For those that don't know the team, it's represented today by Lara Sullivan, their CEO. Thanks for being here.
Thank you for having me.
I thought we'd jump right in. Maybe just to start off, can you give us an explanation of stromal targeting? Obviously, Pyxis is doing something very unique by using 201 to go after EDB. I guess, what does EDB do that earlier generations of stromal targeting maybe may not have been able to achieve?
So EDB is a splice variant of fibronectin, which is a component of the scaffolding of the tumor stroma. Those types of targets in the past, when others had targeted the stroma, had actually targeted (how many times can you use the word target in one sentence?) targeted targets that are on sort of the cell surface of stromal cells, like fibroblasts, etc., whereas this is actually part of the scaffolding. So it's a completely new approach. It's a first-in-concept approach, not just a first-in-class approach. Our belief is that by targeting the stroma and either dismantling it, altering the nature of it, reducing the volume of it, making it less hospitable, that will end up leading to tumor cell death, both through the diffusion of the payload into the tumor cell, but also altering the nature of the stroma that's keeping the tumor alive.
There's sort of a multi-mechanistic approach here by targeting the structural element of the stroma that can take down the tumor.
Got it. That makes sense. Really exciting. Can you then, I guess, walk us through the differentiators of the FACT platform itself, which is what 201 is based in? I guess every ADC nowadays talks about unique payloads, conjugation, linker technology, and you guys certainly have all three. So I guess, what makes FACT special, and how is this ultimately going to play into the differentiation and therapeutic window of 201?
Sure. So we were fortunate to license in the entirety of the FACT platform from Pfizer, along with our lead program at the development candidate stage. The toolkit itself consists of optimized linkers, payloads, conjugation chemistries that Pfizer spent a decade plus figuring out what the kind of best versions of each of those components could be. And as a large pharma, they had the resources and the capacity to empirically test each of these components and which ones match best with one another. That's a challenge small biotechs don't have the resources or capacity to do. And I think in our view, the fact that Pfizer did this and we're able to benefit from that empirical testing had us feeling like we had a leg up from a technical de-risking perspective going into the clinical trial.
So in particular, when we think about the conjugation approach with our particular program, we have engineered or Pfizer did, and we've continued to leverage their work around engineered cysteine residues that allow for a target DAR of about 4. And it's a more homogeneous DAR of 4. In some ADCs that are out there and others that are in development, you might see an average DAR, but a lot of heterogeneity and a lot of variability. And that can lead to lack of stability, lack of clarity around the potency, and PK variability too. So the Pfizer approach has minimized that and created a very homogeneous DAR. The linkers themselves are also optimized Val-Cit linkers. These compare to those that have been out there and marketed products like Padcev, etc., where ours provide more stability. And then finally, the payload itself is an optimized auristatin.
This has improved sort of metabolic properties and permeability of the payload itself. So the total of all of those improvements is embedded within our 201 program. So in addition to the novel approach targeting the stroma with a brand new target, we're coming at it with a more optimized ADC.
Got it. And then maybe slowly moving in the direction of the clinic, which is where you guys are now. I guess as you're running the phase 1 trial, what's been the level of investigator enthusiasm in putting all this together of a new target, novel ADC with a lot of good properties? And I guess, what does that enthusiasm tell you about either the unmet need for an antibody like this or just informs you about recruitment and potential experience going forward as you conduct the trial?
Yeah. So as we all know, we're in a bit of a heyday of ADCs right now with a lot of programs out there and a lot of what I would call incremental improvements to existing targets. I don't think any of us would say we need 92 HER2 ADCs in the world, especially when we've got one that's dominant. And it continues to puzzle us that people go after these marginal incremental improvements. And I think it puzzles investigators and is a source of frustration for investigators and clinicians and frankly, patients as well. As a patient, how would you even make sense of choosing between 92 HER2s if you're approached for a clinical trial? So we felt strongly that to move the needle on the unmet needs in the solid tumor space, that it requires a new approach.
And for the reasons we talked about earlier, I think targeting the stroma could be incredibly impactful. And we've seen that playing out in our clinical trial with waiting lists at every single site. There's, what, 900 ADCs in development? So that means 900 some-odd clinical trials going on out there. And we have waiting lists at each of our sites. It's a global trial. We've got 18 sites. We're in the U.S., Belgium, Spain, and we'll be in the U.K. in the next phase of development. So I think that speaks to appetite for a novel approach and this particular agent. We started dosing in March of 2023. So we're now one year and two months into the clinical trial. And the level of enthusiasm has only increased throughout the trial. So I think folks can interpret that as they may.
If you're an investigator that's got a choice of dozens and dozens and dozens of ADC trials and you're putting your patients on a waiting list for ours, you're probably doing that for a reason.
Got it. As you've continued to dose 201, you've reached what I'd say are fairly high doses for an ADC. So I guess, can you talk about what that means for the profile of the drug, maybe versus what you had expected going into the trial?
Yeah. So with a novel approach like this, we did not have any precedent experiences in the clinic of stromal targeting structural stromal targeting agents. So the best proxy we had when planning for our clinical trial was the Pfizer HER2, which uses the same linker payload as ours. And in that experience, Pfizer was able to dose its HER2 to a range of 3-4 comfortably. They did dose up to 5, but they had a 50% discontinuation rate at 5. They did not specify what the AEs were that led to that discontinuation rate. And they did not publish the full phase 1 trials because they stopped the trial for commercial reasons. They didn't want to be the 92nd HER2, I guess.
And so we don't have the full output of that, but we do have the view into the safety of the first, I don't know, 30, 40 patients. We use that for our planning. We assumed with the same linker payload, we'd see similar behavior in the patient setting. We assumed that probably somewhere around 5 ± would be where we might see kind of the dose maxing out. We announced in March that we had dosed through the 5.4 mg/kg cohort with an attractive safety profile. It was very well tolerated. I believe that was cohort number 6. Then we'd announced that we were beginning dosing cohort 7 at 8 mg/kg. So we're often asked, what's the reason that we think that we're able to dose at higher levels than what you've typically seen with MMAE type of payloads?
I think there's two factors at play that we believe. One is the specificity of the target. So our IHC analysis and the IHC analysis that Pfizer did before us had both pointed to the same conclusion. The target is extremely highly expressed within tumor tissue and negligible in the companion normal tissue. So you have a very differentiated target expression. That's very unusual sort of anywhere in what we all do. And so we've got that sort of going for us, which may be reducing the potential for target-mediated toxicities, allowing perhaps the patient to tolerate more payload, right? Because ultimately, the tolerability for a patient is the sum total of platform toxicity from an ADC and target-mediated toxicity. So if you're able to reduce one part of that equation, perhaps you can tolerate more the other part of the equation.
We won't know all this definitively until we finish our trial and we can go back and analyze sort of all the samples and what's going on. But this is our hypothesis. And then the mechanism. Perhaps there's something about altering the nature of the stroma that's allowing for higher dose levels. I mean, that's a little trickier to understand because to really dive into that, you'd almost want biopsies of stroma every week to see what's happening over time. Obviously, we can't do that. So we're trying to kind of best understand how the stroma itself is changing as we look at tissue samples and understand and tie that to what we observe clinically in the trial. And that can also provide an explanation. So I think those are the two factors that have led us to dosing levels that maybe we just hadn't fully anticipated starting the trial.
Got it. I know you guys reported earnings earlier today. I just wanted to clarify one thing in there. It sounds like you guys are narrowing down on a certain set of doses. And as you mentioned just now, you're dosed cohort 6. You've dosed cohort 7. I guess, what's cohort 8 potentially look like? Can you talk about the doses? You sort of bracketed on either end.
Sure. So we're at the stage in dose escalation getting into increasingly high dose levels, as you just mentioned, where you begin to really think about what are the increments of advancement that you would want to do, right? We're well above where the Pfizer tolerability cutoff was for HER2. So we want to be cognizant of that. We also want to use all the analytical tools at our disposal, like PK analysis, individual patient PK, as well as the PK of the cohort, to figure out what the next dose cut could look like where you could get a meaningful difference in the patient experience. So there's the analytical part, and then there's the qualitative real-world judgment part of what are the patients telling you, what are the investigators telling you, and how do you triangulate this information?
So as we noted today, we've dosed 42 patients in total thus far. We believe we'll be able to dose another 16 before we release data later this year. Those 16 could be spread, frankly, over any of the cohorts we've dosed, right? Because any of the lower levels have already passed their safety cohorts. We entered into the 8 mg/kg dosing cohort back in March when we announced that. So to your point, Leo, I think right now, it's sort of a mixture around what you learn clinically. Sometimes you can allow patient preference. Some patients might show up and say, "I want a clear dose.
I don't want the current experimental dose, and we'd be happy to accommodate them." So at the end of the day, this kind of dose range will help inform ultimately what the dose optimization phase of development would be at the next stage. Going into the next stage there, you want the floor. You've got to go into that with the floor, and then depending upon how you design it, ideally, you'll come to the ceiling. So where we sit and there's a slide that I'm sure you're familiar with. It's sort of a staircase slide in our deck that shows each cohort that we've cleared. So we had, I believe it was 10 patients at 5.4 that we had announced previously. And I think it was 10 patients at 3.6 that we had announced previously. So we know those are really well-tolerated dose levels.
And then we're waiting and observing sort of everybody kind of all the time to see where in the upper echelon we settle at well-tolerated floor dosing levels.
Got it. I think sort of you've alluded to this. And one of the things that's impressive about 201 is that the safety has been very tolerable despite the higher doses. So I guess from the qualitative perspective, I guess, what adverse events might we expect with the EDB mechanism? What about with the auristatin payload? And I guess, how's it looking to the extent that you can say?
Yeah. So I think with any ADC, we all worry about platform-mediated tox and kind of the big three that we keep an eye out for: ocular tox, neutropenias, neuropathies. And I think despite targeting the stroma versus the cell surface, that's no different. I think we still worry about those three. As I mentioned earlier, the specificity of the target has led us to hypothesize that the target-mediated tox may or may not end up being a big component of this program, or it may as we see the next few dose levels unfold. We just don't know yet. We haven't commented on any specific AEs publicly as of yet. But I guess I would summarize it by saying we didn't go in with any particular target-related AE because there was nothing precedent to source from.
We've got an extremely specific target, and then we keep an eye on those sort of big three platform-mediated tox AEs.
Got it. Makes sense. Then I guess the other piece of the picture is obviously efficacy and activity. And I guess, what gives you confidence from what you're seeing so far, maybe from some of the preclinical work, that 201 is an active molecule at the high doses, at other doses, and it's not just reflective of some kind of worst-case scenario where your ADC is going into a sink, or it's just not a very active molecule?
Right. So I guess I would point folks to a couple of things. First, our recent AACR posters, we showed a variety of PDX models that indicated broad tumor activity. And that informed the basket design of our phase 1 trial and why we put 10 tumor types in there. Ultimately, the answer to your question will be the clinical data itself. But between now and then, what we would point to would be the preclinical data, the investigator enthusiasm. I would just find it difficult to believe that if there's 900 choices of ADC trials out there, investigators are putting people on a waiting list for something that doesn't work. But everyone can draw their own conclusions of that since we're not commenting on in-progress efficacy signals or any specific safety issues.
I think the reason for that, frankly, is we just feel our responsibility as drug developers to be able to release information that's fully cleaned and validated. Releasing things in the middle of a trial just isn't helpful, as we saw with Vincera's experience a couple of weeks ago. So I guess I would point to some of those more qualitative clues in terms of the conduct of the clinical trial. I'd point to the PDX models, which I think we had 100+. And when we licensed it from Pfizer, they had 300+. So there's been a lot of PDX models.
Got it. And then just also circling back to something in the press release earlier today, one of the lines mentioned encouraging early responses. And I just wanted to clarify, should we take that to mean that there are responses in a kind of objective clinical sense, or is this more of a totality of what you're seeing?
So I would take it more in the totality sense because, again, we're not ready to comment on sort of RECIST-defined response criteria. And I think it's sort of our way of saying to your earlier question, were people concerned that we have a clean drug that's clean because nothing's happening? And when we say signs of response or whatever the terminology was, that's indicating that this is an active drug and things are happening for our patients in this trial. And our job now is to figure out who could benefit from it the best, in which tumor types, and at what dose levels, right? And it's not an abrupt sort of magical answer appears. We'll get pieces of that information by the end of this trial, and then we'll get further pieces of that information in the next stage of development with dose optimization and cohort expansion.
So I think what we hope to deliver in the data readout is the confidence in the signals of the rough dose level area and the rough tumor types. We'd call that a success. The law of small numbers would not allow us to responsibly go around claiming response rates, that kind of thing, although all those calculations will get done by everybody. That sort of data becomes more meaningful with larger ends.
Got it. You guys have narrowed down the 201 trial to 4, I guess now 5 specific tumor types. So can you talk about what drove that focus, like why hone in on those particular tumor types, and how much of that was based on what you're seeing in the clinic versus some of the preclinical rationale?
Yeah. So the decision on the four initially and now the fifth tumor type that we added, the soft tissue sarcoma, ultimately is a decision driven by that mixture of preclinical or analytically oriented data analysis as well as the clinical experience while we're going through dose escalation. With a basket trial design, you're trying to solve for two things simultaneously that can be at odds with one another: breadth and depth, right? You want to know the breadth of a signal because you don't want to leave any patients behind. But in doing that, you're sacrificing depth of signal because you have to have more tumor types in the trial. It's a constant balance between breadth and depth.
The compromise on this is using the feedback from the clinical trial as it's unfolding to enhance the enrollment of a few of those tumor types so that we still have breadth, but we have a little more depth in some of them for those signals. The addition of soft tissue sarcoma since the March announcement reflects additional clinical insights that we've garnered since the last time we commented on this publicly. That gives you a bit of a clue as to how much clinical insights can actually inform decision-making and inform it pretty quickly too.
Got it. And then I guess we've sort of been alluding to this, but on a related note, how much is EDB expression itself actually a marker for the patients you'd want to be selecting? I know you're not screening in for EDB expression right now, and we have a history of, on one hand, HER2 patients have been traditionally until very recently screened in based on HER2 expression. On the flip side, things like TROP2 and Trodelvy don't screen for expression. I guess, how are you thinking about how much is EDB actually a marker for the patients you want to be screening?
Yeah. So it's one of the million-dollar questions. And as you said, we can't screen for it prospectively going in. And so a lot of the work Jan and his team are doing in assessing the biopsy samples is looking at what are the correlates, whether it's EDB expression or something else, tied to whatever we're learning clinically. And so again, I think we still have the law of small numbers that we're fighting because you don't get fresh biopsies from every patient. And some of them just location-wise, you can't. And we will ultimately, I think, have dosed 58 patients by the end of this trial, X number at whatever the therapeutic dose range is. So we're still dealing with the law of small numbers, even with tissue samples.
I think we'll have some clues as to what some markers could be ultimately for patient selection, but that work will continue through dose optimization and into cohort expansion because it will just need more power of more samples.
Got it. Makes sense. And then sort of to culminate this, as we look to the data in the third quarter or in fall that you'll be presenting, I mean, you mentioned about 58 patients, but thinking in terms of dose, what the dosing regimen is, scans per patient, what would you like to see in terms of response, right? I mean, understanding that we have that whole depth versus breadth equation, but I guess, should we look for some PFS, duration of response data as well? Will it be more ORR-focused? I guess, what would be ultimately an efficacy one for you?
Yeah. So fast-forwarding to data release, I think we'll be sharing a complete readout of efficacy, safety, PK across all the cohorts. We'll cut the data across the study within tumor types and across dose levels. So we'll be able to all understand a lot about what's sort of going on here. Again, I think people will naturally calculate response rates and disease control rates. And we all can decide how meaningful we ascribe that quantitative number versus an understanding directionally. I mean, I think if we had two patients and we had two PRs, I think we'd all hesitate to call that 100% response rate, right, even though mathematically it is. But I think we'd all recognize there's a signal there. And then the question is, is that a signal we want to pursue?
Is it a tumor type that makes sense to develop from a regulatory perspective, a market perspective, a competitive landscape perspective? So we want to come out with that level of insight as well, not just the data, but what it means for the company and for investors for the next stages of development. In terms of progression-free survival and durability of response, looking for data in the fall, really depending on when we end up in this therapeutic dose range, whether we're there now or we'll get there between now and the fall, you don't really know until you look retrospectively, right, because you sort of figure out what those cut points are.
But most likely, we will have some patients who could qualify for that amount of time is my guess, whether we would sort of statistically be able to have confidence behind a progression-free survival assertion or a durability of response assertion, I think, remains to be seen, depending on what level, what dose level. I mean, if it's somebody at the first dose and they're still on, we probably all wouldn't ascribe that a lot of credence to that durability. But if it's at 5.4 or 8.0, it becomes meaningful. So I think we will talk on all of those, and then we'll just have to see what ends up being significant enough.
Right. Well, I think unfortunately, we're out of time. There's so much more to ask, but thank you so much for being here and for the delightful discussion.
Thanks, Leo. It was great to be here.