Okay. Thank you everyone for joining us. My name is Daina Graybosch. My team and I cover immuno-oncology here at Leerink Partners. And I'm excited to talk with Terry and Jen.. and i think we're gonna jump right in.
Okay.
Is that good?
Thanks, yeah. Thank you, everybody, for joining. Thank you, Daina, for.
So I'm gonna start with corporate strategy, then I'm gonna talk about HIF-2α.
Okay.
Then TIGIT, then adenosine pathway.
Perfect.
So there's no way we can cover everything in 30 minutes, but let's see what we can get in. So corporate strategy, you have a unique partnership with Gilead, which you recently have extended, and you're always updating that. I wonder, how does that backing from that large pharma partner impact your portfolio strategy and risk appetite?
Yeah, it's, I'll make a few comments, and then Jen can add to that. It's enormously enabling. So from a strategic standpoint, keep in mind, we're really trying to build a long-term independent company, and we're working on targets that are important. They're in big markets. Our competition is Merck, our competition is Genentech, and it lets us compete in, those, you know, where the opportunity is engendered by our molecules. So we have high-quality molecules, and we get to push them aggressively. From a risk appetite standpoint, I think there's, there's two things. There's the portfolio of programs, and then there's within each program. And I think it lets us, within a program, go at it. And I'll use an example, STAR-221, which is our anti-TIGIT study in upper GI cancers.
We actually were so compelled by the biology that we went right into that. We have a phase 2 study. We'll be updating those data at ASCO, but it put us. It allowed us to be, like, where we're gonna have a first-to-market advantage. We're gonna be a couple of years ahead of Merck. We're gonna be fully enrolled by the middle of this year, and it's because we were able to start that early and take on that risk. Similarly, as we go along, and I think if you look at our targets on the scale of biotech, there we tend to go down the middle of the fairway in terms of validation. They're not generally, other than HIF-2, validated because there's another molecule that's proven things, but they're not bleeding edge.
I think now that we've created a portfolio of those that are more down the middle of the fairway, it allows us to move to things that are maybe a little to the other end of the spectrum, perhaps further to the left. So it's very enabling. It engenders opportunity to go fast in a risk accepting way, and also move into targets that are a little more risky. Is there anything else you would add?
Yeah, like maybe the two other things, you know, one, just building on what Terry was saying that's very unique about us is being able to pursue multiple phase 3 programs in parallel. And I was a banker for a long time, and the pattern that I always saw in biotech is, you know, company gets one product through phase 3, and then everyone starts to ask, you know, what's next? And the next product is years and years behind. And so, you know, if things work out like, you know, we think they can, we hope we can, it could be put us in a position where literally, you know, every year, you know, we're launching a new indication, and probably every two years, launching a new product. So it puts us in a really unique position from a company.
Then the second thing that I'd just add, if you look at the markets that we're going after, you know, they are the very largest markets in oncology, you know, in lung, gastric, pancreatic, colorectal. You know, so both from a development perspective as well as from a commercial perspective, we would probably not be able to do that without Gilead, but it puts us in a really unique position from that perspective.
We're gonna have 4 molecules and 6 Phase 3 studies by early next year, and it's not like each of those was built on a card flip of the last one. We were able to, you know, there's a lot of luck involved in that as well, but we were able to parallel process those because of the Gilead collaboration.
Yeah. Okay, let's start with one of those, CAS.
Okay.
Casdatifan, is that correct?
Casdatifan, and with a su, as opposed to zu, which they tell me because I.
CAS.
I would say CAS.
Your HIF-2α. This, as I understand, is really a differentiation, so a best-in-class hypothesis versus Merck's leading, at least time-wise, belzutifan. And you gave a lot of disclosures in your fourth Q earnings, so I'm gonna dig into that. It's a little technical, so everybody, bear with me, but that is the differentiation of the asset, so we have to go there. It looks to me that CAS is more potent than belzutifan, but why do you believe necessarily that that potency will translate into an observable efficacy advantage?
Sure. So I'm gonna give us technical answers. It won't be an elevator pitch answer since it's a technical question. So the first thing that we've shown, there's a PD marker that belzutifan that Peloton use, Merck use, we use, and it's production of EPO by the kidney. So that has nothing to do with its antitumor mechanism, but it's very convenient, very robust, very reproducible. The other thing that's interesting about that biomarker is it's the on-target mechanism for the AEs, which is anemia. So if you inhibit the production of EPO, you're gonna be inclined to get anemia. But that HIF-2-mediated production of EPO is only about 70, 60, 80% of your total EPO. So what's really nice about that is you have this built-in break.
And nature just happens to put there, that you can hit the target as hard as you want, and you're not gonna create what I'll call catastrophic anemia. So in fact, you're not gonna get the DLTs or the like. So what we've been able to show, Merck uses a 120-milligram daily dose of belzutifan, and that 120-milligram dose basically is just at the cusp of getting that maximal inhibition of erythropoietin production by the kidney. What we've shown is that we can actually get that same inhibition at 20 milligrams of our drug. But then what we're doing is we're actually dosing at 5x that, and we've also shown very clearly that we have linear dose proportional kinetics.
So essentially, we're dosing at 5x the PD equivalent of the Merck drug. So that's what we know. Now, to Daina's question, in terms of how is that gonna manifest itself in terms of improved efficacy? We've actually had a hint of that now that we've seen our dose expansion data. So there's a couple places where you can differentiate on the molecule: ORR, primary progression, PFS, depth of response. So those numbers, to just give you a sense of the opportunity, the ORR for belzutifan has been roughly, trial after trial, including LITESPARK-005, the registration trial, around 20-22%. PFS, about 5.5 months. Primary progression is about, in LITESPARK-005, was 35%, 33-35%. That's something that was actually even more than the control. We've had ad board meetings. The KOLs tell us if you could beat them even on that, that would be a very meaningful difference.
So we've had an early look at the expansion data. We have the escalation data beyond the PK/PD. What we've seen is if you take everybody into the denominator, out of those 30 patients at 100 milligrams of our drug, even though 70% of them have only had 1 or 2 scans. And keep in mind, the median time to response for belzutifan is 3.8 months, and our scans are every 6 weeks, so the 70% of the patients have only either had 6 weeks or 3 months. We are already, basically, if you looked at confirmed and unconfirmed responses, very similar response rate, but we have double-digit patients that have had some tumor reduction, are still early in treatment, we think could contribute to ORR. We also have,l Like, this will not change.
So primary progression, we're already seeing an advantage relative to that 30-some-odd percent. Now, whether that translates into phase 3 is another question, and these are not independent variables from PFS, and we actually think that kinetics, those slow kinetics, anecdotally, we've been getting feedback from these investigators who've worked with belzutifan, love the mechanism, particularly compared to, like, TKI alone, if they are also working with CAS, that they believe we are having an earlier effect. So the totality of data makes us feel pretty confident. The other thing that I would just point out that shouldn't be dismissed: so we're not going after this third-line monotherapy a s our first place.
We're gonna be looking to combine with a TKI and other things ultimately, but with a TKI. Merck combines with lenvatinib. Talk to any investigator, you call any one of your friends, that's probably the worst TKI out there. There's difficulty with dosing it, difficulties with side effects. We're gonna be either using CABOMETYX or Zanza in our combination strategy. So we think the development strategy and the combination partner, because the combination partner creates a single entity, essentially, we think offers us plenty of opportunity for differentiation. Did I miss any of the points?
No, perfect. Let me go back to belzutifan for a second. If you look at, you know, the clinical review, they have patient-by-patient area under the curve and how that correlates to response rate and safety. I think what's interesting is there's a pretty variable exposure with belzutifan, so some patients seem to be getting a lot of exposure. Maybe they're having potency, some patients don't. And I actually didn't talk about that in your earnings. Do you have a more consistent exposure with CAS? And.
I think our data are pretty tight. I still think that's not what's really. I mean, as you know, Peloton had a molecule before belzutifan, and that was also part of the same sort of analysis as what drove them to go after belzutifan, that they saw this correlation, at least at, like, a binary correlation. Like, you had a group that did and a group that didn't, and that led them to say, "We, we need a molecule with better PK." So we certainly have very tight PK. We also have this very nice once-a-day dosing. So we think that that does represent some advantage, but I would actually say with the Merck molecule, I think their ability to hit the target as hard as you want is still probably limiting, not just the interpatient variability.
Got it. They did see some safety dose-dependent increases, so an increase in the toxicity anemia with that variable exposure. So you said you're sort of maxing out on the toxicity, but Merck did see variables. So should I not think that if you're gonna get tighter and higher exposure, you would have higher anemia than.
Oh, it's.
Yeah. I mean, I think actually, like, the best study to look at to determine whether or not, like, the higher doses, at least with respect to belzutifan's results in higher tox, is the LITESPARK-013 study. I think it's 013. You know, where they looked at 120 and 200, and if you look at the tox and they did like the tornado chart, you know, actually, in some cases, the higher dose had lower tox. And the only thing that there was a slight increase in incidence was anemia, and I think the difference was like 75% all grades versus 80%. I mean, so 5% on a 75% base is like nothing. So, you know, we're definitely closely monitoring AE profile.
You know, we provided some data, some pretty detailed data when we did the earnings call, and so far, we're not seeing anything to suggest that we'll see a higher AE rate. But we're monitoring it closely. But like I said, you know, both with the data that we're seeing, as well as just anecdotally, when we talk to advisors, we don't seem to be seeing any differences.
Put a little quantitative nature where actually, you know, all anemias were like roughly at, like, 50-something% in this study with this expansion cohort with 100 milligrams. They see around 70, but we're not—we don't think we're gonna differentiate. We think with time, that'll increase.
Let me push back one second on that. Cause you guys have said, "Look, you can't look at LITESPARK-013 because that's sort of a false study, because those doses are almost equivalent, because they don't get that much exposure on average with the higher dose and the lower dose. What they did show is if you look patient by patient.
In the higher exposure.
P atients that had higher AUC did have higher tox.
Yeah, we need. If I remember correctly, like, the ends were really small, and it was, like, fairly difficult to make that conclusion. So like I said, we're following the data very, very closely. So far, you know, our rates look in line or a little bit lower just because, like, we don't have the same follow-up that they do. So far, we don't seem to be. We think it's because of this compensatory mechanisms that kick in. So even though we're hitting EPO harder, sorry, you know, these compensatory mechanisms kick in that allow your hemoglobin levels and the anemia to plateau.
Here's the way I would describe that. Let's say you could, like, literally look at every single patient.
Yeah
A nd get some correlation. There's two things that come into play that I think are relevant. So the first thing is, when you're looking at clear cell RCC patients versus healthy volunteers, compromised kidney function can actually affect the baseline and the effect that the drug has. So it is the kidney. That's variable one. The second thing is your point about the variability in exposure. I think with, for them, the issue is that, that variability, if you, if you think about the 120 milligram dose, it's, it's more a fluke of nature that that's where their pharmacokinetics plateaued. So to your point about 200 not being dramatically different than 120.
So given that they're right at the cusp of getting the maximal effects on erythropoietin suppression, if you have a patient that has 50% of the exposure, you're actually gonna drop down. We're 5x the PD exposure, so we're always, like, getting the maximal that you're gonna get with the 120 milligram. Just ignoring the tumor on that effect, we're never gonna have a patient that we're not nailing that HIF-2 mediated EPO suppression. So we're gonna, you know, probably be consistently, you know. Other than the inter-patient variability on their own kidney function, we're gonna be nailing the HIF-2 component. Whereas Merck, if they get variability in their PK, is gonna probably drop below the asymptote on their dose response, and they're gonna see less in those patients.
That wouldn't shock me if you looked at a patient-by-patient basis. It's more thinking about your threshold and maximum versus backing down from that, and then we're probably gonna be nailing it all, all the time. It's not like you can get better than 100% inhibition.
Yeah. Got it. I'm gonna move on, although I have three more questions. Maybe we can come back. Let's talk about TIGIT. You have upcoming data, EDGE Gastric at ASCO, which is a follow-up, and what are you looking for? What might we look for that could increase our confidence in STAR-221?
Yeah, so the new data set that will be at ASCO, that we didn't have at ASCO Plenary, is median PFS. So as a reminder, at ASCO Plenary, we reported 6-month landmark PFS of 93% in the PD-L1 high and 77% in the ITT patient population. So, you know, clearly, we're gonna trend north of 6 months, and so now it's a matter of, like, how much, better than 6 months. As a reminder, if you look at the benchmark studies, they range from about 7 months to 8 months. So the CheckMate study, which is the oldest study, was at about 8 months. The most recent study, BeiGene's RATIONALE- 305, was at just under 7 months. So, you know, the further beyond 8 months we are, you know, the, the better off, I think we'll be and the more confident we'll feel in STAR-221. So we're excited. So we, we feel good, and we'll see the data set soon.
The other thing that's gonna happen is, like, that's gonna be right about the time that that study's fully enrolled. So I hope that happens right at the same time. It could be. Love to have that juxtaposition of those phase two data. This is where I was talking about that pre-investment, those data were not generated to, for decision-making purpose. They actually were generated as we were enrolling that study and wanting to open in certain countries that want to see safety and efficacy data with the exact combination in the exact setting. So a nice corollary to that is that we're going to have this data set that we're going to be able to show people on the same day and say, "Hey, by the way, this is fully enrolled. The registrational study is fully enrolled, and here's what this phase two study generated."
And I think it's going to be a pretty exciting data set.
All, I think now, as of stopping ARC-10, all your studies of TIGIT have an addition of chemotherapy, but we actually have minimal randomized data of that kind of triplet mechanistically. I mean, Roche has MORPHEUS EC, which I guess is related in gastric. So what gives you confidence to go forward with a chemocentric strategy across the board?
Sure. So, the EDGE Gastric is one, another one.
It's not randomized, but.
No, no, true, but it still gives us. I mean, you're asking me, that gives me a lot of confidence. I mean, randomized data set is better, but you'll I think what you're going to see over the next several years is the, you know, the late-stage randomized study. So the confidence better have been there some time ago, and now people are going to see if the confidence was justified. But I'll tell you what really gives us the confidence is more the biology.
So people forget, like, when you think about anti-PD-1 and anti-TIGIT and this fact that they work well together, and some people just think, like, oh, that's random. It just, it's happened to be the case. So there were a couple different strategies that companies went after, and including, I think this was part of the collateral damage to like what I think is like, should be a very exciting field, but there was all these mixed early results. And one way, like people went into studies where their sort of canary studies were in, it's a big medical need, commercial opportunity. I'm gonna go in PD-1 refractory. Now, so if you parse back, and I'm sure you're aware of this, if you look at how anti-PD-1 works, it really ends up activating CD38 and CD226.
Now, this is getting, again, getting a little into the weeds, but CD226 is a ligand, a receptor for the same ligand as TIGIT, CD155. So when anti-PD-1 has its effect, to get its full effect, you really want this CD226 that's also activating of immune cells. TIGIT is deactivating of immune cells. So when you disrupt the CD155 TIGIT interaction and now convert that to CD155, CD226 interaction, you're really gonna get the full benefit from the anti-PD-1. So if you think about the settings we're going into and with the chemo addition, the real focus is that you are going into settings where you are seeing an effect of anti-PD-1. You're seeing it in the context of chemotherapy.
I think it's important as you move away from cytotoxic, where efficacy, whether something's useful or not, just sort of happens to be a fortuitous therapeutic index that, you know, if I had a beta- tubulin inhibitor. In this setting, do I kill cancer cells better than I kill your normal cells, and do your normal cells come back? In these type of studies where you understand the mechanism, that's where you, you know, you're really wanting to play. So this chemo that drives this immunogenic response, and then you have anti-PD-1 is effective, it makes all the sense in the world that putting anti-TIGIT on top of it should enhance that effect. And like we're seeing in EDGE Gastric.
In moving from that beta- tubulin-esque type of thing, where you tend to categorize your cancer by the organ, you're not treating an organ, you're treating biology. That's why certain types of cancers, you know, breast cancer, certain therapies are going to be irrelevant, certain are going to be important. In this biology of enhancing the anti-PD-1 activity is going to cross settings. That's more important than the organ. So feel like STAR-121, which is right at the underbelly of Keytruda. That's the PD-L1 all-comer population, which is also going to be fully enrolled this year, feels a great opportunity for.
That was part of the strategic decision that you were referring to, is the high PD-L1, all IO, therapy is getting cannibalized by physicians just saying, "I'm going to put these patients on anti-PD-1 and chemo, even when they're high PD-L1," that's part of the STAR-121 population anyway.
More biology question. So Genentech published in Nature what they had previously talked about, I think, at AACR last year for their TIGIT tiragolumab, that is supportive of an Fc-competent IgG1 design, and that in CITYSCAPE, their frontline doublet lung cancer study, that they saw a correlation with a positive impact on myeloid cells and Tregs in the tumor. Now, of course, you have ARC-7, and I wonder with the Fc-null antibody, you know, how do you interpret that data, and have you been able to look at the myeloid and Treg compartments in your study?
Yeah. So we're not, you know, they did most of that by RNA-seq. We're doing immunohistochemistry and looking more things like CD155 and CD137, various things like that. But I'm gonna answer your question. That paper is a sort of paper that you might you know, it's very academic. You might discuss it at a group meeting in graduate school and spend 45 minutes, and we've done that. What I would just point out, I'll point out a couple of the things about that paper. So the first thing, and probably the biggest thing that you might criticize, as you know, they actually present a lot of these data at the last two cities.
Yeah.
It's a conglomerate of data, and a lot of it still goes back to the mouse work.
Yeah.
And that's where they're using a surrogate anti-TIGIT. And one thing to keep in mind that's never been demonstrated in a human that you get with this mouse surrogate is very robust, intratumoral ADCC, where you're depleting T regs.
Yeah.
And so a lot of their, what's driving their hypothesis comes from those data. And as you know, they're intermixing the data from this analysis with the mouse data. If you look at the RNA-seq data, so for their, their myeloid gene express, you know, gene footprint for this myeloid, it's like 11 genes, but they don't really tell you how they picked it. So you can't go and analyze that. The other thing is a lot of those genes aren't really just selective for myeloid cells. The final piece on that is most of the changes that they saw were less than, or I think maybe even all, were less than equal to, like, twofold.
A final point I would just make is because they were, call it a little bit of confirmation bias, you know, driven by this mouse results, what they argued were, you know, had to do with T regs or myeloids. You could have just argued, like, these patients that had high immune cell at outset, you could have even argued were T effector. So I think it, like a lot of the work they've done, it's, it's high-quality science, it's probing all sorts of hypotheses. I think it's a very much an over extrapolation to say that they're convinced that they remodeled the tumor microenvironment. I think it's a dramatic overstatement and, you know, with time, we're gonna see the data. Make one other final point that, keep in mind, all.
I was thinking about this part as well. All of their studies in the human are done with atezo as the partner. Not only is it different than if you were using anti-PD-1, but you have some of the limitations of atezo that you would wanna control for if you were thinking about an anti-PD-1, for example. Including, like, atezo itself and what happens with ADAs that atezo's generating. Does that affect the relative contributions of TIGIT and PD-1? So I'm not. I'm just talking about, like, in controlling for these things, it's good. You know, they present good data. I take their data at face value. I think their conclusions are probably or almost conclusions are probably overly strong. The nice thing is that over the next several years, you're gonna get data left and right.
I think the one thing that you can say unequivocally, though, about an Fc-enabled anti-TIGIT, is that they deplete peripheral Tr egs.
Yeah.
That does lead to immune AEs. What I would suggest, so BeiGene has actually generated a data set very similar to EDGE Gastric. But what you see is a very high level of immune AEs. When you see the totality of our PFS, our AEs, I would encourage you to compare to that. What's another sort of weedy point about the Fc-enabled anti-TIGITs? So if you go look at Merck's co-formulation, it's actually on the spectrum, the lowest dose of anti-TIGIT. I think they're using, like, 200 milligrams, even though they explored 200 and 700. So you can imagine, I don't know for sure what their strategy was, but the fact that it's in a co-form, you wanna make sure you don't get AEs. BeiGene, I believe, is actually using the highest dose.
Which is 900 milligrams of anti-TIGIT. So it's not surprising, you're probably getting more, you know, T reg depletion in the periphery.
And that also.
That we know so for sure.
E fficacy potentially.
Yeah.
Because if you're having to take patients off of drug, and keep in mind, like.
Yeah
I f you see an immune-related AE, that may be due to the Fc-enabled TIGIT antibody. Unfortunately, you have to withdraw both the, or withhold both the PD-1 antibody and the TIGIT antibody, and so that could have an impact on efficacy as well.
How was that run-on sentence for about that, though?
Yeah. We have 30 seconds. I'll ask one quick question in 30 seconds. MORPHEUS PDAC data is coming at AACR, and how will that impact how we should think about CD73 in pancreatic?
Jen gets the last word.
Yeah, because I can do it in 17 seconds, and Terry can't know.
Yeah.
But yeah, so the studies looking at etrumadenant + atezolizumab + gemcitabine versus gemcitabine. So it's randomized, 35 patients total. So, 20 in the gemcitabine arm, 15 in the etrumadenant arm, will show mature PFS and mature OS. And I think the way to look at the data is it's very confirmatory of what we saw in ARCADE. So even the different molecules, quemliclustat versus etrumadenant, they both block the effects of adenosine, and, you know, when you see the results, you know, we think they'll look pretty similar to what we saw with ARCADE, so.
In the last 15 seconds, ARC-9, which we also hope to share by the middle of the year.
Yeah
Which is etrumadenant and colorectal, third line, will probably put an exclamation point on those data. So it's gonna be a 105-patient data set, where you randomize versus regorafenib. Regorafenib, since we're being webcast, I won't call it. It's not a very good drug, but, it is a standard of care. But you're gonna see a really, I think, profound effect on, OS. And we'll have to deal with contribution of components, as you know, but it's, it's gonna be another adenosine modulation with a different molecule.