C4 Therapeutics, Inc. (CCCC) Earnings Call Transcript & Summary

Hello, everyone. Welcome to the BofA Precision Oncology Conference. My name is Chi Fong, one of the biotech [ EPC ] of BofA. I'm joined by my colleague, Jason Gerberry, senior biotech analyst. Here with us for the next fireside chat is C4 Therapeutics. For those who don't know, C4 is a clinical-stage biotech company focused on developing protein degraders for the treatment of cancers. I'm pleased to introduce Andrew Hirsch, CEO; and Adam Crystal, CMO from C4 Therapeutics. Andrew and Adam, thanks so much for joining us at our fireside chat today.

Thanks for having us.

So Andrew, maybe to start off, can you talk about the company's broad philosophy on designing protein degraders? What are some of the key considerations for selecting a target to pursue in development?

Yes, so I'll sort of take those in chunks. So I'll first talk about our philosophy on how we design degraders, and then we can -- we'll get into then how do we apply that in the oncology space kind of separately. And so it's really an interesting field, and one of the key things that we focus on really when we design a degrader for any therapeutic area is maximizing the efficiency of the catalytic cycle rather than any single step in the process of degradation. And for those of you who aren't familiar with degraders and how they're different from inhibitors, I'll give you sort of a quick introduction of what I mean. So they're different from inhibitors in that their role is really just to connect an E3 ligase to a target protein so that it can be tagged for destruction. In effect, it's catalyzing the destruction of that target protein by kind of tricking the body's natural protein recycling system to recycle a protein that it wouldn't otherwise recycle. And a key feature there is that once that target molecule is tagged for destruction, the greater molecule can go back and make that connection again and again. It can connect another target protein and an E3 ligase. And so we call that the catalytic cycle. And we know that, that can happen anywhere from, I would say, 300 to 3,000 times per minute. But we believe that optimal degraders do this towards the 3,000 times per minute -- in that range. That results both in maximal efficiency of that cycle but also very, very differentiated biology. And the best example we have of that are the IMiD class of medicines. If you go look at like thalidomide or lenalidomide, those are fairly weak degraders and they really only lead to sort of tumor stasis. But as we increase the catalytic efficiency with pomalidomide, with 92480 and, of course, with 7455, we see that those medicines actually induce apoptosis. So we really see different biological outcomes, the more catalytically efficient and potent the degrader is. And so our entire approach to protein degradation is focused on this. It's really a fundamental principle for how we think about creating degrader medicines, and our whole platform is really built around that. I would say the other important component is very early on, we made a strategic choice to focus on cereblon, which is one of the body's E3 ligases. And that's for really 2 reasons. One, it's widely expressed in all tissues and compartments so it gives us kind of wide latitude in terms of target selection, but it's also the only clinically validated ligase that's out there as it's the ligase, right, that's used in the IMiD class of medicines. And so we believe that reduces clinical risk. In addition, we're working with over 45 targets. We have about a 95% success rate using cereblon to degrade those targets, so we don't feel the need to go beyond that. And then lastly, one of the really important components is that what we've learned is that really subtle changes in the cereblon binding component of a degrader can have pretty big impacts on the catalytic efficiency of degrader. So as a result, we've really made a deep investment in cereblon binders. We have at least 15 distinct chemical series of cereblon binders, which we call our cereblon toolkit. Also, that library enables us to have the capability to develop both molecular glues, which we call MonoDACs, or heterobifunctional degraders, which we call BiDACs, to deliver candidates, and that gives us latitude in terms of the different types of target selection. As you mentioned, our focus for our own internal pipeline is on oncology, and I think we want to focus on -- we're focusing on targets where there's a clear genetic link to cancer. And there's a, I would say, a clear degrader rationale for why a degrader against that target.

Great. You mentioned that there's sort of features that make it perhaps more preferable to design a protein degrader compared to, say, an inhibitor. Can you elaborate what are some of these perhaps special features that make it attractive from a target perspective to design a degrader over an inhibitor?

Yes, sure. Adam, do you want to tackle that one?

Yes, absolutely. I think it's an interesting question from a lot of perspectives, but it's worth sort of noting upfront that when you look across the myriad of targeted protein degrader companies now, right, there's not a -- there are a long list of targets. Not everyone is pursuing the same targets. And I think this speaks to the fact that the technology is appropriate against a wide variety of targets as opposed to some new technologies where it really lends itself best to a select few. But that doesn't mean that we don't have very clearly defined criteria for how we think about the targets we prioritize, and I think it's worth saying that our criteria have changed. When we first approached this years ago, we put an emphasis on limiting biologic risk. And you can see this in our lead programs, IKZF1/3, BRAF, EGFR. These are programs that the world knows, and we know that there are -- that good drugs work. Now it hits the other criteria, but it's also worth noting that as we approach what you could call Pipeline 2.0, we've begun to take on more technical risk. And -- rather, we've begun to accept more biological risk, believing that many of the technical risks have already been addressed. So how do we approach target selection, right, to identify those targets where we think TPD is an excellent modality? The first is really not specific for targeted protein degradation but worth emphasizing nonetheless. We want to emphasize targets where we believe that hitting them well would be transformative for patients, where we think it can apply to patients in areas of unmet medical need and be a great advantage to those patients as well as providing commercial opportunity. Andrew mentioned, we like -- in oncology, we'd like to focus on bona fide oncogenic drivers in spaces where we believe that hitting that target will result in bona fide tumor regressions as a single agent. I think the remaining considerations we think about a lot are specific for the targeted protein degradation space. Andrew spoke to this a little bit, but we must have what we refer to as degrader rationale, meaning a reason to believe that a degrader can do something against that target that other modalities, for example, small molecule inhibitors, can't. Scaffolding function is an easy one to point to, and BRD9 provides an example of this. Inhibitors are ineffective, whereas if we eliminate the protein with the degrader, we recapitulate the genetic effect of ablating the target. Second, we think about places where we need to maximize selectivity, and I need to just talk about this a little bit as well. The confirmation of the ternary complex with form -- which forms really does allow exquisite selectivity. And with the targeted protein degrader, you can achieve selectivity against isoforms of the same family or against mutant versus selective that are more challenging to do with other modalities. And I think finally, we think about targets where there is a reason to believe both data and rationale that hitting that target will be tolerable. There are certainly many, many targets where taking it down will kill both the tumor cell as well as normal cells. Those are not targets we are interested in pursuing upfront. Things that allow us to have this, there are few to point to. One is mutant selectivity, so BRAF. It's only expressed in the mutant tumor. You spare normal cells because you don't degrade wild-type BRAF. Synthetic lethality, where the only cells we care about are those with a particular mutation, like the tumor. They are also examples of targets that aren't widely expressed in the adult human, right, but are in development or in cancers. So those are some of the filters we put through as we prioritize, and I think the one that's not on it and it's worth highlighting is traditional druggability, right? I think we are confident that if we are interested in a target by either a BiDAC or MonoDAC approaches, we can probably make a degrader for it. So I think on of the thing that is so exciting is that we do not have that as one of our criteria.

Great. So the company is looking at degraders, developing degraders for several established oncogenic drivers. You've touched upon that already a little bit, but can you elaborate on the potential therapeutic advantage you may get with a degrader over an inhibitor, say, in the BRAF or EGFR space? And conversely, what would be some of the theoretical drawbacks with the degrader approach?

Sure. Adam, do you want to tackle that, and I can come at the end and talk about those, or drawbacks potentially?

Absolutely. There are examples where it's a really clean cut, right? Now the EGFR or BRAF are necessarily those examples because they are kinases, and small molecule inhibitors do turn off those kinase activities. But I think that even with these molecules, we do have very -- we are entirely relying on our belief data supported that a degrader can do something which an inhibitor cannot. BRAF is a great example of that. When you think about the fact that wild-type BRAF is an obligate dimer, right? It is part of a complex signaling cascade, which includes dimerization of BRAF as well as MET downstream of that. And there is a liability of the 3 approved RAF inhibitors against that dimer. They do a great job of, for example, inhibiting the V600 mutant protein, which signals as a monomer, but does not do a particularly good job of inhibiting the dimer. In fact, it can hyperactivate the dimer through a process called paradoxical activation. So we think a lot about the role of dimerization and how a degrader can potentially overcome that. Said simply, when we degrade our target, the V600 mutant protein, it cannot incorporate into the dimer and paradoxical activation cannot be problematic. It provides a very strong rationale to go after mutant BRAF, which can be accomplished with the degrader and not with an inhibitor. I think that there are other aspects of degraders that provide advantages in precision oncology. We've spoken to some of them as well. I don't want to belabor them, but obviously, we think that we can make a degrader against anything that we have a ligand starting point, right? This is not the case with, for example, inhibitors, where you can start with a weak inhibitor or a weak binder, and it's really impossible to develop enough potency to make a molecule against it. We're more confident that with the degrader, we can start with a very weak binder and emerge with a molecule with drug-like properties. And I think the other clear advantage of degraders is the catalytic activity of them, meaning they are fast. These molecules can take down proteins in vivo within 2 hours to maximal levels, and this provides a clear advantage that the modalities we're maximizing the effect on target can take more time. Some proteins, or rather some malignancies, some cells fail to adapt when you take out the protein driver that quickly and they commit to apoptosis. Whereas if you have a slower mechanism, the cells can learn to adapt and they don't apoptose but might growth arrest instead. I think you also asked what challenges there may be in this space, and I think it's interesting. A lot of them have already been addressed, the large part. The ones that were the biggest vectors in the field were, can we make molecules with drug-like properties? Can they be administered orally? Do they get in? Do they hit target? And are they tolerable? And I think that the field has already demonstrated that with the entrance of some BiDACs into the field that the answers to all those questions are yes.

Yes. I mean, I would just add, I think it comes back to what we said earlier that target selection is really important. [indiscernible] a degrader rationale. I think sort of the risk is you don't really have a strong degrader rationale for going after a target, and you spend a lot of chemistry effort to sort of develop a degrader that ends up sort of being not differentiated. So I think one of the things I love about this field is all the benefits we just talked about. But at the end of the day, it's small molecule chemistry, right? It's not -- this isn't anything new kind of biologically. It's not sort of -- it's no more complex than small molecules in terms of manufacturing, et cetera, other risks. You have all the normal risks you have with small molecule chemistry, right? They're no different, and there's nothing really incremental to degrader. It's just a different mode of action to sort of taking -- stopping signaling the spread of cancer.

Okay. Great. I think a question for you, Adam. I think we've talked in the past a little bit about this. With bifunctional degrader, we have lots of hook effect in pre-clinical setting. How do you optimize drugs in clinical setting to avoid the so-called hook effect?

Sure. I think it's a super important question, and just for the audience, I'll sort of explain how I think about hook effect so that we're on the same page. With these bifunctional degraders, you can have maximal degradation in a sweet spot. But at the concentrations of your drug become higher, you see the molecule becomes less effective degrading target. The way I think about this is that when you flood the system with too much molecule, some of it binds to target on one side of the molecule. Some of it binds to E3 ligase cereblon on the other side of the molecule, and so the critical approximation doesn't happen because there's too much drug there. I actually think the most important step in avoiding this becoming a clinical problem is a pre-clinical step. Meaning, we want that sweet spot not to be a threefold range or a tenfold range but more like a 100-fold or 1,000-fold range, in which the molecule does not demonstrate a decay in its activity, going from maximal over 1,000-fold range. And we tend to see that with our later-stage molecules, that we have a wide range during which we see maximal activity. And when we see 100 fold or 1,000-fold range, it actually makes it really easy to think about how to manage that clinically. I think that in the Phase I space, investigators and other clinicians spend all of their time thinking about how to get enough drug onboard safely, right? And that's the challenge, right? And very rarely is it the case that we're getting 10 or 100 or 1,000x the concentrations we need from the maximal effect. So simply stated, I don't think it's going to be a problem with most of our molecules because the struggle will be getting to the maximal effect, not too far beyond it. That said, we'll monitor [indiscernible] to ensure that we're getting maximal degradation and don't see a decay in it as we go up in the dose level, but I don't anticipate it will be a problem. The other thing I would say is that very simply as you dose, it might be the case that at Cmax, you're a little bit into that hook. But as that concentration comes down, you're going to have maximal degradation. So I prefer not to be in a hook effect for any of the 24-hour dosing period. But if you're in there for an hour or 2, I think the overall effect would be trivial.

Okay. So have seen resistance treatment resistance from practically pretty much all the target oncology drugs approved to date. With degraders, what would be some of the theoretical resistant mechanism you might expect? For instance, we really have limited clinical experience [ interrogated ] resistance problem with protein degrader, so curious what have you learned about resistance in a pre-clinical setting?

Sure. So I actually think that there are -- there is a useful set of clinical data to point to where we might see resistance with targeted protein degrader. And it's the experience with the IMiD molecules, which are, of course, degraders. And what emerges and what has been widely published in the literature are really 2 mechanisms of resistance or 2 buckets of mechanisms of resistance with IMiDs, both of which could translate to the BiDAC space or the PROTAC heterobifunctional degrader space rather than the glue degraders like the IMiDs. The first is that in cells or in vivo models or in patients, what is observed with chronic treatment with IKZF1/3 degraders, and I'm speaking specifically of lenalidomide and pomalidomide is down-regulation of cereblon. Now this makes sense, right? The cells are being killed by a molecule, which harnesses ceroblon to degrade target, which kills it. So effectively by selection, the cells learn to downregulate cereblon, resulting in less degradation of target and better survival. The literature is a little bit all over the place, but effectively, active patients who are treated in this way have a 50% or greater down-regulation of cereblon. What this translates to an experiment, the models is resistance. When you have less cereblon, you degrade less targeting, you survive better. Now it's worth saying that with molecules like our CFT7455, because it is so fast and because it is so potent, it retains activity in the setting of lenalidomide or pomalidomide resistance, which is engineered by this mechanism. The other published mechanisms of resistance in patients are really mutations in cereblon, and there are 3 classes of mutation that effectively result in some degree of loss of function. Such mutations are rarely present at baseline, meaning before treatment with len or pom, something like 0.2% of patients have it at baseline. After len, it approaches 15%. After pom, it approaches 30%. And there's some retrospective data that says some of these mutations correlate with oral prognosis when you get pom. For example, if you get len, you have one of these mutations, and then you get pom, you tend to do less well. I don't think it's binary, right? I don't think you are resistant or sensitive, but I think it does cause a relative resistance. We're looking it forward to looking at these mechanisms as we go forward and accrue more patients to our studies, but I think it's important to emphasize that despite the setting -- the presence of these 2 known sets of resistance, the hypothesis that a super fast and potent degrader will be active has effectively already been proven, right? CFT92480 with dexamethasone has an overall response rate in the 50% range after len and pom. So with the hypothesis that such a molecule, and I think our molecule is such a molecule, can overcome resistance has a lot of data to support it.

Great. So for bifunctional degrader, given modular nature of that. From a drug discovery perspective, can you talk about what knowledge transfer you can leverage from one bifunctional degrader to another?

Yes, sure. So basically, what we've sort of seen to date in working with these is that degrader optimation is generally specialized to the individual target, even down to the individual ligand that we're working with in terms of bifunctionals. Because remember, what we're doing is we're essentially tricking a ligase to tag a protein for destruction it doesn't normally tag. So what's really important here is getting the physical confirmation, right, of the target protein in relation to the ligase, right, and that really ends up being very, very different across different target proteins. There are some learnings for sure. I'm not going to share those, because some of those are proprietary and a little bit of our secret sauce. But I would say it's maybe not as modular as people may think since you really end up having a fairly unique structure for the target protein of interest. One of the things we found is that the cereblon binder tool that I mentioned earlier, which really has a high level of chemical diversity and range and binding potency, that's really, really important for our success because it gives us the ability to utilize a modular approach at that cereblon binder to move quickly to find good starting points for rapid but specific optimization up to greater activity in drug properties. As I mentioned earlier, too, it's that those binders differ in very subtle changes in sort of exit trajectory. From cereblon, for example, we find that, that has some pretty big differences in terms of catalytic efficiency. So there are some learnings, I would say, but we think our platform and our toolkit -- really lets us customize kind of the BiDAC for kind of each unique target that we work with and ligand, if we do in the BiDAC space.

So we have a targeted protein, the greatest summit coming up later this month. Does C4 have any planned presentation at the summit?

Yes. No, we do, and we're actually really excited about it. So we're not planning to present any clinical data on our programs at the upcoming summit, but I think this will be the first time our team is going to present on our MonoDAC platform. It's going to highlight our ability to rationally design efficient MonoDAC degraders through our diverse chemical library, combined with a bunch of sort of unique screening approaches that we use. So I think a lot of the focus is early on been on our BiDAC platform. But we're excited to share what we're doing in the MonoDAC space, and we look forward to doing that at the conference in October, or this month.

Looking forward to it. Great. So why do we transition into CFT8634, company first bifunctional degrader in clinical trial. Can you give us a quick overview of the program and where it's at, at development?

Yes. Sure. So as you mentioned, CFT8634, is our orally bioavailable BiDAC. It targets BRD9 for the treatment of cancers that are -- have a dependency on BRD9, which right now includes synovial sarcoma and SMARCB1-null cancers. This has been, as we talked about earlier, considered an undruggable target because bromodomain inhibitors can't treat these. Some of the oncogenicity is driven by some sub-domains. So unlike inhibition, we're able to show effect because we can take out the whole protein in kind of pre-clinical models of synovial sarcoma. We announced in May that we started dosing the first patient in our Phase I/II clinical trial. And that, as you mentioned, is the first buyback degrader kind of in the clinic. Is it traditional sort of Phase I dose escalation, looking at kind of safety, tolerability and antitumor activity in kind of both SMARCB1-null tumors as well as synovial sarcoma. Once we establish a recommended Phase II dose that the goal would be to sort of break out into individual cohorts, one in synovial sarcoma and one in SMARCB1-null tumors.

When can we expect to see initial clinical data from that trial?

Yes. So we haven't guided to that yet. As it is generally our practice in a Phase I dose escalation study, you never sort of really know how many escalation cohorts you're going to need. So what I can say is it's been going well so far since the first cohort we started in May. As we get closer to getting line of sight to when we think we're going to hit that recommended Phase II dose, we'll provide some guidance then.