Welcome to the Morgan Stanley Global Healthcare Conference. I'm Jeff Hung, one of the Biotech Analysts. For important disclosures, please see the Morgan Stanley Research Disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative. For this session, we have from Repare Therapeutics, CEO Lloyd Segal, CFO Steve Forte, CMO Maria Koehler, and CSO Mike Zinda. Welcome, everyone. For those who may not be familiar with Repare, can you provide a brief introduction? Thanks, Jeff, and thanks, Morgan Stanley for having us. Repare is the leading clinical synthetic lethality company. We're focused on DNA damage repair. We have two compounds currently in the clinic that are entirely homegrown. Our more advanced compound is an ATR inhibitor that is expected to have its initial readout pretty soon this fall. Behind it, a PKMYT1 inhibitor that's synthetic lethal with Cyclin E1 and other genetic alterations, and that's a 2022 readout, and a nice pipeline beyond that I don't think we'll get a chance to talk about today. Great. Thanks. Maybe before going through the pipeline, let's start with SNIPRx. Can you talk about the platform and the advantages it offers Repare? Mike? Great. Yeah. Our SNIPRx platform is really a CRISPR-based platform to identify novel synthetic lethal targets to genetic alterations where there's still a significant unmet need in cancer. We really optimized this approach by using isogenic cellular backgrounds, so they're very clean, and it was really to help reduce the number of false negatives and make sure that we could find the maximum number of targets possible for each of these novel genetic alterations. We've been able to, through that methodology, identify targets that couldn't be seen through some of the larger datasets, looking at large cancer cell line panels, and to look at lesions and alterations that aren't represented within these external panels, giving us a real opportunity to bring forward a number of novel targets, including things like PKMYT1, our RP-6306 target that we've talked a little bit about more recently. Great. Then what are the step two screens, and how do those play into your drug candidates? Yeah. Great question. I think the step two screen is where we really leverage that CRISPR platform, but for a different reason. It's looking to understand if there's a network of synthetic lethal interactions that that particular compound might have. Much like with the PARP inhibitors, it wasn't just BRCA1/ 2. There were also these HRD, homologous recombination deficiency genes that have been identified that could provide benefit. This was a screening method to prospectively look for and identify those targets, basically taking any compound, that could've been RP-3500 that we used for our ATR compound, as well as RP-6306, and screening with that CRISPR genome-wide library to identify additional genetic alterations that cause sensitivity and synthetic lethality with those targets. Great. You have a collaboration with BMS using SNIPRx. Can you talk about this collaboration and the update that you recently announced? Yeah. That BMS collaboration was essentially built on the notion that what Mike just described and our ability to discover a fairly rich body of novel targets that we could validate, we always knew that we would have more putative targets than we could ever possibly prosecute into the clinic. In May of 2020, we crafted a deal with BMS, who are, I think it's fair to say, real believers in our synthetic lethal approach, how we found targets, how we validated them, and the DDR sort of sub-focus that is intrinsic to what Mike just described. What was nice is it ensured that if we didn't have the capacity to prosecute something in the clinic, that BMS was likely to pick things that they could take, and we wouldn't have important new insights into potentially important new targets that were sitting on a shelf for others to discover and ultimately prosecute. The recent announcement was that BMS has opted into its first two targets in this relationship, and I think that's important for a few reasons. One of them, I think it's highly validating for us. We see it as validating of the SNIPRx platform that Mike just described and that he and his team have developed wonderfully over the last five, six years. It speaks to a huge accomplishment in this relationship where our scientific teams haven't even got a chance to be together live even once. It's pretty impressive that we got to the stage where they were opting in to two targets. Now they move ahead and do the heavy lifting of advanced validation, medicinal chemistry, and beyond, hopefully getting those things into the clinic at some point. Okay. Maybe moving on to your pipeline, your lead program is an ATR inhibitor, RP-3500. Can you talk about ATM ATR synthetic lethality? ATR is normally a kinase that's sensing replication stress in a cell, sensing when that DNA is being replicated when something's not going right. Under that normal process, it halts the cell cycle and allows those single-stranded breaks and other alterations to be fixed. When you block it with a kinase inhibitor, those single-stranded breaks can now be broken down into double-stranded breaks that would normally be sensed by ATM. In a normal cell, ATM is still there, it's able to sense it, fix those breaks, and the cell can remain viable and progress forward and double appropriately. In cancer, when ATM is lost, it cannot properly sense those double-stranded breaks. Therefore, the cells can progress forward with those breaks and ultimately undergo cell death. There are other ATR inhibitors in development. How do the different ones compare, and what gives you confidence that RP-3500 could be best in class? I think when we were developing RP-3500, we looked back at one of the key factors I think for all kinase inhibitors, that we thought was really critical was looking at the selectivity and getting the most selective inhibitor possible. This meaning that the more selective it is, the clearer you are that any target activity or toxicity you have is due to the target, not an off-target activity. We think we have one of the most selective compounds out there, if not the most selective. We did a lot of work to compare and to look at those. I think that was just the start. We also did a number of other things to help understand the differences between those molecules. In your preclinical studies, you looked at both RP-3500 and Bayer's compound. How did RP-3500 compare, for instance, on tumor growth suppression? Yeah. When we compared head-to-head with Bayer on a continuous dosing basis, we were able to see superior efficacy during that dosing period. Later we went back, and we tested in the three days on, four days off schedule that Bayer had utilized in the clinic. In that case, in that model, we were able to see maximal efficacy, the same as Bayer was seeing. We were able to see a slightly different therapeutic index to myelosuppression that was occurring. Those two pieces of information giving us the belief that we may have a slightly different profile that's allowing us to kind of thread that needle a little bit more finely as we progress forward in the clinic. I think also it allowed us to go back and really test a number of different schedules. We think it's not only the compound, but also exactly how we're going to ask the question clinically, and create those bookends that we're now testing clinically. How does the step two gene alterations differentiate your program versus others in development? How much does the step two alterations increase the patient opportunity for RP-3500? Maria? The ATR inhibitors from multiple other companies, except for Bayer, was developed in unselected patient populations, so it is very hard to say. We are sort of not interested in differentiating from that because it is not the contemporary way of developing drugs. I just wanted to remind everybody that Bayer has 22 patients published and reported data since 2017 when they started the study. It is a very difficult problem to compare to this data, because there is no follow-up on this data. Number two, the data set is very small. It is 22 patients. Of those, they were enrolled based on HRD panel. HRD panel is specific to PARP inhibitors. Fortunately, as you also know, there is some overlap between the HRD panel and our SNIPRx panel. However, these are supposed to be different genes. What we insist on is characterizing the type of genomic alterations much deeper than anybody else, including the zygosity. These are data that, by the way, we don't have yet. We have the samples for the data, and we have partially the data, but they are too early to reveal. There is many differences in not only patient selection in our drug because we believe and you will see it by the toxicity profile. I just wanted to remind you that the grade 3 anemia in the very small cohort of Bayer, so it was 80% grade 3 anemia, and then there were a couple of other toxicities. We are planning to differentiate based on safety profile, based on ability to deliver the drug. The data for efficacy seem to be too early because we started the study formally in August last year. It is 12.5 months of data that will be reported. Thank you. You said that it's hard to compare to the Bayer data that there's, for one, a limited number of patients. I think you said it was 22 patients. I guess, is there any aspects of the data that's helped guide the direction of your development for RP-3500? Definitely. I think that we were extremely pleased with the Bayer data because it was the first attempt to select somehow. While it is not, it at least was selected for the general concept of DDR deficiency. We believe that we have a proprietary selection method. We also need to guide very clearly that in the phase I study, out of the 16 genes that we propose as sensitivity genes, we obviously will not have data for the majority of these genes because these are randomly enrolled and not specifically enrolled. We have, after we establish the dose and schedule, which is sort of imminent. We will then open a statistically guided study to answer the efficacy questions in the ATM population and in[inaudible] populations. This data are not available from the- Okay. You're also looking at the combination with PARP inhibitors. Can you remind us of the data that gives you confidence in a combination? As we mentioned, when we did the step two screens for RP-3500, we found the additional genes for single-agent activity. We went back and did a combination step two screen, where we used both a PARP inhibitor and an ATR inhibitor together to see if there were any unique alterations that caused hypersensitivity to this combination. When we did that, we found four that we fully validated, as well as additional ones that are still ongoing. Basically, the bottom line to that data was that when you took paired cell lines that were either wild type or where that gene of interest was knocked out, you saw little to no activity of the single or combination agent in that wild type setting. In the combination of those two, when that gene was knocked out, we saw a very robust phenotype that was irreversible. It was very rapid, and it was at subtherapeutic doses of both compounds. We think by being able to pick the patient specifically and having a different dose and schedule regimen, that we would be able to move into the patients and see a better therapeutic index with these two agents together, given that they both do cause myelosuppression. In the phase I-B, you're studying the combination with niraparib or olaparib. What would your expectations be, if any, in terms of if there's any differences that might arise with these combinations? You already know that we tested all the trapping PARP inhibitors in our preclinical studies, and we have hypothesized based on the preclinical studies that it doesn't really matter which PARP is combined with our ATR inhibitor, as long as it is a trapping PARP. That is all good in preclinical studies. However, we also know that the PARP inhibitors in clinic are quite different. If nothing else, the half-life of talazoparib is in the 90s. It is given once a day, and some are given twice a day. The toxicity profile is very different. We decided that if we want to sustain the claim that it does not matter which PARP inhibitor is selected, we need to prove it and sort of hold ourselves to what we are saying. We decided to open the various PARPs, but we are obviously going to, for the further development, if any, we are going to choose one and not develop with all. We just want to prove that we are selecting patients. In the ATTACC study, we are selecting patients. In the talazoparib study, we are selecting patients. We just want to show that the efficacy is similar and the toxicity depends on the PARP. Okay. In the fourth quarter, initial results are expected from the monotherapy arm. What should we expect from that data, and when might we see data from the combination with talazoparib? I will quickly open the second question. We guided the data for PARP combinations. It will be probably first half of next year because we opened the talazoparib combination in April and the ATTACC study just last month or two months ago. The data are immature. Regarding what we guided for the release for the monotherapy is we said that this will be a monotherapy cohort that will give you a dose and schedule assessment. We will show how we did it, and we will also show the decision that we made based on this data, ergo the recommended dose and schedule. Further, we will show the activity of the drug, and I'm very carefully using the word activity rather than efficacy because activity can be shown by pharmacodynamic markers, can be shown by surrogate markers such as circulating tumor DNA or tumor marker decreases and not necessarily the tumor shrinkage, especially that the study is very immature. Therefore, the responses are still ongoing. Okay. Great. Your second program is RP-6306. Can you talk about that target and the synthetic lethal relationship of the target with CCNE amplification? As we mentioned, up front when we started the company with doing our SNIPRx screens, we picked multiple different genetic alterations that had a high unmet medical need, and Cyclin E1 amplification was one of those. Cyclin E normally, when it's amplified, is actually driving the cells early into the S phase before all the building blocks for the DNA have been accumulated. You're trying to duplicate the DNA at the same time you're creating the building blocks. This causes a lot of stress during that phase of replication and expands out the S phase. The molecule or the target that we found, PKMYT1, is a sensor in the cell cytoplasm that is sensing some of this replication stress that's occurring due to this event, and it's helping keep the cell in check until that DNA has actually been replicated and then allowing it to go into mitosis. When we inhibit that target now, we actually take the cells while they're in S phase and push them directly into mitosis before they've completed the DNA replication, resulting in cell death, and we think a really robust phenotype to move forward into patients. What do you think differentiates RP-6306, and how does PKM inhibition compare to WEE1 inhibition? It's a very good question. I think a lot of people look at the diagrams that have been written up for years, they see PKMYT1 is right next to WEE1. Although they do have one similarity, they both phosphorylate CDK1. That's really kind of where the similarities stop. Again, PKMYT1 is a cytoplasmic kinase. It's the first inhibitory phosphorylation that occurs on CDK1 in the cytoplasm, whereas WEE1 is a nuclear kinase, and it's actually inhibiting CDK1 usually after it's been activated. They're two different sensors really in the cell, making sure the cell doesn't undergo mitosis at the wrong time. WEE1 also regulates the G1/S transition by phosphorylation of CDK2. At least in our studies when we did our synthetic lethal screen, we only saw PKMYT1 come out as a synthetic lethal interaction with Cyclin E, later data that we showed through a step two screen for FBXW7. We saw WEE1 kill both of those cells equally. When we looked back at that using the pharmacology just to make sure it recapitulated the genetics, we again saw the same thing, a 40-fold window for PKMYT1 and less than two-fold window for WEE1. We think they're fundamentally different, but both have potential therapeutic benefit to patients. Okay. In addition to CCNE1 amplification, you're looking at FBXW7 mutations. What led you to these mutations, and how do the prevalence of these compare? Yeah. We found FBXW7 by doing the step two screen. Again, that CRISPR screen using RP-6306 to look for additional genetic alterations in that synthetic lethality network that might be exciting patient populations. FBXW7 was one of those genes that came out. FBXW7 actually regulates Cyclin E1. When it's lost in cancer, it actually increases the amount of Cyclin E1 as well as a number of other genes in the cell, and we believe that's responsible for that synthetic lethality interaction. The frequency of FBXW7 is actually quite a bit higher than it is for Cyclin E1. In particular, it is a really great opportunity in colorectal cancer, where it's up to 13% of patients that have FBXW7 mutations, as well as other large tumor types, like non-small cell lung cancer, the squamous subtype that has about 2.5%-3% as well. We think those, as well as other synergies in the populations that have Cyclin E are very interesting. Maybe the last key point is they are actually mutually exclusive with one another, so you don't see Cyclin E and FBXW7 in the same patients outside of a small subset of patients with hypermutator phenotypes. You're currently in phase I. Can you talk about that study and when we might see initial data? This study started about 2.5 months ago, three months ago, and we are escalating according to the Bayesian design of one patient per cohort until we reach the next phase. This is the usual very slow start. We guided that we will report the data sometimes later in 2022. Okay. Lloyd, maybe one last question for you, bigger picture, I guess. Are there any aspects of the Repare story that you think that investors underappreciate that they should spend more time focusing on? Yeah. None that I blame them for, right? It's understandable that the Street is very focused on data that's imminent, and we're as excited to present RP-3500 data as the Street I hope will be to receive it. I think if there's something that's underappreciated, it's our ability to generate a very long-term and exciting pipeline. I think RP-6306 and PKMYT1 and that observation and the fact that we're very much playing alone in this space, which we'll hope to prove is very exciting. It's just emblematic of the kind of company we're building and the kind of pipeline we're developing. I think in time, we do expect the Street will realize the real long-term value of that pipeline. It's up to us to demonstrate why that's going to be true. Great. It looks like we'll have to leave it there. Thank you so much for your time, everyone. Thank you. Thanks to everyone at Morgan Stanley.