocker. Can you just walk us briefly through your phase II, EMBOLD design, when we should expect data? Yeah. So, Embold, we should start with where we expect data. Like yesterday, we mentioned mid-year, when we have the results of that study, it's 20 patients with SCN2A or SCN8A gain of function. We're treating them once a day. There are 16 weeks on this study. Four of those are placebo. We haven't disclosed when placebo is happening to mitigate some of the issues that we discussed before. It's going to be quite an interesting one when we actually disclose and we show everyone what happened. What's that difference? Because I think that's going to be a good answer from baseline to the placebo periods and so on. I was very excited with this because there's really nothing for these patients and opens the door to at least 5 to 10 other indications with this molecule. We have very solid preclinical data that would be moving if the market is there, the unmet need is there, and obviously, the therapeutic response is there. That's what we should be doing. So a lot going on this year. Perfect. And with that, we are at time. So thank you very much for a great discussion. Thank you. Thank you for having us. Thank you for having us.