All right. Good day, everyone. Welcome to day one of Cantor's Global Healthcare Conference. My name is Prakhar Agrawal. I am a Biotech Analyst at Cantor, and for our next session, we are very excited to host the team of Shattuck. Representing Shattuck, we have CEO Taylor. Taylor, thank you for joining us today.
My pleasure. Thank you, Prakhar.
Maybe for those of you who are not familiar with the story, maybe just start off with a brief overview of the company and some of the key priorities over the next one-two years.
Sure. Shattuck is exclusively focused on blocking an immune receptor known as DR3. Many people will be more familiar with the sole ligand for DR3, which is known as TL1A. There are a number of companies that have developed TL1A-blocking antibodies, the leaders of which these days are Merck, Roche, and Sanofi. The enthusiasm for the axis has come initially from phase II efficacy data in ulcerative colitis and Crohn's disease for all three of those antibodies that was suggestive that blockade of this axis could be best in disease, better than the other approved therapies there. Now we have evidence from Merck that the class is active, not just in IBD, but in other diseases with the first evidence of activity in a skin condition called hidradenitis suppurativa.
We, to this day, are the only company with a disclosed antibody that is blocking the receptor for TL1A, known as DR3. There are two fundamental advantages of going after the receptor that could lead to higher efficacy. One derives from the fact that DR3 is the more stably expressed target in the axis. The second is due to the fact that all anti-TL1A antibodies have extremely high rates of anti-drug antibodies, and those are secondary to TL1A as a target, not the antibody engineering.
Great. Did want to touch on both these differentiating attributes, but maybe why has nobody else developed a DR3 antibody? Why is everyone going after TL1A?
Yeah. It's a very common question we get. I think there's two fundamental reasons. One is that the academic history of folks that have studied this axis for a long period of time is quite narrow. The first company that developed a TL1A-blocking antibody back in the early 2000s was Human Genome Sciences. When they did that, they sent some of that antibody to two academicians. One was Steph Targan, the other was Fabio Cominelli. Steph Targan was later the scientific founder of Prometheus, and I think it's quite likely that his experience initially with that HGS antibody contributed to the choice to focus continually on TL1A. Also, the proof of concept, although today it seems like almost everybody knows about this axis, it came very recently.
Up until the Prometheus phase II data just a few years ago, all we had was a failed clinical trial of Teva's antibody in asthma. The other reason, obviously, we see this all the time, once something works, there's a sort of a herd mentality to focus on what works. The other reason is that it's a much more significant engineering challenge to build a good DR3 blocking antibody relative to TL1A. The reason for that is that when you're trying to build a TL1A blocking antibody, all it needs to do is provide steric hindrance of TL1A binding to DR3.
When you are building a DR3 blocking antibody, because your antibody can cross-link two subunits of DR3 and DR3 itself can signal, if you get the wrong epitope, the wrong affinity, the wrong Fc engineering, your antibody will cause residual agonism of DR3, and you have to engineer around that very carefully.
Okay. Maybe on the efficacy aspect, why would targeting the DR3 receptor lead to differentiated efficacy, and is it relevant for certain diseases?
Yeah. Our view, first of all, is that everywhere TL1A antibodies work, DR3 blockade will work better. One aspect relates to target stability. I alluded to this, but DR3 is constitutively expressed by lymphocytes and endothelial cells. When a lymphocyte has undergone a fate decision where it decides to turn on DR3, that cell never turns DR3 off again. You can block it and have it stay blocked. TL1A, on the other hand, is not constitutively expressed. It is an inducible ligand by tissue-resident antigen-presenting cells. It gets turned on and off in short pulses. What that means is that when you are trying to block TL1A, you are chasing a moving target.
This is why all of the TL1A antibodies are moving very high doses forward in their phase III clinical trials, because you need to keep a very high concentration of the serum for there to be a gradient that actively drives lots of antibody into tissue to sop up that rapidly cycling target. That is simply a kinetic battle you do not have to fight when you have instead chose to block DR3. The other reason relates to ADA. As I mentioned, all of these antibodies have very high rates of ADA. We have published data that leads to accelerated clearance over time, and with accelerated clearance goes efficacy. This is the reason why none of the TL1A antibodies show accumulating clinical remission from induction to maintenance.
Right. On that aspect, a lot of the TL1A companies will say that they can dose past the ADAs or they can dose high to get to the efficacy level so that around the issue, around the target stability, you can just dose higher.
Yeah.
What is your view on that?
Yeah. I think it is worth considering how the TNF inhibitor space has evolved over the last more than a decade. The reason that the TNF class is a relevant benchmark here is that TNF-alpha circulates as a homotrimeric protein, just like TL1A. Therefore, the source of immunogenicity for anti-TNF and anti-TL1A antibodies is both secondary to creating these large circulating immune complexes. What do we know today? What does a gastroenterologist generally do when they have a patient who has been benefiting from a TNF inhibitor and then starts to lose response?
The first thing they do is they say, "Okay, I can probably double the dose or double the frequency of the TNF inhibitor, and I know that is going to buy me a little bit of time for my patient." But the minute they make that move, they also begin planning for the next class of therapy for the patient because they know it will not last. ADA titers continue to increase, and they lead to other adverse events like IRRs, delayed-type hypersensitivity, all those things. The only other way to get around it is to put a patient on methotrexate. Patients do not want to do that, especially if there is a simple alternative. I think there is no reason to believe that clinical experience won't predict the clinical experience with TL1As, again, because the source of immunogenicity is identical.
Okay. Maybe focusing on SL-325 and the phase I SAD/MAD data that you presented earlier in the year. It's healthy volunteers, but there were some interesting highlights from that data set. Maybe just walk from your perspective, what are the key takeaways?
Yeah. I think first, we had what we believed to be convincing evidence, in vitro and in then monkeys, that this antibody did not have any residual DR3 agonist liabilities. Many people understandably want to see that play out in humans. Our data showed that there wasn't any evidence of this in any patient at any dose at any time. So that question is once and for all put to bed. 325 is a pure DR3 antagonist. The second question is we have in our pipeline, in addition to 325, an antibody we call 425. The only difference between 325 and 425 is that 425 has a YTE mutation to extend the half-life. The reason we did that is we didn't know before generating phase I data what the PK and receptor occupancy profile of 325 would be.
One of the criteria that we used to select these antibodies was that they bound an epitope on DR3 that did not cause internalization of DR3. There was reason to believe that that would lead to highly durable receptor occupancy, but we had to show it. What the phase I data showed is, number one, we saturated DR3 at a very low absolute dose, 0.1 mg/kg . That by the time we got to 1 mg/kg , still a very low dose, that occupancy was stable for three-month intervals. That means that there's a very high probability that 325 is an antibody, even without a half-life extending mutation, that could be dosed at quarterly intervals. These days, we're thinking more about 425 as indication splitting or life cycle management than as a replacement for 325. But we didn't know that before we had the phase I data.
The last thing is the immunogenicity piece. So DR3, as I mentioned, isn't shed, so there's not immune complex formation risk related to ADA the way there is for TL1A. But for any antibody, you still have primary sequence-related immunogenicity risk, and we needed to show what that was. Our ADA rate is 3.7%, the lowest number that any anti-TL1A antibody has put up is 48%, and that's probably an underestimate because of parameters of the ADA assay. Now we know it's 3.7%, and we've shared not just the number, but for any antibody developer, if you want your onlookers to gain a full sense of whether the number you're reporting is accurate, you need to give additional details.
You need to give the sensitivity of your assay. You need to give the drug tolerance level of your assay. You need to tell people that you have assayed ADA at extended time points when you know that your concentrations of antibody always fell within the dynamic range of your assay. That is what we did with the phase I.
Got it. The one interesting thing that you also highlight is the long PD effect without any half-life extension. Why is there a long PD effect with DR3 antibodies? Is it because the receptors are just not turning over?
That is right. The primary cells that are expressing DR3 are long-lived central memory cells. Nobody has yet measured the exact rate of turnover of DR3 on those cells. The phase I data that we have shared now gives a glimpse that it is clearly very slow. That is probably related to just the natural turnover of DR3 itself and the fact that the cells that are expressing DR3 are themselves long-lived. It also just confirms the fact that when this antibody binds DR3, it does not cause internalization. That is not always the case. If you look at vedolizumab, for example, when it binds alpha 4 beta 7 integrin, it causes internalization of alpha 4 beta 7, and that is fine in vedolizumab's case, but it leads to different pharmacokinetics in some cases.
Okay. You chose to go after Crohn's as your initial indication rather than UC, where a lot of TL1A antibodies are going after. Why Crohn's as the first indication choice?
There's a couple reasons. One is that the translational linkage between single nucleotide polymorphisms in TL1A and disease risk has always been much, much stronger in Crohn's than in ulcerative colitis. These days, we have clinical proof of concept that blockade of the axis is active in both Crohn's and UC. The Crohn's clinical trial landscape is quite a bit less crowded these days than ulcerative colitis. When you talk to GI docs, what they agree upon is that if you show activity in Crohn's, you'll generally get credit for UC and have confidence in a signal in UC, but it doesn't always work the other way around. You also don't need to run phase II trials in both UC and Crohn's to then run phase III trials in both UC and Crohn's. You can look at the development of SKYRIZI and see that. That's our plan.
We'll show activity in Crohn's and then run phase IIIs in both. The final reason is that, again, related to the potential for DR3 to drive higher efficacy by virtue of being a more stable target. If that hypothesis plays out, we would expect our ability to detect that to be greater in diseases that have more foci of inflammation.
Right.
More the case in Crohn's.
Okay. Maybe just start with a quick overview of the trial design, and any differences versus all the competitor Crohn's trials.
This is a three active arm, high dose, middle dose, low dose versus placebo, phase II-B trial design, randomized, blinded. The primary endpoint is endoscopic response at the end of induction. Clinical response will be a secondary endpoint. From that perspective, it is sort of a right down the middle phase II design. One place where it is a little different is it is very common these days to do what are known as re-randomization designs in phase II, where the only folks that go on to maintenance therapy are those that achieve clinical response and induction, and those responders get re-randomized. This obscures the benefit that is being seen from induction to maintenance because you report benefit on a percentage basis, but there has been a reduction in the denominator of patients from induction to maintenance.
We are not doing that. We are doing a treat-through design from induction to maintenance, and that has two benefits. One is it maintains power to show efficacy at the maintenance endpoint, and it also facilitates a clean ITT analysis of who is benefiting and do you see accumulating response over time.
Got it. In terms of the formulation that you are using, this is a subQ. If you just expand on the injection volume and frequency for this drug.
Sure. The phase II is an all IV study.
Okay.
We have a subQ formulation. It's about 150 mg/mL . While we are running the phase II study in Crohn's disease, we will also complete a phase I IV versus subQ bioavailability study using that subQ formulation. We've done that in monkeys, and we know that the bioavailability of the subQ formulation is 80% or more of the IV. We haven't commented specifically on the doses themselves in the phase II, but what I can tell you is that we expect to maintain complete DR3 occupancy throughout the dosing interval at all three doses that are moving into phase II. The middle dose is at the high end of what we can feasibly give in a subcutaneous auto-injector in less than a 2 mL volume. The low dose is a way lower dose than that.
The low dose was picked because the trough concentration at the end of the dosing interval butts up right against the concentration that we know is required to maintain full receptor occupancy. We expect that the exposure response data between that low and the middle dose will inform what the maximal schedule is that we can give in the subsequent phase III. The high dose is a much higher dose. It's similar to the high doses that the TL1A developers are moving into phase III. That's just out of an abundance of caution. The worst thing you can do is run a phase II trial and not test high enough, have any risk of leaving efficacy on the table.
Okay. But if the high dose shows differentiated efficacy, could it still be a subcutaneous formulation or?
It would. We'd probably have to go to an on-body device like the TL1As. You could get away with subQ that way.
Okay. Got it. Regarding ADAs and impact on efficacy, do you think induction is the time period where you can show some of that differentiation, or you will need to see the maintenance data because ADAs sort of have this impact on exposure, which takes time, and the maintenance efficacy can fade for some of these drugs.
The impact of ADA is definitely something that accumulates over time. That being said, there is a paper in Lancet Gastroenterology & Hepatology published in 2025 for afimkibart. If you go to the data supplement for that paper, you will find that they broke down exposure levels of afimkibart by anti-drug antibody quartile, and there was an ADA titer-dependent reduction in efficacy for afimkibart. Then they also broke down response at the induction time point by ADA quartile. There was a 50% reduction in efficacy from the first ADA quartile to the fourth ADA quartile. That translates roughly to a 5%-8% placebo-adjusted percentage point benefit being left on the table due to ADA for afimkibart.
Our trial is not powered, especially in a cross-trial comparison, to show that delta relative to that, but there could be trends that are seen. But that delta from induction to maintenance is really where we expect to see a clear differentiation.
Okay. You talked about non-IBD indications as well. Merck had positive data in phase II for HS indication. They've said that they will move ahead with the phase III plan, so probably competitive profile. On the other hand, Spyre has decided not to take their TL1A asset forward in RA, so maybe just talk about the non-IBD indications, which indications are of priority, and when can we hear more?
Yeah. For a long time now, some of you have spoken to us for a while, and HS was at the top of our list of indications that we thought had the highest probability of success, and RA was at the bottom, along with atopic dermatitis. I'm not surprised that it's a negative. The reason that we were expecting this is because the reason that folks went into rheumatoid arthritis is that those folks are known to have very high concentrations of circulating TL1A, but only if you're rheumatoid factor positive. Rheumatoid factor negative patients don't have elevated TL1A. You go back into the TL1A literature and immune complexes are one of the most potent stimulators of TL1A production.
Rheumatoid factor is an immune complex, and so there was always this, in our view, likely fact that the elevation of TL1A in the blood of patients with rheumatoid factor positive RA was effect, not cause of disease. You talk to rheumatologists broadly, and we'd had boards a while back, and what they commented on was that the preclinical data package for the TL1A class in RA reminded them a lot of what they had seen over the years with the IL-1s, the IL-17s, the IL-23s, none of which worked out very well in RA. But they feel quite differently about psoriatic arthritis, and I do too. I think there's a good chance that we see a signal there in HS.
I think the positive data in HS is really just the beginning. I think there will be other indications that work. Psoriatic arthritis would be the next highest probability in my book.
Okay. Where does axSpA fall for axial spondyloarthritis?
It's harder to say because there's much less translational data available in that disease. In an indication like that, you're hanging your hat on other classes of therapy that are tangential to TL1A DR3 that are known to be active.
Okay. I guess you haven't announced any plans to go after these indications. What are you hoping to? What's the next milestone where you make the decision, whether it's HS, whether it's psoriatic arthritis, once we have the data out that you will take your DR3 antibody as well?
Yeah. We're in a position to be part of the lead pack in HS, and that's an opportunity that we don't want to miss. As I alluded to, this is something we've been thinking about for a long time. We've been talking to KOLs for a long time, and it's something that we will have more to say on very soon. Another benefit of HS is, in addition to being part of the lead pack in an indication where I think Merck's going to put up numbers that are better than the IL-17 class, we can move now and really expand the potential of 325.
Okay. Did want to touch on the bispecific as well, DR3/IL23 SL-846. This bispecific, you are going after two receptors. Why go after the IL-23 receptor?
Yeah. Obviously, IL-23 is a validated target in many of the diseases that DR3/TL1A inhibition either is known or expected to be validated. When thinking about IL-23 versus IL-23 receptor, we know through the icotrokinra experience that both sides of the axes can be targeted and lead to efficacy. If you look at the cells that express IL-23 receptor, they all express DR3. You can find DR3 expressing cells that do not express IL-23 receptor, but all IL-23 receptor expressing cells that we look at co-express DR3. So that leads to a cis binding hypothesis for a co-receptor targeted by a specific antibody that provides an avidity benefit. It also raises the question, if you decided to go after DR3 by the cytokine, is it a good idea to create a local concentration of IL-23 near IL-23 receptor expressing cells?
I would suggest that it is not a good idea to do that. These are logical targets that have utility clearly in IBD, but also outside of IBD, in psoriatic arthritis, for example.
Got it. How much is the overlap?
In expression?
Yeah.
IL-23 receptor is a bit less abundant overall than DR3, but as I said, co-expressed with DR3. We think that the phase I data showing the absolute dose levels at which DR3 was saturated are highly instructive for when that will occur with a bispecific. You would expect that if you saturate DR3 at 0.1 mg/ kg in a bispecific, you are certainly not looking at more than 0.3 mg/ kg to saturate both targets.
Okay. How does this bispecific format de-risk some of the agonism risk? We talked about the DR3, but here you are talking about two receptors, if you can just talk about that based on your preclinical work.
Yeah. First, as it relates to DR3 itself, one half of 846 is 325. Same epitope, same affinity, same Fc engineering. But when you're bringing two receptors right next to each other, sometimes funny things can happen. The DR3 by 23R bispecific isn't the only one that we've worked on. We have found examples where you bring DR3 next to another receptor, and all of a sudden, those cytoplasmic domains tickle one another, and you see unexpected signaling. That didn't happen with DR3 by IL-23R, so we don't see any risk from that perspective with the combo. We also just finished our chronic tox study with 846, and we'll be sharing those data at UEGW next month. Everybody will be able to see that for themselves.
Right. Maybe if you can talk about that chronic tox study, how long was the study and is it NHP?
Yeah, this is a cynomolgus macaque, 13-week chronic tox study with a four-week recovery period. Repeat dosing, looking at safety, PK, receptor occupancy, IL-23 inhibition, ADA, all the typical things.
Right. For your DR3 antibody, that NHP study was very predictive on ADA rates. What are you hoping to see on 846?
Yeah. I think we, first of all, believe that sort of that translatability of ADA rates from monkeys to humans for 325 should be helpful when folks look at this data. We had about a third as many ADAs in humans for 325 as we did in monkeys. For any bispecific, you do expect a little bit more ADA than you do with a parental monoclonal. But it shouldn't be much more.
Yeah.
I think this should be compared to the TL1A-directed bispecifics, which, as a group, are even more immunogenic than the TL1A monoclonals. There is no difference in what are being called the second gen TL1A bispecifics from the first gen, the Roche and the Amgen TL1A bispecifics. They all interacted with TL1A in a one-to-one fashion. None of them are two by two or one by two format bispecifics.
Got it. Also, will we have enough data from this chronic tox study on the different lymphocyte subsets, cell proliferation markers, to basically de-risk the residual agonism that we could potentially see with a DR3 antibody? You feel confident that you have de-risked that residual agonism risk?
We do. It certainly translated cleanly for 325, and we've measured things in much the same way for 846.
Okay. Great. What indications could you prioritize for 846?
I hope and expect that Spyre shares data later this year that shows that the axis is active in psoriatic arthritis, because that would be a great indication to go into. It's a disease where IL-23 inhibitors are widely pardon me. Excuse me, widely believed to be underdosed. If it's an indication where TL1A DR3 inhibition is also shown to have additional activity, then you could even consider study designs with active controls in an indication like that.
Got it. I know you're working on other bispecifics as well. What other targets are interesting, and is it going to be focused on the DR3 as kind of your base?
More to come on that. First of all, we have other bispecifics. We believe that this first clinical data with 846 will inform the strategy we've used to make bispecifics. Is it sufficiently non-immunogenic? If so, we'll move some of those other combos forward.
Okay. That's great. That's all the time we have so far. But really appreciate Taylor for a great overview, and thank you so much for joining us.
Thank you.