All right. Welcome to an afternoon session with Surrozen, with Chuck and Craig from Surrozen here with me this afternoon. Thank you guys so much for joining us. So excited to have you. I've got a whole bunch of questions about your programs, about the competitive programs, about the whole landscape in the ophthalmologic space. But maybe I'll start by just saying welcome, and let us know what's most important to you as we come into the end of the year and look forward to 2026.
So we've given guidance that we'll be filing an IND for an ophthalmology candidate in 2026. We'll refine that timing and give some guidance on the trial design and expectations for data sometime in 2026. But we're really excited to have a multifunctional antibody that could be really a step improvement in visual function and retinal health in patients with diabetic macular edema and wet AMD, and potentially even broader than that. So that's really exciting for the company. We're a Wnt biology company by founding. So we're excited that there's now clinical proof of concept for Wnt as a target in retinopathies. And I think building evidence that combining targets in retinopathies can have additive benefit. So it's a really exciting aspect of the program.
So you mentioned Wnt biology. Obviously, this has been top of mind recently because of the Merck acquisition of EyeBio. But can you maybe walk us through your pipeline a little bit and talk about how your assets are distinguished from the pure-play Wnt that Merck acquired?
So we actually discovered, as you called it, like the description of pure play Wnt. We actually have broad intellectual property around the concept, not just for Frizzled 4, but the target in general, and we licensed our first product to Boehringer Ingelheim, so they're not quite in the clinic yet. We hope they will be in the coming year, but that would be a direct competitor mechanistically to EyeBio, Restoret molecule. We think there are some biophysical properties that might be better, ability to concentrate the molecule more, maybe get more protein in the eye, possibly extend the dosing. We think in vitro that pure play that is licensed to BI is a more potent molecule. Of course, we won't know if that translates to clinical benefit for some time, but we do have a pure play molecule that's out licensed that we would benefit from typical milestones and royalties from.
You know, having Chuck was really expert in crafting a deal there that gave us lots of space to continue to work in Wnt biology in the eye. And so we continued to do that even after out-licensing that and saw that evidence was building that there are clearly multiple factors that contribute to the pathologies of retinopathies. You know, now we know IL-6 does, VEGF does, the Tie2-Ang pathway does. And so we made a molecule that combined what we think as of today are the two largest contributors, which are now Merck has shown agonizing Wnt and obviously VEGF inhibition. So independently, those seem to have the largest clinical benefit. And others have shown with other combinations like with IL-6 that you can achieve additive benefit. So that's SZN-8141, where we'll have a 2026 IND.
And then as people who are here for Victor's discussion about Kodiak have seen, and with Roche's data, there's building evidence that IL-6 is a key contributor to the pathology of retinopathies and even more forward in the eye diseases like uveitic macular edema. So we have a molecule 8143 that this is now getting complicated to describe, but I describe it as trifunctional. So Wnt agonism, VEGF inhibition, and IL-6 inhibition. We haven't given guidance on the IND timing for that or exactly how we would think about positioning that. It could be this is just the best molecule, that these are all important contributors to DME and wet AMD, or it could be that that's really the best molecule for the more inflammatory driven diseases like uveitic macular edema.
Makes sense. Let's return back to that trispecific in a moment, but start with SZN-8141, the VEGF and Wnt agonist bispecific. So obviously, VEGF is a very familiar target in these spaces, mechanistically very well understood. But what's the rationale for pursuing bispecific, especially relative to a combination approach or a combination of monospecific antibodies? Is there a preclinical development that supports bispecific architecture?
You mean it's separately injecting the tumor antibodies?
Co-formulation or however, there's a reason to do it in a bifunctional way.
So, we now have preclinical data that suggests there's synergy in combining them into a single molecule, which I'll tell you a bit more about. But there's also a regulatory rationale for not doing co-formulations, which is if you take an approved product and try to co-formulate it or co-inject it with an investigational product, you now have to do a superiority study. You can't do a non-inferiority study. And if you do a power calculation for unless you were to have like a 10-letter vision improvement, if you were to have a couple of letter vision improvement, what would the sample size be to show statistically significant superiority? You know, it's several-fold, if not ten-fold larger clinical trial than you would need to show non-inferiority. So that's the rationale for a single molecule.
There's also a pricing issue if you want to think about like co-formulating. How much better would it have to be to have double the price for the two, in other words, for your investigational agent to have Eylea-like pricing, but now you've doubled the price to the payer for the co-formulation, so there are a lot of reasons to make one molecule. The really exciting preclinical data is that when you put them into one molecule and you compare them in a mouse model of retinopathy with Eylea alone, the Wnt pure play alone, or indeed injecting them separately into the same eye, the one molecule is about three to ten-fold better than either of those approaches, so there's some synergy going on with the actual configuration of the molecule that we need to continue to try to elucidate.
Not mechanistically elucidated properly, but observable.
Yeah, observable. I think one hypothesis is that like, does it matter if you're binding VEGF in the middle of the eye, sort of floating around in the vitreous? What probably matters is binding VEGF at the retinal vessel. So the Frizzled end of the antibody is binding that receptor on retinal endothelial cells. And obviously, it's potentially capturing VEGF in the middle of the eye, but it's also potentially capturing VEGF right at the site of the vessel, which is sort of where the business of VEGF-induced leakage and pathologic neovascularization is happening.
Okay, great. I think this topic comes up in the context of the EyeBio molecule as well, but I'm curious how you would frame the unmet need in the space right now and what the primary endpoints that people ought to be caring about. It seems like the market is very fixated on duration of therapy, on injection burden reduction, but that's not the pitch you're making with the.
Yeah, I mean, I think for folks who are here for Kodiak, I think Victor said this and I'll repeat it in my own way. I think that's not unimportant from a patient perspective, the treatment burden or the injection burden. But I think retinal drying is really of primary importance. And the opportunity there, and it's why the Vabysmo has done so well, is to improve the patients who have no fluid. Because while no one has done 10-year studies, the belief by retinal specialists is that intraretinal fluid, less so subretinal fluid, is ultimately going to lead to vision loss. So you want to dry the retina as completely as possible. If you look at some of the Vabysmo data in DME, I'm going to pick like 48 weeks. I think 22% of Eylea patients achieved no intraretinal fluid, sorry, at 48 weeks.
Something like 37%-44%, depending on the regimen of the Vabysmo patients, achieved no intraretinal fluid. Complete drying of the retina. One and a half to twofold increase in the percent of patients able to achieve a completely dry retina. That's what we think really matters. We think that 40-ish%, 50-ish% still leaves plenty of room for half of patients still not achieving or more.
For an additional mechanism.
Yes, for an additional mechanism. And that's where the Eylea data, sorry, the Restoret data, Merck EyeBio data is really exciting. This is a magnitude of clinical benefit. And apparently drying, we haven't seen a percent of patients with no IRF, but retinal thickness-wise, vision improvement-wise, it's VEGF-like, which no one else has been able to show with a different mechanism.
Yeah. Well, that was going to be my next question, which is supposing that you have good retinal drying results, if a registrational endpoint is non-inferiority on BCVA, how do you communicate the potential benefit to patients? How do you communicate the long-term benefit of dry retina?
I think there's a formal way that it's done, which is through promotional material. And anyone can go to the Vabysmo website and say you're a healthcare professional and look at the charts that they were approved to show in their promotional material, which shows these percent of patients with no intraretinal fluid. It shows the charts for the kinetics of both vision improvement and retinal drying. And so there's a formal way to communicate it promotionally through the secondary endpoints, which you would have that percent of patients with no intraretinal fluid as a secondary endpoint. There are things like percent of patients with 15-line letter gain in vision, which is where others have been able to show quite a large difference in some different approaches. And so there are secondary endpoints that others have included now into promotional material. There's also just the retina specialist experience.
But as you say, right, everybody assumes, but nobody's demonstrated that having any level of intraretinal fluid is going to lead to worse vision outcomes after 10 years, right?
I'm just saying no one's done a long enough randomized controlled study because these studies are run to two years. I think if you had a panel of retinal specialists up here, I think probably 100% of them would say intraretinal fluid is going to translate to vision loss. I think if you ask them why they use Vabysmo, there's been a formal survey that asked them that. I'll just repeat what that said. What do you choose as your preferred agent for both now wet AMD and DME? The answer is faricimab, Vabysmo. Why? Because it dries the retina better. That's their experience with their own practice. It's what the data shows.
You're referring to the KOL surveys at ASRS and.
Correct. Yes, exactly. Yeah.
But didn't those same surveys say that the biggest unmet need in the wet AMD space was duration of therapy?
They did last year. The most recent one didn't have that as the leading. Yeah.
Interesting.
Yeah. So again, I don't think that's unimportant to patients. Of course, that's important to patients, but I think it's secondary to drying.
Makes sense. So let's talk about the development plan then for the bispecific and move on and talk a little bit about the trispecific as well. How are you thinking about study design as you head into the clinic? And for initial studies, I mean here, we've talked a little bit about potential registrational endpoints and what the bars are from a regulatory perspective. But what are the early endpoints that highlight the mechanistic differences and really demonstrate that you have the ability to bend the curve?
I mean, so there are key elements of the trial design that we're still working on and we'll communicate them as soon as we're firm about what both the design and timing are going to be. But here are some critical elements. Are you going to study DME and wet AMD or just DME? Merck just has data in DME now. So that's clearly there's clinical proof of concept there. There's a mechanistic rationale that's very compelling in wet AMD. We don't have time to get into that for the combination of VEGF inhibition and Wnt agonism. But those are key considerations for us. Do we do both of those populations? Do we do just treatment experienced? Their response, if they were to use Eylea or Vabysmo, is very well characterized. And so you would know what to expect. Treatment experienced who are suboptimal responders or incomplete responders or whatever.
There's no well-accepted definition of that.
What patients are not reaching at?
Less well cared, not dry on Eylea or Vabysmo. It could be really compelling to show a benefit in those patients. And that could be 30%-50% of the market. How do you interpret that unless you have an actual comparator arm in the study? That's another consideration for us if we were to do that. But it's going to look something along the lines of what others have done. A dose escalation phase to identify probably two doses to take forward into an expansion phase where you would look at either DME alone or DME and wet AMD, probably in treatment naive patients. And then you have to decide whether you have an active comparator or not. And the endpoints would be. These would be three months studies with another six months of follow-up.
As you're aware from these central subfield thickness charts and visual acuity charts, like the clinical benefit accrues almost immediately. Yeah, within four weeks, you'll know if it was open label. You would look at 12 weeks if it was a controlled study for the visual acuity changes, retinal thickness changes, percent of patients with these other endpoints I already cited.
Can we talk a little bit more about newly diagnosed versus treatment experienced patients and the pushes and pulls on that sort of decision? I would imagine having treatment experienced patients who are inadequately dried, they have a baseline that you can refer to. We know how many shots they've been getting. We know what their response has been to VEGF alone. Then, as you say, it would be very powerful to be able to say, we took this patient who was static at a baseline and we moved the needle for them.
I think you described it perfectly. Those are the benefits of doing it. The challenge of doing that is you have a run-in period where you have to characterize their lack of response. You have to protocolize the lack of response. Is it five letters of vision loss? You can do that. But now you have a run-in period that's probably three to six months, which means your study is three to six months longer. You're going to have a much higher dropout. You're going to have a much higher screen failure rate during that period. So I don't know exactly how long it would add to the study, probably six months to a year to do that study. So that's a key consideration.
Folks in the TKI space and gene therapy space who didn't necessarily have long run-in periods, but took medical history of patients who had been on therapy for a long time and used that as a base.
They had multiple-month run-in periods to make sure people were Eylea responsive or VEGF responsive.
Yes, although they didn't take medical history for that mostly.
Even just that is a several month addition. I mean, obviously we would like to have not the ultimate answer of the full characterization of the clinical benefit of the product, but some indication of the magnitude of the clinical benefit as quickly as possible. So it could be the strategy is treatment naive, sort of a prototypical study, and then also do a treatment experience study, take the time to do it right, take the time to select the right patients, have the right comparator arm, have enough sample size to really be able to draw some conclusions about it. The one other issue with a treatment experience is there's no regulatory path for treatment experience. So I think treaters would love to see that kind of data, but you're not going to get a label for treatment.
It doesn't feel like the first adoption population.
It could be, although when we ask our clinical advisors, if you saw data in only treatment naive patients, how would you use it? If it was the best drying agent, they'd say I'd use it in all of my patients.
Fair enough. The other aspect I want to make sure we return to is that trispecific with the IL-6 in there as well. You alluded to this earlier, but how should we be thinking about this relative to the bispecific? Is this a next-gen? Is this a parallel agent? Are there different indications where you might suspect it could have a different profile?
Chuck, do you want to talk about that?
Yeah, so it may well be that it's the better agent. But I think what we're initially thinking is potentially take it first into UME. They're shorter trials, quicker to market. You don't have to run these 900 patient registration trials.
Right.
Right. But in the phase 2a, what we would likely do is look at wet AMD and DME patients as well and see what happens. And then if the results are robust, then I think we have a question around capital allocation and do we just prioritize 8143 over SZN-8141?
You alluded to other players that have IL-6s, including Kodiak, but nobody's got a trispecific quite like this. We've got multiple other bispecifics, different combinations of these three agents on different sides of the triangle. What's the incremental benefit here? And again, how do I think about the additive effect of having all three mechanisms?
I think there are a couple of facets to that. There's the biology of Wnt, and then there's the making of Wnt antibodies. The biology of Wnt is quite unique, at least in preclinical models, in both reducing vessel leakage, which VEGF does, but also normalizing the vessels. What you see in animal models of both DME and wet AMD is that there are these regions that lose, it's called vessel dropout in human disease. There are these regions that have lost vessels. And VEGF inhibition does not restore vessels in those regions. Wnt agonism restores vessels in those regions. Is that going to translate to vision gain? We'll have an indication of it from Merck. The biology of that is quite distinct from just IL-6 inhibition or VEGF inhibition, which are reducing vessel leakage effectively. We're obviously going to do both.
It's not trivial trying to make a Wnt antibody, anti-Frizzled antibody. So this is not like make the protein, the extracellular domain, immunize some mice or immunize some camelids and you get a bunch of binders. It took the company years of working on this from its founding, from the preeminent structural biologist in this area to even come up with binders to these Frizzled, in particular Frizzled 4 selective binders. So it's not like making an IL-6 antibody or something else or looking in the literature. So we just look in the literature for things like VEGF binders and IL-6 binders. You can't do that with Wnt. And we have this very broad intellectual property. So I think any antibody-based Frizzled LRP targeted multi-valent multi-specific molecule would infringe our patent.
Makes sense. That includes EyeBio's.
That includes EyeBio's, correct?
Okay. I think we're nearly out of time, but I want to make sure we have just a few words on the very early pipeline. There's another agent, SZN-113, potentially in geographic atrophy. Obviously, there's plenty of unmet need left in that indication. What's the easiest opportunity there and how do you demonstrate benefit in an indication where historically it's been challenging to show functional improvement?
Yeah, I mean, we're going to run out of time to talk about that one. But the reason we didn't prioritize that first are probably the obvious reasons to people who have studied the field. It's potentially a very long time to a clinical endpoint. Even if you have a clinical benefit, it might be accruing very slowly. So it's not like retinopathies where in four weeks you get an idea about whether you see something. The animal models for geographic atrophy are also, I'd say, less. No animal models are truly translatable and perfect, but they're much less perfect and translatable probably than retinopathy models. So we're really excited about that data. We don't have the funding right now to move that forward in either Fuchs or GA, but there's really exciting preclinical data showing, for example, that we're able to protect cells in their outer nuclear layer.
These are photoreceptors, Müller glial cells, RPE cells. It might be all three we're protecting, but we're definitely protecting cells in that outer nuclear layer. Of course, you can't ask a mouse whether it can still see well or not, but we know there's a photoprotective effect in a mouse model of specific damage to that layer.
Interesting. Well, we'll wait to see more development along that front as well. And can't wait to see initial results from your clinical trials.
Thanks for the invitation to be here.
Thanks so much, Chuck.