Hello, and welcome to the Deutsche Bank Virtual Investor Conference, dbVIC. This is Zafar Aziz from the Deutsche Bank team. I'm pleased to welcome our next presentation from Belite Bio, whose ADRs are listed on NASDAQ under the symbol BLTE. The presentations today will last for around 30 minutes, and all today's presentations will be recorded and can be accessed via the Deutsche Bank website, www.adr.db.com. I'm very pleased to welcome Nathan Mata, Chief Scientific Officer from Belite Bio, who is speaking to us today from San Diego. Over to you, Nathan.
Thank you very much for the introduction. At Belite Bio, we are currently advancing through phase III development, an oral once-a-day therapeutic intended to treat two different but somewhat related diseases: Stargardt disease, which is a juvenile-inherited macular degeneration, and age-related macular degeneration, which most people know about, which is a disease that affects the elderly. Our drug, Tinlarebant, is intended to reduce the accumulation of toxic byproducts of vitamin A, which have been implicated in disease progression in these two diseases. By way of introduction, here's our leadership team, headed by Tom Lin, our CEO, Hendrik Scholl, our Chief Medical Officer, who's an expert in both Stargardt disease and geographic atrophy in AMD, myself and the CSO, as mentioned, and our CFO, Hao-Yuan Chuang.
If we look at the pipeline overview, you can see here in Stargardt disease, we've completed a two-year phase II open-label study that continued to show a slowing of lesion growth. I'll show you some of that data as we move forward. More importantly, we're currently focused on our phase III program in Stargardt disease, which is called DRAGON. This study has been completed. The database is being locked. The study enrolled 104 subjects aged 12 to 20 years of age, and we expect the top-line data from this study in the next three to four weeks. Another study, similar to the DRAGON study called DRAGON II, is also currently recruiting. This study is recruiting 60 subjects aged 12 to 20 and is designed very, very similarly to the DRAGON study.
So both these studies are phase III studies ongoing, investigating the safety and efficacy of Tinlarebant in the treatment of Stargardt disease. In geographic atrophy, we also have a two-year phase III study. This study is known as PHOENIX. This study is ongoing. It completed enrollment at 530 subjects. So a little bit about the drug. As I mentioned, our drug, Tinlarebant, is a novel oral once-a-day tablet. It's designed to bind a protein called retinol binding protein 4. This protein is the sole vehicle for delivery of vitamin A from the liver to the eye, and it's intended, as I mentioned earlier, to slow the accumulation of toxic byproducts of vitamin A that have been implicated in disease progression in both Stargardt disease and GA.
At Belite Bio, we believe that early intervention directed at emerging retinal pathology that is not mediated by inflammation would be the best approach to potentially slow or alter disease progression in Stargardt disease and geographic atrophy. In terms of the market opportunity, it's important to note that there are no FDA-approved treatments for Stargardt disease and no orally approved treatments for geographic atrophy. We have various designations associated with Tinlarebant, including Breakthrough Therapy Designation, which was recently granted based upon the interim data from our DRAGON study, and I'll talk more about that as we move forward. We have Fast Track and Rare Pediatric Disease Designation in the U.S. and Orphan Drug status in both the U.S. and EU. And finally, I should mention that we do have another designation that was offered by the Japanese regulatory authorities that's called Pioneer Drug Designation for Stargardt.
This is very similar to Breakthrough Therapy Designation that the FDA grants. In terms of our patent protection, we have 14 active patent families. The composition of matter patent is not expected to expire until 2040. That's without patent term extension, and these are all very robust coverages for Tinlarebant across various geographies globally. In terms of the market opportunity, Stargardt disease is an orphan disease, but it's the most common inherited retinal dystrophy among those diseases in about 1 in 8,800 people. The breakdown for the population in the U.S. is roughly about 53,000, and then you have about twice that amount in China. We mention China because we did recruit pretty heavily from DRAGON in China, so there's a very large population of Stargardt disease in China.
This is a really excellent market opportunity when you consider that, as I said before, it is the most prevalent inherited retinal dystrophy among those. If you look to the right, you see the prevalence of advanced AMD. Most people know that AMD, of course, increases in prevalence with age, and our focus really is the advanced form of dry AMD, which is known as geographic atrophy. There's roughly about a million patients in America with geographic atrophy. So a little bit about the disease biology. I mentioned that our drug, Tinlarebant, is intended to slow the accumulation of toxic byproducts of vitamin A. They're implicated in disease progression in Stargardt disease and geographic atrophy. I want to show you here how those molecules are actually derived and accumulate within the eye. This is a schematic of the back of the eye.
It's a visual cycle, and it shows three different layers. On the very bottom, you have the bloodstream. In the middle, you have this layer of tissue, which is called the retinal pigment epithelium, and right above that are your photoreceptors. So the visual cycle is initiated by the introduction of vitamin A to the basal surface of the retinal pigment epithelium, and it's presented as a complex you'll see in the lower right-hand corner, comprised of retinol binding protein 4, at-ROL, that's the chemical abbreviation for vitamin A, and this larger protein called transthyretin. So this is the complex that gets liberated from the liver into the circulation, and because of its large molecular size, it resists filtration in the kidney.
So this is a mechanism whereby you can maintain a high steady-state level of vitamin A in the blood because, again, this complex is very large and resists filtration in the kidney. But what's important to note in terms of delivery of vitamin A to the eye is that only the eye requires delivery of vitamin A bound to retinol binding protein 4 because it uniquely expresses a retinol binding protein 4 receptor that's not present abundantly in other extrahepatic target tissues. Consequently, the eye has reliance for uptake of vitamin A when it's delivered to RBP4. Other tissues do not have that reliance. So when this complex docks onto the receptor, the retinal is admitted into the back of the eye, and it goes through a series of enzymatic reactions where it's eventually converted to rhodopsin within the retina.
You see that at the very top of the screen. Light activation of rhodopsin liberates the aldehyde form of vitamin A, which is known as all-trans-retinal. This is a very toxic form of aldehyde, of vitamin A rather, and it must get out of the retina through some mechanism or actually start poisoning and damaging retinal membranes. So there's a biological mechanism in place to remove that aldehyde, and it's mediated by a protein called ABCA4. ABCA4 is essentially an enzymatic flippase, which grabs the aldehyde from inside the retina and flips it outward, availing it to another enzyme for further detoxification back to alcohol and re-entry back into the visual cycle. So this is the normal processing of vitamin A in a healthy, unaffected eye. But in patients with Stargardt disease, there are genetic mutations that affect the function of the ABCA4 protein.
Consequently, the aldehyde is not efficiently removed from the retina, and it will linger and accumulate within the retina where it condenses upon itself, forming these dimers of vitamin A that are known as bisretinoids. These are the toxic byproducts that I spoke about earlier. The most abundant bisretinoid that's been identified in human tissue is known as A2E, and this molecule has been shown to kill retinal tissue through diverse mechanisms. So in Stargardt disease, it's generally well accepted in the scientific and clinical community that the reason for initial pathology and eventual blindness in Stargardt disease is because of the accumulation of A2E and related bisretinoid molecules. So what does Tinlarebant do? Well, what Tinlarebant does is work in the liver to compete with retinol for binding to RBP4, and it doesn't allow the subsequent binding of that larger protein, transthyretin.
Consequently, what gets liberated into the circulation is a complex of RBP4 with our drug bound to it. And because this complex is so small, it gets readily filtered through the kidney, that is, through renal filtration. So the net effect of our drug really is to reduce the concentration of retinol binding protein 4 in the circulation. And by doing that, we reduce the amount of this complex that is normally used to deliver vitamin A to the eye. As I mentioned, the eye requires delivery of vitamin A bound to retinol binding protein 4. So if we reduce retinol binding protein 4, we're going to reduce the amount of vitamin A going into the eye. And of course, since vitamin A is the substrate for the bisretinoids, ultimately, we will reduce the accumulation of those compounds over time.
An important aspect about these bisretinoid compounds is that they actually fluoresce. That is, they emit light, and that's because of their vitamin A composition. An ophthalmologist can actually follow these compounds over time by using a specialized imaging camera called the fundus autofluorescence camera. Here you're seeing two case studies of disease progression in Stargardt disease and GA. The Stargardt disease patient is shown on the top. The GA patient is shown on the bottom, and you're looking at serial retinal images taken over about four and a half years. If we start with the baseline image in the Stargardt disease patient in the upper left-hand corner, you see these two central areas of blackness. These two black blotches that you see in the center of the image are dead retina. That's atrophic retina that is basically never returning. It's dead.
But if you look peripheral to adjacent to those lesions, you see on all sides, there's this border of autofluorescence that encompasses those lesions. And now, over time, you can see annually the lesions, the black areas spread into the autofluorescent zone, and the autofluorescent zone continues to expand outward in a centrifugal manner to accommodate that dead retina. So what this is telling us is that wherever you see the autofluorescence, that is, the bisretinoids, you will soon see retinal pathology. And it stands to reason that if you can get rid of those bisretinoids, you will slow or even halt the progression of this pathology and therefore preserve vision. If we look now to the patient with geographic atrophy on the bottom left, you see this baseline image with a small central area of atrophy. It's dead center in the middle of the image.
But peripheral to that blackness, you see these little punctate areas of autofluorescence, sort of like satellites circulating a planet. They're a little bit more obvious at the 12-month mark. You see these little punctate areas of autofluorescence. Those are the bisretinoids. And again, as you follow over time through 24 months, 36 months, and eventually 55 months, you see that those areas that were previously just autofluorescent have slowly become atrophic retinal lesions. And so by the time they get to 55 months, all of those autofluorescent areas have converted to atrophy, and they all have little zones of autofluorescence around their perimeter. So once again, wherever you see the autofluorescence, you will soon see retinal disease geographic atrophy.
I should point out that there is a proof of concept study that was conducted to demonstrate that reduction of retinal delivery to the eye would, in fact, have a slowing of lesion growth. This study was not done with Tinlarebant, but rather a different drug called Fenretinide. And this study was conducted approximately 15 years ago by myself when I was a CSO at a different company. This drug, Fenretinide, was actually developed as an anti-cancer drug but had the side effect of reducing retinol binding protein 4 because of its chemical structure. So I repurposed Fenretinide into a phase II proof of concept study enrolling 246 patients with geographic atrophy to evaluate the hypothesis that reducing retinal delivery to the eye would have an effect on slowing lesion growth.
Here on the right-hand side, you're seeing a histogram showing you the extent of lesion growth over the two-year study in patients in our study, so the black bars show you the extent of growth expressed as a percent increase from baseline in the placebo group, so there's about a 50% increase in the size of retinal lesions in the placebo group without treatment. The two green bars show you the data from our high-dose Fenretinide group at 300 milligram. We saw something very interesting in this dose cohort. There was a group of patients that was very responsive to treatment and had at least a 70% reduction or more of retinol binding protein 4. In those subjects, we saw a reduction in lesion growth that began at month 12 and persisted through months 18 and months 24.
In that same dose cohort, subjects that did not attain that profound reduction of RBP4, that is, it was less than 70% reduction, there was no change in the lesion growth whatsoever. It looked just like placebo. So what these data are telling us is that there is a requisite level of reduction of RBP4 that you must achieve systemically before you can start affecting some change on lesion growth within the eye, and that number appears to be 70%. Another important finding of this study is that during the same time period when the lesion growth was slowed, that is, from 12 months to 24 months, there was also a stabilization of visual acuity. So by month 25, if you look at the lower right-hand corner here, you see this graph that plots visual acuity change from baseline.
Up to 12 months, all subjects are losing about the same letters, about anywhere from four to six letters. But during the second year, from 12 months to 24, the responding group was static at six letters. That is, they didn't lose any more letters. Meanwhile, the placebo group and the non-responding 300 milligram cohort continued to lose letters. So by the end of 24 months, we have a pretty good visual acuity benefit for these responding subjects and a profound lesion growth reduction. The problem with this study was that only one in every three subjects had this very robust response to the drug, and that was owing largely to two deficits of the Fenretinide drug. One was its bioavailability. It has terrible bioavailability, and it had to be given with a high-fat meal in order to increase exposure in blood.
Many patients complied with that guidance during the first year, but during the second year, that compliance fell off, and we could tell that because we could monitor retinol binding protein 4 in the blood and see inflection points upward. The second problem is its potency. The drug is only as equally as potent as the native ligand, vitamin A, because its ability to act as a retinol binding protein 4 antagonist is based upon its structural similarity to vitamin A. So it has the same affinity for the target as the native ligand. Therefore, it's continually competing with the native ligand for occupancy on the binding site. With Tinlarebant, we've overcome those deficits. First of all, Tinlarebant is a purpose-built retinol binding protein 4 antagonist. It's not repurposed.
It was designed with 3D in silico chemistry just to attach to the residues on the retinol binding protein 4 target site. Secondly, it has great bioavailability, so it can be taken without regard to food. And finally, it has extremely good potency. It's a hundredfold more potent than it is compared to Fenretinide. So with these attributes, we feel confident going back into geographic atrophy with this purpose-built drug and believe that we will get a response in all subjects, not just a handful, because these deficit features have been addressed. Moving forward now into Stargardt disease, our clinical trial data. Our first clinical trial data to test safety and efficacy of Tinlarebant was realized or evaluated in a two-year phase II trial enrolling 13 adolescent subjects with clinically and molecularly confirmed Stargardt disease.
I should mention that the genotyping from these subjects showed that they all harbored very severe pathogenic mutations that would predict pathology over a two-year period. The important thing about these subjects is they came in with an early stage of disease where they only have the autofluorescent lesions, which are referred to as QDAF. These lesions have not yet converted to atrophic lesions, which are referred to as DDAF. In this study, we want to watch the conversion from one lesion type to another. Then in those subjects that convert to these DDAF lesions, that is, atrophic retinal lesions, we want to measure the growth of the lesions because that's the endpoint for approval to slow the growth of the atrophic retinal lesions.
You can see there in the various rows, the criteria, the sites for the study, the duration, and the assessments in the very bottom, the key inclusion criteria for this study. And as I mentioned, all of these subjects were clinically and molecularly confirmed to have Stargardt disease. We knew from prior dose finding studies that a five milligram dose would be effective to reach or beat this target threshold of 70% or greater that was identified in the Fenretinide study. And what you're seeing here is the pharmacokinetic and pharmacodynamic profile of the five milligram dose in the phase II study. The blue line shows you the increase of Tinlarebant in blood, and the red line shows you the decrease of retinol binding protein 4 expressed as a percent decrease relative to baseline.
You can see there's a very nice correlation between the increase of Tinlarebant in blood and the decrease of retinol binding protein 4. The retinol binding protein 4 stays reduced at approximately a mean 80% reduction during daily dosing until we withdraw the drug at month 24. There's a rapid clearance of Tinlarebant from blood and a rapid rebounding of RBP4 back toward the baseline value, demonstrating the reversibility of this pharmacodynamic effect. I mentioned that we wanted to watch the conversion of lesion types and also the growth of atrophic retinal lesions. That data is shown here on the left-hand side. It's important to note that in terms of the first metric, that is, the conversion of autofluorescent lesions to atrophic lesions, five of 12 subjects never spawned an atrophic incident retinal lesion. In fact, their disease was essentially static over time.
There was no change whatsoever. That could not be attributed to milder benign mutations because, as I mentioned, all these subjects have very severe pathogenic mutations. I should also mention that although we recruited 13 subjects for the trial, we did lose one to follow up. The efficacy analyses are just including the 12 subjects. In the seven subjects that actually spawned these incident atrophic lesions, their growth rate is shown in the red line. We wanted to understand whether or not we were seeing a clinically meaningful result. What we did is we compared this growth to the growth of lesions from a natural history study obtained from a large study called ProgStar. ProgStar was the largest natural history study of Stargardt's ever conducted. The study recruited both children and adults.
We went to the database and extracted just that lesion growth data from subjects aged 18 years or less and with similar aggregate lesion sizes as we saw in the phase II Tinlarebant treated subjects. Those growth rates are shown in the blue line. If you compare the blue line from natural history to the red line, what you see is about a 50% reduction in lesion growth in the Tinlarebant treatment group. On the right-hand side, you can see the numerical outcome from these data. Moving forward now to our phase III trial called DRAGON . I do want to talk about an interim analysis that we did recently because the data were very, very positive. Before I get to that, I do want to outline the similarities in our two Stargardt phase III studies, that is, DRAGON and DRAGON II.
These studies are designed identically with the exception of three variables on the very top: the number of subjects being enrolled, the global geography from which we're recruiting subjects, and the randomization. It's a two-to-one randomization in DRAGON versus one-to-one in the DRAGON II study. Other than that, all of the criteria and study design for these trials is exactly the same. Here are the baseline characteristics and demographics for the subjects enrolling in the DRAGON trial. I mentioned before there were 104 subjects. On the left-hand side, you can see the age and baseline height and weight of these subjects. These are teen and preteen individuals of average height and weight. On the right-hand side, you can see the distribution for sex, and right below that, the distribution for race.
Getting back to the interim analysis, this interim analysis that we had for the DRAGON trial was designed as a sample size re-estimation exercise, and the way this was conducted was that we pre-specified a conditional zone of efficacy power. What that means is that there was this zone that's statistically referred to as a promising zone that would indicate a treatment effect between placebo and the Tinlarebant treatment group, and if, in fact, the data fell within this promising zone, which indicated a treatment effect, we would be permitted to add 30 additional subjects to the database to enhance the conditional power during the second year of the study, so we had an independent DSMB that was tasked with the responsibility of looking at the unmasked data at the interim and tell us whether or not we fell within this promising zone.
I should mention that by the time this interim analysis was conducted, we had data out to about 15 months. So the interim analysis included data up to 15 months, and that represented roughly 70% of the total available data from the study. In terms of the promising zone, there were two other options that basically we would not be able to add or increase sample size. And that is either for futility, that is, we fell below the promising zone, or being overly efficacious, that is, our estimate for the efficacy was more than we anticipated. And in that case, we would not be able to add subjects. So when the DSMB looked at the unmasked trial data, they recommended no modification of the study and that we should continue the study without a sample size increase.
In the absence of any other information, we knew that we were not in the promising zone, and we were either on the futile side or the overly efficacious side. We would not have known any different except that the DSMB provided a very important recommendation in their report form that they recommend we submit the data for further regulatory review for drug approval. They clearly would not have offered that recommendation if the data were futile. We believe we were on the overly efficacious side. Of course, since then, we've sent these data, the unmasked interim data to the FDA, and we've been granted Breakthrough Therapy Designation for that data. The FDA only does that when they see a statistically significant difference between the placebo and the treatment group. We believe we are seeing that, in fact, at the interim analysis.
In terms of safety, it's important to note that the overall withdrawal rate in the study was less than 10%, so only 10 of 104 subjects withdrew for any reason, and the withdrawal rate due to ocular adverse events was less than 4%. Only four of 104 subjects withdrew due to ocular adverse events. We are very pleased by this because we anticipated somewhere between 10%-20% withdrawal due to ocular AEs because this drug, again, is reducing vitamin A delivery to the eye, but apparently, these subjects are tolerating this drug very, very well because the dropout rate due to ocular AEs is very slow or very low, rather, and finally, the visual acuity was stabilized in the majority of subjects with a mean change from baseline of less than three letters under both standard and low luminance.
Here's an overview of the treatment emergent adverse events during the trial. The ocular AEs that we anticipated are two. The first is a form of chromatopsia called xanthopsia. This is triggered by exposure to bright light. Typically, these patients are reporting in their diaries that when they wake from sleep, they see this startling color in their visual field. This is caused by photoreceptors in the eye that are not getting vitamin A quickly enough. These are called cone photoreceptors. And before they fill up with vitamin A, they will electrically misfire when they're exposed to light and produce these transient artificial hues of color in the visual field. In this case, xanthopsia indicates yellow. This AE lasts seconds to just a few minutes, and it's reported as mild. And on the right-hand side, you can see the number and percent of patients experiencing that AE.
The other AE is sort of the opposite effect. This is called delayed dark adaptation, and this is triggered by onset to dim light. So when patients transition from a very bright environment to a dimly lit environment, they will have a delay in their ability to accommodate to dim light. This is called delayed dark adaptation. This is not night blindness. And that delay is roughly two to three times longer than normal. In subjects where it's even more exacerbated, it's called night vision impairment. And this means it can last up to approximately 20 minutes. But eventually, these subjects will all regain their dark adapted sensitivity. We also see some headache reported by subjects. This can happen when subjects try to use their visual acuity while experiencing these AEs. It can produce what's called a visual migraine. But again, this is a mild condition.
Systemically, there were no drug-related AEs with the exception of acne. This can happen in teenagers because vitamin A is important to clean out pores. So if you reduce vitamin A content in the skin, there could be some occlusion of the pores. But that was really the only systemic AE. There were no clinically significant findings in relation to vital signs, physical exams, cardiac health, or organ function. So the drug is being extremely well tolerated in this patient population. Moving forward now to the visual acuity. Here, you're looking at the change from baseline under standard luminance. And basically, what you're seeing is data from the study eye and the fellow eye of subjects. For people not aware, in an ophthalmology trial, you have to select a study eye for reporting data. The fellow eye is just followed in the same manner.
Of course, this is a bilateral disease and a systemic treatment, so we expect the same effect in both eyes. The data are blended, so we don't know who's on Tinlarebant and who's on placebo, but the fact that 66% of the subjects are on Tinlarebant because of the two-to-one randomization suggests that these lines are being dictated largely by the Tinlarebant treatment group, but as you can see, there's really no change from baseline in terms of visual acuity, so this is very important because you're never going to improve visual acuity, but if you stabilize visual acuity in the context of slowing lesion growth over time, this will predict preservation of vision for all subjects, so this is a very, very positive finding from the study.
Finally, now going into our phase III trial in geographic atrophy called PHOENIX , it's important to note that we can use the same exact dose of five milligrams that we used in our adolescent subjects in Stargardt because that dose produces the same pharmacodynamic response in terms of reducing retinol binding protein 4. So here, you're seeing the data from a PKPD study performed in healthy elderly adults meant to match the higher age range and BMI of subjects with geographic atrophy. We gave them a five milligram dose. And you can see here during the dosing period, there's a mean 80% reduction of retinol binding protein 4. And then when we withdraw the dose, there's a rapid rebounding of the RBP4, just like we saw in the Stargardt patients. So thankfully, we can use the same dose for GA. It's the same endpoint.
In fact, all of the elements of the trial design are very, very similar to what we're doing in the DRAGON studies for Stargardt disease. The only differences in the GA trial are the indication, which of course is geographic atrophy, and the enrollment at 530 subjects to reflect the higher prevalence of the disease in the population. The importance of having these trials designed similarly is critical because if you have a positive readout in one of them, like we're seeing in DRAGON , it predicts a positive outcome in the subsequent trials and actually de-risk them to some extent if we're seeing a positive outcome. In fact, we are seeing a positive outcome. Collectively, we're having very good success in DRAGON . We hope to see the same successes in DRAGON II and GA as we move forward.
I want to thank you for your attendance and appreciate your time here. If you have any further queries or you'd like to learn more about our company, please visit our website at www.belitebio.com shown on the lower right-hand corner of the screen. Thank you very much.