All right. Hi, everyone. Welcome to our fireside chat with Belite Bio. My name is Jennifer Kim. I'm one of the biotech analysts here at Cantor, and I'm excited to welcome Dr. Nathan L. Mata, CSO. Thanks for being here. Thank you. Maybe to start off, can you introduce yourself and give a quick overview of the company and your lead asset team? Yeah. As you stated, I'm the Chief Scientific Officer for Belite Bio. Belite Bio is a biotech company based in San Diego, California. Our lead asset is called Tinlarebant. This is an oral once-a-day small molecule drug that's intended to reduce the accumulation of byproducts of vitamin A in the eye. These byproducts have been implicated in disease progression in Stargardt disease and geographic atrophy. We believe that by reducing the amount of vitamin A going into the eye, we can reduce the accumulation of these toxins and slow disease progression in Stargardt disease and the eye. Awesome. Nathan, you have actually had a very close role in the development of visual cycle modulators as a strategy for Stargardt and RND. Can you just walk us through some of the history that's led up to Tinlarebant's development and why mechanistically it makes sense? Yeah, that's a really good question. I think in order to really understand the context of where we are today, we have to go back to 1999, believe it or not, when we developed the first animal model for Stargardt disease. That was when I was a postdoc at UT Southwestern Medical Center in Dallas. Once we developed that model, we understood exactly what the biochemical pathology was. I was the lead in identifying these compounds in the retina and retinal pigment epithelium that were very unusual, these highly autofluorescent compounds that seemed to be associated with retinal pathology. We then embarked upon an effort to ameliorate the accumulation of these toxins by using visual cycle modulators. Now, I want to say here that we don't consider Tinlarebant a visual cycle modulator. Visual cycle modulators, that was a coin that I actually termed when I was with a different company, developing a drug called Emixustat. Emixustat is a true visual cycle modulator because it inhibits an enzyme, a very important enzyme in the visual cycle. We learned through many trials with Emixustat and other drugs like it that intervening in the visual cycle is much too aggressive for these typical indications. We went outside of the eye, and I actually developed a therapeutic approach where we target a protein called Retinol Binding Protein 4 that is this exclusive protein that delivers vitamin A from the liver to the eye. This drug, Tinlarebant, does not affect any enzyme or protein in the visual cycle. In fact, it grabs the protein and causes it to be eliminated in the urine and has an indirect effect on slowing the amount of vitamin A going into the eye. There is a distinction there. This was a culmination of roughly 25 years of both R&D and clinical development experience that brought us to Tinlarebant, which is a highly potent and very, very unique antagonist of Retinol Binding Protein 4. Awesome. Let's dig into it. You have a completed phase two trial. One question we often get asked is does it offer limited read-throughs, either because of the small trial size or disease heterogeneity? What would you say to this? Can you walk us through what you saw in the phase two and what helped get the team comfortable going into phase three? Another good question because that phase 2 study was an open-label study that enrolled 13 adolescent patients with Stargardt disease. The important feature about this particular study and the patient population was that, one, all of these patients had very severe mutations in the gene ABCA4. We expected that there would be a very rapid disease progression over two years, like the first thing. The second thing is they came in with an early stage disease. In Stargardt disease, sort of the chronology of the pathology is that lesions begin in this sort of autofluorescent nature. Basically, these are those vitamin A compounds that fluoresce in the back of the eye. These autofluorescent lesions then convert to atrophic lesions over time. In this study, we were monitoring the progression of one lesion type to another. Importantly, in those subjects that spawn these atrophic retinal lesions, which is basically dead retina, we wanted to look at the progression of those lesions in those subjects. What we found is at the end of the study, five subjects, that's 42% of the cohort, never developed any atrophic lesions. They never progressed from one lesion type to another, which is very unusual considering the genetic profile. As I said, they had very severe pathogenic mutations. That was one note that we noted that this is something special. We seem to have altered the disease progression in these subjects. The other seven subjects where we had lesion growth data, the growth rate of lesions in that cohort was significantly lower than what would be predicted by natural history. We used data from the FROG-START study, which was the largest natural history study of Stargardt disease ever conducted. By looking at that data, both the lack of conversion from one lesion type to another, and then in those subjects that converted, the slower progression rate or growth of lesions over time compared to natural history really gave us the impetus to begin on the phase 3 development. How are you thinking about, I guess, the unmet need and the opportunities on the Stargardt side and the GA side? Yeah. In Stargardt disease, obviously, there's no treatment whatsoever. This is a high unmet need. We're dealing with the adolescent patient population, even though the disease affects both adolescents and adults. In the adolescent patient population, there seems to be a higher degree of severe mutations, so their disease progression is much more rapid. In these subjects, they get diagnosed typically by the age of six or eight, and then by the age of 20, they're legally blind for the most part. This is a very important unmet need to address in these patients—kids basically going blind before they reach adulthood. This is clearly an unmet need. In geographic atrophy, although there are two injectable therapies, treatments that have been approved, there's no oral treatment for geographic atrophy. This is very important because we believe that an oral treatment, first of all, will reduce treatment burden. We think that the absence of having an invasive injection in the eye will also create a better safety profile. We think these two things are advantages over the intravitreal injections, and we believe we'll have a better safety profile overall because of it. Okay, awesome. Maybe we can dig into the phase 3 in Stargardt, the Dragon study. Maybe to start off, remind us what that study is designed to show. Right. This was actually the largest study of adolescent Stargardt disease ever conducted. The largest interventional trial enrolled 104 subjects, two-year study. Subjects were randomized two to one, Tinlarebant to placebo. The objective really here, in addition to looking at tolerability and safety, was to look at the change in lesion growth rate in treatment versus placebo. That really was a primary endpoint of looking at the difference in lesion growth rate over two years, and of course, safety and tolerability. As we'll probably get to, we're having a very, very good outcome in that trial so far. Oh, awesome. Maybe, yeah, let's get into it. Earlier this year, you had a very interesting outcome from an interim analysis. I think it was late February. Can you just walk through what that was and what were your key takeaways? Yeah, it's a little bit nuanced because I do have to explain what that interim analysis was. It was essentially a sample size re-estimation exercise. Basically, once the last person in the trial had his or her 12-month visit, that triggered the interim analysis. By that time, because of the staggered nature of enrollment, we had patients that were at 12 months, some at 16, some even later. The interim analysis actually included patients up to month 16. The objective of this sample size re-estimation was to determine whether or not we needed to add additional subjects to enhance the conditional power during the second year of study. We designed what's referred to as a promising zone. It's basically a zone of conditional power that tells us if we fell within this zone, that tells us that there's a trend for efficacy. In order to sort of enhance that treatment effect, we would add 30 additional subjects for the second year of the study. We asked the Data Safety Monitoring Board. We have an independent Data Safety Monitoring Board look at the data, the unmasked data. They looked at the data and they felt that there was no reason to add additional subjects. We don't know whether or not we were on the positive side of that promising zone, which would mean overly efficacious, or on the negative side, which would mean it's futile. They added a very important note that they said that they recommend that we seek regulatory review for drug approval. That comment would not have been made unless there were some treatment efficacy that they noted. That was corroborated, of course, by the clinical Breakthrough Therapy Designation that was awarded by the FDA recently. Yeah, that was on the back of the interim results. Absolutely, yeah. The FDA looked at the same data that the DSMB looked at. It's basically essentially a corroboration of what the DSMB saw. Can you walk us through what the feedback has been from different regulatory bodies? Based on your back and forth, what is your understanding in terms of, let's start with the FDA, what the FDA wants to see in Dragon? You know, the FDA is the most, how shall I say, unclear in terms of exactly what they're looking for because they never want to play their hand and tell you like what sort of treatment effect they're looking at, what sort of physical power that they're looking to achieve. Basically, what they say is their default response in degenerative retinal diseases like Stargardt disease and geographic atrophy is that they would require two well-controlled trials. That's basically their standard line. We've been to the FDA on multiple occasions in type C meetings and finally came away with sort of a win, we think, because what they said was that you can submit on one study. It will be a review issue to determine whether or not that one study will be sufficient. Whereas other territories outside of the U.S. have said very firmly that a single phase 3 study, at least the way ours is designed, would be sufficient for drug approval with, of course, appropriate efficacy and safety. The FDA is the only outlier. I think with the clinical Breakthrough Therapy Designation and our excellent safety profile that we received today, we think we have a really good chance at arguing for a single study approval. What other regulatory authorities have you had discussions with? Many. Japan, you know, PMDA, EMA, MHRA, the Chinese NMPA. Basically all over the world, Australia, Swissmedic, and Switzerland. The problem is that those other ex-U.S. territories are very slow, one, to grant you an appointment, a meeting, and then secondly to execute that meeting, whereas the FDA was relatively fast. Even though we proceeded with all of those efforts sort of at the same time, the FDA sort of came through first because we got an earlier review date from them, and of course the DTD from that review of the data. Okay. If the FDA is saying it might be a review issue, and depending on the totality of data, that is a package of efficacy and safety. I think you've also talked a bit about the safety you've seen in Dragon. Can you talk about that? Oh my God, I was really surprised because again, I've been doing this for many years, my whole life, really about 20 some years, and working with various types of drugs that are designed to target various proteins or enzymes to ameliorate the disease. You typically have dropout rates in the 20% in that range. Our dropout rate is not going to exceed 10%. We don't know what the shakeout is between treatment and placebo or active and placebo, but to have something that's actually 9.8% or not 9.6%, so roughly 10 subjects out of 104 withdrew. We don't know the reasons, that's a very small number when you consider, you know, these are adolescent kids taking an oral drug that does reduce vitamin A metabolism and reduce vitamin A levels in the eye. Yet the ocular AEs from the trial was less than 4%. When you have less than 4% on target AEs and less than 10% overall withdrawal, we are just a bit confused and very ecstatic about that outcome. Any theories on what drives that low discontinuation rate? Is it the lack of approved treatments or what kind of feedback? It's a couple of things. One, I think it's so critical is PI and patient education. That's one of the learnings we got from the phase 2 study. We didn't really do a lot of mitigation or management of ocular AEs. We wanted them to be reported to us and understand what they are so that we could use that information to understand how we educate patients and PIs for the phase 3 study. I think that was the biggest difference, that we really made a concerted effort to inform PIs and patients of what the anticipated ocular AEs would be. We implemented mitigation and management tools to reduce the incident severity of these AEs, basically things that they can do on a day-to-day basis to help lessen the severity of these AEs. We think that was a big difference. These patients are highly motivated vis-à-vis their parents. There's another sort of, we can't control that, of course, but these parents are obviously highly motivated to have their children see better, right? We think that there probably could be some encouragement on the parent side to tell Johnny or Susie, "Hey, look, hang in there, you know, this could be a really good thing for you." All these different things, I think, culminated to produce this great safety profile and tolerability profile we're seeing. Great. For people who are less familiar, can you just walk through what the on-target AEs look like in terms of severity? Right. Because we're reducing vitamin A content in the eye, your eye, your retina has two types of photoreceptors, which help you see. One lets you see in the daytime. Those are called cones. The other ones help you see at night. Those are called rods. They both need vitamin A. The AEs that are triggered, the one that affects cones, is called chromatopsia. When patients transition suddenly from a very dim light environment to bright light, this activates cone photoreceptors. They'll demand vitamin A. With our treatment regimen, that vitamin A will not be supplied as rapidly as in a normal context. Those cone photoreceptors will electrically misfire and produce these artificial transient cues of color in the visual field. They're not real. They're just sort of electrical manifestations. They last seconds to minutes. That's a very well tolerated AE because patients understand, "Okay, it's only going to last for a few seconds. Let me just stand here and it'll go away." They'll see yellows sort of fade away and then their vision's fine. The other one is delayed dark adaptation, which affects the rod photoreceptors. It's sort of the opposite effect with respect to the light exposure. When patients transition from a bright light environment, let's say a bright sunny day, into, let's say, a movie theater, just as many of us do, you have trouble finding your seat initially because you're accommodating to that dim light. In patients that are getting Tinlarebant, that delay is even longer because, again, the vitamin A takes time to fill up into those photoreceptors. That delayed dark adaptation can last two to three times longer than normal. Those are the two AEs, the chromatopsia, the delayed dark adaptation. If you learn how to accommodate your transitions from light to dark and dark to light, that's one of the mitigation tools. You can significantly reduce the incidence and severity of these AEs. On the other side, the hopeful promise is to slow the lesions. That's a really important point because, you know, in our GA study, for instance, we have the same ocular AEs. These patients are much older. Their retinas are much more delicate and diseased, unfortunately. They may have a harder time dealing with these AEs. We want to make sure to give them tools to help them manage those AEs, the delayed dark adaptation, which, by the way, I should mention, is part of the disease process. These patients, in many cases, are actually used to the delayed dark adaptation because it's part of their own disease that they've been dealing with for many, many years. Okay, awesome on safety. On the efficacy side, do you have an idea of what the potential regulatory paths might look like just depending on the strength of the final data? Are there scenarios or ranges that you should consider? In terms of the regulatory path, I think, you know, obviously with the FDA, I just said we're going to use that phase 3 data. We're going to file all of our drug approval applications based upon the outcome of the phase 3 data. We think that it's going to read out just like we saw in the interim. We think it's not going to change. The safety profile is not going to change. The efficacy trend, we don't believe will change. When that data finally reads out, the study is going to end. End of study is this month, later this month. We will take that data. We will submit it for drug approvals to all regulatory authorities. As I mentioned, those outside of the U.S. have expressed a greater willingness to accept a single phase 3 drug approval. We will proceed on the normal drug approval path to, you know, put the data in front of the regulators and see what they see. Yeah, I mean, from the FDA, I think we saw a recent approval for a drug based on open-label quality trials. Yeah, it's funny because, you know, all of the divisions in the FDA, they all operate differently. Even though we have one coherent set of ICH guidelines, every division applies them a little bit differently. The Division of Ophthalmology, I think, is one of the most stringent because you get away with a lot of stuff in oncology, open-label studies, short-term studies, surrogate markers, that type of stuff. The Division of Ophthalmology is much more strict on those sort of different measures. We are drawing close to the study completion. I think you've got it to 4Q25. Can you just walk through what is left between now, the last patient visit, and then top line description? Right, yeah. Those things are incomplete in phase 3. That was in August. Right now we're in the follow-up period. That follow-up period will end, as I said, probably sometime end of September. At that time, basically what will happen is that we're going to announce the end of study. Right after the end of study, we will be, over the next probably two to three months after that end of study, locking the database, analyzing the data, and getting the top line data out in a press release. That's probably what we'll do first is provide a press release both for end of study and for top line data. In terms of the format for detailed analysis, that will be presented at ophthalmology conferences, the Academy, RVO, ASRS, and the LIHAT. Those will be sort of the two different formats. At first, you'll see the press releases come out. If you take the circuit of ophthalmology conferences, then you'll see us presenting that data at the various conferences. Ultimately, I would like to prepare a manuscript to communicate these data. Is it reasonable to expect the detailed presentation first quarter 2023? Yeah, I think so, yeah. Actually, I wanted to ask this. Since AAO, I saw that the interim analysis is being presented at the subspecialty day for retina. Should we expect any new data there? Nothing more than I'm telling you here because we don't even know the treatment effect. I should mention that in order to get the Breakthrough Therapy Designation with the FDA, we did have to submit, the sponsor has to submit the application. We unmasked two members of our team who are not part of the study team. We needed that because someone on our side had to actually look at the data, write the application, and then submit it to the FDA and have the correspondence with the FDA regarding acceptance or not. That was done, right? That's something that was necessary in order to move forward with the FDA. You also have a second Stargardt study going on, Dragon II. What role will that study play for registrational purposes, and how might the results of Dragon impact any decisions around Dragon II? Another good question. Dragon II is another phase 3 study in Stargardt disease designed nearly identical to Dragon. The biggest difference was that Dragon II was initiated to take advantage of a recent designation we got from the Japanese regulatory authority, PMDA. It's called Sakigake designation. It's very similar to Breakthrough Therapy Designation in the U.S. It basically gives you a rapid path towards acceleration. What the Japanese authorities told us was that if you enroll at least 10 Japanese subjects in your phase 3 study and show that the drug behaves the same way in our Japanese patients with proper efficacy and safety, you could get approval based upon a single phase 3 study, and that would be the Dragon II study. Dragon II was designed to get approval in Japan quickly based upon the criteria that they wanted. It also has a secondary purpose because in the event that the FDA does want to see a confirmatory study, the Dragon II study would fill that purpose. The primary role really was to sort of take advantage of the Sakigake designation to get approval in Japan. It would be among the first countries to provide approval of the drug. Yeah, I wanted to ask, what are the regulatory timelines in Japan? It's really tough because, oh my God, PMDA is very difficult to work with in terms of their scheduling. They have a lot of pre-formal meetings and then formal formal meetings. It's just, it's a mess. It's very hard to predict the schedule. I think that it'll probably stretch into probably a third quarter of next year, unfortunately, but that's just the way the timelines are. In between that, we'll have, you know, UK, MHRA, EMA, everyone else sort of lined up. Even though there's a little bit of a delay, I would say overall between, say, Q2 and Q3 is when a lot of the activity I think is going to be occurring. How is progress in Dragon II going? Really good, yeah. The enrollment, I think we're at about 19, something like that. We're shooting for 60. It was a little slow to enroll initially because we were very focused on Dragon, and we didn't want to sort of cannibalize. Remember, we did a sample size re-estimation exercise. There was a potential that we may need 30 additional patients. We didn't want to basically compromise our Dragon study by putting all our patients in Dragon II. These patients are not easy to recruit. These are all global studies. We put a pause on enrollment in Dragon II until we found out whether or not we needed to add additional subjects. Once we found out we did it, we went full steam ahead and really focused on the enrollment in Dragon II. That one's going fine. Just remind us, what is the pathway to go from adolescent Stargardt to maybe the broader population? Yeah, and we'll be making that argument. I think that argument is probably going to be more relevant for the FDA. Other regulatory authorities have not even brought it up. Clearly, the disease affects both children and adults. Our rationale for getting the broader label is that it's the same disease. It's the same genetic mutation. It's a spectrum of the pathology from children to adults. The biggest difference is that, as I said before, the children seem to have more severe deleterious mutations, whereas in adults, they're more mild. This is why there's a longer delay before the disease is actually evident or manifest in the adult population. It's the same disease, right? I think we can make that argument. I think what the FDA may require is to show that the pharmacokinetic and pharmacodynamic profile of the drug is the same in children as it is in adults. We already have that data because we did it for the GA study. Okay, maybe we can move into the GA program. Perfect. You also completed enrollment in the first phase 3 trial in GA Phoenix. Can you walk through the design of that trial, what its power to show, and what you believe the efficacy and safety bars are? Yeah, okay. That study completed enrollment with 530 subjects. It is designed, again, very similar to the Dragon and Dragon II study. It is a two-year trial design where we're looking at the change in lesion growth over time, over the two-year period, safety and tolerability as well. The study is designed for an 80% power to detect a 25% change or treatment effect in lesion growth over the two-year study period. I said that I missed something. What do you believe the efficacy and safety bars are? Right. For GA, if you look at what was approved in GA, you have sort of a smattering of treatment effects. You have as low as probably 14% to 18% and maybe upwards of 20%. Those got approved, even with a sort of questionable safety profile. We think that a 20% treatment effect with an oral drug that has a better safety profile will be a winner and will be highly preferred over an injectable should we get that approval. We think that there are a lot of advantages to this oral treatment. When you consider that the Apellis drug, the Syfovre, and the Izervay are sort of the benchmarks for what the FDA would consider clinically meaningful data, I think we can achieve that bar even higher, get greater than 20%, and have a better safety profile. I apologize. I should have asked this before. How is Dragon powered? Yeah, we powered a little higher. 80% power for a 35% treatment effect over two years. The reason why we expect a little bit better treatment effect is because the bispharonoids are playing a larger role, we think, in Stargardt disease than in GA. They are clearly playing a role in geographic atrophy. In Stargardt disease, they are the sole reason for the disease. We expect to have a greater treatment effect because our MOA is directed specifically at that thing, the underlying disease pathology in Stargardt disease. That was where my line of questioning was going. For Phoenix, can you talk about the tolerability profile and the discontinuation you've seen? It's looking as anticipated. We just closed or finished enrollment, so we don't have a lot of safety data. Of course, we're collecting it as it comes in, but it's basically as expected, is what I can say. We're seeing the anticipated ocular EEs. The withdrawal rate is not anything more than what we look for. We're on track in terms of the safety and tolerability profile, the anticipated profile anyway, in the GA study. Is there a way to compare the profile you're seeing to the approved therapies? Not really, because those therapies are invasive therapies. Many of the safety outcomes have to do with the inflammatory response or some of the things that are caused by the injection itself. Whereas, of course, we don't have an injection. We're an oral therapeutic. Their safety profile is going to look a little bit different than ours. It's very hard to compare overall. At the end of the day, what people are going to look at is what is the dropout rate and what is the nature of those ocular AEs? How severe are they? Are they reversible? That's another important difference. Those AEs with the injectable therapies, I haven't seen any data of reversibility. Whereas with our drug, the ocular AEs that are manifest can be completely reversed by withdrawing the drug. Over a 30-day drug period, the drug essentially washes out and all of the ocular AEs are essentially mitigated. They go away. You can't do that with an injectable therapy. We think there's a little bit more flexibility with our drug in terms of monitoring and modulating, managing those AEs. What are your latest plans for an interim analysis? Yeah, we haven't worked out the design or timing for the interim analysis, but clearly it would be advantageous for us to have a peek at the data because, and I feel your next question will be leading to, you know, the second GA study, right? We're only doing the one GA study now, and clearly in geographic atrophy, there'll be no argument. We will have to have a second confirmatory study. We will probably want to look at the interim analysis of the ongoing GA study to determine whether or not it's sort of a go/no-go decision, right? Whether or not we initiate, because it's a lot of money to initiate another confirmatory study. It'll probably be designed like we did in Dragon, where it'll be a sample size re-estimation to determine whether or not we add additional patients if we see a trend for efficacy. Again, that falls within this sort of promising zone. Would you anticipate a similar framework for the interim analysis? Would there be a futility outcome or how are you thinking about that? Yeah, another good question because, you know, for Stargardt disease, especially in the adolescent patient population, the FDA highly recommended that we not stop the study either for futility or efficacy. That may not be the same for GA because, again, there are approved therapies. This is not like a sort of an unknown disease or mechanism. It may be a little bit different than in Stargardt disease. We'll just have to wait and see. When can we expect more details on plans for the interim? Oh, yeah, we'll have to finalize those probably by the end of the year. End of the year? Yeah. How should one think about potential pricing between these two indications because the market size is so different? Yeah, so for those who are not aware, Stargardt disease has roughly, you know, 50,000 to 60,000 patients in the U.S. Our study in Stargardt disease is going to be ahead of the GA by about four years. If we get approval in Stargardt disease, we'll have premium pricing in Stargardt disease for at least four years. We're going to look at the precedent data, the markers for orphan drugs and rare drugs and what their pricings are. It's going to lie somewhere between $100,000 to $300,000, I think, per year. Of course, once, if and when GA gets approved, that number, that pricing will have to come down considerably to match, you know, sort of what the comparables are for the Ixivir and the Siphovri treatment. Okay. What work have you done or are you doing now to prepare for eventual commercialization and rolling? That's an ongoing effort right now, getting the financing going, searching for marketing and salespeople to do the commercialization. We think that at the end of the day, we'll be able to do the U.S. commercializations ourselves. Of course, for ex-U.S., we'll likely have a partner, a big pharma partner, because that's a much more daunting effort to go ex-U.S. Within the U.S., because it is a relatively small patient population, 50,000, 60,000 subjects, patients, we think that we can handle that ourselves. That's sort of the plan, that we would do U.S. ourselves, ex-U.S. we would partner out. For GA, we'll have to wait to see what comes once the data come out. How quickly after the phase 3 data do you think you could file? You mean with... For the U.S., yeah, for Dragon. We're going to file probably before, you know, Q1, Q2 of next year. That NDA filing will happen certainly before Q2, but it could be as early as Q1. That would be sort of the window of time. Yeah, we're full steam ahead on preparing our NDA. Awesome. I think we're at the end of our time. Maybe one quick question. Can you just walk us through your balance sheet and runway and key catalysts? Yeah, we have a great CFO. He does a wonderful job of making sure we have a steady and lengthy cash runway. Right now, our cash runway is about four years. We have all the money we need to complete all the studies we have ongoing. We haven't gotten money for the, if we need a second GA study. For everything we have ongoing right now, we are fine. In terms of going forward, obviously we're going to be out doing some fundraising later, but there's no need for it right now. Certainly, as the commercialization efforts ramp up, we'll need to get more financing coming in to fund those up. The key catalyst, I think it's the data coming up soon and then... Right, right. So close, so announcement of end of study for Dragon. And then, of course, top line data for Dragon as well as the data. Then perhaps after that, close of enrollment for Dragon II. I think within probably 18 months, we may actually pull the trigger or be in the window of the interim analysis for the Phoenix study, the GA study. Those are the milestones for up to 18 months from now. Awesome. Looking forward to it. Thank you, Nathan. Thank you, Jennifer. Thanks, everyone. Thank you all. Thanks, John. Thanks, John. Thanks, John. Okay, That was amazing. Hello everyone, and welcome to the Cantor Global Healthcare Conference. I'm Samantha Schafer. I'm with the Cantor Biotech team. Today we have with us Belite Bio, a company that Pete Stavropoulos covers, and I'm pleased to introduce Dan O'Connell, the CEO, and Jim Doherty, the Chief Development Officer. Let's get started off with an introduction of yourselves, followed by a description of the company, and why and how Belite Bio was founded. Yeah, great. Thanks, Samantha. Thanks to you and the team at Cantor for having us here. I'm Dr. Nathan L. Mata. I'm the Chief Scientific Officer of Belite Bio. I've been in the life sciences industry for over 25 years as an early-stage investor, and more recently as the Chief Scientific Officer of Belite Bio. I've been focused on working with neuroscience companies for an extended period of time, principally on the business and leadership side, but excited to share some updates on Belite Bio today. Fantastic. I'm Jim Doherty, the Chief Development Officer for Acunin. I'm a neuroscientist by training. I've worked in pharmaceutical R&D for my whole career, 25+ years now, in both large companies and small, and been with Acunin for a little over a year and a half now. Perfect. Thanks, Jim. Let's just jump into it. Your lead compound is Tinlarebant. It's an anti-amyloid beta antibody with a unique selectivity profile. Can you take us through the history of the molecule, how it was developed, and how it came into use? Sure. Soburnitug is a humanized monoclonal antibody that has high selectivity for a soluble A-beta aggregate, which is principally referred to as oligomers. Soburnitug actually goes back to the founding origins of Acunin, which was started as an academic spin-off from Northwestern and USC, where some of the scientific founders of the company had been investigating and interrogating the toxic properties associated with these oligomers or ADDLs at the time that they were working on it. That early phase research and discovery effort turned into a research and development collaboration between Acunin and Merck. In fact, Merck was the primary lead on inventing Soburnitug, which previously was referred to as AC193. The unique properties of Soburnitug are really based on its selectivity profile, that it has a high affinity and preference for soluble aggregates that have been shown to be toxic, which is different than amyloid A-beta monomers or A-beta plaques in most lives. Over a course of efforts internally to Merck, they made a strategic decision to advance a different Alzheimer's candidate in the 2010 timeframe, and Acunin was able to reacquire the entire data package and worldwide exclusive rights. Soburnitug, as it sits here advancing through a sizable phase II study, and what we'll talk about is a wholly owned asset of Acunin. We're really excited about some of the phase I data that was generated a couple of years back, and now how quickly the phase II has been executed. When we hear about amyloid beta antibodies for Alzheimer's disease, we think of lock reducers of the approved drugs like aducanumab and donanemab. However, the therapeutic hypothesis for Soburnitug is quite differentiated from aducanumab and donanemab. Could you give us a synopsis on the oligomeric hypothesis of Soburnitug? Sure. The oligomer hypothesis is one that holds these soluble aggregates of the A-beta peptide. If A-beta is cleaved in a couple of different places and can produce a peptide of varying lengths, A-beta 42 being the one that's most prone to aggregation and so forth, oligomers are a consequence of one or more of those A-beta peptides coming together. The underpinnings of the oligomer hypothesis hold that these oligomers, as evidenced in labs, founding labs and labs elsewhere, have shown that these aggregates have a propensity to bind to neurons at synapses and cause deleterious and disruptive effects, including calcium influx into cells, disruption of long-term potentiation, which is a surrogate for cognitive or memory processes, and hippocampal slices, as well as induced tau hyperphosphorylation. The strongest linkage between A-beta and tau, we believe, is mediated through oligomer toxicity. Could you go into some of the specific molecular properties about affinity and selectivity for the A-beta oligomers over A-beta monomers? Sure. At a minimum, we really don't, we have high selectivity versus monomer, which is really the important element to maintain functional target selectivity and potency of the antibody. Monomer is orders of magnitude more abundant. The extent that your peer agent, your antibody, or other agent interferes with monomers, we sort of characterize that as target distraction. Right? I think our selectivity for oligomers over monomer is 10,000-fold. It's very, very, very high in various different experimental setups. For fibrils or plaques, it's more in the order of 90-fold selective for oligomers over those other fibril/plaques aggregates. That really is, you know, we're testing the oligomer hypothesis with Soburnitug. It was originated as an oligomer-directed antibody, and we had great phase I results, which we talk a little bit more about. That has really helped facilitate the speed and rapidity of the enrollment. We will jump into those data points. Very quickly, what were some of the key scientific data that enhanced your conviction in the oligomeric hypothesis? It's 30 years of research that really continued to call out and validate the toxicity associated with these different species. There are a broad spectrum of toxic species. I think it's to be determined which in particular confirmations are the quote-unquote "most toxic of the toxic." These are things I think we're very confident in Soburnitug's profile for hitting a broad spectrum of aggregates from a few of the dimers, like low molecular weight to mid-molecular weight oligomers. Having that diversity of mechanism and coverage, we think, is likely to be beneficial in a clinic, which yet to be elucidated precisely, which if there is one or more, I mean, at least from taking view is unlikely, it's probably a variety that function in numerous pathologics. Maybe just to add a little bit to what Dan is saying, obviously what's so successful about monoclonal antibodies as medicines is their specificity. You'll often get people talking about amyloid targeting. Not that that's wrong, but it's probably way too broad because, as Dan's been explaining, amyloid biology, these proteins are very sticky. They tend to glob together, and they have multiple, multiple forms. Different antibodies are recognizing individual forms. To answer your question around the oligomers, what caught my eye in the data, and as Dan was saying, there's 25+ years' worth of studies that are out there. It's a lot to summarize, but I would say there's both binding to synapses. Now you're talking about the sort of business units of cortical function, and these toxic oligomers are actually intercalating into the synapses and disrupting synaptic function. You're seeing consequences when it comes to the functional measures. Things like synaptic plasticity is no longer as efficient, and even up through into the modeling, there's disruption of ongoing cortical activity. I think that's really what is the differentiating piece for the hypothesis, because by targeting those forms specifically, we think we're going to have a differentiated effect on the pathophysiology. It's very fascinating. Okay, let's jump into some data. In July 2023 at AAIC, we had the first look at the P1B program for Soburnitug in early Alzheimer's patients, and then some patient-level data at CTAD a couple months later. Could you give us a brief overview of the study design, including the dosing levels of the drug? Sure. You'll take us through some of that. For the intercept study, it is a phase I study, but it was a phase I study conducted in Alzheimer's patients. The population that we're studying with Soburnitug is early Alzheimer's disease, either MCI or mild dementia. That's measured through a number of different clinical scales. The important point is, even in the phase I study, we're looking at patients, and that really gives you the opportunity to do a number of things. In addition to the tolerability and pharmacokinetics, which stand for phase I, the team was able to include a number of metrics that look at function along the way. From a pharmacokinetic perspective, the drug is well behaved and consistent with once-monthly dosing. That's why we continued with that into the later study. From a safety perspective first, generally well tolerated. The issue for any amyloid-related therapeutic is this ARIA and ARIA E specifically. We did see a few cases of ARIA E in the phase I study, a total of five across all the dose groups. Multiple dose groups, I'll just talk about the MAD cohorts, a low dose of 10.5 and 50 milligrams per kilogram, sorry, 60 milligrams per kilogram in phase I. A good dose range, and we saw a total of five cases of ARIA, which works out to about 10%. Although it's a small sample size, that compares nicely with the available data sets for the marketed drugs. It's worth talking about the efficacy side of things, since we were talking a bit about the mechanism of action. A number of the measures were designed to really give a feel for the effect that Soburnitug may be having on amyloid levels. We did look at amyloid plaque distribution by PET scans. Although that's not the primary effect for Soburnitug, we did see a nice reduction in plaque size over the three-month dosing period for the MAD cohort, about 20% reduction at the top dose, which is nicely in line with what was seen at a similar time point in the clinical program for lecanemab. What's probably more interesting to us is we looked at a number of soluble biomarkers, and those are biomarkers directly of disease processes. Looking at A-beta 42-40 ratios, sort of the next step downstream, looking at tau hyperphosphorylation, but then also going to the next level of integration and looking at synaptic markers. The consistent story is that you see a reversal of the aberrant levels of these markers, which is exactly what you predict if you're having a beneficial effect on amyloid levels. We're very encouraged by this data, and it really is a nice launching pad to take us into the current outside data usage. The only point I'd add is the target engagement, phase I too. I mean, the safety, the biomarker activity was robust, and target engagement was really, I mean, given that we're targeting oligomers, there had not, you know, uniform been established oligomer assays. The team developed a novel ELISA-based method of confirming target engagement of Soburnitug bound to oligomers as an omblex. We see those related effects, and we also see sort of maximal target engagement at the 25 mg/kg and 60 mg/kg in that MAD dose cohort. That was really another element to the totality of the phase I data that helped inform and underpin the phase II exchange. Definitely. I Definitely. I do want to spend a minute on ARIA. You know, it's a concern of the accrued lecanemab and donanumab. There, as you pointed out, Jim, there were some observations of ARIA, but it was at a much lower rate than lecanemab and donanumab. Do you think that the lower rates were due to lower affinity for certain amyloid species, or possibly driven by the fact that Soburnitug is an IgG2 molecule with reduced effector function? I think it's both possibilities. I can tell you a little bit about, yeah, the IgG2 isotype was a development decision actually taken back in the Merck phase, and that was intentional to sort of minimize microglial effective function and response to the inflammatory elements. I do think that given the preference for oligomers over plaques, it's equally likely that Soburnitug doesn't see as much plaque as some of these other plaque-directed agents. It's also important to establish that, with this dose range in phase I, we're delivering a lot more Soburnitug drug relative to the other plaque-directed agents that have an ARIA ceiling. In fact, when the phase I was designed, that 60 mg/kg SAD cohort was very much intentional, given that at the time, aducanumab was being developed in phase I. Their SAD study had a 60 mg/kg dose cohort, and three out of eight of the subjects that were dosed at that level experienced symptomatic ARIA hospitalization. They sort of can't push the dose higher for those other agents. As we found, particularly with clinically delivered antibodies, the higher dose equals greater central exposure. This dosing window that we have for Soburnitug is safe and demonstrates our engagement. We think it's a valuable thought. Do you think these ARIA observations affect your view of the oligomer hypothesis, or do you need to wait for some functional data? In terms of safety overall, I think there are other non-A-beta agents that have elicited ARIA, right, come to and so forth. I think the biology is probably pretty complex. It's hard to pin down precisely how these mechanisms play. As Jim mentioned, we did move the PET signal 20-25% at the higher dose level. There could be some elements of that, although there's no correlation between the plaque reduction and the ARIA, which is another question to be answered in time. No, it's a puzzle. Okay, moving on to the P2 Altitude study. In March, you announced enrollment completion in the study with a lead out of 4Q26. Could you talk about the P2 design and how the Intercept data informed study design? Sure. Maybe I'll start, and Jim, you can comment. The phase II is a fairly well-powered phase II in Alzheimer's patients. This is the early AD population we employed in phase I, and this is kind of the standard population. This is mild cognitive impairment, and then dementia is the early Alzheimer's disease. The study was designed with two active dose levels, one at 35 mg/kg dosing every four weeks before W, and then a higher dose at 150 mg/kg dosing before W. It's an 18-month study, which is standard for primary clinical outcomes. We're using the IDRS as that primary outcome measure, and then we'll have CDR sample boxes, 8 assays, elements of the IDRS, as well as all of the fluid and imaging biomarkers, both the amyloid patch as well as a substudy for the study. We did move quickly on the study. We have a first patient dose that was announced in May 2024. We thought we might have to, we exceeded our expectations in terms of our cohort. We enrolled 542 subjects, so roughly 180 per cohort in roughly 10 months. As a small biopharma, kind of in a space that's traditionally kind of the larger players' base, I think we're really encouraged to do a remarkable job. I think it's a combination of the quality of the team, the ability to execute, as well as the value proposition of the suburban research based on phase I data. It's probably worth saying, as Dan said, the study enrolled in 10 months, so there was obviously a lot of interest in the study. It's certainly gratifying to us because the study is being conducted in the U.S., in Canada, in Germany, and in Spain. It's across a number of different countries. What I would say is that the level of interest that we saw was matched pretty much in all the geographies, which is great. There are certainly approved options out there now. That was a question that we often got, was that going to be an issue for us? Clearly it wasn't. We were able to enroll 582 subjects in less than a year. How about baseline characteristics? Do you see this cohort as being similar to the lecanemab and donanumab populations? The lecanemab and donanumab populations are actually a little bit different. If you look at Clarity and Trailblazer II, because Lilly employed the tau measure as part of their stratification in that study, those populations seem to be a little bit more advanced than Clarity AD. I think Clarity AD is one that is pretty consistent with what we saw in phase I and what we would anticipate for phase II. You mentioned, Dan, the IDRS as the primary efficacy. How is this scale different from CDR sample boxes? Is this a strategic choice? It is a strategic choice, and it's one based on statistics. The IDRS is a composite scale that takes elements of the ADAS-Cog and the ACDS ADL, or activities of daily living measure. It's basically been established as a more sensitive measure for this early ADL population. It was developed at Eli Lilly. Our Chief Medical Officer is a former medical lead on the Solanezumab program at Eli Lilly, so he's very familiar with the measure. If you look at donanemab, which is Lilly's drug, in the phase II that they ran, Trailblazer primary, they had IDRS and CDR-SB. They hit on the IDRS. They didn't hit on CDR-SB. In the phase III, they maintained the IDRS as the primary. They hit on both the IDRS and the CDR when you do a larger study. We think that just from the basis of reported outcomes and statistics, the IDRS is an appropriate one for the stage that we're in. I think it's the right choice. The doses you've chosen for Altitude are 50 mg per kg and 35 mg per kg? Correct. Every four weeks. That's IV dosing every four weeks. Yes. For that low dose, 35 mg per kg, what are your expectations in terms of ARIA and plaque reduction? I think in the phase I, we had a total of five cases. Three were at 16 mg/kg. We had one each at 10 and 25. I do think that there's the potential for the 35 to differentiate from the 50 mg/kg dose cohort on ARIA. We'll have to see the data tell. I think for the dose sort of strategy in phase II, we employed a good degree of dose modeling and leveraging a lot of that target engagement data to support the notion that at these levels, which you have even at trough levels of drug, adequate coverage of target, such that we're essentially the continued data on optimized treatment effect. For patient screening, you implemented a very unique strategy. It's a two-step process that uses plasma pTau217 biomarker testing. Could you describe this process and the outcomes? Yeah, absolutely. As you know, there's been a tremendous explosion in the use and expansion of plasma-based biomarkers. Eric and the team did a great job in recognizing that, and in particular, pTau217, which has turned out to be quite a sensitive marker for amyloid levels, interestingly, in AD patients. One of the challenges historically in Alzheimer's clinical trials has been, if you use clinical diagnoses alone to patients, what people found, and I'm sure everybody knows these data quite well, is that a number of subjects that were enrolled in earlier phase Alzheimer's trials probably didn't have Alzheimer's disease. They certainly had dementia, but they had dementia for other reasons, and the amyloid levels were showing that those folks were inappropriate. That's the last thing you want to do in a trial, have a significant population of folks who are not appropriate for the therapy. Certainly bringing in PET as a screening pool was beneficial. PET also has its own limitations. There is radiation exposure. It's much more time-consuming, invasive, and all those sorts of things. What you can see with the plasma-based biomarkers is they offer the opportunity for much simpler, much more rapid, and frankly, much more cost-effective screening. Eric and the team put effectively a prescreen in place where, during the first screening visit, subjects were tested for pTau217 levels. We established an internal cutoff, and you only moved forward to later screening if you had a pTau level above that screening cutoff. The effect of that was that in our phase I study, consistent with the earlier studies, a substantial fraction of subjects who went for a PET scan turned out to be amyloid negative. That changed quite a bit in the presence of the pTau217 screen. You go from double-digit percentages without the pTau217 screen to a 15% to 16% failure rate in the presence of pTau. That had major consequences for efficiency in the study. It means that many more of the patients who are going for the final screen of either PET or CSF markers were positive for the study. It saved money, it streamlined, and it did a number of positive things. It really just is an interesting highlight for, I think, what will be a much larger use for plasma-based biomarkers in the future. Yes. You entered into a license and collaboration agreement with Halozyme to develop a sub-Q formulation with its enhanced technology. Thank you. What is the overall strategy and development plan for this program, and when can we expect some updates? We think there's a lot of value and flexibility for patients for having a sub-Q formulation for Soburnitug. As you say, we have the data from the phase I study, which is our initial look at the pharmacokinetics of delivering with the enhanced technology. We use those data to plan for the future. We were able to generate the PK data we wanted to be able to see to continue forward. I think at this point, it really comes down to something we've already been talking about, which is what's the appropriate dose level of Soburnitug to target for a subcutaneous formulation. Given how much work there is involved in delivery devices and that part of the process, we don't want to move forward with next steps until we have a better sense of the appropriate dose to be targeting for subcutaneous. Having said that, the team is very much already working on plans and what the pathway would look like and what devices we would want to qualify for the purposes. We think it's going to be an important part of the program, but it's tied into everything. Do you think you could possibly switch a proportion of patients from the altitude study to a sub-Q version? is much more likely that we would do something like that as part of a phase 3 confirmatory study than the current active study. Since we enrolled so quickly, the study's moving very quickly, and we really want to get those data on the doses at 35 and 50 before they just move. Is there anything else you'd like to mention about Phase II before we move on to? No, we've got it to top line results late 2026. Yes. The study's going really well. Retention is good. We're really encouraged with the progress of the study. The last thing we want to touch on is something that we're really excited about, a recent agreement you entered into with JCR Pharmaceuticals for their blood-brain barrier crossing technology. This is a technology we believe will be very disruptive to the neuroscience field for CNS drug discovery. Can you tell us about this tech and how you plan to leverage it for Soburnitug and possibly any other could be placed? Sure. We agree with you that this technology is really fascinating and potentially very, very interesting for neuroscience. As fantastic as monoclonal antibodies have been in the treatment of disease, one of the key challenges for brain disorders is that only a very small fraction of the systemic dose actually gets into the brain. For context, the general number that gets tossed around is about 0.1% of the systemic dose actually gets into the brain. Anything that one can do to meaningfully increase the fraction that gets into the brain, you would predict is going to have pretty meaningful effects on programs overall. With that kind of thinking, we started doing a scan of the landscape, looking at the various technologies that are out there. In addition, we've chosen, as you say, to work with JCR around their ISCARGO transferrin-based system. There are other systems out there that people are interested in. For us, the decision was relatively straightforward. There is the most available clinical validation for transferrin as the appropriate carrier into the brain. More than that, JCR has a marketed drug that uses this technology for Hunter's syndrome. For us, that was very attractive as a clinical level of validation for moving forward with this. Your technology in collaboration with JCR is called the Enhanced Brain Delivery Tech. Yes, we've characterized our terminology. Our terminology Our terminology is Enhanced Brain Delivery, or EBD, and we, as Jim just described, we're using the JCR carrier technology with Acumen proprietary Cargo Soburnitug and Soburnitug-like monoclonals that again have kind of a high selectivity towards our favorite target being oligomers. The focus is going to be oligomers. Yes, it will continue to be oligomers. We do like the way the arrangement works with JCR. We do have the ability to advance as many as two candidates under the option term. All right, in the very last minute, our last question is, you know, if we're sitting here 12 months from now, what are the accomplishments that you'd love to tell me about from this past year? Sure, as I look out over the next 12 months, it's going to be an amazing run here. We'll have preclinical candidates out of the EBD program early 2026. I think characterizing kind of that strategy, as we do, again, agree with your assessment that that technology long-term is going to be really compelling in the space. In 12 months, in September 2026, we'll be working on final patient visits and other elements of closing out the study and locking the database and getting very close to data. It'll be an exciting time. Amazing. Thank you so much, Belite Bio team, for joining our Cantor Healthcare Conference, and thank you to our audience. Thank you, Samantha. Thank you, Matt.