All right. Good afternoon, everyone, and thanks for joining us at the Morgan Stanley Global Healthcare Conference. I'm Mike Goltz. We're the biotech analysts here, and it's my pleasure to introduce Marty Huber, President, from Xilio Therapeutics. Just a reminder, the format for today is a fireside chat, but before we get in the Q&A, I need to read a quick disclaimer. For important disclosures, please see the Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative. With that, Marty, thanks for being here. Maybe just to start off since, you know, not everyone might be familiar with your platform, if you could talk about GPS or your Geographically Precise Solution and how that actually works.
Thank you. It's good to be here. Just to outline, what we believe we've developed is a third-generation approach to solving the problem of immunotherapy. The first-generation approaches were cytokines or antibodies that affect the immune system, but we all recognized there were problems with having effects in the periphery as well as in the tumor. Multiple people have tried to solve this, and the way they've solved this is kind of in these second-generation approaches, they've either applied a mask or some other conditional activation mechanism, or they've modified the primary molecule to have better PK, modified PD, et cetera. What we've taken the approach of is both. How do we make a more potent, more active molecule and then combine it with the GPS so that it's only activated in the tumor using MMPs?
Maybe just you mentioned MMPs, so maybe talk about those a little bit and why those are sort of well suited for this sort of precise solution.
Thank you. There's lots of data in the literature over decades now, showing that matrix metalloproteinases are important. The dysregulation of them is important for tumors for invasion and metastasis. The way that tumors can spread is by using increased activity of MMPs to invade into other spaces. The challenge has been how do you get the molecules to specifically link to those? We think multiple people have tried it now, and while the perfect molecule may not be out there yet, multiple people have shown that you can in fact get activation of MMPs in tumors. We look at some of the data, even the XTX202 data coming out of ESMO showed tumor-specific activation.
Our own data on this topic, while we haven't got our phase 2 data yet, what we're very excited by is we have collaborations with research hospitals that send us tumor samples. We use those tumor samples ex vivo, put our product candidates in with those tumors, and we're able to show activation in most patient tumors, activation of our molecules.
Gotcha. Maybe you can just talk about, you know, your initial two targets or mechanisms. It's a CTLA-4 as well as an IL-2. So maybe, you know, why those two mechanisms for this specific platform?
Well, with the new platform, one of the things we didn't wanna take on board as risk was target risk. You know, that's always challenging if you're engaging the target, but it ultimately doesn't result in antitumor effects. Where we chose CTLA-4 and IL-2 for our first one is both of those have clear clinical validation. We know that ipilimumab at 10 milligrams per kilogram has superior survival to ipilimumab at 3 mg per kilogram. You have a clinically validated need for higher dose and exposure for better outcomes. For interleukin-2, for aldesleukin, we have evidence that you can cure a subset of patients with high-dose aldesleukin. Once again, of both of those targets, there's strong clinical data saying that if you get it into the tumor, you're gonna have just this really profound positive impact on outcomes.
It's now a matter of getting that without the toxicity.
Yeah. Maybe you can just touch on what those toxicities look like and how severe those are and maybe how it's limited the
Yeah
current treatments.
When we look at anti-CTLA-4, if you think of ipilimumab, the main toxicities are you get these immune-related adverse events. Those are endocrinopathies. You get hypoadrenal type activity, severe skin rash and other reactions. The one that we most often think about is you get colitis. It's an inflammation of the colon. In fact, these immune-related adverse events with ipilimumab, for example, in the 10 mg per kilogram dose, a third of patients were unable to complete treatment with monotherapy. Whereas half that or the lower dose of 3 mg per kg, you had about a 15% discontinuation. There we have clear data that those immune-related adverse events resulted in the inability to deliver the maximal dose.
For IL-2, it's a different profile in that we get this thing called vascular leak syndrome, which is essentially the capillaries get leaky, the albumin comes out, and you get hypotension and pulmonary edema. The way you get patients through this is, first of all, you start with very healthy patients, and you treat them in essentially the equivalent of an ICU setting. If you can eliminate that peripheral toxicity so that you don't have to put the patient on vasopressors and treat their pulmonary edema, et cetera.
Gotcha. Two validated targets. You know, targeting the tumor specifically, you know, has the potential to improve the safety profile. Is the goal here to sort of match the efficacy that we've seen with the current products, or is it to try and take it another step further and sort of maximize the efficacy as well?
The ultimate goal is to improve efficacy. We think for example, with CTLA-4, if you could treat patients at that high dose equivalent, we know you'll have better survival. With IL-2, is there potential to go beyond aldesleukin? If more patients can be treated with a more sustained regimen, there is that potential. For now, we're focused on can we get at least that efficacy? I think ultimately we believe there's the opportunity to improve efficacy.
Gotcha. Makes sense. Maybe we can just shift gears to the IL-2 program. That's XTX202. You know, so many companies have tried to develop next generation IL-2s, but, you know, data so far has been somewhat disappointing, you'd probably argue. Maybe talk about, you know, how you can overcome some of those challenges.
That's been one of the. To be frank, this has been a challenge for us. People have looked at the bempegaldesleukin data, I mean, it was presented recently at ASCO, so I think we all are familiar with it. It has sent this kind of negative tone of, well, IL-2 doesn't work. The original aldesleukin data and the excitement around high-dose aldesleukin was, to be frank, back in the nineties. A lot of people have forgotten about that excitement that occurred with those first IO patients. You know, we kinda have to sometimes remind people is high-dose aldesleukin was 0.5-1 mg per kg, you know, mg per kg ranges over four days for a patient.
This is very toxic, but if you got a complete response, which occurred about 6 or 7% of patients, they were cured. When we look at the second generation molecules, and I'm gonna use bempegaldesleukin as an example, but there are other examples out there. The dose they were giving was 6 micrograms per kilogram. So fractions, I mean, orders of magnitude lower doses. As a consequence, you didn't get the toxicity, you didn't see severe hypotension, et cetera, but you also did not see monotherapy efficacy. Just to remind, the response rate for monotherapy for bempegaldesleukin was zero. We did not see PRs or CRs. So the way we're thinking about it is you need to get that high dose intensity to get that anti-tumor effect. With XTX202, our goal is to get into that mg per kg range again.
Mm-hmm.
What we're very excited about from our ongoing clinical trial is, as we recently disclosed, we're already at 0.5 milligram per kilogram. We're already in the lower aldesleukin range.
Yeah.
We're only substantially higher than the doses with any of the second generation molecules.
Gotcha. Maybe just quickly touch on, you know, the not alpha choice versus wild type and, you know, why you decided to go in that direction.
This is another area that creates a lot of interesting debate in the immunology community. The way we've. You know, if you think about it, what alpha is the predominant receptor on IL-2, and the whole purpose of alpha is to engage and essentially activate Tregs. The goal of the alpha is essentially to block IL-2 from causing a profound inflammatory reaction. Otherwise, if you had circulating beta-gamma, you would have toxicity, but without a break. Wild-type is all about the break. Now, the reason high-dose aldesleukin works is because you not only bind the alpha, but you get such high concentrations that you start getting a beta-gamma effect, which gives you CD8 effector cells and NK cells. This is the way you gotta think about the molecule holistically.
Since our goal is to have this thing only activated in the tumor, instead of having the brake be the main signal coming out of it with the alpha, we wanna put the foot on the gas for inflammation with CD8 effector cells and NK cells. We've shown in our preclinical data that we can activate CD8 effector and NK cells without activating Tregs, and that results in meaningful clinical antitumor activity.
Gotcha. Maybe you can just describe some of the preclinical data you've
Sure
Generated and sort of what's most exciting there.
First, what I would like to do is send a quick shout-out and thank you to all my colleagues back in Waltham. We have our protein engineers, protein scientists, and lots of biologists that you did not see hypoalbuminemia. Hypoalbuminemia is important because that is the clinical marker of vascular leak syndrome. It gives us a high degree of confidence. One, it helped the FDA align on giving us a starting dose of 0.27 mg per kg and why we're very confident we're gonna achieve that goal of 1 mg per kg. Then get unmask in the tumors. We recognize that the ultimate test of that is the phase 2 that we're gonna be starting next year. We do have patient data.
We have taken tumors from patients, fresh tumors, have them sent to our center, and then we basically essentially disaggregate them and then incubate them with our molecule, and we measure the rate of cleavage. We know that if we put that in normal human plasma, we see very few samples that will cleave the molecules or at least achieve what we call meaningful. There's always, I mean, there's a little background 1% or 2%.
Yeah.
We don't see meaningful cleavage in the vast majority of plasma samples. If we take samples from the same patients, tumor samples, you can show that the vast majority of solid tumors will cleave a meaningful amount of drug. This is for XTX101 as well as XTX202. That 202 data is really critical in convincing us that once we get the levels high enough in the patient, there is sufficient MMP activity to result in cleaved and active molecule in the tumor.
Makes sense. You touched on the ongoing phase 1/2 study and sort of where you are in dosing. You know, maybe just talk about next steps there and when we should expect the next updates and just how to think about that.
Yep. In our recent disclosure, we kind of spelled out some of our upcoming milestones. For XTX202, the dose escalation is ongoing. Our goal is to be at that milligram per kilogram dose by the end of this year. Then we're planning on reading out the phase 1 dose escalation, 1A, dose escalation data in the first half of 2023. That would be recommended phase 2 dose, PK/PD, schedule, safety, and available antitumor activity. We do recognize this is kind of an all-solid tumor dose escalation, so we're really not putting a primary efficacy readout, but we will present what's there.
The other important piece of data that we're in the process of starting, which is looking for a subset of patients with biopsiable disease so that we can get measurements of PK/PD in the tumors after treatment to show activation of the tumors. This data should be in the second half of 2023. Then the final piece, the definitive proof of concept for efficacy, is a phase 2 study in both melanoma patients and renal cell patients. We plan to initiate that in the first half of 2023 once we have a recommended phase 2 dose. This is critical because this is the proof that you're clearly differentiated from second generation and that you have a monotherapy objective response target in those cohorts.
We have not given guidance when we'll read those out, but we do intend to start those in the first half of next year.
Gotcha. I know there hasn't been sort of CR rates shown yet across, you know, other IL-twos in development. You know, that'd be clearly differentiating, you know, I guess your level of confidence achieving that.
We are confident that we'll be able to deliver objective responses. When you get to the specific CR rate, remember, even with high-dose aldesleukin, it was 6%-7%.
Yep.
It becomes a bit of a numbers game. The other thing is what's shifted a little bit in the world is everyone with melanoma and RCC gets a PD-1. There are data showing that you can get responses in post-PD-1 patients, including CRs. But I think we do wanna be a little cognizant of the world of the 1990s pre-PD-1 CR rates may be a little different than a post-PD-1. But clearly, we expect to see clear objective, at least partial responses. Ideally, if you treat enough patients, we would anticipate seeing CRs as well.
Makes sense. Just, you know, big picture in terms of the strategy, you know, is there a place for monotherapy treatment, or do you think it's more likely to be, you know, combination?
Yeah. What we see monotherapy as a critical step of proof of concept. We do not want to invest in large randomized phase 3 registration trials until we know that.
Yeah.
We do think ultimately the benefit of our approach is going to ultimately be combination. If you think about it, high-dose IL-2 cannot combine with PD-1. You had to go to very low dose. That was one of the challenges Bempeg had, low doses of IL-2 with the nivolumab. Our approach would envision that you should be able to give a full dose to the tumor with good exposure in the tumor of an IL-2 platform, and then combine it with a PD-1, or combine it with a CTLA-4, or even combine it with an IL-12, for example. That's where the masking really once we've shown that monotherapy proof of concept, the ability to put multiple IO agents together is something that's been a serious limitation of most IO therapies to date.
Makes sense. Maybe we can just shift to the CTLA-4 program. That's XTX101. You recently announced some promising phase 1 dose escalation data, and maybe you can just highlight some of the takeaways there.
The details are in disclosure, so I don't wanna spend a lot of time on it, but just kind of the highlights are our goal was to get to a 60 mg dose. Why 60 milligram dose was our molecule in preclinical data is 10 times as potent as ipilimumab. The goal for the best dose to date for ipilimumab, or the most active dose was 10 mg per kilogram. In a 1 mg per kg range for us is on par with that. We cleared 60, and we even went. We're at 180. We're three times what we need to be at. We were very pleased that we were able to achieve that high dose level.
Our current effort is kind of focused on, now that we have the PK and some PD data, really refining that dose to get the right dose somewhere in that range to take forward into phase 2.
Gotcha. Maybe you can talk about sort of next steps from here and when we can expect updates and.
As we're refining the recommended phase 2 dose, that will enable kind of two next steps. One is we have already initiated with our current 60 mg dose, the 1B, which is a PD cohort, where we're gonna get tumor biopsies. That's ongoing, and we anticipate sharing that data mid 2023. In addition, once we have a clear recommended phase 2 dose, we will start the dose escalation in combination with pembrolizumab through our collaboration with Merck. The goal of that will be to show that we can give a solid dose of 101 in combination with PD-1, which has been, you know, with Ipi, we've not been able to do that. That would lead us into our phase 2 programs.
We have not given, you know, a specific phase 2 design yet, but one of the things we're exploring is based on kind of this evolving concept that's happening at CTLA-4. With Agenus, for example, has shown data that with these more potent CTLA-4s you could potentially have activity in MSS colorectal, for example. Ipi and Pembro, et cetera, have not shown meaningful activity there to date, but Agenus in the combination with PD-1 showed a 24% response rate in MSS colorectal. The problem is they are a systemically active Fc-enhanced molecule. With our Fc enhancement, but the masking, our goal would be to have a meaningful response rate with a mitigated toxicity profile in that population to show that's something Ipi can't do. They cannot show a 20+% response rate in that setting.
Okay. Makes sense. Maybe we can shift to 301, that's your tumor-selective IL-12. To start, if you can just talk about, you know, how IL-12 works, and maybe why its use there has been even more challenging versus an IL-2, for example.
I was very fortunate along the way. I got to be the clinical lead for the subQ IL-12 at Roche, and I hate to say how long ago that was. But that drug never made it into registration enabling. The fundamental issue was you could see responses, but the doses necessarily result of responses also caused hepatotoxicity. That's the challenge of IL-12. If you think about why is IL-12 interesting is IL-12 triggers interferon gamma as well as. Essentially, if there's no inflammation going on and IL-12 gets there, it will trigger an inflammatory response. Wherever it touches, it triggers. The reason IL-12 is so exciting is you can literally make a cold tumor warm or a warm tumor hot because you'll stimulate an immune response.
The problem with that for an unmasked IL-12, and that's what we learned in the clinic, was wherever it touches, if it goes into liver, you're gonna get inflammation in the liver and you're gonna get hepatic inflammation and inflammation. Our focus was, this is really the great molecule for our approach is we don't want it touching normal tissue. We only want it touching in the tumor, and our platform seemed to be the right way. We've generated a molecule which is in process. We've now been able to show efficacy down into the microgram per kilogram dose range with this molecule, but at the same time, in non-human primates, be able to get to HNSTDs in the milligram per kilogram range. We think we're seeing the opportunity for a fairly wide therapeutic index.
Based on that, we are in the process of getting ready for an IND, which we'll file in the fourth quarter this year. Once we have that, then get going in the clinic.
Yeah. Makes sense. I know it's early, but do you have a sense of, you know, what tumor indications, you know, IL-12 might be, you know, well suited for at this point?
We haven't really given specific guidance, but what we're looking for those tumors where there is some activity from IO treatments today, but not necessarily really satisfactory. If you think about those. For example, PD-1s work in head and neck cancer, but have we really, really gotten those big, robust treatment effects downstream with in head and neck cancer that a lot of us anticipated? There's a few other tumor types out there that have similar, where they're, you know, IO is active, but there is room for improvement. A lot of it will be based on we'll start with the phase 1 in all solid tumors, show that we can give a dose, then, you know, we'll have a little time to refine the exact patient population.
Gotcha. Just in terms of the relationship of, you know, 202- 301, if you can demonstrate proof of concept in 202-
Yeah.
Do you think, you know, that's read through to 301 or there's some reason why, you know, we should be cautious there?
Well, I would like in a perfect world to say if 202 works, then every single one of our molecules is gonna work, and just-
Yeah.
Sign up for the next 10 targets, which we already have them queued up in our heads where to go. I think one of the things that's unique about our platform is we are not just a generic blocker mask. Each one of our molecules is the combination of making sure you have the right effector molecule, like for IL-2, changing it to be beta-gamma. For 301, we actually didn't need to make it more potent because IL-12 is just ridiculously potent. Then once you have that effector molecule, how do you get the exact right mask to have this very high affinity and then solve the problem for, well, how do we get the linker to hold the mask next to the effector but still cleave with MMPs in physiologic conditions?
I think each one of these is its own entity. I think conceptually it's helpful, but I think we do recognize that each molecule does stand on its own.
Yeah. Makes sense. Maybe just, you know, last question in the last few minutes here. Looking forward, just, you know, we touched on a lot of this already, but just, you know, if you could map out the next catalyst.
Sure.
Sort of in order for us over the next year.
Sure. I'll do it in a calendar format. The end of this year, we should be at. I think the main thing we're gonna do is get for XTX202 the IND in. Well, let me start from IND for XTX301 and get the 1 mg per kg dosing done for XTX202. For XTX101, by the end of the year, get the combination with pembrolizumab going. When we move into the first half of 2023, XTX101, we would expect to see the 1A dose escalation data readout, the XTX202 dose escalation readout, XTX301 starting patient dosing.
When we move kind of into the midyear, for 101, it is the 1B pharmacodynamic cohort, and that's where we start getting that real read on what's going on in the tumors. For 202, that same event happens in the second half of 2023. I think then maybe the final thing is the startup of the confirmatory phase 2 for 202 in the first half. Those are probably, I think, most of the major milestones.
Okay.
Lots of data.
Yep. Yeah. Great. Sounds like lots to look forward to here over the next year. Why don't we end it there? Thanks so much.