Welcome, everybody, once again for the 45th Annual TD Cowen Healthcare Conference. I'm Yaron Werber from the Biotech team, along with my colleague Jane Ahan. It's a great pleasure to introduce the Fireside Chat with Stoke Therapeutics. With us today, really needs no introduction, Ed Kay, CEO. Ed, great to see you. Thanks for coming.
Thank you.
Appreciate it. Lots to talk about. We did a panel yesterday on epilepsy and the KOLs. We actually ran a physician survey. The KOLs were we spent a lot of time on STK-001 or zorevunersen. There were really glowing reviews about the data, the phase III trial design. It was chosen also in the physician survey as the most promising therapy in development right now for all epilepsies, essentially. It was really great to see. Can we talk about you announced that you selected Biogen as a partner. What was the reason behind that? You also have a unique you're going to be running the studies. They're going to be contributing. Maybe talk about that decision, too.
Yeah. One of the first decisions we had to decide upon was, should we do it alone for global commercialization? Kind of looked at what are the precedents and talked to several colleagues. Even thinking back in my own experience being at Genzyme, it took us about 15 years to get into about 60 different countries. It was a lot of money and a lot of time. I think if you look at the precedent, I have talked with former colleagues from Bluebird or from Medicine's Company. They said, you know, in retrospect, we wish we would have never done the rest of the world ourselves. It just took too long.
I think one of my concerns had been is we believe that once this medicine gets on the market, there's going to be such a clamor and demand for this that we wanted to make sure that it could get to patients. Once we decided that we really wanted to have a rest of the world partner, we were very determined that we were going to keep it in the U.S. ourselves. We feel very confident. Jason Hoitt, who's sitting here, he and I had worked at Sarepta. Jason did a great job in commercializing Exondys 51 in the U.S.. We knew we could do that. The question is, who's the right partner? We were fortunate that we had a number of different suitors that were interested in the rest of the world rights.
We chose Biogen primarily because of their experience in neurology, and especially their experience with Spinraza. That was very successful. They are in 71 countries. They have had a great launch and a lot of success with that. It made perfect sense for us to partner with somebody that already knows how to do it and really understands intrathecal antisense oligonucleotide therapies. The focus was really, though, we wanted to make sure that we wanted to keep that U.S., but one important thing that I think people are appreciating is that it also gives us cash all the way a runway down to the middle of 2028. That allows us to not have to worry about financing, focus on the study design and the execution of that. It allows us also to focus on the commercialization in the U.S., and our pipeline.
How do you, in terms of the phase III trial conduct, how do you split the role between you and Biogen?
We are in charge of the global execution. They were involved in the approval of the global development plan. That was early in our discussions for the partnership. No, we are responsible. Stoke's responsible for doing all of the work for the execution on the trials.
Do they have any responsibility on the regulatory side?
One of the advantages, of course, that they have is tremendous regulatory experience, but also pricing and market access. Once it gets to that level, they will be very much involved in the rest of the world. That is where they have their expertise.
Right. And you're solely in charge of the U.S.?
That's correct. Absolutely.
There is a joint steering committee and a joint development committee?
Yes. There will also be a joint manufacturing committee because they will be supplying drug in their territories.
They will be supplying drug from their plants?
From their plants. That's correct.
You're going to be producing in the U.S., through a contract manufacturer?
Yeah, we do. We have a contract manufacturer, and then w e have both for drug substance and drug product. We also have backup manufacturers.
For the phase III trial, it's all going to be coming from your vendor?
That's right. We already have supply. We have manufacturer and we have supply waiting to go on the trial.
Just remind us, the phase III trial design was discussed with the PMDA. Biogen will be able to file in Japan with this data?
Absolutely. Yes.
Let's maybe just talk a little bit of some of the highlights in the phase I/II . You just showed us the updated eight months data that showed an 87% reduction in durable seizures. This is the latest time point. This is the patients who had two to three maintenance doses, right, of 45.
That's correct.
Those patients got one or two doses of the 70 mg?
Yes, that's correct.
In the phase III, you decided to do two 70 mg, right, and not three at the end?
That's right.
They're going to get it at the very least, everybody gets one more dose. I'm just trying to.
Yeah. So what we'll have is we had done in the study in the U.K., we had done two or three doses of 70 mg. We didn't see any significant difference between two or three loading doses. So we will be using two doses. It will be times zero and then two months later. And then every four months, they will be receiving 45 mg.
The phase I/ II data that we saw, patients, I mean, in the phase III, before you do the primary endpoint at 52 weeks, patients are going to get two 70s and two 45s?
That's correct.
With no time lag between the doses. With the phase I/II , can you talk about just how that dosing worked out?
Remember, this was initial dosing. It was both we had a single ascending dose in the U.S., and then we had multiple ascending doses in the U.K. What we had is a six-month timeline after the last dose. That was more for a safety review. In this dose, we will be recapitulating what we expect to do from a commercial standpoint. That will be two loading doses. Every four months, they will be receiving the 45 mg maintenance. We will do exactly what we plan to do outside of the clinical trial in the commercial realm.
The data also had a censoring requirement as part of the phase I/II , right? Excuse me. When patients actually needed to add an anti-seizure med, that had to get censored. How would that work in the phase III?
It would be very similar. One of the things that obviously what we found is that by really having careful inclusion and exclusion, we had a population that for the most part, the physicians tried very hard not to have to change the medications. Obviously, in this population, it's a very sick population. They have intractable epilepsy. There's going to be some patients that make some changes. We have, with a 150-patient study that we've announced, we will be very well overpowered for the primary endpoint of reduction in seizures. We're not too worried about censoring the data. That would be we want to make sure that the effect that we're seeing in reduction of seizures is not due to the addition of another medicine or changing another medicine, but it's actually related to zorevunersen.
The phase I/II , when you look at the data, what was your main conclusion about the seizure effect and how fast is the onset?
It's different than all of the other anti-seizure medications. I think, again, this is the first genetic therapy that's addressing the underlying cause. One of the things that happens is that most anti-seizure medications have an effect literally within days to a week or two, right? They have a very rapid onset. That's great. Our effect is very, very different. We're upregulating the NaV1.1 protein. We know that it takes a few weeks before that level of protein gets upregulated into the cells. It has to have an effect. That effect is really there's a reorganization of the brain that occurs once that NaV1.1 gets up to its normal level. When we look at the effect, it takes about three months before we really see a significant effect.
We see the maximum effect over between three and six months. It is very different than all of the other anticonvulsants that are there. The other thing that is really important is that if you look at Dravet patients, and I can speak from my own experience years ago when I would take care of these patients, you were always recycling medicine one after another, adding, subtracting, increasing the dose. It was really challenging to take care of these patients. One of the things that I was very pleased about is that what we are seeing is once these patients get on the medication, they actually get better over time. They are not coming off the medicine. Most of the patients are staying on. They have that long-term effect.
The longer they're on the medicine, it seems to work even better, which is completely opposite of all of the current therapies that are available. I think what we were even more pleased about and what we had hoped to see is that the improvements in cognition and behavior, again, starting around nine months- 12 months. That was, I think, to us, that we had assumed that this would happen because we were addressing the underlying genetic cause. I think one thing that was really that we learned a lot from the phase I/ phase II, the assumption was initially that the younger patients would do the best, right? Because they had the disease for less time. They would be more responsive. There's more plasticity. I think Dravet syndrome is really a disorder of neural connections, of synaptic connections.
What we realized is that over time, when those synaptic connections kind of reformed and that remapping of the brain is occurring, even the older patients were having very significant improvements in cognition and behavior. That was very, very helpful for us because that population also needs to be addressed. We know that a significant portion of patients are adults. What we've heard from the community and from the physicians and from the patients, they want a medicine for all of the patients, not just for the young children. I think what we're seeing is that we're seeing a really nice response in young patients, but we're also seeing a very nice response in the older patients. We will be doing studies in adults specifically because we feel that that's a population that needs to be addressed.
We are going to be looking at that as part of our development plan. We are working with Pedco to do infants from less than one year, less than two years of age. That is part of our development plan also.
What about the one element that we got questions on is we did see CSF protein elevations. Can you talk about what those are and how was there any concomitant NfL elevations? Were they inflammatory at all or any clinical sequelae?
Yeah, no, that was an important question. I think if you look at a lot of other intrathecal medications, whether it's Spinraza, tofersen, this has been seen with intrathecal administration of ASO. This is not unique to our product. It has been seen. We have talked with other companies about this. Occasionally, it is also seen even if you give placebo, artificial CSF. There seems to be something about the irritation of the meninges surrounding that intrathecal administration that has an effect. We have looked very carefully. There is no evidence of inflammation. There are no cells. There is no IgG. There is monoclonal gammopathy. Nothing to suggest that this is an inflammatory response. There seems to be some leakage that occurs. This is really blood proteins, primarily albumin that we see. The important thing to remember, though, is this at this point is a laboratory finding.
It's not associated with any clinical syndrome. We've not seen any evidence of hydrocephalus or anything significant. No headaches, nothing. It's quite variable, the response. Sometimes it'll go away. Sometimes it'll stay in. We've not seen any significant problems. I think very early in the course, we had one patient that had discontinued because it was unclear what the significance was. Patients have not discontinued. We've continued to dose patients. We know that sometimes it goes down spontaneously.
Shana, over to you.
Great. Thanks, Yaron. You touched on it earlier, but I wanted to really dig into kind of the details of your phase III trial design. You know the phase III EMPEROR trial. You're going to test two loading doses, two maintenance, or two loading doses of 70, two maintenance doses of 45 versus sham with the primary endpoint at 28 weeks on motor seizures and then some secondary endpoints per Vineland-3 and behavior at 52 weeks. I was curious, why did you decide on a sham control for this trial? What are the advantages and disadvantages of kind of using sham versus placebo?
OK. It was really with discussions with the regulatory authorities that had requested that it would be a sham. There really is no difference between a sham and a placebo per se. One of the things that was critical that we had to do, though, when we had discussions with regulatory authorities, we wanted to make sure that we wouldn't accidentally unblind the study. One of the ways to unblind the study is that if you don't do a lumbar puncture and remove some fluid, the most common adverse event that we saw in the study was related to the lumbar puncture and the post-lumbar puncture syndrome, which is often headache, sometimes vomiting. If we didn't do a lumbar puncture in the placebo group or the sham group, the problem with that, we could easily unblind that because you'd have a very different side effect profile.
People would know whether or not they're on a sham or whether they're actually getting the drug. The regulatory authorities agreed with us that that was important because we wanted to make sure that we really tightly controlled this study. I think it was a lot of discussions certainly with the patient groups and with investigators. They initially had hope that, well, can't we just—you did a natural history. Can't we compare it to the natural history? We had the discussions with the regulatory authorities in all three regions. They did not agree. They wanted to have really a placebo group, a control group at the same time. We explained that that was absolutely essential. The other thing in my experience is not only do we have to get regulatory approval, but we have to get reimbursement.
Let me just say, I've tried it both ways. I've tried having accelerated approval or with historical controls. That's fine for approval. Let me tell you, it is a hard road to get reimbursed because the payers want that placebo control. The best way for us to get the medicine to people as quickly as possible is not only to get it approved in the country, but also to get it paid for. The best way for that is to have either a sham. You have to have a control group. Even though it'd be great to do it faster and better if you could get it out to patients, if you don't have the quality data at the end, you'll never get reimbursement.
Right. That makes a lot of sense. I know you said that the study is sufficiently powered kind of to see an effect for zorevunersen. Can you give us a little more detail on the powering of the primary and secondary endpoints and kind of what placebo rate you assumed and how you arrived at that?
Sure. I would say we're extremely overpowered for the primary endpoint of seizure reduction. The reason for that is it's powered for the secondary endpoints. We believe very strongly that we need to have the secondary endpoints in the label because that will be critical for reimbursement. What we looked at is the nice feature about we had announced alignment with the three regulatory agencies for the study design. Everybody agreed that the Vineland-3 would be an appropriate endpoint. That was a big win for us that we were able to get that. What we're including is key secondary endpoints for the Vineland-3. This is both for the composite and also looking at various different domains.
What we've done, and again, trying to reassure that we would have a positive study, is if we look at previous studies that have been done in Dravet syndrome, they have anywhere from a 5%-25% placebo effect. We chose the highest placebo effect of 25%, even though we feel that it's highly unlikely for a one-year study we would actually have a 25% placebo effect. We've used the least responsive domain in the Vineland-3 for that 25% placebo effect. We've made sure that we're powered the study that we reached statistical significance on those key secondary endpoints.
Great. I want to get a little deeper into the secondaries. Just before that, where do you kind of see the threshold for seizure reduction, the primary endpoint in phase III for kind of approval and uptake?
Historically, many drugs for anti-seizure medicines have been approved with a reduction of about 30%-40%. With our 70%-80% reduction that we're seeing, we think that is obviously very significant. I remember it's important that the 70%-80% reduction in seizures that we are seeing is on top of the best medications available. We've seen, and as what we're seeing is there are many patients now that have very long periods of seizure freedom. That's, again, not something that's typically been seen. We expect that given our historical reduction in seizures and the duration of effect that's happening there, we feel very confident that that reduction in seizures for the primary endpoint will be hit. The question is, is making sure that we hit the secondary endpoints.
Right. On these secondary endpoints, I know the Vineland-3 has a number of subdomains that you were mentioning, kind of communication and personal relationships, coping, personal skills. Among these different domains, which do you think are actually the most important in showing that zorevunersen has kind of disease-modifying properties? What level of improvement do you think you need to show on this endpoint?
One of the things we did before we actually were planning the phase III trial is we looked at what was the clinical significance of some of the changes that we were seeing. It looks like a two to three-point change in these domains in the Vineland is considered to be clinically significant. What was interesting is if you talk to certainly the investigators and also the patients and the advocacy groups, the things that are most important, communication that came up number one, both for expressive and receptive. That is going to be key for us. The coping skills, the personal skills. We knew that because many of these patients function at about a two to three-year-old level, there are certain domains that they would never be able to do just because intellectually they are not at that level of ability.
I think the important thing to remember is what was important to the community as far as improvement is exactly what we're testing in our phase III.
Right. Are you going to be evaluating these domains hierarchically or just as a composite overall for Vineland?
It will be both ways. It depends on the country. Sometimes they want it different ways. Basically, we've designed the study to reach significance in both from a composite standpoint or individual domains.
Right. Is there any reason to think that some of these domains are more likely to change with zorevunersen just based on kind of the bioavailability or kind of the route of administration?
Not that we can tell. I mean, it's an interesting question. One of the things that was, I think, interesting when we first started studying Dravet syndrome, there was a lot of debate. It was like, well, where does Dravet live in the brain? What are the important structures? There's been a large focus on the inhibitory interneurons, which are GABAergic. That is very important for seizure control. Our feeling was, though, if we are going to treat the entire patient, the entire syndrome, we're going to have to get into many other neurons within the central nervous system.
What we learned from studies and looking at our PK modeling in non-human primates, and we studied about 150 NHPs to understand that model, which was confirmed in humans, is that we really needed to get a loading dose to get those deeper structures because we think we even have to get into the midbrain. We get into the basal ganglia. You need to make sure that you get the entire brain bathed in ASO. There seems to be a certain amount that is necessary. That 45 mg dose appears to maintain a level of NaV1.1 production. I think we were fortunate by really studying and really trying to understand upregulation of NaV1.1 because the nice thing is we do not really care where it goes in the brain because we can only upregulate in those cells that normally have the transcript.
One of the concerns about gene therapy, you can't control where that vector is going to go. If you get it in the wrong cell and it's upregulating, you can cause more damage than good. In our case, we can only upregulate if it's naturally occurring in those cells. That gives us a huge advantage from a safety and also from efficacy.
Right. You have it to starting this phase III trial in the second quarter of this year. What activities are gating for starting this trial? Once it starts, what do you think will be the cadence of enrollments?
Yeah. One of the things that's been really helpful is people, the investigators, are very excited. I think we had presented data at the American Epilepsy Society in December. We were the bell of the ball. I mean, people were so excited about the data that everybody wanted to be participating in the study. For me, it was nice because probably about three quarters of the investigators I've either worked with at different academic institutions, many of them were my trainees, and they wanted to work again, which was fun. We have had so much of an interest that we have a number of sites in the U.S., that are competing with each other to try to be the first to enroll and first to give it to patients. There has also been a huge interest in the U.K. We have already done phase II studies there.
We've also had a lot of interest in Europe because they weren't participating in the initial phase I, phase II. Now they want to be part of it. We've had a lot of interest in these studies. I think what we'll see is that it'll be a very competitive process. What we're seeing from the Europeans is they're concerned because it takes longer to start up the sites in Europe than it does in the U.S., or in the U.K. They're worried that they won't be able to enroll enough patients. They're competing against each other to enroll. We'll give guidelines as we go through. Right now, our guideline is that we expect we'll have about a one-year enrollment rate. It could be better, but we'll have to see.
Just curious, why does it take longer to open a site in Europe?
It's just a whole process. We've already started. You have to do a combined process both for the clinical trials, but then also for the ethics committee. There are certain times when you can commit it. There's a phase I. There's a phase II. We're starting the process, but it just takes longer. The U.S., is faster.
Right. Do you think there's any possibility of filing in the U.S., before you finish EMPEROR?
No, because basically we need those 150 patients, and that will be global. Also, one of the things that was interesting that people were pushing us is, well, why don't you just do an accelerated approval and file for seizures and then later amend, do a supplemental NDA and get the cognition and behavior? The problem with that is then we would be considered as just simply an anti-seizure medication. It would be priced and reimbursed at that level. You can't go back and change and then say, oh, now we have a disease modifying. It would be foolish of us and really wouldn't be appropriate for patients because patients are looking for a disease-modifying therapy. That's been the holy grail for epilepsy for years.
What we've heard, when I was in training, I was told the most toxic drug that you can give to a child other than chemotherapy is an anticonvulsant. It's a neurotoxin. Having a genetic therapy that's targeted to the primary cause is really, I think, that this is the future of epilepsy treatment and what's going to happen. Remember, 50% of all epilepsies are genetic. We used to think 20 years ago, we thought we used to refer to them as idiopathic. We used to tease and said, idiopathic means the doctor's an idiot and the patient has a pathology. I think this is where we're headed. Having targeted genetic therapies just makes more and more sense because you reduce the toxicity. One of the problems you have with all of these anti-seizure medicines is they're very sedating.
You're causing a lot of complications related to the medicines. One thing that we have not seen with our medication, we're not seeing any sedation. These people seem to tolerate it very well and are not having sedation. That's a major advancement.
Right. And then assuming approval, do you think that zorevunersen's label is going to be limited to patients who are refractory already to anti-seizure drugs, or do you think it'll see more broad use?
Our expectation is the uptake will be very broad. I think it would be very surprising that we would limit. If you think of the natural history of Dravet, these children are perfectly normal at birth. They suddenly around nine months start to have a febrile seizure. It is not a very brief febrile seizure. It is a complicated febrile seizure. It is long. They often have to go in the hospital. They have to receive medication to stop the seizure. They quickly have more, and they are cycled on medicine. We do expect that these kids will all be on some anticonvulsants before they get put on our therapy. What we have heard from the investigators is they want to use this medicine as early as possible because they are trying to prevent a lot of the cognitive decline that is being seen.
We expect that there's going to be a big uptake. The nice thing is the effect that we're seeing is on top of all the other medicines. What we're trying to record is are patients coming off medicines that they were previously on because very likely they weren't doing much. At some point, kind of the rule of thumb in treating epilepsy is after about three medicines, whatever you're adding is probably not doing anything. At some point, it's just not effective. You can't just, and some of these patients that we're seeing in our study are on six or seven different medications. You can almost guarantee that many of those medications aren't doing anything, but they are causing side effects.
Yeah. What do you think your kind of estimates for ultimate market penetration?
Yeah, I would say that we're not giving specifics, obviously, until we see it. But from what we've heard, certainly from the investigators, we expect that the uptake is going to be very, very pronounced just because there's such an unmet need. What we've heard is that it's not just the seizures. Even though we're seeing profound changes in seizure control, it's the cognition and behavior that people are looking for. That's from the patients, and that's from the physicians. We expect that this would be a very high uptake.
Right.
Quick question. The primary endpoint is at 24 weeks, 24 or 28 weeks, right, on seizures.
That's right.
It is a 52-week study. The study stays blinded the whole time. You analyze the primary endpoint when you hit it at 28 weeks. Do you communicate it to us at that point while the study continues to be blinded?
No, it continues to be blinded because we really need to. One of the things, again, this was designed specifically to kind of enhance the likelihood of success. One of the challenges that we heard about was that it might be difficult for control, other use of medications for longer than six months. That was going to be a bit of a challenge. In order to make sure that we did not have too much of an effect related to medication changes, we had that primary endpoint at 28 weeks. The regulatory authorities wanted to know also what is going to happen at the end of the study at 52. We really have two endpoints. One is probably the least likely to have any change in medications, but then a real-world experience, what happens after 52 weeks.
We need that 52 weeks to be able to see the cognition and behavior.
There's no cognitive official secondaries before 52 weeks?
No.
They start at nine months?
No, we'll be looking at it, but it'll be at the end of the study.
It is at 12 months.
Yes.
Okay. Hierarchically, is it the subdomains at first, or is it the composite first?
It kind of depends on the region. Some prefer one versus another. Basically, we've designed the study to reach statistical significance either way.
Okay. What's the most sensitive endpoint in this population? It's expressive. I guess it depends on age, too, or is it receptive?
Yeah, it is communication, which is the most. It kind of depends on your age. Expressive and receptive language would be considered to be the most sensitive.
They go hand in hand mostly.
They go hand in hand for the most part. That's right.
Terrific. Ed, thanks for coming. I think we're at time. This is really terrific. Congratulations. We'll follow closely.
Thanks, everyone.
Thank you.