Belite Bio. Today I'm happy to have with us Tom Lin, the Chairman and CEO, Nathan Mata, CSO, and Dr. Michel Michaelides from University College London Institute of Ophthalmology and Moorfields Eye Hospital. Thank you all for joining us.
Thank you.
So I've laid out the overall agenda. We'll just share this. The overall agenda for the webinar. As we go through these topics, people in the audience can feel free to submit questions via the chat. We'll try to address them if time allows. Quick disclosure, the speakers have confirmed that they will not disclose any material, non-public information. The audience should frame their questions appropriately. So maybe we can dive right in. Dr. Michaelides, maybe we can start with you. Can you give a quick introduction and overview of your practice and background, including how many Stargardt and GA patients you treat?
Sure. So, I'm an attending at Moorfields Eye Hospital. I've been an attending for about 12 years. I spend half my time seeing patients with retinal disease, both inherited retinal disease and AMD, and the other half of my time seeing in clinical research, predominantly clinical trial design in developing novel endpoints, structure function associations, genotype, phenotype correlations. In terms of how many patients I see with Stargardt disease, it would be the second commonest disease that I see in terms of inherited retinal disease, that the first would be RP. I see about 50 adults with inherited retinal disease a week and about 30 children with IRD a week. And of those sort of 80 patients, 15-20 will have Stargardt disease a week.
And then, in terms of how many patients with GA do I see in a typical week? I'll see sort of five to 10 patients a week.
Okay. Since this discussion is going to focus both on Stargardt and Geographic Atrophy, can you walk us through how diagnosis looks like in those different patient populations? And also the kind of variants you might see in terms of disease progression that one might expect for those different patients.
Sure. Absolutely. So Stargardt Disease is a monogenic disease. There is one gene that causes Stargardt Disease called ABCA4. There are three main presentations of Stargardt Disease in terms of age of onset. You've got by far the commonest presentation is childhood onset. The second commonest will be in adulthood, in sort of 20s, 30s. And the least common and mildest form of Stargardt Disease would be the late onset foveal sparing form of Stargardt Disease. In terms of diagnosis, there's often a delay of all three of those presentations. In the early stages, especially in the childhood form and to some extent the adult form, the findings can be really quite subtle and could be readily missed if the appropriate investigations, especially retinal imaging, are not performed early.
And in the late onset form, it's not unusual that may be misdiagnosed as age-related macular degeneration, especially the dry form, so GA. The way it's confirmed is with genetic testing, which is now very readily available in most developed countries, and 70%-80% of patients with Stargardt Disease, with clinical diagnosis, you'd anticipate to make a genetic diagnosis. In terms of geographic atrophy, again, very common condition that we see. It depends on where you are in the Western world. In the U.K., often patients are diagnosed and not regularly followed up and then may be monitored by their optometrist.
I think in the United States, they're more readily followed up, long-term by a retina specialist, in terms of giving them some sense of the rate of change that's occurring over time, warning them and looking out for any potential complications, such as the development of wet macular degeneration, which clearly is treatable. And in terms of the diagnosis, again, it's very imaging based. Fundus fluorescein imaging is very helpful, as is optical coherence tomography.
In terms of available treatments, how do you treat your Stargardt patients, and how do you treat your GA patients?
Sure. So there is no approved therapy for Stargardt disease. And so we provide them information on healthy living that we really extrapolate from some of the AMD studies, suggesting that could be helpful for patients with Stargardt disease. Arguably, the most important two things we tell patients with Stargardt disease is not to take vitamin A supplementation. There's concern that taking vitamin A would accelerate the rate of progression. And we also tell them to avoid excessive exposure to natural sunlight. Again, suggestion that excessive UV light may be harmful. So again, ask them to wear good UVA, UVB blocking sunglasses. We also provide them with supportive measures, so low vision aids, spectacles, magnifiers, lots of different assistive technologies, and then provide registration, which provides additional support.
There is no, no treatment or cure to slow, halt, or improve vision.
And then GA?
For GA, in the U.K., we don't have any available therapies for GA. If they have GA, confluent drusen will advise them to take AREDS 2, that may slow rate of progression. So, there's quite a high pickup of taking AREDS 2 supplementation. And again, it will be similar lifestyle in terms of avoiding excessive exposure to bright sunlight with good sunglasses. We'll also add in the smoking, which is a decent, good evidence there that people should stop smoking. And then we tell them about the therapies that you have on the slide that are available in the United States. As you know, SYFOVRE was not given EMA approval, primarily based on lack of functional improvement, with some concern about safety.
Izervay is currently under assessment by the EMA, is my understanding.
If you had to assess any limitations or ongoing unmet needs in each of those patient populations, what are sort of the key things that patients are still looking for?
Well, certainly both will be looking for a slowing or a halting of their disease progression. It always progresses, and so they're certainly looking for that. The children in particular, in addition to having difficulty with central vision, also start losing peripheral vision. And so it's sort of a, they have both of those aspects, GA, clearly central vision. But in childhood-onset Stargardt, they have both central and peripheral loss of vision over time. So they're even more visually compromised. Of course, once one's able to slow or halt progression, the human species wants more. It always wants more, so it would want improvement of vision. And after improvement, it will want full restoration.
But it's a great place to start with slowing or halting progression, and certainly all patients ask me, is there anything they can do that might slow progression? And, and they allude to lots of other therapies that certainly aren't approved and aren't advisable. So there's a huge appetite for slowing or halting progression.
May I ask a hypothetical? Because like you said, the complement inhibitors aren't approved in the E.U. They are approved in the U.S. If they were available on the E.U. market, are there still unmet needs that you see with those treatments?
Sure. Absolutely, there are unmet needs. I mean, looking at the data, it looks as if it, it slows progression by about 15%-25%, something like that. So there's clearly scope for greater levels of, disease modification, slowing, progression. There clearly are concerns about SYFOVRE, certainly in the U.S., and just got back from a U.S. meeting, and I'm aware some of my retina colleagues in the U.S. are sort of somewhat reluctant to use SYFOVRE because of the, the retinal vasculitis and the inability to predict who will develop retinal vasculitis. And when you do develop it, it can be sight-threatening.
And then there is the association of both of these with the development of wet macular degeneration, which does seem a bit counterintuitive to be treating dry macular degeneration, GA, and inducing wet macular degeneration, although we do have very good treatments for that. So I think there is some concern, and of course, we're not really sure yet whether Izervay will have some of the similar side effects of SYFOVRE. Of course, I hope it won't have that retinal vasculitis concern. Very different doses of those two drugs, so maybe that different dose will come into play.
And maybe turning back to how this all falls into the Stargardt a little. The mechanistic rationale for Tinlarebant, and maybe this question fits for Nathan, if you like. Some people in the audience might be less familiar with how Tinlarebant works and its history. Can you walk us through the mechanistic rationale of that asset?
Sure, sure, sure, Jennifer. Yeah, I think, excuse me, the first point to make clear is that in both Stargardt disease and geographic atrophy, the accumulation of toxic vitamin A byproducts is implicated in disease progression. These byproducts are called bisretinoids, and they're derived from vitamin A, that is retinol, that is circulating in the blood. And so—and you see it here as a schematic. This is a schematic of the visual cycle. The retinol, abbreviated ATrol, that's vitamin A. In the liver, vitamin A binds to Retinol Binding Protein 4, and then another large accessory protein called transthyretin binds to that complex. And what gets liberated into the circulation is this ternary complex of Retinol Binding Protein 4, retinol, and transthyretin. It's a very large molecular size complex, so it resists filtration in the kidney.
This is biology's way of maintaining a high, steady state level of vitamin A in the blood. But what's very important to note is that only the eye has a need for delivery of retinol bound to Retinol Binding Protein 4. That's because of the presence of a Retinol Binding Protein 4 receptor in the back of the eye, which is not present in other extraocular target tissues. In the eye, there is a unique dependence for delivery of vitamin A bound to Retinol Binding Protein 4.
When the complex docks onto the receptor, the vitamin A goes in through a series of enzymatic reactions, where it's eventually converted to rhodopsin, and then light activation of the rhodopsin liberates a retinaldehyde species of vitamin A, which is very toxic to membranes, and it has to get out of the retina or it'll actually start damaging retinal tissue. That's where the ABCA4 protein comes in. This is the protein that's encoded for by the ABCA4 gene, which, of course, is mutated in Stargardt disease. Without that mutation and with a functional protein, that retinal can readily leave the retina and reenter the visual cycle. But in the case with Stargardt disease, where there are mutations that affect the function of the ABCA4 protein, the retinal cannot efficiently leave the retina.
It lingers within the retina, and it complexes upon itself, essentially forming these dimers of vitamin A, so two pieces of aldehyde bound together. And basically, these compounds, which we call bisretinoids, are highly toxic. The one that's been identified in greatest abundance in human tissue is called A2E, and this molecule has been shown to be very stable, and it kills retinal tissue through diverse mechanisms. So there's nothing good about this compound. It really is the whole reason for retinal pathology in Stargardt's. In GA, these bisretinoids accumulate as well, but for a different reason. They accumulate because there's debris beneath and below the retinal pigment epithelium. This interferes with nutrient exchange, transfer, and trophic sort of influences that the RPE has on the retina.
It also causes the dysfunction of the enzymes within the RPE that normally metabolize and process vitamin A. What happens when they run amok is these bisretinoids form locally, right within the retinal pigment epithelium. So in both diseases, they form, but in different ways. For Stargardt, it's a genetic insult, and in Geographic Atrophy, there's a whole series of events, including the accumulation of drusenoid bodies within the RPE that cause the RPE to become sick, and the bisretinoids will form locally there. There's an interesting study, right here. This is actually not a study, actually, it's like, it's sort of a case history of two subjects with disease, a Stargardt on the top, Geographic Atrophy on the bottom. And you're looking at a series of retinal images. Professor Michaelides looks at these all the time.
These are fundus autofluorescence photography images, and they allow the ophthalmologist to visualize what's happening in the back of the eye. The black blotches you see in all the images are atrophic retina. That's dead retina that's never coming back. But the whole point of this, these images here is that if you look chronologically from baseline out to 57 months, so roughly about four and a half years in each series of images, what you'll see is that the appearance of autofluorescence precedes the onset and spread of atrophic lesion growth.
So in the Stargardt patient, you see at baseline these two large areas of atrophy surrounded by autofluorescence, and as you move forward in time, out to 57 months, you see that the darkened areas spread into the autofluorescence zone, and the autofluorescence zone continues to expand centripetally to accommodate that dead retina. You see the same thing in GA. It's a sort of a different pattern. The autofluorescence there is sort of in speckles, little punctate pieces of light, if you will. You can see them a little bit more clearly at 12 months. And then as you go forward from 12 months out to 55 months, what you see, those areas that were previously just autofluorescent have now become retinal atrophy, and the new areas of retinal atrophy have little zones of autofluorescence around their perimeter, just like the large lesions.
So in both diseases, we believe these bisretinoids are implicated, and these clinical data sort of confirm that. There's also some very interesting histological data that show us the presence of these bisretinoids within the tissue of geographic atrophy patients. So what you're looking at here in slide, sorry, panel A, is a cross-section through the back of the eye. Of course, this is a postmortem tissue from a patient that had geographic atrophy, and they're using a fluorescence microscopy to visualize where there's autofluorescence. And the reason they're doing this is because these bisretinoids fluoresce. Because they contain vitamin A, if you shine the appropriate light at them, they will shine back light, and that's what you're seeing in panel A, and we're focusing really on this pink band that you see throughout the retinal pigment epithelium.
This pink band actually has a spectral, what's called a fingerprint, if you will. Just like humans have fingerprints on their fingers, molecules have spectral fingerprints that are very unique to each molecule. So what you see in B is the spectral fingerprint of that autofluorescence that's shown in panel A, that pink band. And on panel C, what you see are authentic spectra from a Stargardt patient, postmortem Stargardt patient, showing the A2E molecule and its precursor, called A2PE. And now, if you overlay the two, Jennifer, you can hit a Q stroke, you can see that the spectral fingerprint I obtained from the GA tissue is identical to the spectral fingerprint obtained from the Stargardt patient. So once again, this is telling us that these molecules, these bisretinoid molecules, do in fact accumulate in GA tissue.
Again, as they're doing in Stargardt Disease, they're ravaging tissue, and so there's nothing good about them. They're derived from vitamin A, and so that's our approach, really, is to limit the amount of vitamin A going into the eye as a means of reducing the accumulation of these nasty toxins. Next slide, Jennifer. So that's what Tinlarebant does. So Tinlarebant is an oral available small molecule that goes into the liver and competes with retinol for binding to RBP4, and it doesn't allow the larger protein transthyretin to bind to it. Consequently, what gets liberated into the circulation is a relatively small complex of RBP4 and our drug bound to it. And because it's so small, it gets readily filtered through the kidney.
So the net effect of that function then is to reduce the level of Retinol Binding Protein 4 in blood, which of course, would reduce the native complex. Thank you, Jennifer. The native complex of RBP4 retinol and TTR, that would go down, and of course, once that goes down, the amount of vitamin A going into the eye goes down, and of course, all the retinoids cascading downstream, including those end product, bisretinoids, would also decrease. So this is the mechanism whereby we intend to slow lesion growth by going after these bisretinoid toxins, which we know are implicated in the progression of lesions, spreading of lesions in both Stargardt's disease and geographic atrophy.
Great. Thanks, Nathan.
Sure.
Dr. Michaelides, what are your thoughts on Tinlarebant's approach from a mechanistic standpoint? Does your enthusiasm differ from Stargardt to GA, or does it make sense in both diseases, in your view?
You know, it makes sense in both diseases. I think Nathan made the argument very compellingly. I like the fact that it is targeting what we think, believe to be the underlying central pathogenesis of disease in both. I like the fact that it's also reducing the amount of retinol available to the eye specifically, so there's no reduction in retinol availability to any other tissues. There's no systemic deficiency. It's targeted to the target organ. And also, I like the fact that it's reducing the amount of retinol, but it's not affecting the rate at which the visual cycle operates. And that's sort of quite different to some of the other agents out there. So yeah, you know, mechanistically, I like it, and let's see if it works.
Yeah, so that takes us to our next focus, the data to date for Tinlarebant. I know a lot of the data we've seen have been from the phase II study in Stargardt. There are some interesting new analyses that Belite presented at ARVO recently. Nathan, would you mind walking us through that data?
Sure, absolutely. Yeah. So this is the overview first to give everyone a sense of exactly what we're doing. The phase II study that you see boxed there, that was the open label study that enrolled 13 adolescent Stargardt subjects. Again, open label study, two years in duration. This is the one we recently completed, where we've had a lot of very interesting data come out of. Very, very promising efficacy and safety data. Just an overview of the study. Important thing to note, these 13 subjects coming in had a very early stage of the disease, where they just have the autofluorescent lesions. So they haven't yet converted to the atrophic lesions.
Based upon natural history and the size and location of these autofluorescent lesions, we predicted that the majority of them would convert to atrophic lesions over a two year period, which was the study period. So the autofluorescent lesions are called questionably decreased autofluorescence. That's where you see that abbreviation, and they will convert or transition into definitely decreased autofluorescence, which are the atrophic lesions. So in this study, we're monitoring that transition, and in those subjects that actually grow incident atrophic lesions, we wanna measure that growth rate. So I'll show you that in a moment, and then you can see the Pharmacokinetics and Pharmacodynamics of Tinlarebant at 5 milligrams daily. This is from the phase II study.
So you can see here, there's a very nice correlation between the increase of Tinlarebant in blood, that's shown in blue, and the decrease of Retinol Binding Protein 4 in blood, that's shown as a percent of the baseline value. So by the time we get to what we call steady state, we're getting about a mean 80% reduction of Retinol Binding Protein 4 in blood. And then at the end of month 24, we'll withdraw the treatment. You can see there's a rapid removal of Tinlarebant in blood. You see the blue line goes straight down, and the Retinol Binding Protein 4 bounces right back toward the baseline value. So this is very nice to see this very rapid reversibility of the pharmacodynamic effect.
Should there be any untoward concern, where you need to bring a patient back to their baseline status, you can do so quite readily within a month. These are the data from the patients in the study. We did lose one subject to follow-up at month 12, so we're talking about five subjects here in this efficacy analysis that you're looking at. We're looking at, first, as I said before, the transition of the autofluorescent lesion to the atrophic lesion, and strikingly, five of 12 subjects, that's 42% of the cohort, never converted to an atrophic lesion over time. Again, as I said, we would've predicted that most of those kids would've converted, but to see 42% not convert, despite having very significant autofluorescent lesions, is actually quite promising.
The other interesting point is that in those subjects that have actually converted to atrophic retinal lesions, most of them occurred after month 12, so they were sporadically month 12, month 18. But the growth rate of those lesions was considerably lower than the growth rate that has been seen in natural history from subjects that have similar baseline characteristics as subjects in our phase II study, and that's shown in the blue line. This is from the largest natural history study of Stargardt's conducted today, called ProgStar. Largely enrolled adult subjects, but there was a cohort of younger subjects that matched our subjects in the phase II, and we looked at just at the growth rate of those subjects. You can see in the blue line, there's roughly a doubling of the growth rate compared to what we see in the Tinlarebant treatment.
You can see the numerical data on the right-hand side. So this is a very, very promising. We're slowing the transition of the autofluorescent lesions to the atrophic lesion, and then once the atrophic lesion is formed, we're seeing a slowing of that growth rate. These are all consistent with the MOA that I just explained earlier.
And then you showed some new data at ARVO. Can we go into that?
Yeah. So we did a genotype, phenotype relationship analysis. We hadn't done this previously. Of course, the genetic data we've had, but we were waiting for it to get analyzed by Dr. Rando Allikmets. He's the individual who identified the gene associated with Stargardt's disease, that is the ABCA4 gene.
So he read these genetic data for us, and he provided the outputs that you see here. The primary score is called the CADD score, the combined annotation dependent depletion score. This is basically a numerical value that tells you how severe a lesion is or, sorry, a mutation is. Values that are 20 and above are predicted to be among the 1% most deleterious. And so what you can see here is that 11 of 13 subjects have very high CADD scores. That means they have very severe biallelic mutations. There's only two subjects that don't have severe biallelic mutations, subject 3 and 5. You can see they have an allele that that is moderate on each for each subject.
But if you look at the in vitro testing, you can see that both of those moderate alleles were, in fact, found to be pathogenic in in vitro testing. So suffice it to say that overall, this cohort is very largely severely biallelic mutated. That means they're gonna have pathogenic lesions. But despite having all these pathogenic lesions, five of those subjects I told you earlier never transitioned from an autofluorescent lesion to an atrophic lesion, and their autofluorescence was quite stable. So in these subjects, we've really stopped, we believe, the progression of the disease. That's one new piece of information that we've had. There's another one as well regarding siblings.
Before we go into that. Dr. Michaelides, when you look at this type of data, how does it play into your interpretation of the disease progression you've seen and the kind of data you've seen for Tinlarebant?
Sure. So it's very impressive. I mean, of course, you know, obviously, it's 13 patients, okay? So the patient number is small. All right? But if we just park that inherent limitation of small numbers of patients, this is a very severe group of patients. So childhood-onset disease is the most severe form of Stargardt disease. It progresses most rapidly. It generally has generalized involvement, as we described, central and peripheral visual loss. So you could argue they're the hardest to prevent progression. I've always advocated for them. Because they progress rapidly, one should be able to see something early and have a bigger impact on this patient group, on these children. So to see this, to actually prevent the inexorable progression of QDAF to DDAF is significant.
And then even the ones that did progress did have new DDAF, the rate of reduction of 50% is huge.
You also presented this data? Nathan?
Yeah. So, the whole point here is that, Well, first going back to those five subjects, so the slowing of transition could not be attributed to mild or benign mutations. They're very severe. The other finding is very interesting, that we have two pairs of siblings that have the exact same genetic mutation, and that's subjects 9 and 10, which are two brothers approximately one year apart, and subjects 12 and 13, which are brother and sister. I think they're about six years apart. Why this is important is because there's been some claim that identical mutations predict an identical disease course, and this is not something we've seen in our data set, and we're using these four patients as sort of an example of that.
As we go through the data, you'll see what I'm talking about in just a moment. Here it is, actually. This is sort of a separate analysis, but it's sort of two in one, if you will. One, we looked at the visual acuity outcomes prior to coming into the study. We're looking at visual acuity in the subjects prior to enrolling into the study, and we're looking to see which subjects have bilateral vision loss. That is, vision loss in both eyes prior to coming into study. Turns out there was a total of six subjects that had a mean of 10 letters lost per year in each eye.
So that's significant, and over a two year period, which is our clinical study duration, two years, we would predict you'd have roughly 20 letters lost. Put a note in that, because that's not what happened, but it's significant to note that these kids were losing significant vision prior to coming into study, and they don't have atrophic lesions at this point. They only have autofluorescence. So this does tell us initially that autofluorescence encroaching into the fovea does in fact impact vision. More on that later. But the other point now is, if you look back at subjects 9 and 10, which again, are identical mutations, they have the exact same mutations on each allele. They have a similar disease duration, approximately one year, but they have very different visual acuity outcomes.
You see, subject 10 was losing letters over time before coming into study, but the brother, subject 9, didn't lose any, in fact, fared quite well. And again, subjects 12 and 13, they have this different disease duration, but if you look, subject 13, again, identical mutations as subject 9 and 10. Subject 13 only had two years disease duration, but lost a mean of 12 letters in each eye. Meanwhile, the sister, with eight years disease duration, did not lose any of, any visual acuity, at least not significantly. So these data are telling us, just from the genetic perspective, that visual acuity outcomes at least are not predicted by, identical genotypes.
And another question maybe for Dr. Michaelides. Does that analysis fall into what you see in your patients? Like, does one sibling, is one sibling predictive of another, in your experience?
No, almost never. So, it's very common to have significant within-family variation and between-family variation, even when they harbor exactly the same disease-causing mutation. It's actually a characteristic of inherited retinal disease. So it would be very unusual and undesirable, not recommended, to be trying to predict disease progression based on what happened to a sibling. So I'm not surprised by this data. You know, we have our own data at Moorfields that would be in keeping with this. We've published it. I know that the Dutch groups have published similar findings in siblings, showing how those siblings are discordant in terms of their rate of progression, even when their phenotype at baseline might look reasonably similar, but the progression is very different.
Nathan, do you want to keep going?
Yeah, these are the visual acuity outcomes, broken out into different groups. On the left, you have the visual acuity outcomes in all subjects, and you can see here, over two years, there's a mean letter loss of about five letters. That's 2.5, 2.5 letters lost per year. This is well within a standard error or deviation of the BCVA assessment. So we call this stabilized vision, and we're very happy to see it, because it is telling us that things are slowing down, right? The lesions are slowing down. Visual acuity is stable, so this is a very positive outcome. On the right-hand side, you see a similar analysis, but this is just for those six subjects that were losing 10 letters per year before coming into study.
While on study, they only lost about 1.9 letters per year. So we've significantly slowed visual acuity loss in this small cohort of subjects. Again, it's a small group overall, but the trends are going the right direction. And again, everything is in keeping with the mechanism of action of the drug. The slowing of the lesions encroaching into the fovea, you would predict would have some effect on slowing or stabilizing visual acuity, and that's exactly what we're seeing here.
I want to turn it back to Dr. Michaelides as again. Can you confirm, in your experience, does the pre-enrollment mean bilateral loss of 10 letters? Would you expect that to be predictive of future loss rate?
It's certainly within what I would predict would be the visual loss. These are children that lose visual acuity really quite rapidly throughout their teenage years. The magnitude of the change, I think, is significant. Even if it hasn't kept it dead stable, it's significantly slowed the rate of visual acuity loss, and the fact that you have structure and function going together is also reassuring. Again, of course, it's 13 patients, but, you know, the data is what it is. It is promising, it's compelling. It looks plausible.
Nathan, what should people know about QDAF and DDAF?
Yeah, I think, you know, what, what's important on this slide really is to sort of the take-home message is that visual in the upper right-hand side. The DDAF lesions largely form within the area of QDAF, and so the overall size of the lesion doesn't change. So you can see here the baseline image. This is from the subject 10, by the way. This is the actual data. This is the autofluorescent lesion at baseline, and then at month 24, you can see within that autofluorescent lesion grew some atrophy, which we call definitely decreased autofluorescence. And the graphic on the bottom is basically a histogram to show you there's a clear proportionality between the decrease of the autofluorescent size and the increase of the atrophic lesion size. In each subject, you see this very interesting correlation.
Only in subject 5 was there a lesion that was identified outside or away from the initial lesion that was identified at baseline. And we believe this is really attributable to an image error or, or sort of a-- the lesion was there, but it didn't get picked up by the, by the region-finding software or the readers that were grading this image, 'cause it essentially appeared out of nowhere, and it was probably there all the time. So this is very important because it sort of leads us to this next idea about, you know, new imaging modalities for assessing atrophic lesions in patients. And so, as I mentioned to you before, before coming into this study, all of the subjects had only autofluorescent lesions.
The region-finding software that comes with the Heidelberg instrument that is used to visualize the lesions in the retina, basically is used with a reader. So basically, the software identifies areas of changing autofluorescence, and then the reader goes in and essentially draws a boundary or identifies really where the atrophic lesions is. So there's really some subjective input, or you could call it bias, for doing this type of reading. Our reading center, [OIRCC] in Palo Alto, has developed a new grading algorithm, and it really does remove the subjective reader bias because it's based on a pure mathematical classification of lesions. It's looking basically at pixel density, single pixel densities, and it's grading them based upon a grayscale overall from really healthy retina to the darkened area, which is the optic disc.
So it's looking at a continuum of gray levels, it's removing subjective reader bias, and it's highly sensitive and highly reproducible. Moreover, it's focused purely on the 6-millimeter area, which is essentially your macula. Of course, we can look at the 1 -millimeter and 3- millimeter and 6- millimeter, 6-millimeter ring, but this is focusing right now on just the 6-millimeter ring. And when they use this new imaging or grading modality to look back at those images from baseline and forward, we saw that, in fact, there was 12 eyes of eight subjects that actually had atrophic lesions within the macula at baseline. And again, this was only accessible through this new grading method. So Jennifer, you go to the next slide, I'll show you the growth of those lesions.
So on the left-hand side are the data from those subjects, the 12 eyes of eight subjects that had the atrophic lesions identified within the 6-millimeter ring at baseline. You can see that atrophy tends to grow into the macular area up to about month 16. Then from month 16 to month 24, it's essentially halted. By the way, the dotted line, you see the dotted red line, is a third-order polynomial. It's a function drawn through the data points, so you can see the trend line. So you can see very clearly there is a stasis of growth into the macular area from month 16 to month 24 in these subjects. On the right, we look at the percent of the lesions within the 6-millimeter zone. So if all of the 6-millimeter zone was occupied with lesion, it would be 100%.
But you can see in our subjects, it starts at four, goes to seven, and really doesn't change much more over time. It's 7% involvement into the macula of the 6-millimeter zone. So this is really, again, it's consistent with what we showed, what we showed previously, that is slowed lesion growth, stabilization of visual acuity, because you got to know that if you're slowing lesion growth into the macula, you will definitely be slowing visual acuity loss, or at least preserving whatever healthy vision is there.
So, is my interpretation of it correct that using this new imaging method allows you to better detect the treatment effect in the completed phase II trial? Is that what the sort of goal was?
Well, I wouldn't say it's about treatment effect. It's a better way to identify the atrophic lesions. That's, that's the most important element. That's the key to this, and it's very fast. So in cases where two readers disagree about a certain single image, typically that has to go to an arbitrator, and the arbitrator is sort of the tiebreaker. That process can take up to two-three hours at some reading centers. This imaging modality, this grading algorithm, does it in seconds. So it quickly can discern this at a very high level, and as I said, at the single pixel level. So we're moving forward, hopefully to get this validated. We'll be speaking to the FDA about that because we're really confident this is really a superior grading method to use forward in our phase III study.
In the phase III trial, one design difference is that patients at baseline do have atrophic lesions, correct?
Absolutely. That's important, Jennifer, because the primary endpoint for approval in Stargardt's and GA is slowing the growth of the atrophic lesion. So you have to start with some atrophic lesion at baseline as a reference point to measure subsequent growth.
Dr. Michaelides, do you have any thoughts on this new imaging approach, and does it feed into how you think about the phase III program?
It makes a lot of sense. It's gonna be more reliable, more repeatable, more sensitive. We'll be able to have greater faith in the results. So yeah, it makes perfect sense. There's no doubt there is variability both between and within individual graders, for all metrics, but arguably maybe more so for autofluorescence.
Nathan, do you wanna talk about the safety you've seen so far?
Yeah, we're very, very proud of the safety profile of this drug, and it really is a testament to how the drug was designed. It's a purpose-built Retinol Binding Protein 4 antagonist, which means it just goes to the residues on the Retinol Binding Protein 4, a protein that bind retinol. So we predicted it would have very little systemic drug-related effects. In fact, we saw none. So over two years of treatment, there's not one drug-related systemic AE whatsoever. The kids are doing wonderfully. What we do see are the anticipated drug-related ocular AEs we need to see because they're telling us we're having the intended biological effect on the retina. The first is a form of chromatopsia called xanthopsia. This is a color vision aberration caused by onset of bright light.
So basically transitioning from a very darkened environment to a bright environment would act, would sort of induce this phenomenon. And basically what happens is that when you experience bright light suddenly, your cone photoreceptors will demand chromophore. And under our treatment regimen, that chromophore will be only slowly supplied again, 'cause we're lowering the amount of vitamin A in the eye. So it'll take longer for those cone photoreceptors to fill up with chromophore. During that period of time, while they're filling up, they will electrically misfire and produce these hues of color in the visual field, in this case is yellow. That's Xanthopsia. It lasts seconds to minutes, and it's, you know, of course, transient and reversible. Everyone's tolerated it quite well. You can see the majority of subjects experience it. Not one subject left study because of Chromatopsia.
Same thing for delayed dark adaptation, sort of the opposite direction in terms of the lighting extreme. So when patients transition from a bright environment to a darkened environment, just like you would going from a bright, sunny day into a movie theater, you have a time where you don't have full dim light sensitivity. Well, if we slow the chromophore regeneration or chromophore supply in these subjects, they'll have a longer delay, so maybe two-three times longer than normal. It's called delayed dark adaptation, and again, most kids are getting it. No one's left study because of it, because, again, it is reversible, and it is transient. Night vision impairment is when the delayed dark adaptation is exceedingly long, up to 20 minutes in this one patient. Increasing error score on the FM-100 is a more severe exacerbation of the chromatopsia.
And we believe these two subjects with intermittent headaches may get them if they're struggling to use their visual acuity while experiencing these AEs. It can induce what's called a visual migraine, so that may be what's happening here. But overall, very clean safety profile for two years' duration, where we've reduced the Retinol Binding Protein 4 by a mean 80%, and you can see no untoward effects over this long daily treatment period.
Nathan, the next set of questions I think are gonna be geared towards you, Dr. Michaelides. So we've seen the data to date. I guess, overall, what are your thoughts on the data, both on the efficacy and safety side? And then thinking about a larger phase III program, what do you think Tinlarebant really has to show to be viable, moving ahead?
Sure. Sure, so I think given there are the multiple structural metrics that are in keeping with a significant slowing of disease or even halting of transition, allied to that functional improvement in terms of visual acuity, I think that, that's very promising. In terms of safety, you know, as Nathan was saying, you know, you're expecting that there are going to be some of these, short-lived, reversible side effects. And we warn our patients who are now in the DRAGON phase III trial about these. And they're all been happy to continue, and in fact, they're keen to go into DRAGON two if they're able to, the second phase III after washout. So I think it's very promising. I think it bodes well for the phase III.
And, you know, it tells me a lot when patients wanna stay in studies and are asking about, you know, how can they get more drug? Is there some sort of compassionate access to drug? Can they stay on drug? Because they can tell if they're on active drug because of these changes they're experiencing.
In terms of efficacy, if we're looking at lesion growth rate, is there a number that you think that has to be achieved?
Is there a number?
Yeah.
I mean, look, start up, there's no treatment at all. There, there's nothing. There's no benchmark. There's no precedent. If we say, well, look, let's look at a related disease, geographic atrophy, you know, that's 15%-25% slowing. So if it achieved that, I'd say that we're meaningful. Of course, I want that 50% we're seeing in this phase II. It's a small study, but, you know, 15%-25% would be great, and I know patients would be wanting to take that medication.
You've touched on the safety. If the safety sort of mirrors what we've seen with the phase II trial, or what is sort of the wiggle room on that side?
You know, if it looks similar, I think that's gonna be just fine. You know, I think there are modifications, lifestyle type modifications patients can make. In my discussion with the children and their families, and adults who keep asking, you know, "Can I come into the Belite trial?" "I am afraid not. You're not an adolescent." You know, you have those discussions about, look, it's about trying to preserve central vision, and maybe you'll have some slightly longer dark adaptation. You may get these bright lights when you wake up, but they're not things that prevent people wanting to take part. So I'm not overly concerned about those AEs.
Okay.
Yeah, Jennifer, if I could just add one point, this is important. So we have 104 subjects in our phase III DRAGON study. To date, we've lost five subjects due to AEs, okay? Five subjects. And this study is more than halfway completed. So we have... Most subjects have completed one year, and our dropout rate is less than 5%. That's phenomenal from my perspective, having worked with other drugs like Emixustat and Fenretinide, where there's a much higher dropout rate with those drugs. So I'm very pleased to say, you know, cross fingers it's gonna keep up, but 5%, less than 5% dropout one year in is really phenomenal.
It's good to hear. Another question that I get just about clinical trial design. Some people ask about lesion sizes at baseline and how that affects the expected lesion growth rate. I wanna be specific here because I think in DRAGON, Nathan, correct me if I'm wrong, I think the average lesion size is around 2 -millimeters squared. I believe ProgStar actually defines that as a larger lesion. I don't know if that's fully appreciated, but Dr. Michaelides, could you talk to us about how baseline lesion size predicts growth rate of lesions?
Sure. So, you know, my understanding of the literature and, both in Stargardt and in GA, is that, the larger the lesion, the more slowly it will, it will, progress. So you actually do want to recruit patients with smaller lesions, in terms of having a bigger impact on slowing of progression. And so certainly in the Stargardt, phase III world, Acucela, Kubota, their phase III, which was Emixustat, so an RPE65 inhibitor, very different to Tinlarebant in that it, it's not inhibiting any enzymes. They didn't meet their endpoint, but when they actually did a post-hoc analysis, they did find in patients who had small lesions, they were able to slow progression, notably. So small is better in terms of slowing the rate of progression.
When you talk about small lesions, is there, like, a clear, I guess, defined number for that or a sweet spot, or do we know or not know?
Yeah, I think we don't yet know. But trying to enrich for towards the smaller end of the spectrum is certainly beneficial. I mean, at the moment, most people use sort of 1 -millimeter squared as a cut-off. I think actually using the technology Nathan's talked about, we might be able to get down to 0.5, and it would be far better to intervene early and prevent loss of vision.
In the phase III, if DRAGON data sort of falls into what you would like to see, how would you envision incorporating Tinlarebant into your patients in Stargardt?
Sure. So, if it's safe and it works in the way we've just discussed or in keeping with the phase II, I would be offering it to all my patients with Stargardt Disease, all that had molecularly confirmed Stargardt Disease. At least one disease-causing and typical phenotype, so in keeping with the trials. I would be offering it to all of them, and they would be taking it potentially lifelong or until something more efficacious comes along.
I think you touched or you hinted at this near the beginning of the call. But when you think about some of the other later-stage competitors, there's a deuterated vitamin A product as well that's also, well, that's in late stage. Do you have any thoughts around that approach and how it might compare to Tinlarebant, or are there limitations or concerns, or-
Sure.
How do you think about that?
Sure. I mean, it's been around significantly longer, which always sometimes makes me wonder why it's been around for so long. Of course, there are confounding things, you know, funding is obviously a big issue in running clinical trials. But the sort of two main things that are on my radar when it comes to the ALK-001 or the deuterated vitamin A. One is that they seem to have multiple studies, and it's quite hard to make out exactly what the patient groups are recruited to the different studies. They seem to have different characteristics, different follow-up, different regimens. There doesn't seem to be a concerted plan there, in my opinion.
But arguably more so is the fact that there is, at least theoretically, a concern that by having high doses of vitamin A, chronically, you may actually have vitamin A toxicity. You know, there is literature out there raising concern about taking vitamin A supplements, both in terms of neurotoxicity. And I think, I do believe that they did have one or two patients that developed papilledema in their trial, which might be in keeping with neurotoxicity. And also there are concerns about bone disease, liver disease. So, you know, having chronically elevated vitamin A that is in the order of four or five times recommended daily allowance is a potential safety concern.
What about, I think Nanoscope has an optogenetic therapy, MCO-010. Do you have any thoughts on that program?
Sure. Very different approach. So that's now really more about patients who've got advanced visual loss and trying to give them a degree of light sensitivity or vision based on trying to make some of their inner retinal cells, I believe Nanoscope's targeting the retinal ganglion cells, to make those light sensitive. So to act as in place of the light sensitivity you would get from cone photoreceptors, to get that from retinal ganglion cells, using a modified opsin delivered intravitreally. I know they presented some data suggesting some improvements in visual acuity.
I guess we just have to see how that progresses, but I see that really more for very advanced patients who've sort of fallen off the chart, if you like, no longer see anything on the visual acuity chart.
Is there anything else on your radar in terms of the Stargardt pipeline?
I mean, I know there are multiple other groups working in the same optogenetic field as Nanoscope, Restore Vision, Ray Therapeutics, Gyroscope. So, several players looking at that advanced visual loss. In terms of other things, of course, there are genetic therapies, which are, have been being worked upon for quite some time. The challenge is the gene is enormous, and it doesn't fit into the preferred viral vector, AAV. And so people are using all manner of different approaches, you know, including dual vector approaches, so sort of halving the gene. Cells have to be infected by both viruses, has to be a recombination event. So multiple sponsors are looking at that, dual vector approaches. Others are looking at sort of exon editing, others are looking at CRISPR-Cas9, others are looking at antisense oligonucleotides.
So there are lots, but they're really very much at very early phase. Very early phase. Some are not in clinical trial yet, but others are really at the very beginning of that journey.
Turning to Geographic Atrophy.
So, Jennifer,
Yeah, Tommy.
Jennifer, may I come, coming back to, Dr. Michel Michaelides' discussion on the vitamin A toxicity. So, from, from my understanding is that the deuterated vitamin A is a lifelong treatment as well. So it's a exposure to lifetime of, vitamin A toxicity. So I probably would have some concern given that, you know, it's a lifetime treatment and having, daily doses of four times over the limit, is certainly concerning. And given that, you know, out of the 30 subjects that's recruited on the TEASE trial, one or two subjects have, incident of, of papilledema, which is secondary to increased intracranial pressure. I think, given that intracranial pressure is rare and you have such a high incidence in that trial, that, that just speaks to, about the vitamin A toxicity, the dangers of vitamin A toxicity.
Yeah. Maybe for the last part of this call, geographic atrophy. Dr. Michel Michaelides, what are your thoughts on, I guess from what you've seen so far for Tinlarebant, your enthusiasm around its potential in geographic atrophy? And same question here, what is the bar, for the PHOENIX trial?
Sure. You know, I am enthusiastic for a couple of reasons. I guess I like the fact that it's targeting potentially in an early part of the pathway rather than a late one, or a response to something else. So I like the fact that it's early. I think it will have a role in potentially being used early in disease to prevent progression. I think it could, of course, have a role throughout, but I like that it's early in the pathway. It could be applied early in disease. I like the fact that it's oral, it's treating both eyes. This is a bilateral treatment, sorry, bilateral condition.
There's certainly a degree of fatigue about intravitreal therapies that would need to be currently monthly for years, with little knowledge of the patient of whether it's working or not, and the burden on patients and healthcare systems with an oral medication, of course, is markedly less than a serial intravitreal one. In terms of what does it need to achieve, you know, there are approved therapies. So, you know, it will need to be at least as good as those that are approved. Of course, you know, ideally, it would be better in terms of efficacy. And in terms of the safety, again, one would hope it will have a similar profile to the Stargardt patients. And you know, time will tell.
How would you see? I think you pointed to earlier disease, but can you just talk about how would you incorporate Tinlarebant into your practice alongside current treatments, understanding that those treatments aren't approved in the E.U.?
Sure. So look, again, it would be early. I don't think we'd need to be waiting for significant atrophy. And it would be throughout life. So if it's well-tolerated, no other drug interactions, then, you know, it'd be throughout life.
There are other treatments in development in GA as well. I think someone in the audience was asking about optogenetic cell therapy. Are you familiar with that approach, and do you have any thoughts around that program?
You know, not specifically. I'm aware that there are groups developing RPE transplantation approaches for GA, as they have also done in the past for Stargardt disease. You know, I understand the rationale. I'm skeptical about whether it will be effective, and whether it's really feasible and scalable. But I am aware of it. I wonder whether there'll also need to be some photoreceptor transplantation in addition to the RPE, which of course, will complicate matters. And there are also genetic therapies in development for dry AMD, again, generally targeting complement components. But I'm a little bit concerned about complement targeting, given the development of wet AMD.
And there's, again, kind of an unclear relationship between complement and wet and dry AMD, so less excited about that.
Are there any other approaches that you find exciting in GA that you're keeping an eye on?
So, you know, I really like the drug therapies. I think we are, as a specialty, getting a little bit tired of intravitreal approaches. It'll be good for one and done. You know, if there are intravitreally delivered gene therapy, that would be scalable and high throughput. And if it's not associated with inflammation with the intravitreal approach, then that, I could see that that would be very desirable as a one and done high throughput type of approach.
Maybe to close off, Nathan and Tom, do you want to highlight sort of the next steps in your programs? Dr. Michaelides, thank you so much again for joining us today.
Great pleasure. Great pleasure.
Yeah, I'll let Tom-
Oh, you want me to take it?
Yeah, go ahead.
Sure, sure, sure. So, as you know, we've completed the DRAGON1 study. We have just initiated the phase, the DRAGON2. This is more so for Japanese subjects, even though we're extending this to the U.S., where as many sites have came to ask us and requested that they join the study because they didn't have the opportunity to put their patients onto the DRAGON1. So there's many other subjects that would like to join this study, and we've kind of extended that to the U.S., certain sites in the U.S. as well, and I believe in the U.K. as well, where Dr. Michaelides have more subjects that they are keen to join the study.
So the DRAGON2 study has now initiated, and as for the AMD study, the GA, I think we've got more sites to initiate, but so far we're making good progress. We've just basically recruited around 100 subjects as of today, and the rate of recruitment is picking up quite quickly. So we're making tremendous progress and of course I've mentioned on the other, in our earnings call, that we're expecting our interim analysis for the DRAGON1 study to read out in December or maybe early January, February, given when the data cleaning completes and when the DSMB can convene.
Okay, great. So we are at the top of our hour. I wanna thank all of our speakers here today for joining us. And to those in the audience, if you have any extra questions, feel free to email me, and I can reach back out. All right. Thanks, everyone.
Thank you so much. Thank you so much. Thank you so much, Dr. Michaelides.