All right, excellent. Good morning, everyone. Thank you for joining us. I'm super excited to kick off our 2024 conference with the first company. Cameron, thanks for joining. I know you guys had data recently, so why don't you kick it off and we'll jump right into it?
Sure. Yeah, so maybe just to orient everyone, so Spyre Therapeutics, we're working on novel products for IBD. We're starting with optimized versions of what we think are the three best targets in IBD, which are Alpha-4 Beta-7, TL1A, and the P19 subunit of IL-23. And we're advancing these antibodies in a parallel fashion. So there are kind of three to four months apart between these antibodies. As Umar just mentioned, we just reported phase one data on our Alpha-4 Beta-7 molecule. We just dosed our TL1A molecules in our first-in-human study, and we'll report out data in the first half of next year. And then our IL-23 program will start dosing in Q1 of next year and report out data in the second half of next year. So all three programs are progressing in parallel.
Then I think it sets up what should be a very exciting phase two approach here, where we're testing these drugs not just as monotherapies, as extended half-life versions of molecules that we know work, but also developing combinations of these molecules on this extended half-life basis.
Excellent. Definitely want to dig into the data on the Alpha-4 Beta-7 that you guys put out recently. But just ahead of that, on the TL1A, can you remind us? I think you guys had some updates yesterday. Could you remind us? Is it one or two of your TL1A both going forward?
Both. So we chose to take forward two TL1As. We know this is an incredibly valuable target with opportunities both in IBD and outside IBD. And we know that this can also be an immunogenic target as well, antibodies that can develop a higher rate of ADAs. And so we really wanted to ensure that we have a viable molecule going forward. And immunogenicity is something that's hard to de-risk preclinically. So we took two molecules into tox. Both of them had the highest dose tested was the NOAEL for those molecules. And so we are moving both into phase one. We'll look at those data in parallel and then decide which one we want to move forward.
To be clear, there's one name, but there's two programs within it.
That's right. So it's SPY002 is the TL1A program. It's SPY002-091 and SPY002-072 are the two separate molecules, both phase Is.
Got it. OK, excellent and in terms of half-life expectations, I think the slides are now saying greater than 2x, but the preclinicals we're tracking more like in the 3x range. Is that reasonable?
That's right. So we saw 24-day half-lives in non-human primates for our TL1As. That's about double what's seen with the Prometheus and Merck molecule, more than triple what's seen with the Roivant Roche molecule, and substantially more than the Sanofi Teva molecule. I think what we would expect in.
3x Teva, you said.
Yeah, I think in that range, yeah. We didn't actually test that molecule in non-human primates head-to-head against ours, but we know its human half-life is less than 10 days. And whereas the Merck and Roche molecules have about 19-day half-lives in humans, we would expect with a 24-day non-human primate half-life for our molecules, it should translate to more than triple any of the others.
24 days in non-human primates, but in humans, we know Merck is at right around 19-20 days. You guys are expecting something closer to 60 days, something like that?
That's right.
OK, got it. So quarterly dosing possible, as you guys pointed out.
Yeah, we set expectations for the kind of what thresholds we need to accomplish quarterly dosing, which is about a 40-day to 45-day half-life for the molecule. And then we think with about a 65-day half-life, we would be able to dose our TL1As on a twice-annual basis. So those are kind of the thresholds that we think we would need to enable four times a year or twice a year dosing. And we'll see, we think with a 24-day non-human primate half-life, 45 days should be readily achievable. And then the 65 days, we think, should be especially given that we're taking two molecules forward.
60 days, I guess to me, 60-day half-life is 3x Merck. So to me, it sounds like a quarterly regimen, not biannual. I guess how do you guys think about that?
So in general, we have both kind of these extended half-lives, and we also have very high-concentration formulations of these as well. So we have 200 mg/mL subcutaneous formulations, meaning we can put up to 400 milligrams in a single 2-mL autoinjector. And we're looking at kind of the amount of target coverage that we need to get relative to the previous TL1As, also recognizing that our molecules aren't just longer half-lives. They're also substantially more potent, for example, than the Merck molecule. So it has a 19-day half-life, but our molecules also have about tenfold the potency of the Merck molecule. So we don't even need to hit the same level of concentration where we expect to have the same level of target engagement.
Got it. So maybe just digging into that a little more. I remember on the Teva molecule, there's some asthma data suggesting on one induction dose or two induction actually, I take it back. It was three induction doses. And somehow, even though that's a very short half-life, the PD, meaning the TL1A knockdown, was durable for almost 60 days- 80 days out. I guess, are you guys relying on a PD driving this extended dosing interval?
Yeah, it's a great question. So I think what we see in the induction setting for TL1As, we don't really see a dose or exposure response for the TL1As that are out there. And that's across the Roche, Merck, and the Teva molecules. They all move forward with doses that saturated their PD measure. So in the case of Merck and Roche, it was using soluble TL1A levels, and they dosed to where they saw saturation. And then they just went slightly above that in terms of what they took forward into phase two. But then when we see the longer-term data from IBD, which is where we've been predominantly focused on this, in the maintenance setting, they actually do see dose responses where the higher doses tested for both the Merck and the Roche molecule showed that the highest doses tested were the most efficacious.
When you do a dose response and you see the highest dose tested is the most efficacious, you're never sure if greater target engagement could lead to actually better efficacy than that. That's something that we intend to test. Again, we have molecules that we expect to have triple the half-life and, compared to Merck, tenfold the potency. We think that gives us kind of a pretty wide range to be able to test could we get greater target engagement, greater efficacy in that maintenance setting as well.
Excellent, excellent, excellent. I guess since we're on topic and there's an important TL1A readout coming up as well, do you guys have an expectation based on everything you've seen on the Teva molecule? Does it look more or less the same as the Merck and the Roche molecules?
I think what's interesting to us, when we look across the three, they're very different molecules across all three. The Merck one we mentioned, it has the lowest potency of the three, about a log-fold less potent than the other two. But it has a good half-life. It has good bioavailability. It has low immunogenicity. And they see a good result. It's also the only one that binds to the monomeric or trimeric forms of TL1A. The Roche molecule, trimer binder, quite potent, higher rate of immunogenicity, higher rate of ADAs with that molecule, lower concentration formulatability, very different epitope than the Prometheus molecule. Both had great data. So very different molecules, both excellent data. The Teva molecule, yet again, distinct. It has kind of a different epitope. It has good potency. It has a very short half-life as well. So we'll see.
I think we already know two very different molecules work. Ours kind of have what we think is the best of both worlds of those two that work, and so we think we have high expectations that ours should work kind of regardless of what we see from the Teva molecule. I think what's interesting about the Teva study is their dosing is the most distinct. They dose a lot in induction, as I think you've pointed out, over two grams of drug in the induction setting, but then they have relatively short half-life, so they have really high Cmaxes, quite low Cmins. Over a monthly dosing interval, kind of with a seven to 10-day half-life, that's three to four half-lives of the drug over a monthly dosing interval, as much as a 16-fold peak to trough ratio.
You get a lot of coverage at Cmax, a lot less at Cmin. That, I think, is the main question is kind of whether that's the same.
But if the Cmax is orders of magnitude above EC50 anyways, do you think it matters? And could that explain this lack of dose response in the early trials?
I think it's a really interesting question. I mean, there are other targets. I mean, we spend a lot of time on Alpha-4 Beta-7, for example, where even at very low doses of vedolizumab, they saturate the PD measure. And yet they have a really striking exposure response showing that you need a lot more drug than to saturate the PD measure to get maximal efficacy. We don't know that for TL1A yet. So PD is what we have in terms of selecting our doses. But in the long run, you really need these larger data sets to be able to show what the opportunity.
I guess the only data point we do have on TL1As would be the three doses on the Pfizer program, which were relatively flat on dosing.
Which did not separate exactly in the induction setting, and so I would argue from an efficacy perspective, dosing 2.3 grams, 2.25 grams , 2.2 grams is not necessarily enough, or I would doubt you get increased efficacy with that, but we'll see. I mean, it's a much different level than where Roche and.
Got it. I want to transition past TL1A, but just two very quick ones. One, placebo response has been an important question lately. We were sort of noodling with this idea that Teva has a lot of Eastern European sites. Would it matter? But one of the points also that's coming up is Prometheus had 50% plus Eastern sites as well. And they're like a 1% placebo response, which you almost never see. I guess, how do you guys think about that? Because it'll be very relevant for your studies into next year as well.
Yeah, so when we look at the data, which I assume you did a very similar thing, kind of where patients enrolled and what parameters correlated best with placebo responses, the strongest predictors of placebo response in our analysis were really kind of the disease severity and kind of naive level of prior biologic use in patient cohorts, which does correlate well with geography. In Western Europe and the U.S., smaller proportion of naive patients that you enroll in those geographies, higher number of prior biologic or advanced therapy use. In Eastern Europe and other geographies, lower. And you end up with different placebo rates in kind of Western Europe and U.S. versus Eastern Europe. That said, there is quite a bit of variability even among that. And I think you point out the best counterargument to that.
And we've hired a decent chunk of the clinical team that ran the Prometheus studies on both the ClinDev and ClinOps side. And it's more than just where you enroll patients. It's how you train the sites. It's how you talk to patients about what they're putting into their diaries. So we think it's more than just where you're enrolling patients that matters. And hopefully, we can replicate what Prometheus did, which is kind of a leading placebo rate in this field.
Right. OK, excellent. Last one, 200 mg/mL. I think this is the highest concentration you're using for any of your antibodies. Isn't that a very viscous and thick injection?
Yeah, so I think when we're saying what kind of concentrations we've established, that is at a viscosity that we can use in a standard autoinjector. So we have.
Is it Dupi-like viscosity?
I mean, it's sub-20 centipoise for all of these. So meaning that we can get into kind of a normal autoinjector without substantial challenges on kind of injection force required. And that's all six of our products that we've had at that. So that includes our three monotherapies in this kind of low viscosity normal autoinjector utility, as well as our co-formulations as well. So kind of the one-to-one mixes of our antibodies together, we also get at that level.
Excellent, excellent. All right, well, I want to get to Alpha-4 Beta-7 as well. But maybe I think it'll be important just because I'm thinking what's coming next year. There's obviously an important trial you guys have in mind, the Master Protocol on the combination setting. It will have a TL1A. It will have an Alpha-4 Beta-7. It will have an IL-23. The prior combo data we know off of J&J is TNF plus IL-23. I guess, how do you think about the combinations? Makes a lot of sense. But what specific things combine better synergistically, we don't really know. How do you think about that, knowing that you guys don't have a TNF, but you're combining different things? How do you think about that?
Yeah, so we, in general, try to use as many different data sources as we can to kind of support the usage of different combinations and some of the same types of data that were used to support the TNF IL23 combo at J&J. So I tend to start with human genetics, where we see that kind of variants that mimic these drugs have effects that cause this disease, and we see additive effects when you have multiple variants across these different targets, so we know for all three of our combinations, as with TNF and IL23, there's genetic evidence that multiple pathways contribute to the disease, same as seen in animal studies. We also do cell-based models as well, so this is kind of the preclinical or human genetic data that supports each of these combos.
And then I also think it is much more practical to also think about, look, we look at the VEGA result with a TNF and IL-23 and imagine replacing the TNF with an Alpha-4 Beta-7 in that combo. Alpha-4 Beta-7 beats TNF head to head in this population in the VARSITY study on both safety and efficacy. I think that's a logical swap. Same thing is true with replacing a TNF with a TL1A, as you mentioned, it's in the same superfamily as TNF. And the data to date would suggest that TL1A perhaps has even greater efficacy than TNF. And we haven't yet seen the safety effects that we know of for TNF as a class.
Excellent. So OK, got it. And then I guess as I think about sort of the J&J readout into next year, is it reasonable to expect that the type of efficacy doubling we saw in the DUET sorry, in VEGA would actually replicate in DUET as well? I guess, what's the minimum threshold we need to see to say, you know what, combinations do make a lot of sense, like 40% +? Is there a number in mind, or?
Yeah, so I don't think using the same absolute numbers in the refractory population that's being tested in DUET makes sense relative to VEGA. So VEGA was done in patients that were completely naive. They didn't have prior biologic use. And so you had about 25% on each of the monos, almost 50% on the combo. I think in the DUET studies, which are now refractory patient populations, we know what's going to happen with the monotherapies. I mean, this has been tested before. We know the monotherapy remission rates are going to come down substantially. And I would hazard a guess that the combination rate is going to come down substantially as well. So I think the chances that you're over 40% in the combo arm is very low. That said, do I still think we'll see some level of incremental efficacy compared to the monotherapies? I think yes.
Then I think what ultimately we care about is what's the mix of that. We're expecting to run a study that includes both naive patients as well as refractory patients. We have a pretty good understanding from VEGA of what happened in the naive. DUET will give us a good understanding of what happens in the refractory population. And then we'll see kind of if you had a 50/50 mix, what would you expect the delta to be.
Right. I guess the one thing on the Vega trial that did confuse me a bit was there was a slight imbalance in prior immunosuppressant use. And so if you want to be a real skeptic on combos in UC, you could say, you know what, Crohn's was sort of mixed data. UC was the only thing where we saw a signal. Within UC, there's two monotherapy arms and one combo. It's that combo arm with the immunosuppressant imbalance. That's what might have drove the efficacy. I guess, how do you think about that? And what if that were not to be the case?
Yeah, no, I mean, I think one study is one study, and I think that's why everyone in this space is looking at exploring additional combinations in this space with different patient populations and other data. Though I wouldn't say that VEGA was the only thing we have. I mean, there's a long history of using combinations clinically in this space. You can find hundreds of case examples of physicians using it. I think the KOLs and top physicians in this field have experience using combinations in their practice as well and are believers that this will be the future too, so VEGA, though it's the only randomized controlled study that shows it, and you almost doubled the response rates, I doubt an imbalance in something like immunosuppressants would kind of fully remove that benefit that was observed in that study.
I think there are plenty of other data sources suggesting that this is likely going to be the future of this space.
Got it. Excellent. I can't believe we're almost at time. But just in the last two minutes, since it happens to be your lead program, let's talk about Alpha-4 Beta-7. I guess just the thought process behind using the 600 mg dose versus 300 mg dose going forward, how do you think about that? And realizing, obviously, the PK exposure on Entyvio as well.
Yeah, so why we think so Vedolizumab uses a 300 milligram IV dose. They dose it twice at week zero and week two. And what's been observed both in their clinical studies and in the real world is that patients that have the greatest exposure from that dose have the best response. This was highlighted in the FDA review of those data where they said you should test a higher dose functionally. Since our molecule has the same epitope and same potency as Vedolizumab, same selectivity as well, but just a substantially longer half-life, we think it's rational for us to do that, to test that higher exposure. We don't need to do that by actually pushing the Cmax. Vedo has dosed up to 750 milligrams and shown it's safe at that level.
We don't need to go that high to get all subjects in our phase 2 study into the range where Vedolizumab had the maximal efficacy, which is this fourth quartile that we talk about. It's north of about 40 micrograms per mL at week six. That's really what we're targeting to get all of the subjects in our phase 2 study to that level of exposure. Let's see if we can get an improvement in efficacy. We think that it's likely to not have a difference in safety compared to Veto.
Got it. Is it reasonable to assume, Cameron, that if we only look at your like-for-like dose, 300 milligrams from the SAD portion, the PK was still tracking in that Q3, Q4 versus I realize I'm taking mean and comparing versus quartiles. But how do you think about that?
Yeah, so none of our doses that we showed yet in the phase 1 study directly reflect what we're planning to do in phase 2. So it's a little bit hard to say kind of a 300, 300 dose. Did we get everybody into the fourth quartile? The answer is actually yes. In our MAD dose, where we had two 300 milligram subQ doses, you can look there. And we see that all patients in that cohort were actually in the fourth quartile at week six. But that's actually not what we're planning to do. We're planning to use a higher dose and dose it IV. And so we do expect to be able to get everybody into that level of exposure.
Q4 plus, and I guess my last one, we all looked at the mean data. It looked very good. Half-life data was really good. The question I had was, I'm pretty sure you guys looked at patient-level data. How consistent were the PK curves? Because confidence interval is more a reflection of the mean than, let's say, one-offs.
Oh, the confidence interval, I mean, there's six subjects per cohort. So if you had even a single patient that was dropping off dramatically, you would see it in the error bars. So I think at this point, we think we have kind of the normal level of kind of IgG variability between individuals in the cohorts, and kind of we're certainly not seeing something like an ADA effect in anyone.
Excellent. Anything we missed on Spyre just before we decide to wrap up? Could you remind us the timing again for next year's catalyst? I know there's TL1A, but there's also Combo Start, which is after the TL1A readout.
That's right. So we'll have the TL1A data first half of next year, IL23 data second half of next year. And we'll be starting that phase two study mid-year, which will include the monotherapies and the combos in that study.
That data won't be till 2026, the combo data.
That's right. We'll be starting those studies in mid-next year. We expect it to take around 18 months to enroll.
U.S.-centric, the combo trial? Or is it Eastern Europe?
I think that study will be a global study.
It'll be global.
It'll be a global study. It's a large study. To enroll that study, we'll have to go globally.
Excellent. Outstanding. Well, unless there's anything else, thank you so much for making time.
Thank you.