Revolution Medicines, Inc. (RVMD) Earnings Call Transcript & Summary

Welcome back to the Cowen Health Care Conference. Next session here, we're very glad to have with us Mark Goldsmith, the CEO of Revolution Medicines. We're going to do some Q&A in a minute. But we'll start off with Mark [ giving you ] a couple minute overview, a status update of the company, and then we'll dive into specific questions. I think he's also going to share a couple of slides as he does that. So with no further ado, Mark, the floor is yours.

Thanks very much, Marc. We really appreciate the chance to talk with you today. I just thought I'd start by reminding everybody that targeted therapies are still needed for very common, serious and genetically defined RAS-addicted cancers. And this slide attempts to convey the segmentation of those RAS-addicted cancers, representing in total roughly 230,000 new cancer cases in the U.S. alone each year. And with this segmentation, it's evident how genetic stratification can define subsets of patients with distinct RAS bearings that drive their cancers, and this also provides a road map of what types of RAS inhibitors might be most useful. As many of you may know, we've taken a strategic approach to building our pipeline of programs directed against really strategic -- biologically strategic targets that are key drivers of both RAS addiction and resistance to treatment. We've organized these into RAS companion inhibitors, which are inhibitors directed towards signaling nodes like SHP2, SOS1 and mTORC1 that are intended primarily to be used as combination agents in order to complement and support RAS inhibitors themselves; and the second basket being the RAS inhibitors themselves. We refer to these as RAS(ON) agents because they do bind to the activated form of RAS or the RAS(ON) form. And as this slide depicts, we now have 5 development-stage assets. In addition, 2 lead op programs have been disclosed and a very rich pipeline of discovery programs behind that. But with this complementary and cohesive set of therapeutic assets, we think we have the greatest opportunity now to begin really trying to defeat various segments of RAS-addicted cancers and patients who experience those cancers. So that's an overview of the approach the company is taking, and I hope we'll have a chance to delve into both the RAS companion inhibitors as well as the RAS(ON) inhibitors.

Great. Thanks, Mark, for that overview. Maybe we'll start with the shift to where -- you've gone through a number of dosing regimens to kind of find what you believe to be the right kind of exposure levels that limit the on-target AEs that might be seen but also have a potent inhibitory effect that will have efficacy hopefully. So I guess, maybe can you review some of the evidence there and why you think that you've found that window, why that's the right amount of exposure for these pathways?

Yes. Sure. So RMC-4630 is our inhibitor SHP2. It's a highly potent compound, very selective for the SHP2 target. It's orally bioavailable, and it behaves very well in vivo. And we've seen excellent translation from the extensive preclinical work we had done with RMC-4630 into the clinic over the last 2 years in which we really conducted an extensive program to figure out how to optimize dosing. And as you pointed out, we believe very much that the biology of the SHP2 target is such that the determinants of antitumor effects can be separated pharmacologically to some degree from the determinants of tolerability, and that's because SHP2 regulates RAS signaling even in certain genetically defined cancer context. And we believe that it's possible through driving very high levels of SHP2 suppression for part of the period of time to maximize apoptosis induction in the tumors and also to convert tolerability by permitting a brief rest period during a 7-day cycle. As I mentioned, all of this was predicated on and then predicted by extensive preclinical work we did. But then in the clinic, we've quite systematically invested, along with the gracious participation of our patients, to explore these dosing regimens with the goal of optimizing antitumor effect and achieving antitumor effect in a tolerable way. And we think about it in that order. The first goal is to optimize antitumor effect, and the second goal is to assure that it's tolerable. And we're confident now that we've pushed the inhibition of SHP2 in the right directions to achieve that. So with our intermittent dosing, which is now a day 1, day 2 schedule, so day 1 of each week followed on day 2 of each week, with that schedule, we do drive very high concentrations of the SHP2 inhibitor in the plasma and then in the tumor for the first roughly 48, 72 hours. And then it gradually declines in plasma concentration throughout the week but maintains itself above the EC 50, the effective concentration dose, of which 50% of the target is covered. We're well above that for the remainder of the week. And it's really only then in the final roughly 24 hours -- 12 to 24 hours of the week that the compound concentration in plasma starts to reach the EC 50 or fall just below that. And that seems to be a very optimized kind of thread the needle approach. And frankly, that's translated into clinical activity and tolerability. So both of the goals have been achieved. We've demonstrated that, provided a clinical evidence for that in various public scientific presentations. And so we're quite encouraged. We've seen antitumor activity now and reported it across 5 different RAS pathway mutant cancer-causing genotypes. And included in that has been both monotherapy -- preliminary monotherapy activity and preliminary combination activity, which I'm sure you're going to want to dig into a little bit. But that's -- I think the bottom line is that extensive preclinical work and now confirmed by empirical in-human evaluation, I think, really strongly supports our contention.

So you mentioned that you want this window where you can kind of push cells into apoptosis, right? And one of the attractions here is that SHP2 inhibition has potentially very broad applications, right, across many different tumor types, many different genetic context. When you look across that breadth of opportunity, are there particular either tumor types or genetic mutations that you think are more or less sensitive to -- a little bit easier to push into that apoptosis where the probability of success just seems higher?

Yes. So we're very much a science-driven company. We do quite a bit of basic and translational work to really understand those fundamental mechanisms, and then we've had a history of reporting that work publicly. So in fact, we've described various genetic contexts that were not a priori assumed to be sensitive to SHP2 inhibition but empirically are sensitive. We reported that in a paper published a couple of years ago, and we've talked about it extensively publicly. It tends to be the tumors -- the RAS-addicted tumors in which there is ongoing cycling between the RAS(OFF) and the RAS(ON) forms, typically, meaning those tumors in which the genetic lesion is somehow dependent upon that RAS cycling in order to drive its oncogenic effects, that cycling can be suppressed by suppressing SHP2. And that mechanism, I think, opens up certain RAS-specific mutants. It opens up NF1 loss of function mutations and others, for example, BRAF Class III mutations, all of which we've described as conferring a higher likelihood of sensitivity of the tumor cell. And remarkably, we have seen most of those observations that were made preclinically manifest in the clinic as well with SHP2 monotherapy inhibition. We have demonstrated in lung cancer, for example, evidence of activity against 4 different RAS genotypes that really look almost exactly like what we saw preclinically. So the translation has been a pretty high fidelity so far, and we're very proud to be able to share that information publicly. Now in terms of histotypes, I was talking about genotypes, but in terms of histotypes or tumor tissue types, we're still evaluating that. It's pretty clear that some tissue types are just generally more sensitive to monotherapy suppression of the RAS pathway, lung cancer, non-small cell lung cancer being the most obvious example of that. And in fact, that's where we've seen -- an example of where we've seen monotherapy activity with the SHP2 inhibitor, just like that there's been monotherapy activity with the KRAS G12C(OFF) inhibitors. Other tissues are just harder to suppress other even RAS-dependent tumors like colorectal cancers, and we know that achieving objective tumor volume reductions is much more difficult generally in colorectal cancers. And we think that has to do with the details of the biological pathways that are activated and engaged, that there are multiple endogenous pathways that essentially confer resistance to these monotherapy treatments, whether they be SHP2 inhibitor or a RAS-specific inhibitor. But that sort of leads to the general framework that we believe should be -- will prove to be useful and that is strategically designed combination approaches that can really more effectively suppress RAS signaling in a tumor. That should allow us to start extending efficacy out of lung cancer and into other tumor types, and we already have some work underway beginning to demonstrate that.

And given that somewhat varying sensitivity, do you think you ultimately have kind of one optimized dosing protocol? Or is this really going to be -- or are you going to define different protocols for different tumor types or different tumor mutation combinations?

Yes. So I think you're asking about a sort of a treatment or dosing strategy, and I think there are 2 components to the dosing strategy. One is how frequently you dose the drug, and the other is at what dose you use it. And I think we're convinced as of today, always subject to new data in the future but based on what we've seen so far, that the 4 -- RMC-4630, the day 1, day 2 dosing regimen on a weekly basis, provides a really optimized pharmacokinetic profile, as we talked about earlier, for high peaks and a short trough period at the end of the cycle. So I think the dosing regimen of day 1, day 2 is likely to remain intact across most, if not all, of the applications. Now the second piece, though, is the dose itself, and that is likely to vary depending upon the combination of partner. In the settings where we're combining RMC-4630 with a mutant-selective inhibitor, for example, a RAS mutant-selective inhibitor that doesn't really have on-pathway effects in normal cells, we ought to be able to push the dose up to the maximum level, the dose of RMC-4630, which we believe is roughly 200 milligrams day 1, day 2. So for those kinds of combination settings, we would expect to get close to or even at 200 milligrams day 1, day 2, which is the monotherapy recommended Phase II dose. For those situations where we're treating in combination with a second agent that also has broad pathway suppression effects in normal cells, that becomes trickier. And rather than adjusting the dosing schedule, we would use the dose itself, a dose reduction, as a way to achieve those combinations in a tolerable way. We've already demonstrated that, I think, quite thoroughly with the cobimetinib combination. Cobimetinib and RMC-4630 both can cause on-pathway normal tissue toxicity that's fairly similar to each other. And so when you put them together, you do get additive toxicity. And so we spent quite a bit of time over last year, 1.5 years, figuring out exactly the dose that can work for each of those 2 agents in combination. And now we have a combination study underway in colorectal cancer, one of those more challenging cancer types carrying RAS mutations. Similarly, with an RTK inhibitor, unless the RTK inhibitor is highly selective for a mutant target and fails to hit the normal target, unless it's got that characteristic, it's likely to have some on-pathway side effect profile. And an example is osimertinib. Although in vitro it's very selective for the mutant, in reality, when patients are dosed, a major metabolite of osimertinib is also active on the wild-type receptor. And hence, there is on-pathway toxicity that's been described and well recognized when patients are treated with osimertinib at an effective dose. Combining that with RMC-4630 will also then create that combination challenge because of the additive toxicity. And we just don't know yet today whether it's even possible to combine osimertinib with RMC-4630 in people. That's what we're evaluating. So I think those are the 2 kinds of extremes: cobimetinib, TAGRISSO, et cetera. And then the other end of the spectrum would be the mutant -- very mutant-selective inhibitors, including sotorasib, AMG 510, which is highly selective for KRAS G12C and has very little effect on normal pathway. And our own RAS inhibitors, our RAS(ON) inhibitors, for example, RMC-6291, are highly selective for the mutant. So I think that's the range we'll be working in is overlapping toxicity versus nonoverlapping toxicity, and we'll simply make a dose adjustment, if we can, to accommodate.

Yes. Okay. And given kind of all those pushes and pulls that we just kind of walked through, is there a combination that you're most excited about, to make you pick among your children?

Yes. Well, that's a tough one. We have -- again, sort of going back to our philosophy, we believe that this is a very big challenge to defeat RAS-addicted cancers, and it's a really important challenge. And the only way to be successful is to systematically test a series of different, somewhat related but distinct clinical hypotheses. And as a result, right now, we have 5 different clinical hypotheses being tested: 1 monotherapy hypothesis and 4 combination strategies. And there are more to come, without any doubt. And so while I think one could go through and in some high resolution argue about each one of these hypotheses and what are the pros and cons of it, it's pretty hard to rank them. They're all quite interesting. And I think there's really not extensive read-through from one to the other because they are sufficiently distinct mechanistically. So we're just pursuing them all. That's part of the reason why we entered a partnership with Sanofi, that we have the financial muscle to do so a couple of years ago when we were a younger, earlier stage company. And we'll just pursue these one at a time. But in parallel, some multiple programs running in parallel. So I think you're not going to back me into a corner to prioritize one over the other. I know that various investors have opinions about that, though.

You mentioned the desire to expand upon the combinations that you're already running. We've seen you announce deals for -- or agreements for collaborations on multiple G12C inhibitors. When you say more, is that continuing the kind of same hypotheses but just exploring more deeply with more agents on the same pathway? Or is it also breadth in terms of there are other pathways that you want to do combinations with?

Yes. Well, I think it's all of the above, to be honest. I mean we have a very robust preclinical program that is constantly exploring both our own ideas as well as things that come in from the outside. So we get contacted by various companies and other organizations with agents that they'd like to combine with ours or that we've -- our own ideas that we thought of. And we test them preclinically, and we do it very thoroughly. And where it makes sense and when we have the opportunities, we'll move those forward. So in some cases, it will be creating a larger [ N ], for example, the KRAS G12C inhibitors. We think RMC-4630 could be and maybe should be a backbone therapy for KRAS G12C(OFF) inhibitors, so we'd like to test it against multiple KRAS G12C(OFF) inhibitors. And every time we do that, we learn something that can help us make the next study even more successful. So we are certainly open for business in that regard and happy to have now 2 formal collaborations identified, one of which is already underway and the other of which is pending. And there's interest from others as well. But there are other hypotheses -- there are other RTK hypotheses, for example. Even though we have osimertinib in the clinic right now, there are multiple other RTKs with different mechanisms of action, different targets for which a SHP2 combination -- SHP2 inhibitor combination could make sense. We also have to keep in mind that RMC-4630 has immunological effects, which we think can be beneficial, and we've reported those both preclinically and now clinically. And so one can imagine other ways of thinking about the combination that goes beyond purely suppressing the RAS pathway and suppressing resistance mechanisms, for example, combinations with other immunologic agents. And we do have a combination study underway with pembrolizumab, as you know. And that's interesting in and of itself, but it also potentially sets up a triplet combination between a RAS inhibitor, a SHP2 inhibitor and a PD-1 inhibitor or some other checkpoint inhibitor. So lots of different directions to go here. I'm not saying that we'll do everything, but we certainly will do as many of the compelling ideas that make sense and continuously learning as we're going to figure out what are the most interesting things, what are the things that are beginning to get traction and to double down on those.

Okay. We do have a question sent in from one of the investors on the line. I just want to step back to when you laid out earlier some of that translation from preclinical models of which tumor types are sensitive versus what you've seen in the clinic. If you could review some of -- what some of those models were. Like were they just mouse patient-derived xenograft? Just kind of what was some of that evidence and how linear was it?

Right. So we do everything from testing cell lines in vitro. We do large panels. We're constantly looking for panels to evaluate. We do 2-dimensional cultures, 3-dimensional cultures. We do xenograft studies using established cell lines. We use a patient-derived or PDX xenografts as well. So we do all of the above, and we have a very extensive program, so very large data sets. And we look for patterns. We look for how to optimize in any given setting. For immunological studies, we will use syngeneic graft context mouse mirroring tumors engrafted into mice in order to do those sorts of studies where you have an intact immune system. We do all sorts of things. And at the end of the day, it's a sum of the evidence that points us to a particular hypothesis that then bears testing in the clinic. And some of those hypotheses will -- some of those tests will surprise us on the upside. Some will surprise us on the downside. I mean we don't expect absolutely everything to be predictive. But so far, the translation has been pretty good. I mean let me give you an example. So NF1 loss of function is a genotype that nobody has ever tried to drug before because it's a loss of a function. How do you drug that? It turned out we identified that loss of function -- NF1 loss of function mutants are hypersensitive. Many of them are hypersensitive to SHP2 inhibition. We tested that across many, many preclinical models. Well, in the clinic, our most dramatic clinical response has been in a patient with uterine -- advanced uterine adenosarcoma, stage IV, with an NF1 loss of function genotype driving the cancer and monotherapy delivered a complete response in that patient who has remained in complete response for quite a long time. So that's great. I don't know, though, whether that means that's 1 out of 1, 1 out of 10, 1 out of 100 or 100 out of 100. We now have to go do an expansion cohort of gynecologic tumors with that genotype to see really quantitatively how well that plays out, and that is currently enrolling as we speak.

Okay. Great. And maybe in the -- let's turn to the KRAS(ON) inhibitors that you had unveiled earlier this year. One, can you just quickly explain the difference between a KRAS(ON) versus a KRAS(OFF) inhibitor? Maybe we can get into the exact profile of [ these settings ].

Yes. So the only thing that they have in common is the word RAS, but everything else is different. The chemistry -- the chemical compounds are structurally dramatically different, really almost no similarity at all. They bind -- our compounds bind to the activated RAS(ON) form, that's the GTP-bound form, rather than the RAS(OFF) form. They bind to a specific site on the RAS protein that's different from where the RAS(OFF) inhibitors bind. Their action is also very different. The RAS(OFF) inhibitors trap RAS in the off state but completely depend upon the ability of RAS to cycle from on to off in order to capture them at that moment in time when they're in the off state, and that creates an opportunity for the cell to respond to KRAS(OFF) inhibition by shifting the [ pool ] of RAS from off to on. And that confers generalized resistance to those RAS(OFF) inhibitors. The RAS(ON) inhibitors, on the other hand, are sitting there waiting for RAS(ON) proteins. So as soon as there's an active molecule of RAS(ON) protein, really almost instantaneously, it's bound by the RAS(ON) inhibitor and it's shut down. And we can see those kinetics happening in a cell. We've measured them in a variety of different ways. And it really causes rapid, dramatic, deep and very sustained inhibition of the RAS pathway. So I think they're dramatically different even though they converge on the RAS signaling cascade. We think the RAS(OFF) inhibitors are remarkable. They have blazed the trail, and we're really excited to see sotorasib and others head towards approval. I think that's really important for patients. But the clinical results have also defined for us, from the approval package, for example, for sotorasib, not only what it's useful for but also what its limitations are. And those limitations certainly relate to the mechanism of action rather than to the compound itself. And we believe the RAS(ON) inhibitors and have shown, I think, a pretty strong body of evidence that the RAS(ON) inhibitors can overcome some of those resistance mechanisms, cause deeper antitumor effects and more sustained antitumor effects. So we really do believe the RAS(ON) inhibitors represent a next-generation that build upon some of the conceptual background from the RAS(OFF) inhibitors, but open up new territory. The other point to make is that the RAS(ON) inhibitors, beyond their mechanism of action, also give us chemical access to a wide range of RAS mutants. And I started this talk -- this conversation with you by showing the very large number of patient subsets with different genotypes. G12C only represents roughly, let's say, 1/10 of those. So 85%, 90% of all RAS-driven tumors can't be touched by a KRAS G12C(OFF) inhibitor. We've now proven that we can reach with the RAS(ON) inhibitors across a broad spectrum of different RAS mutations, and that's really therapeutic white space. That's an unmet need for which there's really no targeted therapy and no prospect for any targeted therapy in the near future other than an approach like our RAS(ON) inhibitor collection, which is quite extensive. So we think it has the potential to be a game-changing approach to RAS-addicted cancers.

Thinking about G12C where we do have off inhibitors, right? Do you see it more like in the clinic more likely to give you a higher initial response rate? Or maybe just take -- or take the same patients that are responding today and just give them a more durable response?

I think there are a lot of possibilities, and in the preclinical modeling, we've seen evidence for pretty much any hypothesis you want to put forward around that or any idea. We've seen deeper responses, more frequent responses. We've seen responses in cells that are less responsive to a RAS(OFF) inhibitor, and we've seen more durable responses. So really, all of the above are possible. And right now, the way we think about entering the clinic is with this broad set of possibility in mind and designing the clinical experiments to give us the chance of identifying which of those sort of predominate in patients. But it is quite possible that a compound like RMC-6291 will prove to be superior in at least one, if not multiple, dimensions like those that you just mentioned and could either serve as some sort of salvage treatment for those who fail the RAS(OFF) inhibitors and/or displace them in an earlier stage treatment as front line ultimately.

And given that those G12C(OFF) inhibitors are either on file or nearing filing, are you expecting to have to start that trial in people who have already failed those agents? Or do you think you can access G12C inhibitor-naive patients?

Yes. I think there are a lot of possibilities. Of course, from a clinical point of view, if you have a KRAS G12C mutant lung cancer, which is where the approval will be, the initial approval, it will make sense to make sure that you get on one of those treatments. And whether somebody would choose an experimental versus the approved therapy is probably a longer discussion and it might depend on various factors. We'd like to be able to identify biomarkers that can project or predict whether a patient is going to respond to one or another, which would make it really attractive and very clinically relevant to be able to stratify patients into those who might benefit from an off inhibitor versus an on inhibitor. We're certainly working on things like that. I think there are a lot of different possibilities here, and it's a little bit early for us to sort of lay out the specifics of the clinical program. But I can assure you, it's a topic of intensive work right now. And of course, by the time we enter the clinic, we'll lay out that strategy with more clarity.

I know we're running up on time here, but maybe just quickly on the multi-RAS inhibitor that was also unveiled. You've also talked about continuing to work on -- that covers a large number of mutations. But you also, in your slides, talk about continuing on some of those. Maybe just lay out the strategy there, why it can kind of play both sides.

Yes. Well, that's exactly what we're doing is we're playing both sides at the same time. So the RMC-6291, highly mutant selective for G12C; 6236, broad family inhibitor of RAS. So these represent the 2 bookends, if you will, for a pipeline of assets. And then in between those are compounds with varying degrees of selectivity for different mutants, and we're quite excited about those as well. Our thesis is this. There are advantages to inhibiting multiple forms of RAS in a given cell, including wild-type forms, advantages in terms of suppressing the tumor because of resistance mechanisms. But that also carries a certain liability, which is if you suppress those too much in normal tissues, you'll have on-pathway toxicity. And so what we don't know today is whether it is clinically beneficial and superior to use a mutant-selective inhibitor combined with a node inhibitor, one of our companion inhibitors, or to provide a broader multi-RAS inhibitor like 6236 that captures those resistance mechanisms but has some sort of ceiling as to safety and therapeutic index. We don't know. And our prediction is that, that answer will vary depending upon the genotype and the histologic type of the tumor. Some tumors may do very well with a targeted mutant-selective inhibitor combined with a companion inhibitor, and others may do better with a multi-RAS inhibitor. Since we don't know all those answers yet, we're developing assets on both sides of that question, and we'll test them. And we will be the ones who will determine in the clinic through actual data what is best for various patient subsets and then have on-the-shelf available for each patient the regimen that is optimized for them. And we're committed to doing that. It will take resources to do so. But we think ultimately, that will provide the most benefit for the most patients.

Great. Unfortunately, there's plenty more to talk about but we don't have the time for it today. Thank you very much, Mark. And thanks, everybody, for joining online.

Thank you, Marc.