Arvinas, Inc. (ARVN) Earnings Call Transcript & Summary

Great. Good afternoon, everybody. Thank you for joining us. I'm Terence Flynn, the biopharma analyst at Goldman Sachs. I'm very pleased to be hosting Arvinas this afternoon. Joining us from the company is Ron Peck, who's Chief Medical Officer. Ron, thank you very much for joining us today, and thank you to everyone for tuning in. Maybe just to start, Ron, you could give a brief introduction to the company's platform for those who aren't as familiar with it, just to help set the stage.

Sure. So we have a platform of a class of drugs that broadly are considered targeted protein degraders. And the platform that we're specifically working on are referred to as PROTACs. These are molecules that are bispecific molecules that one end -- and just to be clear, these are small molecules. One end will bind to the target protein of interest. And the other end will recruit E3 ligase and create a triad complex that enables protein degradation through the ubiquitin proteasome system. So this is the natural -- the cell's natural function for degrading proteins. No one has ever really figured out how to sort of exploit this cellular function, but it's quite a very -- it's a very elegant mechanism. It's a very targeted and specific mechanism as well and potent. And one of the great features of this is that you can apply this not to the traditional druggable targets, but also to what is estimated to be about 80% of targets that are judged as being nontargeted because you don't need to find the functional pocket. And these are very potent. One of the other interesting features of them is that they can -- they essentially have a catalytic mechanism. So with an inhibitor, you might have a 1:1 ratio for a drug to have very often a competitive inhibition. But here, these will actually recycle and continue to degrade multiple proteins. So we have actually estimated that one PROTAC can degrade as many as 200 different disease-causing proteins. So it's a very powerful mechanism and modality, and it could be applied to any therapeutic area. And we're just excited to be at the forefront. We're the first ones to really have a PROTAC or PROTAC-like molecule enter the clinic, and now we have had 2 in the clinic for over a year.

Great. And maybe one other question we get a lot from folks is just is why did you choose to go after AR and ER as your first targets of interest here?

Yes. So I think this gets to the point that with a powerful technology that you can apply to nondruggable targets, and there's plenty of them, there's a feeling like, well, why don't you start with one of those? The reason why we started with ER and AR, so estrogen receptor, androgen receptor, these are among the most validated targets or the longest validated targets in oncology. And the feeling was that this would be a cleaner way to establish proof of concept. That's number one. But #2 is that in both cases, there is still -- the way I put it is there's still a lot of mileage left in trying to target these receptors coming in with a very different modality. Instead of an inhibitor, come in with a degrader because this is something that could actually work in patients who have developed resistance to the existing forms of endocrine therapy. And also, it has a great potential for working in earlier lines as well.

Okay. Great. And maybe one of the other questions we get a lot is, obviously, you guys are kind of frontrunners here, but there are some other companies now private and some of the larger pharma companies that are looking to get into the space as well. So maybe how would you differentiate your approach maybe versus some of the approaches taken by the competition?

Yes. So I think a lot of it is just that we've been doing this for a while. And there is a definite -- let's put it this way, like a learning curve. There's a lot of pharmaceutical chemistry and also preclinical testing that is unique to this class of drugs, and there is definitely a learning that is acquired. We also have a lot of resource applied to this. We've been doing this since 2013. We're a company of about 150, a very large part of that are medicinal chemists and biologists. And we also have a large group of contract chemists as well. So we've been doing this a while. We've taken a lot of learnings from this experience. We've been able to design molecules that have degraded that are orally bioavailable, which has been one of the big hurdles to make them drug-like in properties. And then also drugs that for our other therapeutic area of interest other than cancer is neuroscience, drugs that can actually cross the blood-brain barrier and degrade a lot of the targets that the field has been really looking for the best modality to do that because the existing modalities are just inadequate.

Great. And what are -- obviously, with any new platform or technology, there are hurdles or risks. So as you think about that side of the equation here, what are some of the key things that are hurdles or what are some of the outstanding questions that you see just from a kind of platform basis?

Yes. I would say that from a -- just from a clinical side, really focused on making sure that for a set of molecules that are going to be larger than the traditional small molecules and this concept of Lipinski Rule of Five and can you really make these things absorbable, have proper pharmacokinetics, and reliable and predictable pharmacokinetics? Can these be drug-like? So as the Chief Medical Officer, that's what I'm looking for. I'm looking for also that from the preclinical package that we have to make sure that there is a clear therapeutic index. So just looking at what we have for 110 and 471, we're just -- I mean I think the -- and especially 471 -- I'm sorry, 110 that we can talk about today because that's the data that we just presented. There's continued signs that this behaves like a regular drug and the good thing is that we now have efficacy shown with our first PROTAC, and it really supports the concept of this platform as a whole.

Okay. Great. Well, that's a great segue maybe to the ARV-110 data. The first full Phase I data presented at ASCO about over a week ago now. Would you just love some perspective on that data, kind of what were the key takeaways in your view as we think about the data now that we're a little bit past the presentation.

Yes. Certainly. So -- and again, I didn't mention this, but thank you for including us to provide this opportunity to speak to the data. So I think, first and foremost, which is what I was alluding to before, is we actually with our first PROTAC, we're able to now say that we have proof-of-concept for the fact that a PROTAC can demonstrate efficacy. So it's a Phase I study. It's not really intended to -- it's not designed specifically to evaluate efficacy. But the fact that at this stage in development, as we're continuing to dose escalate, the fact that we can show efficacy is a testament to the platform and the work that we've been doing since 2013. So that's number one. And then just thinking about the drug, specifically 110, and part of that is to understand the patient population. And I think this is maybe something that the investor community may not be as, I guess, familiar with, I mean this is a -- the population that was enrolled in this trial was a very advanced population. And I would say probably more advanced than I think a lot of the folks that we've been talking to over the months would have expected, and I think requires a little bit of, I guess, illustration, and that comes in the form of the fact that the median number of cycle or regimens administered in the castrate-resistant setting was 7. That's a lot of regimens that these patients have received. They've received within that, a lot of AR-directed therapy. So even there, most of our patients have received anthracycline -- I'm sorry, had received enzalutamide and abiraterone as well as the traditional androgen-deprivation therapy. So these folks had every general modality to target AR and in addition to that, chemotherapy. So this is sort of like a seventh line population. And in our review of the literature and talking with experts in the field, there's really not an expectation that another AR therapy that exists today could really have any meaningful effect here. Yet we have 2 -- we have a very clear efficacy signal. We have 2 patients who had a confirmed PSA response. And one of these 2 actually was evaluable for RECIST. This patient had tumor that was surrounding his major vessels, and he had a confirmed partial response. This patient had an 80% reduction in this tumor mass by CAT scan, and again, a PSA response that was confirmed. And these patients are both ongoing. One of the patients was ongoing up to 30 weeks at the time of the cutoff and the other patient was 18 weeks. And they are very deep responses. One of the PSA responses was 97% reduction in his PSA. So again, very clear signal of efficacy in a Phase I trial in a heavily pretreated population. I think the other important takeaway is that the data already provide us signs that there may be certain populations that could benefit even more from this therapy. So preclinically, we already knew that there was one androgen receptor point mutation that would not be degradable by 110. 110 degrades the other common AR mutations. But the one that was not degraded actually turned out to associate with really no response, and that was sort of expected. So it's actually an evidence of our preclinical data translating. Also AR-V7, which is a form of AR that is sort of truncated by a transcription error, we know that we don't degrade that because the binding pocket is kind of locked off. And not surprisingly, we did not see much activity in the AR-V7. However, if you look in the relevant doses, and when I say relevant doses, I mean, those that are achieving exposures that align with exposures reached in the preclinical models that show activity, we show that of 7 patients who have AR forms that are degradable by 110, we have 2 confirmed responses, again, including the RECIST response. So 2 out of 7. And -- so that's -- that gives us a really good opportunity as we continue the dose escalation and for expansion. And I think one other thing I'll mention is that of the 2 responses, they both had a very similar AR mutations. The question that we're getting, of course, and we've asked ourselves is there some reason why the mutation patients may benefit even more? It's not because of 110. We know that 110 degrades the degradable forms comparably across the different types. However, there may be some biological explanations for it. It's sort of speculation, but it may be that some of the mutations may be more drivers in these cases. So what that is doing is to inform us about how we can adapt our Phase II study. We have a Phase II expansion that's already written into the protocol. We'll start that once we get a recommended dose, but we are going to be refining that design so that we can get better information and more complete information about some of these subtypes, like the mutations, because that may enable us to have a much quicker registrational path if it turns that these patients are uniquely responsive to 110. And then last thing I'll just mention is that safety has been really, I would say, quite clean, especially compared to a lot of other cancer drugs, and that enables us to go up in dose. And so we did mention that we've gone now up to the 420-milligram dose and it also allows us to think about taking this drug into earlier settings as well.

Okay. Great. Maybe just a couple of follow-ups. I guess, one would be in terms of the baseline PSA levels, maybe you could speak to that a little bit. I know that was a question coming out of ASCO. Just is that -- does that matter either way, if a patient has low PSA or high PSA as likelihood of having a response? And then I guess the other question -- well, maybe we'll stop there and I'll give you my other question.

Yes, sure. Yes. So the 2 responders, they were actually were on some different ends of the spectrum. So the one patient who didn't have RECIST evaluable, so his disease was in multiple bone sites. He had something like a 16 -- PSA of 16 at baseline. The patient with the RECIST response had a PSA of around 275 at baseline. I may have actually stated something different in the call that we had 2 weeks ago, but it was 275. So on both ends, which is -- to us, is good because at least we know that, that's going to be -- that it's going to be broadly active across that spectrum.

And where did most of the other patients fall in that spectrum? Like if you look at the mean or median, like what was the scope of baseline PSAs?

Yes. It varied. I think at the 140-milligram dose, it was about 200. We did have one patient come on with a PSA below 1. In Phase I, we are having a more sort of open enrollment criteria. And that patient was one of the -- was one of the 2 who are not PSA evaluable. But otherwise, it was a pretty good spectrum of PSA.

Okay. Great. And I guess the other question is just in terms of the dose response or lack of a dose response thus far, I know you guys talked about this on your conference call. But I think it would be good to kind of just review that and kind of how to think about that aspect because I think that was one of the pieces that investors are still asking about is just, as you've mentioned, you're dosing higher here. But why don't you think you've seen a dose response at this point based on what you know?

Yes. Thank you for asking the question because, yes, this is one of those questions that, I guess, not surprisingly, we're getting. This is also one where I can help, I guess, provide some context here. So really -- so we have 7 patients at this dose. The last 4 came on towards the end of the time leading up to ASCO. And really, our conclusion is that we really have insufficient data to assess. So it's not even that we don't have a dose response. It's really insufficient data. And I'll explain that a little bit. So again, the last 4 patients were enrolled just prior to the ASCO cutoff. They had one month follow-up. And so it -- according to the standard guidelines, 12 weeks is really the best time to assess for best PSA response. We know that in terms of PSA kinetics for drugs like enzalutamide, where there's been a lot of research on this while response can happen quick, they also can -- you can have a patient who has no response and then -- at 4 weeks and then turn into a response. So that's the last 4 patients. The first 3 patients are really not informative about the 280-milligram dose relative to efficacy. So one patient was the one and only dose-limiting toxicity. This was a consequence of an interaction that we identified early in the trial. And that patient came off after just 2 weeks. It was actually even before the drug could reach steady state. So that patient was not evaluable. We had another patient who had the mutation I mentioned earlier, which is the one that 110 does not degrade. It's this L702H. So that patient actually came off after one month, not surprisingly. And so that doesn't really give a good test of what 280 can do. And then the third patient of the first group came off -- actually was dose reduced 3 weeks into treatment, out of caution because of the other dose -- because of the one dose-limiting toxicity, it was before we had the explanation around this interaction. So that patient really quickly went down to 140. Incidentally, that patient is still on therapy. He's received several months of therapy. The patient started last fall and actually just recently dose increased to 280. So we actually -- I guess the other part of that is, number one, insufficient data. Number two is we actually really believe that there will be a dose response relationship. Number one, the 110. There's a dose proportional efficacy relationship preclinically, very clearly. Number two is that we know in patients that 280 has dose proportional increase in exposure above 140. So from a pharmacokinetic perspective, even with these large molecules, we are getting dose proportional increase in exposure and getting into -- closer into the range of where we have seen efficacy preclinically. And we have a slide in the ASCO presentation that shows 2 sort of benchmarks based on preclinical data. So that gets us into higher range and more confidence about efficacy. We also have -- we had one patient where we had paired biopsies to show that we had AR degradation. That was actually in a patient at 280. And then lastly, we know that the dose is safe. So it's quite possible that with enrolling additional patients, which is our plan at 280, that we may end up even thinking that this as a dose to study in Phase II. So we believe that we will show dose proportion -- a dose-dependent effect, and it's just a matter of just getting enough patients in its Phase I. There's always challenges with small numbers.

Okay. That's helpful. I guess just one follow-up. Is any additional color on those patients of the 280 cohort, like in terms of baseline characteristics? Like is there anything different about the PSA levels? Is there anything different about the mutational profile? You mentioned maybe there's this subset of patients that has a driver mutation. Is there anything different about those 280-mg patients on further respect to baseline characteristics that you've teased out at this point?

Yes. I think -- I mean, so one thing that we're really focused on is really kind of prying through the genomic data. So our focus for the presentation and what we're really centered on was around AR genomic profile that was in the presentation. But there's -- there are other mutations that are part of the panel that we're looking at. We are using a research use only foundation platform. And so we're actually digging through those data because they may set -- establish hypotheses for other ways that we can think about enriching or selecting patients. So yes, we're very much in tune to this. And we'll -- as we get more data, we'll continue to investigate this and have more to talk about later on.

Okay. And is there any possibility that you think you might have to go higher than the 420-mg cohort? Or do you think that's kind of -- you're going to be in a good spot after that? Or do you fully need to explore to get to dose-limiting talks essentially to answer that question so there's a possibility you could go higher just from that perspective?

Yes. We're right now focused on getting to maximum tolerated dose. Again, the safety profile is conducive to that. And so as long as we have that and understanding that there's always extrapolation considerations for going from animals to humans, we will continue to dose escalate. And we're also moving -- evaluating or establishing a new dosing strength that makes this easier because we're now up to a fair number of pills. Yes, and we'll use all the data that we have at our disposal to help in defining recommended dose. We're doing a lot of modeling, looking at PK/PD analyses and looking at different PSA metrics, just -- more than just a decline of 50% or more. So as we continue to go on, I think we'll have a lot more information to help us on dose selection.

Okay. Great. And maybe the other one, again, I think we will get some perspective on is just that the liver function elevations that you saw with Crestor. I mean one of the questions that we've got from folks is I guess, confidence that this isn't kind of a platform issue that is going to be limited to Crestor only. And wouldn't impact anything more broadly. So maybe just any more details you could share on what you know so far about it and confidence level that it won't be seen more broadly?

Yes. We're confident that it won't be a platform issue, number one, preclinically in terms of the transport testing. We don't have -- I mean, the results are not equivalent between different PROTACs. Number two is that, of course, going into this, we're thinking, I mean, these are -- well, they share the same backbone structure, if you will, for a PROTAC with the 3 domains. These are very different molecules. The warheads are very different, of course. So we're not -- we would be surprised if there would be a platform effect. And then I guess one other little sort of tidbit is that in the 471, our breast cancer program, we did actually have one patient on rosuvastatin, who had no liver enzyme elevations. And for that matter, really very little toxicity of any sort. And that compares with 2 out of 2 patients who had the Grade 3, 4 liver enzymes. So we have a good level of confidence on that.

Okay. Great. Good to hear. And I guess the -- in terms of the update, I guess we're going to get an update later this year. Again, I'm assuming that would include additional patients follow-up from the current doses. But maybe anything more in terms of some of the genomic analyses that you guys are conducting? I mean how much more data should we expect in kind of some of these deeper endpoints and other analyses you're doing at that point? Or is this more going to be an update kind of similar to what we got to ask, kind of focused on follow-up on the current patients and then the next kind of cohorts of patients?

Yes. I mean, we haven't yet provided any specific guidance on numbers of patients and specifics about how much biomarker data, but our intent is to provide a good update towards the end of the year to catch everyone up to speed on how the program is progressing. And we also -- for the 471 program, we do have a plan for a 4Q disclosure there. And so we think that it will be a good update for everyone at that time.

Okay. Got it. And any -- I guess, it's a little early, just the last one before we go on the 471 is in terms of the enrichment strategy. Maybe any preliminary thoughts there on what that might look like?

Yes. Yes. I think how we're looking at this is that we have sort of 3 groups of patients based on their AR genomic data. So we have -- we'll start with the AR mutation group. So we have the 702H, which is this one form that's nondegradable. But there are, obviously, a number of other AR mutations. The frequency of AR mutations is roughly about 15%. There's not a lot of specific data in this exact population. So we'll certainly be learning more here. And this is the group that, at least as an early signal, there's an intriguing association between these 2 responses and these 2 very deep -- these 2 deep responses in these 2 mutations that are seen in both patients. So our interest is to really refine the study, the Phase II study, so that we can really get enough of these patients to really follow through on the signal because if we can -- if we're able to see that this holds up in more patients, then this becomes a very intriguing way to target a group of patients for a quicker registration with the concept that this is -- will be compelling data that you can establish a good, strong case based on a single-arm trial in an advanced population. So that's one group. The -- on the other end of the spectrum are patients with, again, 702H, where our priority, we think, well, it may not work so well. And then we have AR-V7. That's a little bit more complicated because AR-V7 is truncated form, that's been -- there's been a lot of interest in this over the last 10 years or so. This is one that we know that 110 does not degrade. It's been a little bit of a debate about is this a driver or a passenger for resistance. So as we get additional data, we'll have a better understanding about that. And if it turns out that the drug does not work in the setting, then it helps us really refine the population to know exactly who's the most likely to benefit from this. And then it leaves the middle group, which is all the wild types, the amplified patients, which is sort of -- becomes a much larger group of patients. And so the idea is that we can sort of focus our analyses on these different groups, make sure that we have enough patients in Phase II within these individual cohorts. So that's the thinking right now, and we're acting very quickly to expedite those changes so that we can sort of adapt to the science.

Okay. And maybe just one follow-up is just on the 878 and the 875 mutations that you may mentioned or maybe they're more responsive is do you only think about the structure of those in terms of those mutations, like where they fall in the protein? Would there be some kind of better interaction with your drug or something? And then the second question is, did those also show better activity preclinically? Or is there just not a model for them?

Yes. It's a good question. So I would -- with regard to does -- I mean, does 110 have some sort of special type of degradation kinetics or whatever with this group, we actually have a very sort of comprehensive package of data. That's one thing that I was very impressed by the way when I came to the company is the extent of research on 110 and 471. And that includes a lot of degradation work on different mutations. And what we know today is that there isn't any sort of preferential degradation on 878 or 875. And yes, so really, as we think about this, it may be more about the biology of prostate cancer and what happens in these patients where -- and maybe 878 and 875 are a little bit more sort of associated with the driver type of signaling from AR. But it's really just speculation. I mean, this is -- these are things that we'll learn as we go through the clinical program and -- but the good thing about this is we have a -- we already see a pattern very early on, and that's what you want in Phase I oncology is to sort of see the patterns. And I've been doing this and the medical oncologists have been doing this for 25 years, including starting as a Phase I oncologist, and it's -- we always hear the stories about sunitinib was discovered that it had 2 kidney cancer responses in Phase I. It was before they made the connection with VEGF. So it's the same thing. We have to act on this signal and just really -- our responsibility as a follow-up on that, and that's what we'll do.

Okay. Sounds good. Maybe just in the last 10 minutes or so, I would love to cover ARV-471. So a Phase I program going -- ongoing here. Again, you mentioned data in the fourth quarter. So just maybe remind us what we know already about the drug, what you've disclosed? And then kind of same question, like how much data are we going to get in this update in the fourth quarter?

Yes. So we -- the first peak at clinical data was back in October. So we -- there was a targeted protein degradation meeting in Boston. We took the opportunity in that forum to present the initial data on both programs, 471 had started later. So we only had data from the first cohort. And in that cohort -- excuse me, we -- it was actually a nice story back then because, with both drugs, we showed that these had drug life properties in terms of the pharmacokinetics. We are able to get absorption, which was sort of not a trivial thing at that time since these were the first PROTACs in the clinic. And we were able to establish exposures that correlated with exposures seen in the animal models that showed activity. So that was all really good. And safety wise, at that time, things really uneventful, really no adverse events to speak of. So that was back in October. Since then, and John Houston in our call with investors 2 weeks ago had just sort of given a very quick verbal update as part of his pipeline update. We're very pleased in the progress of that trial. So we have not had any dose-limiting toxicities to date and continue through the dose escalation process there. And we are continuing to see dose proportional exposures. So obviously -- and so that might be just at least a quick allusion to that. And we also mentioned that we have had the first signs of ER degradation in that trial. We tried not to get into specifics. We did not want to take away the attention from 110. But we just wanted to make it clear that this was not a sort of a -- just as pure as -- not a one-trick pony, I guess, in terms of 110 that we have 2 drugs in the clinic. We're, obviously, working very hard on other programs in discovery to continue to advance our pipeline.

Great. And how do you think about this patient population? I mean you mentioned that part of the 110 story, just how late-stage these patients were with median 7 prior lines. They've already received AR therapy. So as you think about the patient population you're exploring here, is it similar dynamic where these are late-stage patients and there's something maybe underappreciated about them? Or do you think these patients might be more responsive to therapy? Just want to understand kind of the air bars, I guess, around how to think about...

Yes. And that's an important point here. And obviously, for us, most people look at the SERDs as our competitors. I mean, the SERDs are sort of a completely different sort of mechanism, I guess. But when you look at this -- the setting and where we're testing our patients, the one important point to highlighting here is that all of our patients must have received a CDK4/6 inhibitor. So this is drugs like palbociclib. And this is actually a little bit different than what we have seen in terms of the SERD data because a lot of these have been done in mixed populations, where prior CDK was not a requirement like it is with our trial. So that's number one. We did have allowed up to 3 prior lines of chemotherapy. And in breast cancer, that can be -- that's a fairly significant amount of prior treatment and looking just sort of to compare with the SERD studies, you'll see that sometimes 3, sometimes 2, sometimes 1. So in that regard, it's also going to be a heavily pretreated group. So -- and that's important for us in terms of how we think about benchmarking these data. The other thing is that there's not a lot of data in the post-CDK setting. Now at ASCO, there's -- some of the SERD compounds have been tested or have disclosed data in the setting, and now the data are starting to be reported by activity after CDK versus not. And yes, so that's going to be important. And I think -- yes, so I think when we have our disclosures, it's going to be sort of putting in the context of those results.

Okay. Got it. Understood. And then maybe just in the last few minutes, just remind us in terms of kind of where you stand for your CNS programs. I know that's the other big focus. Obviously, cancer is in the clinic now, but CNS is kind of going to be the second part of the story here. So any updates on that front that you can share? Or what are the next milestones we should look for?

Yes. So we have -- we've actually been staffing up on the neurology side. We have a really strong effort. We've got really exceptionally talented folks who are involved there. We had a lot of great experience in neuroscience before they've come to Arvinas. And yes, so we're continuing to work through our tau program as the lead neuroscience program. We have a line of sight to clinical candidate. We're estimating that we would have a candidate sometime in 2021 and potentially start clinical program in the following year. That always, of course, could change, of course, as we're continuing to make discoveries and make different advances there. And we're especially excited that we are able to show that these -- the PROTACs that we are developing are -- have the ability to cross blood-brain barrier to a very significant degree. We did have data disclosed in October that shows that even in a murine tauopathy model, we were able to -- by administering the drug peripherally, we're able to degrade pathologic tau. And so that has really excited the neuroscience community. And we really see that there is a real potential great advantage for PROTAC here versus an antibody that can't get into the intracellular space that has to be administered IV. The penetration in the blood-brain barrier is going to be very limited. And then the antisense molecules that have to be administered in the spinal cord, which is really -- just really a big challenge for patients, so. I mean we're very excited about the neuroscience program.

Great. Well, I think we're right up against time, Ron. But thank you very much for answering all the questions. Appreciate the time and best of luck over the coming months and take care.

Thank you. And thanks for giving us the opportunity to speak with you.

Thank you, everyone. Have a good day.

Bye.

Bye.