Hi, welcome to the 45th Annual TD Cowen Healthcare Conference. I'm Marc Frahm from the biotech team here. Next up we have René Russo, Chief Executive Officer and President of Xilio, who will be talking about their tumor-activated technology, which has been being applied to some cytokines and other antibodies, and now it's increasingly moving towards bispecific. I'll turn it over to René. He's going to kind of give an overview for maybe about 20 minutes, and then we will jump into some Q&A that I've come up with. Anybody in the room is certainly welcome to put in some of their questions as well. With that, I'll turn it over to René. Thank you, Marc. Good morning, everybody. Can you guys hear me? Excellent. First, I'm fighting, I don't know, the third cold of the season, so bear with me. Hopefully, I'm not going to start a coughing fit. We'll get through it without too much tissues here. All right. As Marc alluded to, at Xilio, we've been working on a technology for about, I'm going to say, the last seven years since the formation of the company that allows molecules to be preferentially activated inside the solid tumor microenvironment. That makes a lot of sense because we can really push the activity and the efficacy and immune stimulation in the tumor and have minimal immune stimulation out in the healthy tissue. The goal is to increase the activity, decrease systemic toxicity, allow us to go to higher doses, longer durations, keep patients on therapy, and get into combinations that would otherwise be really difficult, and sometimes even go after targets that would be really impossible without masking. That's really the premise. At this point now, we've treated over 200 patients across our clinical programs, and this is across a number of diverse molecular architectures. We've gotten really good at masking and understanding, I think, importantly, the cycle of the design, moving that design all the way through NHPs into the clinic, and then going back to the bench. Now we've done that in three cycles with quite a lot of clinical data. I think that's really key. I think we've learned how to do that, and we're very efficient at it now. All of our molecules have the same general premise, including a masking domain that shuts down the activity when the molecule's circulating outside of the tumor, linkers with a protease cleavage site, and we can really tune that protease cleavage site to be activated by a number of families of MMPs. We have wide activity, broad activity across a wide range of solid tumors. Also importantly, and this is a very important component, the 3D conformation of how that protease cleavage site is presented to the proteases in the environment of the tumor is everything. The same linkers, cleavage site, and mask in a different molecule, 3D structure may not work at all, and it may work really well in another molecule. It is definitely not plug and play. We also have done something I think that is unique in that we do have a bank of over 1,000 human solid tumor samples in-house. We test 3D conformations, our final molecules, in multiple conformations in actual human tumor biopsies for activation, which I think is also key. That is kind of where we've been. As Marc mentioned, our pipeline has really sort of evolved and shifted to more complex molecules over time. We are going to cover some of these today. We will not cover all of them. Our most advanced molecule is our tumor-activated Fc-enhanced anti-CTLA-4 in Phase II, along with atezolizumab, which is Roche's PD-L1. We are studying MSS colorectal cancer, which we will talk about. That next readout will be coming mid-year this year. That is in a co-funded clinical collaboration with Roche. We actually have three pharma partnerships, which we'll talk about as well. Second clinical program is our tumor-activated IL-12. And this is a good example of a target that, without masking, it's untouchable. It's just too toxic. That is in a collaboration with Gilead. We're in Phase I dose escalation currently. Behind that, we have our first bispecific as we start to move the masking technology now into more complex molecules, bispecifics and trispecifics, including our PD-1/ IL-2, and then moving into cell engagers, both in the bispecific format with our ATACR format and in a trispecific format with SEECR. We'll go through some of those new programs today. We have a number of programs we've just announced in a collaboration with AbbVie that are undisclosed but include masked cell engagers. Okay, so we're going to start with vilastobart, which is our tumor-activated Fc-enhanced anti-CTLA-4. Sorry. I have backup in case I go down in the back of the room. So this molecule is one of two that we're aware of in development that is highly potent. So it's about 10 times the potency of ipilimumab and brings Fc enhancement. That's kind of how I categorize it. There's a lot of CTLA-4, so I'd put those ours, and there's one other in that category with the Fc enhancement high potency. Only ours includes the masking technology, and we'll get back to that. I would say the others are generally ipilimumab-like at their core. We've done a lot of work on this molecule. We've studied it in the clinic, monotherapy, combos, and about 100 patients to date, actually more than that now. We've demonstrated a couple of important things. In actual patient biopsies, we measure the activation of the molecule. Consistently, it's over 70% activated in the tumor, up to 96%. Consistently, it's less than 15% active in peripheral circulation. That's part one, right? The tumor selectivity is working, and it's consistently working across a wide range of solid tumor types, even in metastatic lesions. We continue to measure that activation. That's key. We were able to demonstrate really improved safety profile over what would be expected for a molecule like this if it was systemically active, both in monotherapy and in combo. We'll go through some of that. Most importantly, we saw monotherapy activity in Phase I in a colder tumor setting, a PD-L1 negative non-small cell lung cancer. What was really important here is we saw eradication of innumerable liver mets in this patient, which are notoriously difficult to treat with immunotherapy. We did see activity in combo with atezolizumab in Phase I- C in cold tumor types like ampullary carcinoma, so pancreatic duct, and also in MSS colorectal cancer with liver mets, which has kind of been really hard to crack. We'll come back to that. That was at a dose of 150, so keep that in mind. We decided to move our CTLA-4 first into MSS colorectal cancer because it is a huge growing unmet need. You probably all heard about it. It's become very mainstream knowledge that for some reason, MSS CRC is growing rapidly in incidence. It's now the second leading cause of cancer-related death in men under 50. It's growing in young people, which is kind of one of the more concerning trends that is not really understood. Because young people are getting diagnosed, they're not getting screened. They feel relatively healthy, and they get diagnosed in late stage very often. If you get diagnosed early stage, surgery, radiation can be curative. If you get diagnosed in later stage, stage three, stage four, you're really down to very few options other than chemo with not very good results. This is concerning. The 95%, which you see here in the yellow and orange of these later stage CRC patients, are diagnosed as MSS, microsatellite stable, meaning not sensitive to prior immunotherapy. That is why we're really stuck with chemo. Excuse me again. I mentioned this is where we're starting. It's a really big challenge, but the bar for efficacy is pretty low because we really don't have good options. These patients, when they're diagnosed later stage, we have two types of chemo usually that are tried. Sometimes TKI can be added, VEGF, EGFRs, but overall, they have about a six- to nine-month OS when they get to this later stage or if they're diagnosed in this later stage. Prior immune checkpoint inhibitors have been studied extensively, monotherapy, combo therapy, and the response rates in MSS are 0%-3%. That's kind of the bar. More recently, that other CTLA-4 I mentioned, the only other one we're aware of that's got that FC enhancement and potency, has kind of cracked this a little bit. They were able to show response rates 8%-19%. It's dramatically different than 0%-3%. The problem is they didn't have masking so that they're seeing a very high level of immune activation outside of the tumor leading to toxicity, mainly colitis, and a discontinuation rate of 30% due to AEs. What we're trying to solve is maintain or improve that efficacy level of 8%-19%, but really dramatically improve the safety profile so patients can stay on. That's what we were aiming for. I'll mention that this is only in the non-liver mets population. They have seen no activity in the liver mets population, which continues to be really tough to crack. We initiated the study back in August, again, with Roche's atezolizumab. They're co-funding the trial with us. We did an early look at the data at ASCO GI last month or two months ago now. It's early, but very exciting data. We saw a preliminary 27% response rate. If you do not know MSS CRC, you might say, "27, I do not know what to do with that." Given what we are dealing with with the disease and given the prior 0%-3% activity of immunotherapy combinations, this kind of blew us away, frankly. Additionally, we had some patients that had responses that were deepening over time. That is the other key piece, and we will go through that. Seeing that initial response, seeing a deepening, and seeing the patient's symptomology resolve, all of that is really key. We will go through some of these data. Just to be clear, we had 18 patients that were response evaluable in this data set, 11 without liver mets. Of those, three hit that PR criteria at the first scan. One additional was a near PR. We're watching that patient with a 24% reduction. In the liver met side, we have not seen a PR yet in this early data set, but we are bringing a higher dose forward now of 150 mg, which is where we saw responses in liver mets previously. Hopefully, that higher dose can make a difference here. This is the waterfall. Again, a really nice waterfall to see in very late-stage MSS CRC. You can see here all of the patients that have a green dot are continuing on therapy, including all of the responders in that near PR on the right. We had, at this point, 13 additional patients who had not yet had a first assessment. Around mid-year, we will share an update on this data set. There will be more patients with a first scan and follow-up on these patients. I really quickly want to go through just the three PRs just to give you a couple sort of components that are really key. These are meaningful PRs in the fact that these are late-line patients. The PRs were deep and deepened over time. They had reductions in serum tumor biomarkers, all of these patients, and symptoms. When you have all of that together, that's like a really good sign. The first patient, we saw a 47% reduction in some of the tumor lesion sizes by RECIST. This patient came into the study in fifth line. As you can see here at the top, those are baseline lung metastases, and you can see them at the bottom at the 13-week scan getting much smaller in size. Over on the left, you can see here the serum tumor biomarker, CEA and ctDNA, both decreasing and improvement in clinical signs and symptoms, especially cough. Second patient here had been through a third scan, a 57% reduction. Again, meaningful and deep. You can see lung mets for this patient similarly on the top at baseline and see how those are shrinking by the first nine-week scan at the bottom. This patient over on the left, their CEA went from 192 at screening down to normal by cycle five. Also a significant decrease in ctDNA now is nondetectable. This one I always find very striking. This patient came into the trial seventh line. We have a single scan, so it's an unconfirmed PR so far, a 35% reduction. This patient had over 20 metastatic lesions in the lung. You can see here the patient's one lung is completely clouded in the CT scan due to a pleural effusion, so full of fluid around the outside of the lung. Within that first scan, we started to see an improvement. You could see the lung space opening up, reduction in the tumor volume, and getting that not only importantly for this patient, symptomatology decreasing, but also serum tumor biomarkers. Very good to see. These are really tough patients, and we're just very excited we're doing something for them because they have very limited options. We talked about the other side of the equation. How do we keep or improve that response rate? Again, that was 8%-19%, which looks like we are maybe improving on, but not have 30% of the patients coming off due to systemic tox. We were really happy to see the side effect profile here in combination with atezolizumab. Importantly, colitis. Colitis with systemic high-potency CTLA-4s can be 20%, 30%, even 40%. We saw a rate of 5%, which is phenomenal. That is usually the reason for discontinuation with CTLA-4. We only had three patients that DC'd due to the toxicity of the combo. Overall, just a very differentiated safety profile. You could just see here, especially grade three is where you start to be concerned. Very few grade three AEs in this combination. That is one of the reasons we're bringing a higher dose now forward with the CTLA-4 in combination with atezolizumab, in particular for patients with liver mets. Our plan for this molecule, we love this molecule. We would love to just hold on to and bring it forward. Based on what we have seen with other MSS CRC therapies in the past, the regulatory path is a registration trial with overall survival as your endpoint. We expect this is going to need a partnership for a company of our stage and size to really bring this forward. We think there are many other indications that should be pursued, in particular lung as an example where we've seen activity in the past. We are seeking a partner to take this forward beyond the data readout of the first 40 patients in this proof of concept. We'll talk about this later in the Q&A, but we're open to and considering many different forms that this type of partnership could take. Okay. All right. I'm going to move into T-cell engagers here. I'm going to take a sip of water so I don't have a coughing fit. Okay. I mentioned kind of in the beginning, we applied our masking technology at first to more simple—they probably weren't simple at the time, albeit, more simple molecular formats. Straightforward antibodies, cytokines, more complex cytokines. Then we started venturing into bispecifics like PD-1/IL-2. Now we're getting into more complex bispecifics like cell engagers. I would argue, and probably all have opinions on this, but cell engagers have really been a breakthrough in immunotherapy, mostly in hematologic malignancies, but we think this can really work broadly in solid tumors as well. The problem with doing this is you have to find a tumor antigen that is pretty sort of localized to the tumor and not in healthy tissue. Otherwise, you're activating T-cells in healthy tissue. You get toxicity. Generally, you get cytokine release syndrome with these generally, but you can also get that kind of out-of-the-tumor-on-tissue toxicity. This is where we think masking is not only helpful, but is going to be required to really open up this whole class. We have seen some early evidence of this from other companies, which is very exciting. Our CSO, Uli, is in the back, has been working on this with his team for about 18 months. We recently unveiled our programs. We have a couple of different approaches. I am going to introduce you to two formats, and we can get into more Q&A on this. There are really two approaches we take. When we think philosophically about how we approach masking, simplicity is always the most important thing. You always go for the simplest design you can. Our more simple and efficient design that we can move very fast, so it's really built if we want to go fast as we can, is our ATACR format. That is a simple—I put this in quotes—a simple masked cell engager. Now, with our format, we have the ability to mask the tumor antigen targeting domain, the CD3 targeting domain. Of course, we also can dial in that half-life where the half-life extension domain is on when it's masked, comes off post-masking so that you get short half-life once it's active. All of that helps to really drive that activity into the tumor, get that T-cell activation in the tumor, and limit activity outside the tumor. We are taking it a step further. What I think is going to be really, really key in the future is not only doing that, but bringing along co-stimulatory signaling into that molecule in a trispecific format so you can avoid that T-cell exhaustion. You could just keep this going over and over again. That we think, depending on the co-stimulatory signaling, again, requires masking to do this well. We also have this second format we call SEECR that has all the components of ATACR but brings in trispecific co-stimulatory signaling. You can imagine with these kind of molecules, the idea is to see that clinical monotherapy activity early in clinical development. When you're bringing all that to the table and then focusing it in the tumor with masking, that's what we expect to see. The first three internal programs we announced are all clinically validated targets that we think are interesting, that we think can clearly benefit from masking. The first is an ATACR format, PSMA. This has, again, been in the works for a while, but we were very happy to see there's other data out there from other companies showing that masking has been, I think, validated as really meaningful with this target. Second, CLDN18.2, which we're also bringing forward in the ATACR format. Then finally, STEAP1, which we're bringing forward in the SEECR format along with that co-stimulatory signaling. It was important to us, alongside announcing our own internal programs, that we were able to announce a really meaningful partnership with large pharma that has deep, deep expertise in oncology and is really dedicated to the cell engager space and also believes masking is required. That was really everything that came to fruition in this collaboration we announced recently with AbbVie. We announced this a month ago, two months. Time's flying recently. This is a multi-program collaboration. It includes masked T-cell engagers. it includes masked antibody. It included a $52 million upfront, which is now in. That was $42 million in cash, $10 million in equity at a premium, and has the potential for up to $2.1 billion in follow-on milestone payments. Some milestone payments can hit in the next 12 months. Just kind of put that in perspective. Of course, normal royalties that you would expect. Okay. My last slide I'll end with is we have a lot of things going on. We have a lot of milestones coming up over the course of the next 12 months. We can really see line of sight beyond that with the platform. In terms of what's happening near term, mid-year this year, looking at that full proof of concept data for the colorectal cancer study, we're really excited to be sharing that data around mid-year. Behind that, we will be sharing data on the Phase I/II for the masked IL-12 in collaboration with Gilead. We're not guiding to that because we need to do it in concert with Gilead, but we do expect a milestone payment of $17.5 million from Gilead when we initiate the next Phase of that I/II study. In addition, we expect to have our IND for our PD-1/IL-2 coming up mid-next year. We start to get into INDs for the masked molecules in 2027. When we're talking about the T-cell engagers, the next really key things for us are DCs for the PSMA and CLDN18.2, which will be coming this year. Then we'll be moving those toward IND. All right. I'm going to stop there. Hopefully, that was a helpful overview. We'll go to Q&A. All right. Yeah. We'll do a little bit of Q&A now. Maybe you can come over here. Yeah. I won't need to over here by yourself. That's fine. I'm used to it. My associates definitely don't want to hang out near me. Maybe just to start off with the most recent update, which is kind of the disclosure of the targets for SEECR and ATACR formats. Just how did you settle on PSMA being the optimal first target? Because as you mentioned, later this year, you're going to develop a DC, and then you still got to get to an IND. And there are a couple of programs that certainly have exciting data, but the already clinical data. Why start with PSMA versus maybe trying to just go to an alternative kind of equally validated in terms of other modalities, something like CLDN18.2 or STEAP1, where it's even a bispecific, just not a masked bispecific? Yeah. Yeah. I I mean, I would say our strategy on PSMA, which again, has been in the works for a bit, is really about kind of is a fast-follower strategy. We see clear validation that masking can improve the outcomes for patients with this molecule, but we still see there's room for improvement. We think that the components of masking that can be improved is how well masked the molecule is. Can we really eliminate CRS further? Can we improve the ease with which you get to the dose you want to get to? Can we then do that but maintain or even improve the efficiency of the unmasking in the tumor microenvironment? I think there's some room for improvement. It's still a very large kind of cancer space, and we think there is definitely room for more than one well-masked PSMA molecule. I'll mention one other thing, though, that I said briefly, like our SEECR format, which brings in that co-stimulatory signaling, is built kind of on ATACR. All of these molecules will kind of be brought forward in both formats. You start with ATACR, then you build SEECR. If we get there and we decide it's getting really crowded or there isn't enough room for differentiation with just our improved masking, then we can flip it into the SEECR format and bring along that co-stim signaling. I think we will be flexible to the environment and to the treatment landscape. Right now, we think speed and efficiency and room for improvement there is worth going for. Okay. You mentioned there's a lot of nuance in saying masked and that these are all masked molecules. Just in terms of your approach, how would you kind of compare and contrast it to what Janux and Vir have been working on in the PSMA space? What's similar in terms of how you mask? How do you think you're going to have a little bit better masking that maybe drives less CRS? Just kind of what's that nuance that may drive that difference? Yeah. I would say our approach is more similar to Janux, more different than the Vir original Amunix masking. We're big fans of that original Amunix team. In fact, they give us a lot of advice. That was, I would say, one of the earliest forms of masking. It's steric hindrance. The mask itself doesn't interact with the molecule, but it could work. It's working. I think our approach is more Janux-like in that we build a specific mask that actually interacts with the molecule. When it's cleaved, it disassociates. You don't need two masks because the CD3 is very well masked. PSMA is a target that's really not expressed all over the body, so we don't need to mask that part. That efficiency in design, but really improving the interaction of the mask with the molecule, but still maintaining that efficiency and activation. When it gets in the right space, it's activated. We think there's some room there. We've done some head-to-head work to ensure that we think there's enough room to improve it. We do think that's there. Okay. What non-clinical work kind of still remains to be done since you were ready to unveil the target and a little bit of preclinical data the other day? What work still is outstanding to get to a formal DC nomination? Yeah. I think one of the things that is probably unique to us, we have found over the years, and again, because we've done this cycle now all the way through the clinic multiple times, that you need to actually build multiple constructs and take them all the way through NHP studies. You can't pick one and then just take one through NHP studies and think you got it right. Look, that happens because it's a big investment to do this. Our next phase, even pre-DC, is NHP work. That's what we're doing now to really ensure we have the best possible design to take forward. Okay. I mean, clearly, this would not be the initial clinical trial. Given that Janux and Vir are already in the thing, do you anticipate that at some point you'll likely need to run a head-to-head against them, or do you think there's enough space to get places where they just won't have gotten yet? Yeah. I mean, our strategy would be the latter. I think there's a lot of space here. We also think even if we went with the straightforward ATACR PSMA molecule, we could probably uniquely combine that with masked cytokines that are in our pipeline that can improve a regimen. There's different places you can go. I think it's unlikely we would really be boxed into a head-to-head given all the space here, but something we'll be keeping a close eye on. If we felt at any point there wasn't enough space there, again, we would likely move into that SEECR format. Okay. Before we leave the TC, just the next one with the STEAP1 target that you've talked about, that one, you do say your likelihood is a SEECR, not an ATACR. I guess, walk through that decision because the Amgen STEAP1 bispecific has shown what looks to be fairly meaningful activity in prostate cancer, but also a lot of toxicity. Why is it best to do both the masking and the SEECR for that versus just get rid of some of that tox? Yeah. Yeah. Yeah. I think it's a really good point. I think if we were only focusing STEAP1 in prostate, we probably could do it in ATACR and be in a really good place. We do see opportunity for this tumor-associated antigen outside of prostate in other cold tumor types, again, like colorectal cancer. If we want to go there, I think bringing that co-stimulatory signaling is going to be key for those colder. Okay. With the last part, there will be an update this year. That's the clinical update, the next one. You mentioned there's 13 patients that were on trial as the last cut, but not yet available. What's kind of the trigger to releasing that data mid-year? Is it getting kind of two scans on those patients so they've had the chance to potentially get to a confirmed response? Is it a total follow-up on the population? Just kind of how are you thinking about when to show data? Yeah. We are expecting that we should have the ability to demonstrate confirmatory scans on most of these patients by mid-year. That was kind of the original goal of choosing mid-year. Of course, if we can, we always try to target a medical meeting. Okay. I think you probably, well, you had some discussions with potential partners when you even led to starting the trial because Roche is participating in it. What do you think kind of the key is to get somebody over the hurdle to kind of want to sign up to take this further since that is kind of the plan from a Xilio perspective? Yeah. Yeah. We're having, obviously, a lot of conversations, and we're open to different approaches for this. It could be as simple as giving our CTLA-4 to other companies that are running trials with their combo. We're just providing the clinical supply. Given the data we're seeing, there's interest to do that. Additional combos where all we're doing is providing drug is an easy way to expand the information here that can lead to that next stage of the partnership. We're also open to licensing, also co-development. We've become very good at developing in colorectal cancer because there's just not a lot of options. You quickly get very close with those major institutions treating these patients and with the patient advocacy groups. It could be a co-development where we use our expertise in the CRC space to move forward. Ideally, we want a partner that has agents in their pipeline that make sense in combination and wants to see this molecule really fully explored in CRC and beyond. Okay. Unfortunately, we're out of time. We are going to have to cut it off there. Thanks a lot for joining, René. Thanks a lot for everybody in the room as well as on the webcast. Thank you.