Hi guys, I'm Ellie Merle. I'm one of the biotech analysts here at UBS. Very happy to have Arvinas here with us at the UBS Healthcare Conference. With us from Arvinas is John Houston, President and CEO, and Ian Taylor, Chief Scientific Officer. Very happy to have you guys here today. With that, we'll jump right in and maybe high level, can you talk about the benefits of a PROTAC versus a small molecule inhibitor and the vision that you see for your platform over the next one to two years, as well as long term and the applications of PROTACs? Hi, Ellie. Great to see you. Yeah, I can start off, and I'll hand over to Ian to give more of kind of the rationale for differences between degradation and inhibitor approaches, which we do believe drives a lot of the differential biology that we see pre-clinically and now we believe seeing in the clinic. Maybe just to start off, just again, a brief statement of where we are as the company today. We have two programs in phase II in the clinic. ARV-471 is the first one, which we believe has the potential to be a best-in-class estrogen receptor targeting therapy. So far in the clinic, we're seeing profound ER degradation, good tumor responses, outstanding safety profile, and our VERITAC phase II dose expansion trial is ongoing. ARV-110 is in patients with late stage metastatic castration-resistant prostate cancer, and the ARDENT Study is in the dose expansion stage and also progressing really well. Beyond those two lead programs, we have another program, ARV-766, which is another ER degrader, which will be in the clinic in the first half of this year. Around 20 other programs in oncology and neuroscience with the idea of getting to one IND per year from that pipeline. We should have another four or five IND filings between now and the end of 2023. It's an exciting stage for the company. A lot of excitement in the protein degradation space in general, and we are happy to be still very much in the leading edge of this technology and this platform, and we'll be happy to talk about all aspects of our pipeline today. With that intro, maybe Ian can give you the answer to the specific question you're asking is, differences between PROTACs, protein degradation, and then inhibitors. Sure. Thanks, John. We always look at how PROTACs and protein degradation in general have advantages over small molecule inhibitors. Obviously, one obvious one is just from the name. Small molecule inhibitors inhibit the protein, in the enzymatic function, whereas a PROTAC fully degrades the protein. That's important because a lot of targets have more than just their enzymatic function. They serve as, for example, scaffolding proteins, so other proteins bind to them, and that adds to their activity. By just inhibiting the enzymatic function, you're leaving behind that scaffolding function, but by degrading the protein, you're not only removing the enzymatic function, but also the scaffolding function, which can be a big part of its activity. We have a number of programs that fit into that category. Of course, also there's the factor of the iterative or catalytic mechanism of action of the PROTAC that leads to an event-driven pharmacology. One PROTAC molecule can degrade multiple, we've measured up to 200 molecules of a protein by a single molecule of a PROTAC in some cases. You have an event-driven pharmacology, the event being just the binding and then the formation of the trimer complex between the E3 ligase and the target protein that leads to degradation, as opposed to occupancy-driven pharmacology, which is an inhibitor. You need an excess amount of a small molecule inhibitor to have antagonistic or inhibitory activity. Actually, PROTACs work substoichiometrically because of that iterative mechanism of action. We believe that leads to requiring a lower exposure of a PROTAC relative to inhibitor, which can, in the long run, lead to fewer adverse events in the clinic. Of course, you have the whole undruggable class. The undruggables are labeled as such because there's been no small molecule inhibitor that can bind or function. A PROTAC, the warhead that binds to the target protein can bind anywhere. It doesn't have to bind to an enzymatic pocket. It can bind to an allosteric site. It actually doesn't even have to bind very tightly, because really what you're trying to do is optimize that trimer complex formation, that protein-protein interaction between the E3 and the target protein. That's the key parameter, not so much where it's binding or how tightly it's binding. That's another advantage, that you can get into that undruggable space, which has been estimated to upwards of 80% of target proteins are in that undruggable class. Absolutely. Thanks. That's a helpful overview. Just in terms of the differences between different types of degraders, such as particularly molecular glues versus heterobifunctional, can you talk about how these are different and what the potential advantages and possible disadvantages are of PROTACs versus molecular glues? Sure. A PROTAC is a heterobifunctional molecule, as you mentioned, so there's two ends to the molecule, held together by a linker. One binds to your target protein of interest, and the other to E3 ligase, so you're dragging an E3 ligase to a target protein and actively forcing its ubiquitination and then ultimately degradation. Of course, the linker is a key part of the molecule because it helps orient the target protein and the E3 ligase and really optimize that interaction, optimize ubiquitination, and optimize degradation. A molecular glue is also sometimes called as a MonoTAC. Really what it does is it binds to a protein, in this case for degradation, it would be an E3 ligase, changes the conformation of that protein a lot, Basically exposing new surfaces to which a protein which normally wouldn't bind to it and be degraded, now does bind and is degraded by it. The glue is sort of helping bring those two proteins together, mainly indirectly, mainly by causing those new protein-protein interactions. The thing with the glues is it's hard to prospectively design them to target a protein of interest. Usually you figure that out from a phenotypic screen. You kind of figure out bootstrapping backwards what protein is being affected. That certainly was the case with the IMiDs, you know, thalidomide, pomalidomide. I think it's still, I think people recognize it's still very hard to design a glue prospectively for a target as opposed to a PROTAC. You know what target you want to degrade, you add a warhead to it, you have an E3 ligase ligand, as I said, you bring the ligase to it, and so you can prospectively degrade. I just find it a much more efficient, if you will, form of degradation than a glue. Obviously, glues are smaller, and that's what people point to and say, well, they're going to be more drug-like. I think our data with ARV-110 and ARV-471 has clearly demonstrated that PROTACs can be just as drug-like as a traditional small molecule. Yes, they're on the larger side, and they're larger than glues, but that doesn't mean that they can't be drug-like. Again, we've shown it for 110, 471. That advantage-disadvantage argument I don't really buy into because of the clinical data that we've shown. Again, I think PROTACs just again, more efficient because you can more prospectively go after degradation as opposed to glue approaches. Maybe in the future there'll be a breakthrough there, but right now PROTACs clearly have the advantage in that area, and that's a big advantage in my opinion. Mm-hmm. Absolutely. Thanks for that color. Just thinking from a kind of platform perspective and all of the years of know-how that you guys have built up and your internal work around assays, the unique PK/PD. Can you describe some of the processes that you've developed and I guess why it's so hard to make a PROTAC? I mean, in principle, it sounds like a simple concept, but I think in application it's proven to be much harder than maybe the concept of a jab. Maybe what are the things and internal know-how that you've developed, and if you could talk a little bit more about that? Yeah. We've done a lot of investment into the platform, if you will, of PROTACs. I guess I'll change your question around a little bit. I think actually what you've seen, because so many people are doing it, so many companies, academic groups, a lot of people can make PROTACs that degrade. That part actually is not difficult. What's really difficult is turning them into drugs. Like they have the drug-like properties, they can work in vivo, oral bioavailability, cross the blood-brain barrier. That's where we spent a lot of time, really from 2015 onwards in particular, focusing on to turn what Craig Crews used to call a molecular chemical biology parlor trick into drugs. To your point, we had to revamp some of the assays that are traditionally used in SNAP cascades. Because there's no two ways around it. PROTACs are class 4, BCS class 4 molecules, meaning they have low solubility, low permeability. When we were doing permeability assays, we had to revamp them to tailor them to PROTACs. The traditional assays weren't really being informative. Same with plasma protein binding assays, these traditional ADME PK type assays that are more designed for traditional small molecules. We had to revamp those. The PD, as you mentioned, and obviously spent a lot of time on oral bioavailability, but because of the catalytic activity, that intermittent mechanism of action that I talked about, you do get extended pharmacodynamics relative to the plasma PK. Learning those parameters, that extended PD, making the PK/PD correlations to efficacy required a different way of looking at it, different calculations, different data sets, all of those things. In our deck, I don't know if you have it available, on slide seven, I believe, we have the PROTAC discovery engine laid out. We have in those two buckets, we have ligase selection, ligand identification. We also have rapid PROTAC design, where really we focus on optimizing the zone of ubiquitination. For a while, that was more empirical, right? Make a compound, see how it degrades, make another one, sort of the traditional medicinal chemistry SAR, if you will. Over the last several years in particular, we've gone into more computational modeling, CADD approaches, to be able to model that, using a new generation of linkers that connect the molecule, again, to optimize that zone of ubiquitination. Using artificial intelligence now to do that. These are all advances that we've done to make the process more efficient, more prospective, more predictive, which E3 ligase will pair which with target protein better. All of those things we've built in to make that process more efficient, again, around PROTACs. Of course, the drug-like properties, the oral bioavailability, blood-brain barrier penetration I mentioned as well. Those are all things that, of course, we view as our Arvinas know-how that we built in. Again, to make the PROTACs, these tool molecules into drugs. Great. Helpful. Thanks. A question for John, just on kind of strategy and prioritization. There's obviously a lot of areas of potential growth for the field across novel E3 ligase identification, novel ligand identification for the quote on druggables, or even just creating PROTACs for the druggable molecules where you see that you can improve on a traditional small molecule inhibitor. How do you prioritize investment across these various sort of spectrums and thinking about kind of the strategy for the company and for their growth? Yeah, great question. Clearly, the company's been highly focused on the oncology part of our pipeline and over the last few years, our neuroscience pipeline. As you could argue, that's a nice balanced risk within our portfolio. We know that neuroscience is tougher. There's a higher risk there. Oncology, we believe that there's a lower risk, but still a huge unmet need. We've balanced the portfolio that way, and we realized those were good choices to make. We've added to the oncology piece by having some immuno-oncology programs. We're also looking at aspects of the neurological rare disease. We're adapting even that core strategy. We are focused on those areas as being the ones that will drive the biggest value and where we think PROTACs can have the biggest impact. Having said that, we also have a significant part of our resource focused on continued platform development. Certainly, the platform we started the company with has moved on considerably since 2013. That PROTAC discovery engine now has quite a significant ligand discovery capability, ligase evaluation capability. As Ian talked about, artificial intelligence and other aspects of lead optimization that allow us to put in these drug-like properties into PROTACs with the necessary requirements either for oral bioavailability or blood-brain barrier penetrance. That capability has grown and grown. It's a balance of significant portfolio activity to keep building the pipeline in core diseases, also building the platform out. We do keep an eye out for other potential scenarios in different diseases. I wouldn't necessarily say it was opportunistic, but if there's a target or a very obvious disease area that we think could be positively impacted by a degrader, we will assess that. Also potentially looking at partnerships as it relates to that. We do have some significantly strong partnerships with Genentech, Pfizer and Bayer. We have a joint venture with Bayer also in the agrichem space, just to show you the facile nature of the technology is it can be applied outside of human disease. We're looking at all the different ideas and opportunities. We want to keep laser-focused on our pipeline, but there are opportunities outside that we think we could still grow into over the coming years. Mm-hmm. On the E3 ligase, I'm sure you're probably not going to answer this, but I have to try. How should we think about when we could potentially see an asset with a novel E3 ligase sort of enter IND enabling work or potential IND filing? I know you guys have spoken a lot about your E3 ligase identification work. Any more color that you can give us beyond that? Well, we've now published our first two structures for ARV-110 and ARV-471, you see certainly with those two programs, we hijacked a cereblon ligase, which I think for some people was surprising. I think the guessing was it was VHL. When we look at our portfolio, we do still have an array of different ligases that are, again, deployed. The answer to the question will be yes, over the coming years, you'll probably see INDs with different types of ligases. Which ones they are and when they will occur, we won't say. We're certainly focused on using the appropriate ligase for the particular target and for the type of properties we want to have with the molecule as well. Yes, we're still taking a more diverse approach to ligases than just one single ligase. Mm-hmm. Got it. I see a question from the audience here, just in terms of the partnerships with Roche, Pfizer. Can you talk a little bit about sort of the progress that you've made with those collaborations, say, in the past couple of years since those were formed? Yeah, no. We believe they're being very successful. We're not able to talk about them under the contractual agreement between the different companies. They don't want us to talk about the targets or how well they're progressing. The only way we can say they're progressing is pointing to our cues and showing the milestones that we're getting from those different companies. Programs are progressing, and they are, we believe, very successful. The fact that Genentech added to the collaboration that was set up in 2015, and they added again to it in 2017. Bayer, we did a discovery deal with them, but also we did that joint venture. I think, yeah, these companies have been very interactive, and it's a very good relationship with them. Mm-hmm. Got it. Thanks. That's helpful color. Just big picture maybe for Ian, or I guess I'll turn it to both of you guys for this one. Just in thinking about long-term, the mechanisms of resistance against degraders. I know there's some emerging literature around that. How are you thinking about what the key drivers of resistance could be and how you're sort of developing your strategy around that? Ian did experiments on it, so over to you, Ian. Yeah, sure. We've certainly generated cell lines that are resistant to 471 and 110 in the lab by culturing them with increasing concentrations. The traditional way to do it long-term, to try to identify mechanism resistance. I think there's going to be multiple ways that cells become resistant to PROTACs, and they will become resistant to PROTACs. Cancer cells are very good at avoiding and eluding therapeutic approaches. We certainly are aware of the publications that have identified resistance mechanisms through the E3 ligase, either through mutation or translocation, because a lot of those publications have been using some of our PROTACs that we had published on earlier, particularly with BRD4. I suspect it's going to be a promoter of the E3 could get methylated, expression gets turned down, gene gets rearranged, et cetera. In ARV-110 and 471-resistant cell lines, we've not seen that yet. We've not seen the ability of the degraded ER and AR to be lost, or other proteins to be lost in the cell lines that we've used to generate those resistant cell lines. We're still characterizing what the resistance mechanisms are, they seem to be more ER and AR independent. It could be that in other cell lines that if we were to create them, we would see an effect on the E3s. That's still to be determined. It makes sense that the cell could figure out which E3 ligase is being used, as long as it's not essential for growth, that would be something that could then be shut off and prevent degradation. That's part of our criteria for selecting new E3 ligases, to your question about strategy going forward, which ones, particularly in the cell types that we're interested in, may be essential to growth, and therefore, the cell couldn't shut it off and still survive. Those are the kind of things we're looking at, those kind of things we're characterizing in our current resistant cell lines. Like I said, it's likely to be multifactorial in the long run. Mm-hmm. Got it. No surprise there. I think just in terms of the ability to measure degradation and sort of the proof point, I know we've seen some initial biopsy data, both from ARV-110 as well as ARV-471. I guess from a mechanistic and kind of platform approach perspective, how do you measure the difference between degradation versus inhibition in patients, and particularly across various tissue types, and in terms of being distributed across the body? Oh, well, in patients, that's the practical matter. That's really tough to do, right? We've been using for an oncology indication, obviously, tumor biopsies pre- and post-treatment. With ARV-471, we've gone to the AQUA method, the quantitative immunofluorescence, over the traditional chromogenic staining, just because that method's clearly very subjective. It requires a pathologist to score the samples and provide an H-score. Usually, you require multiple pathologists because you will get pathologist-to-pathologist differences in the scoring. Whereas with the AQUA method, it's more objective. It's a computer, basically, a computer algorithm scoring based on the quantitative immunofluorescence. It kind of takes that pathologist out of the equation. In patients from a practical standpoint, especially in a clinical trial, especially in a phase I dose escalation where we've generated our samples. The degradation's always going to be a secondary endpoint because you're trying to get to the recommended phase II dose. You're trying to identify any DLTs, the maximum tolerated dose. You don't want to slow down that aspect of the trial to measure degradation, which is why we always make the paired biopsies optional, at least during the escalation phase. That's from a pragmatic and practical standpoint, that's where we're limited. It's going to be impossible to show a difference between degradation and inhibition in patients because you'd actually have to run that trial, and that's just, again, not practical. Unfortunately, the tools are limited. We're using the ones that we have to the best of our ability in the context of a clinical trial. That's in terms of tissue distribution, again, we can do that much easier pre-clinically just by using mass spec analysis or even radiolabel studies down the road. Again, that's really hard to do in the human clinical trial setting. We're doing the best we can with what we have, and I think the data that we have so far with 471 in particular, because we've gotten more paired biopsies than we did in the prostate cancer trial, just as we predicted, has shown that in all the samples that we have, ER levels are going down, again, with that quantitative assay. We're very pleased with that, but recognize that really going beyond that with any other methodology is really not practical. Mm-hmm. Yeah. Just to follow up on that, there is some variability across patients. I understand the assays and sort of the availability of the data and it's limited also, dose escalation study. I guess, how are you thinking about the feasibility of getting up to that 90% degradation level in all patients? Is that something that's feasible at higher doses? How are you thinking about the ability both, I guess, with 471, but also, broadly, when you think about the PROTAC platform? Sure. Well, we've already shown that we can get 90% degradation, even for ARV-471 trial, even at the lowest dose of 30 mg, we had a paired biopsy that was 90%. Not all of them are there. That's why our averages was around 62%. We feel like as we get more samples, go to higher doses, that number will go higher, and there will be certainly some paired tumor biopsies where we have it 90% and probably some that'll be lower. Whether we need to be at 90% or lower, that's part of the experiment we're doing in the clinical trials. We've certainly seen pre-clinically that 90% plus gives us tumor regressions, and that's why that was our target, but we certainly knew that having lower levels of degradation also led to robust tumor growth inhibition. Where that range is required to be in humans is something that we're going to try to see as part of this trial. Again, with limited sample numbers, that's just going to be inherent to the trial design. I'm not sure how rigorously we'll be able to determine that, but we're certainly going to try. Certainly, there's going to be differences in ER levels, patient to patient, tumor to tumor. That's been shown historically with all like, for example, the fulvestrant studies that have been done looking at degradation. We see the pre-treatment range of ER levels. It's all over the map. What you hope to see is that whatever level you're starting with, you're getting down to much lower levels post-treatment. That's what we've actually shown when we look at our graph with just the handful of samples we've shown, is that four out of five of those samples are getting down close to the lower limit of detection. In one case, I think basically really close to the lower limit of detection of the assay. That's kind of the trend, and we want to see that for as many samples as possible, regardless of what the starting levels are. That to me is really the proof of mechanism that will be the strongest and obviously that hopefully will come along with a higher percentage. I wouldn't focus so much on the percentage of where is it getting down. It's getting close to the lower limit of detection. I think that's the goal, right? The goal is to eliminate as much ER in the tumor as possible. Of course, with tumors, you're looking at fine needle core biopsies. There will be tumor heterogeneity, of course, not just with the amount of tumor that's in that particular biopsy, from biopsy to biopsy, but the amount of ER in those tumors. There's always going to be some variability just from the biology of a tumor. A human tumor is obviously much different than a cell line xenograft in a mouse, whether it be a PDX or a cell line xenograft. That's just the case. Those are much more homogeneous than human tumors. We're not the first to show that. That type of variability is there, but we feel like we've taken the variability out of the assay with the AQUA assay, with the quantitative immunofluorescence. Mm-hmm. Got it. That's very helpful. I think that's a good point around looking at not the percentage, but sort of the ending level and what proportion are sort of near that below level of quantification. I guess maybe just pivoting to the neurology side of things, maybe, John, can you comment from a strategic perspective, maybe why you find neurology a compelling area for PROTACs and sort of your vision there in terms of target selection and how you view sort of the validation of some of these targets like mutant huntingtin, tau, alpha-synuclein? We were very excited when we saw some data now several years back where the team showed they could degrade tau. This was in a preclinical setting. They were able to show degradation of tau. They then injected that PROTAC into the brains of mice and showed you could degrade tau there. That was the first sign that maybe we could move beyond just an oncology focus into an area like neurodegeneration. Since then, the biggest challenge was can you make PROTACs brain penetrant? That problem was cracked several years back by our team, and now that gave us confidence that we really could build an exciting portfolio of neuroscience opportunities. Why neuroscience? Yes, once you've broken through that kind of technical barrier, then you're right into are these the right targets? Are these the right diseases? Obviously in neurodegeneration, massive unmet need. As you can tell from the portfolios of many other companies in this space, there's been significant amount of failure or disappointment that particular types of drugs haven't managed to show the effects that they hoped for. The question you have to ask is that because the targets were not the right target or the targets weren't drugged appropriately? If you take that latter view, we do believe that PROTACs will actually drug these targets very effectively. They will get into the brain. They'll get into the right compartment of the brain. We will show that we're able to bind and show some degradation in these settings. It leaves the big question, when you do that, does targeting tau or alpha-synuclein or huntingtin, does it actually give you a clinical benefit? The literature has been saying all along that these are the right targets. I don't think they've been appropriately drugged. I think our next phase is showing that you can drug these targets appropriately, then we'll see how well they translate into the clinic. There's still a risk associated with that. We are mitigating some of that risk by also not jumping right into Alzheimer's or not right into Parkinson's, but looking at tauopathy-based diseases or synucleinopathy-based diseases where we know those particular diseases are driven by that dysfunction in the target. If we can see a clinical outcome there, that will build confidence that you can move into these more broader diseases. Having said that, Alzheimer's is clearly a heterogeneous disease where there's multiple targets or mechanisms in play. The idea that some kind of monotherapy was going to be a cure-all was naïve. I do believe you need a multiple combination approach in that space. We'll be thinking about that as we move forward. Neurodegeneration and neurology in general, huge opportunity, massive unmet needs, a real requirement out there in many populations for drugs, and we believe that PROTACs really could have a position to play over the next several years. INDs hopefully over the next year to two years that will get us into that position to test the theory that PROTACs can do that. That's exciting. Makes sense. Just in terms of the challenges of crossing the blood-brain barrier with the PROTAC being sort of a larger small molecule, I guess, what were some of the key ways that you were able to achieve this without sort of giving away your secret sauce, and maybe some of the risks that we should think about as you move from preclinical into the clinic? Yeah, obviously people want to know in the same way that when we worked on making these larger than average small molecules orally bioavailable, how did you do it? Well, we kept that secret. There's more people understanding how to do it. Similarly with brain penetrance, we've worked on the design of PROTACs to allow us to do that. Yeah, we haven't shared any of that. No doubt over the next few years, that will come out, too. Yeah, that technology breakthrough we made several years ago, we'll be able to build on that insight quite considerably. Yeah, I'm not going to tell you everything. Obviously, it's a competitive advantage. In terms of challenges, the challenges are going to be the same challenges that every company that's in the rare space is facing. Getting into the right patient population, understanding whether this is truly a monogenic-based disease, which most of them aren't. What are the true clinical outcomes you're trying to get to prove that your molecule is having an impact? We're taking that all into consideration as we move molecules to the IND stage and think of the patient populations that we move into for proof of concept. Mm-hmm. Got it. Makes sense. Turning, I guess, just to the clinical programs, maybe just big picture for ER and 471. Obviously, a competitive space. A lot of players in development in both the oral SERD space, but also breast cancer, and requires much larger studies to reach the earlier lines of therapy. How are you thinking about this from a development perspective and in terms of segmenting different patient populations as well as longer term, potentially a strategic option in terms of collaborations or things like that, in terms of funding these probably potentially large trials? Yeah, I'll touch on that, and then I can hand over to Ian. Clearly, certainly with the ARV-471 in the metastatic breast cancer space, we recognize it is a highly competitive space. When we moved into this whole area several years back, the view was fulvestrant works as an ER degrader, not a targeted ER degrader like a PROTAC, but the mechanism of degradation had already been proven. The question was, could you come up with a better version of a fulvestrant that was oral targeted degradation? In other words, getting more degradation, and the principle being, if you get more degradation, you're likely to get a better clinical outcome. That was our major driver. Of course, then there's all the SERDs that are out there, which is a mixture of degrader, destabilizers, disintegrators. They all have different names in reality, but they were all ahead of us. The question was, could we come up with the profile of a PROTAC degrader that would be seen as potentially best in class? We believe with 471, we've got that. Our trials are in really late-stage patients. They're 100% post-CDK4/6, which is very different from a lot of the trials you see with SERDs. Over 70% of them are post-fulvestrant, and they've had a median maybe of four therapies. It's a very late-stage population. We went into the trial really trying to show a safety profile. We had no expectations of seeing efficacy or a CBR rate that was particularly high. We were incredibly gratified to see not only a safety profile that looked best in class, but also efficacy in terms of a CBR rate and responses, which was remarkable for such an early trial in a very late-stage patient population. Our game plan is to move, obviously, the program forward aggressively in phase II. We're also looking at combination studies. We've already initiated a combination study with palbociclib, and there'll be other combinations that we can look at, and really to aggressively pursue the program so we can at least close the gap in some of those SERDs, but also hopefully maintain that best-in-class profile, so when the time we get to the market, it's going to have a significant position to play. Got it. Maybe just from a scientific perspective, what drives your conviction that ER degradation would be superior to, say, other forms of ER modulation, such as inhibition, and how do you think about this and the key advantages of a PROTAC? Great question, and Ian can wax lyrical about degradation and inhibition, which our PROTAC does. That's right. We've now published the structure of 471, so people can see that the warhead is lasofoxifene, which is a SERM, essentially, so it has antagonist activity. We've certainly shown that when we make a cereblon-dead, so the molecule's not able to engage cereblon and not able to degrade, you still see very potent antagonist activity. When we compare that to degradation, proliferation, other assays, we see a clear advantage of degrading over simply antagonizing. We've seen that against other antagonists as well. We've certainly seen it relative to fulvestrant in vivo, relating to the points that John had made. All of our in vivo experiments have gone against fulvestrant. We know that not only is degrading better than inhibiting from those experiments, we also know that more degradation is better than less, because we've always been superior to fulvestrant, both in terms of degradation as well as tumor growth inhibition, really tumor regressions. So far from our studies, comparing against the other SERDs that have reported data at the same stage of development as we are, phase I dose escalation, 471 does have the best degradation profile. As well as the safety and the Clinical Benefit Rate, et cetera. That's why we really have conviction that in the long run, even though we are behind and we're working hard to catch up, the 471, that profile will continue. When we get into earlier lines, first-line metastatic combination with palbociclib or CDK4/6 inhibitor, getting into the adjuvant setting, that will win out in the long run. 471 as the best degrader will be the endocrine therapy of choice. That's the strategy. That's helpful. Maybe just to think about the second half of the year update for ARV-471, to the extent that you can comment, how should we think about what we can expect in terms of patient numbers, data updates, and maybe what you're looking for? I understand if you're not inclined to answer, but I figured I'd try. No. In fact, if you look at slide five, it gives you a sense of what we're going to be talking about through this year for both ARV-471 and ARV-110. If you look at the rest of this year, for 471, we'll actually share the end of phase I complete data by the end of the year at a scientific conference. We were still in dose escalation as we started phase II, so we'll give the full story on the phase I data at some point later this year. Our game plan is also to share interim data on the CDK4/6 combo study by the end of the year as well. In terms of initiations for that program, initiate a window of opportunity study and other potential combination studies. A fair amount of activity around ARV-471 in 2021. For 110, similarly, share the completed end of phase I data at the end of the year, and then an interim readout of our ARDENT phase II study at the end of the year, and we'll also be initiating a combination study for 110. As I mentioned earlier on, ARV-766, we'll initiate phase I in the first half of this year. Quite a busy time in terms of data release in the second half of the year and initiations. If you look at 2022, you see similarly, based on how we're progressing, we'll have interim VERITAC phase II data for 471, completion of the CDK4/6 combo data, and we'll share any interim data we have for other combinations. At that point, we'll also share the full ARDENT phase II data, again, any combination data, and also by then, ARV-766 should be in a position where we can share phase I data and initiate phase II for that program. Also by that time, Ian will have pushed a number of the rest of the discovery pipeline over into INDs. Yeah, quite an exciting period of time coming up over the next half year and year with data releases. Mm-hmm. Yeah. Very exciting time. Just a question from the inbox on sort of the safety profile of PROTACs. I guess given that we're sort of quote, "hijacking" a natural part of the body's protein disposal system, how are you thinking about the safety of this long term, particularly given some of the resistance mechanisms that might develop? Yeah. Ian, do you want to take that? I know we've discussed this in detail. Sure. First thing is, yes, certainly, we're hijacking the ubiquitin proteasome system with a specific E3 ligase in the case of 110 and 471 cereblon. Again, because of the potency of the molecules, and they are quite potent, and the iterative action, we're really just using a fraction of the total cereblon that's in the cells. Certainly looked at that pre-clinically, and really a fraction of the capacity of the ubiquitin proteasome system. The mere fact of hijacking it really, even theoretically, is not a problem. Certainly, we've done the appropriate tox studies in rats and dogs, 28-day GLP tox studies. Really in both, 471 profile was remarkably clean, 110 was also very clean. There was nothing that was a PROTAC specific toxicity that showed up. Most of them were target mechanism related, particularly with 471, basically reproductive organ toxicities, basically exactly what you expect if you're degrading ER. There was no overlapping tox between 110 and 471. Everyone sees that the ligase we're using is the same for those, yet the toxicities were actually completely different, mild in general. None of the data that we've had, and certainly in the clinic, we've seen, again, similar types of observations, really no toxicities are really overlapping that would point to, "That's a protein degradation-related toxicity. That's a cereblon hijacking toxicity. That's a PROTAC toxicity just in general." None of that has emerged. I think that's going to continue. Obviously, we're the first PROTACs in the clinic, so it's early data. Other data can emerge, but I think there's really nothing that we've seen that points to a PROTAC general toxicity. In terms of resistance, again, we'll see how that plays out in the clinic. We're going to try to collect the appropriate samples to try to characterize the mechanisms of resistance. Again, that's not as easy to do in human trials as it is pre-clinically with cell lines, but we'll look. Mm-hmm. Got it. No, that's very helpful. I think we have three minutes left, so I'm going to open it up to the audience if anyone wants to put some Q&A in the question box here on the webcast. Meanwhile, just taking a step back, how do you, I guess, well, strategic, but also kind of maybe execution on target-based, in a couple of years from now, where do you see the company in terms of developing and potentially commercial resources as well as sort of investment in R&D, and what are your priorities in terms of, as a company, where you want your focus to be, particularly as some of these assets head towards later stage and maybe some of the genetic subtype populations? In terms of the company plan, obviously we want to grow Arvinas to being a fully integrated biotech company. We have a pipeline that is growing based on our platform, and we have the opportunity to take a number of those assets all the way through to launch and into the market. If all things go well, over the next two years, we'll be in a position where we have line of sight to thinking about what does a commercial plan look like and building that part of the organization. If you get to within two years of a launch, you need to really start mapping out that whole organization. Bizarre to think about it, because just a few years ago, we were a pre-clinical private company, and now we're planning what you need to do to think about commercial plans. That's the exciting part of what a pipeline like this can do and a platform can do. Both for 471 and 110 and the rest of the pipeline behind it, we are really using that to scale the organization. We're over 215 employees here now in the New Haven area. We just announced that we'll be moving into a new building in New Haven, where it'll really allow us to expand significantly. We're building for the long- term, and we have a great substrate to do that. Mm-hmm. Great. With that, I think we'll close it off. John, Ian, thank you so much. Great speaking with you as always. I could ask you questions all day, but unfortunately, we have only 45 minutes, so I'll stop. Thank you so much. Really appreciate it, and thanks for attending the conference.