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

Good morning, everyone, and welcome to the C4 Therapeutics webcast to review new preclinical data for our EGFR degrader, CFT8919, being presented at the Keystone Symposium on Targeted Protein Degradation. With me on the call today with prepared remarks are Andrew Hirsch, our Chief Executive Officer; and Dr. Adam Crystal, our Chief Medical Officer. Dr. Stew Fisher, our Chief Scientific Officer, will join for Q&A. Before we get started, I would like to remind everyone that some of the statements we make on this call will include forward-looking statements. Actual events and results could differ materially from those expressed or implied by any forward-looking statements as a result of various risks, uncertainties and other factors including those set forth in the most recent filings with the SEC and any future filings that we may make with the SEC. Following our prepared remarks, we will have time for a Q&A session. [Operator Instructions] With that, I'll turn the call over to Andrew.

Thanks, Kendra. And thank you all for joining us this morning. We recently announced that we've decided to advance CFT8919, our selective degrader of EGFR L858R to IND-enabling studies and are targeting IND submission by mid 2022. We made this decision to advance CFT8919 based on the encouraging preclinical data we've generated to date and the unmet need in EGFR-mutant patients with the L858R mutation. We're excited to share more about CFT8919 and preclinical data with you today. CFT8919 is an orally bioavailable mutant-selective degrader of epidermal growth factor receptor in non-small cell lung cancer. Importantly, CFT8919 is an allosteric degrader of EGFR L858R with in vitro and in vivo activity in models with secondary mutations, including T790M and C797S as well as in models of the frontline setting where it has the potential to avoid the emergence of resistance causing secondary EGFR mutations seen with currently approved EGFR inhibitors. In addition, CFT8919 has demonstrated intracranial activity preclinically, indicating the potential to be effective against CNS metastases, which is important, given 30% to 40% of these patients develop brain mets. It's important to note that patients with the L858R mutation are not adequately addressed with current EGFR therapies and show inferior clinical outcomes. Adam will walk you through our preclinical data that gives us confidence that CFT8919 has a compelling profile compared to the approved EGFR inhibitors and suggests CFT8919 has a path to registration in patients who progressed on osimertinib. As a reminder, CFT8919 has achieved development candidate status and is one of the 4 programs we expect to have in the clinic by the end of next year. While CFT8919 is the focus of our call today, I'll remind you that we have a broad portfolio of degrader medicines across various oncology indications, as shown on Slide 5. As we look out over the next 18 months, we have multiple milestones across our 5 most advanced programs. For CFT8919, we're advancing into IND-enabling studies this year, and expect to file an IND by the middle of next year to enable a Phase I trial start by year-end 2022. Adam, I'll turn the call over to you to walk us through the preclinical data we're sharing today at the Keystone Conference.

Thank you, Andrew. I'll be presenting the rationale and background on this molecule, why it is we're so excited about advancing it towards clinical studies and presenting data which demonstrates exactly why we think it has a very real potential to make an impact in this space. On Slide 8, you can see the background on EGFR driven non-small cell lung cancer, which is so relevant for this program. EGFR is a fortified oncology target, which drives approximately 10% to 15% of non-small cell lung cancer in Western countries, including the United States. In Asia, the incidence of EGFR mutation in non-small cell lung cancer is even higher, approaching 40%. What's less widely known about EGFR-driven lung cancers is that there are really 2 driver mutations responsible for the vast majority of cases. These are exon 19 deletions, which accounts for a little bit more than half of such cases, and L858R mutation, that's an activating point mutation, which is responsible for about 35% of EGFR-driven lung cancers. Generally speaking, this doesn't impact management of the disease. However, there has been data which demonstrates something particularly relevant for this program. And I'll speak to this in a little bit more detail on the following slide. But what we're highlighting here is the driver mutation, whether it's exon 19 deletion or L858R is predictive of response to osimertinib. And patients with the L858R mutation do inferiorly. While the median PFS of osimertinib in the front-line setting is about 19 months, those with Exon19del tend to do much better, whereas those with L858R do worse with a median PFS of only 14.4 months, which is to say patients with L858R receiving osimertinib have an inferior clinical outcome. And finally, in the non-small cell lung cancer space across all of the relevant treatments targeted or otherwise, it's very important that these molecules are active in the CNS. And that's because in the neighborhood of 40% of such patients will develop brain metastases and using a therapy which has activity, but not in the CNS, presents very real clinical problems. Here, we present data demonstrating the phenomenon I just highlighted, which is present not just in patients treated with osimertinib, but also with first-generation TKIs. The trend remains where patients with L858R do less well. This doesn't appear to be because L858R is a more aggressive disease, but may in fact be because there is a liability for inhibitors treating this particular mutation. And I think it's reasonable to reach a conclusion that such patients are underserved by current EGFR inhibitor therapies. This slide also speaks to the fact that resistance develops. So in L858R patients, after about 14 months, patients progress in the setting of acquired resistance. So osimertinib, while a wonderful tyrosine kinase inhibitor, does not solve the problem of EGFR-driven lung cancer. Resistance emerges and as is often the case with tyrosine kinases, tyrosine kinase inhibitors treating RTKs and solid malignancy, one such resistance mechanism is mutations in EGFR. One such mutation, the most common such mutation, is C797S mutation, which actually destroys the binding site of osimertinib. We believe that there is an opportunity to treat those patients with the degrader and overcome those emerging resistance mechanisms such as EGFR mutation and C797S. And on the following slide, we present data, which speaks to the activity of our molecule in that space. This slide presents several pieces of critical data, beginning to frame out why we're so excited about this molecule. On the left, in panel A, what we are demonstrating in vitro, in several cell lines, is the potency and depth at which our molecule, CFT8919 degrades mutant EGFR. What's important to note here is that in the colored cell lines, red, blue and purple, what we are measuring is the degradation of mutant EGFR. Each of these cell lines has an EGFR mutation L858R. In one case, that mutation is by itself, at the purple line, H3225. And H1975, a gatekeeper mutation, which causes resistance to first-line inhibitors, is also present. As you can see, degradation is relatively not impacted by that mutation. And perhaps even more importantly, that resistance mechanism I spoke to of osimertinib is dialed into the red cell line in which C797S is expressed in conjunction with L858R and T790M, yet you can still see the potency and depth of EGFR degradation, which is achieved. Panel B presents similar, but -- similar data, but subtly different and critically important, same cell lines, again, the A431 cell line in which EGFR is wild type. In this cell line, it is neither EGFR wild type, is neither inhibited nor degraded, but in all of the cell lines we presented before, there is deep inhibition of phospho-EGFR indicating near-complete inhibition of signaling really with deep inhibition achieved at that 10 nanomole and beyond dose level. This highlights both the ability of this molecule to target secondary mutations and the ability of the molecule to spare wild-type EGFR. This, of course, is important because with nonselective EGFR inhibitors, such as the first generation molecules, which inhibit both wild-type and mutant EGFRs, there are toxicities which emerge as a result of wild-type EGFR, including rash, which can be dose-limiting. And on the right side here, in Panel C, is basically a tabulation of the data that we've presented on the left for our molecule, quantitating the potency of this molecule against wild-type and mutant EGFR isoforms. Also presented here is the degree to which these mutations render osimertinib and erlotinib ineffective. So what you can see, for example, in that middle column, osimertinib is inactive in the setting of that C79S (sic) [ C797S ] mutation, whereas our molecule retains activity. Similarly, first line, as one would expect, are inactive even when T790M is present. Slide 11 presents additional data, which really extends the data I presented on the previous slide even further. What we're using here are not human cancer cell lines, as we did in the previous page, but Ba/F3 cells in which various EGFR isoforms are expressed. Using this model allows us to test rare mutations, which are not commonly found in human cancer cell lines, and test the potency of our molecule against those mutations. What you can see on the left is a subset of the mutation's dose, which we've already looked at, again, demonstrating the ability of our molecule to spare wild-type EGFR, while it is potent and active against not just the L858R mutation in purple, but that mutation when T790M, C797S or both of those resistance mutations are present. And on the right, the data on the left is also presented, though quantitated, and contrast to osimertinib and erlotinib. In addition, other secondary mutations in EGFR, which can cause resistance, such as L718Q and L792H are also presented, demonstrating the activity of our molecule against these mutations which cause resistance to osimertinib. So what I've shown you to this point is data demonstrating that this molecule is active as a degrader against L858R, regardless of whether or not secondary mutations that cause resistance to EGFR inhibitors, including osimertinib, are present. What's on this slide is different. Here, I'm -- here what we intend to demonstrate is the selectivity of this molecule for the intended target mutant EGFR. On the left is basically a kinome assay, sometimes referred to as an [ Ambit ] profile, which surveys the potency of our molecule against the kinome. This is probably a plot which many of you have seen before, but effectively what one observes are red circles on each of these nodes when the molecule inhibits a kinase. And across the universe of approved kinases, some are selective, some are less selective. But in most cases, at least a small handful of kinases are inhibited. Here, what we're highlighting is that this molecule that's quite selective, really the only kinases which were detected as inhibited in this assay are EGFR L858R, which we've spoken to in detail as well as the second rare activated mutation L816Q. This molecule is exquisitely selective for a mutant EGFR. On the right is similar -- well, it's different set of data with a similar conclusion. What we show here is our molecule, a heterobifunctional degrader or BiDAC is specific in degradation for EGFR and does not hit what we consider to be antitarget molecules or proteins, which are often neosubstrates of [indiscernible] degraders. Here, I'm speaking specifically of SALL4, which is the target of the IMiD class of molecules and responsible for thalidomide syndrome. Our molecule does not degrade SALL4. It was screened against as an antitarget. In contrast, as we know, lenalidomide does degrade this target. And similarly, a panessential gene known as GSPT1 is considered an antitarget for this molecule. And here, we're demonstrating that the molecule is inactive against GSPT1. On Slide 13, we are presenting some of the details of the proteome profiling of this molecule. We intended to make a molecule which was exquisitely selective for the degradation of mutant EGFR. And we assess our molecules such as this molecule across a panel of approximately 8,500 to 9,000 proteins, really the proteome to see what proteins are downregulated in cell line assays by our molecule. And here, we've used 2 cell lines. H1975, a cell line we've presented earlier. As we've discussed, in this cell line, L858R is expressed as is T790M. The mass spec assay we used was able to detect about 8,900 proteins and 2 proteins were called as downregulated significantly greater than for 50% protein decrease. These were EGFR, obviously, mutant EGFR, as this is what is expressed in the cell line and detected; and the second protein known as CCND1. This is an exquisite degree of selectivity. Furthermore, we do not believe degradation of CCND1 is taking place here. We believe that this is downstream -- this is a downstream effect of the inhibition of EGFR, and this is supported by the observation of osimertinib, which is not a degrader, has the same result. Inhibition of EGFR results in the downregulation of CCND1 protein levels. Of the other 8,800 proteins in that assay, no others were detected as downregulated in H1975. A431 is a second cell line, which was assessed using this approach. In this cell line, EGFR is expressed. It's actually over expressed, but it's wild-type. And we would not expect it to be degraded as our molecule spares wild-type EGFR. But the data presented, in fact, no protein is demonstrated to be downregulated in this model, including EGFR, consistent with our statement that this is a molecule, which is a selective degrader of L858R mutant EGFR. Here, data is presented in a xenograft of the H1975 model, demonstrating the degree to which CFT8919 inhibits EGFR and degrades EGFR. And this is compared to osimertinib at a clinically relevant dose of 25 milligrams per kilogram. Our molecule is tested for dose dependency, testing 10 milligrams per kilogram, 25 milligrams per kilogram and 50 milligrams per kilogram and a time course is also run to demonstrate the speed at which inhibition and degradation are achieved. What you can see on the left side of this panel is that osimertinib achieves complete or near-complete inhibition of phospho-EGFR, as one would expect in this xenograft assay, active in the L858R, T790M setting. And as one would expect, the protein, the middle row here is not itself downregulated, only the phosphorylated isoform. However, when we use CFT8919 starting at 10 milligrams per kg, what's noted is rapid at 6 hours, degradation of the protein itself, shown there in the middle row and inhibition of phospho-EGFR equal to osimertinib. On the right, you can see our 50-milligram per kg dose level at 4, 6 and 24 hours. And I think one can conclude from this experiment that the inhibition and degradation of EGFR is rapid, near maximal at 4 hours and sustained for the duration of a 24-hour dosing cycle. The data I just spoke to presented in the middle panel is quantitated in the bar graphs below, but I won't speak to them in detail. Slide 15 presents in vivo data demonstrating the efficacy, activity and tolerability of CFT8919 compared to osimertinib. On the left is the workhorse model we presented before, H1975. This is a reasonable model to consider as model resistant to first-generation EGFR inhibitors. L858R is the activated mutation. T790M, the gatekeeper mutation. In the upper left panel, you can see the activity of our molecule compared to osimertinib. Obviously, vehicle results in no effect and robust tumor growth. Osimertinib at 25 mg per kg results in a tumor regression. And our molecule at doses as low as 25 mg per kg p.o. b.i.d. results in a robust regression comparable to osimertinib. On the bottom left, it's worth noting that our molecule at all dose levels was tolerated, resulting in either no change in body weight or slight change in body weight. On the right is critical data, this done in Ba/F3 model into which L858R plus T790M is knocked in and critically to that doublet mutation is added the C797S mutation. So this is a triple mutation containing not just the gatekeeper that results in resistance to first generation inhibitors, but the C797S mutation that destroys the binding site of osimertinib and results inactivity of osimertinib. Indeed, what you can see in the top panel on the right in this Ba/F3 model xenograft is that osimertinib is now inactive with really minimal or no measurable effect on the growth of this cell line. In contrast, CFT8919 retains activity with treatment resulting in a robust regression. And similarly, on the bottom right is data demonstrating the tolerability of our molecule at all dose levels tested. Slide 16 presents one final piece of critical data. In this slide, data is presented from a model in which the H1975 cell line engineered to express luciferase is injected intracranially in order to model brain metastasis and test the hypothesis that this molecule is active in the central nervous system or brain. Three pieces of data are presented. On the left is pharmacokinetic data, demonstrating the concentrations achieved in red in plasma in the mouse. And below, in blue, the concentration achieved in the brain tumor itself. And what you can see is that robust levels are achieved in plasma and that a significant proportion achieves penetration into the brain tumor itself. It's also worth noting the half life that's measured here in mouse, plasma is 3.1 hours. The middle panel presents data demonstrating efficacy of CFT8919 in this model. Here, presented are really just 2 conditions. The luciferase expressing tumor is treated with either vehicle, resulting in an increase in the amount of luciferase expressed, really going off of the scale after day 10 and continuing upwards thereon. In contrast, our molecule at doses of 50 milligrams per kilogram p.o. b.i.d. results in a robust decrease of the luciferase assay, indicating that this molecule is active in the CS and capable of shrinking these tumors which express luciferase. On the right is data, again, demonstrating the tolerability of this molecule. Most relevant is the experimental conditions shown in green, body weight change with our molecule is negligible. In contrast, here, actually, the vehicle-treated mouse loses weight. This is likely because as the tumor burden increases without active treatment, the mouse becomes ill and loses weight. CFT8919 is a potent, orally bioavailable, allosteric mutant-selected degrader of EGFR L858R. The data that we've presented demonstrates the activity of this molecule, not just in the setting of L858R modeling the first-line setting, but also in the setting of resistance -- acquired resistance mutations in EGFR, which commonly result in resistance to known EGFR inhibitors, including mutations such as the T790M gatekeeper mutation as well as the C797S mutation, which destroys the cysteine residue to which osimertinib binds. We've also presented data demonstrating the intracranial activity of this molecule, supporting the hypothesis that this molecule would be active in both the prevention of and treatment of CNS metastases. This data together leads us to conclude the clinical evaluation of this molecule is warranted, beginning with patients who have progressed on osimertinib, where we believe there is a clear path to registration in patients who currently have few treatment options. Also worth noting is that this molecule is an allosteric degrader. As a result of this fact because it's allosteric it doesn't bind the active site, there is an opportunity to vertically integrate on EGFR by combining the molecule with an active site inhibitor, all of the approved EGFR inhibitors being active site inhibitors. This may provide the opportunity to deepen responses upfront or perhaps more significantly prevent the emergence of EGFR mutations, which commonly cause resistance. The final point that I'd like to make is that this molecule is active in models of the frontline setting of EGFR L858R-driven non-small cell lung cancer. There is an opportunity in this space to improve upon existing therapies, where, for example, osimertinib results in a median PFS of less than 15 months. This molecule is a specific degrader of L858R, which works in an allosteric manner. And we believe their exists a bonafide opportunity to improve upon standard of care where our molecule could be effective either as single agent or potentially in combination with a third-generation EGFR inhibitor. In terms of timing, we are on pace to submit an IND for this molecule in mid 2022, with plans to initiate a first-in-human study by year-end 2022. I'll conclude there and pause to allow the audience to ask questions.

[Operator Instructions] Okay. First question is, do you expect that CFT8919 induces resistant mutations differently compared to EGFR inhibitor [Technical Difficulty]

So I think that brief audio interruption has resolved. So I'll answer the question which Kendra put forward, really. Do we think the resistant mutations to our molecule would be different than the existing inhibitors? I think the speculative answer is, yes, they would be different. This is for one primary reason. It binds to a different site. It is an allosteric site binder and many or most of the mutations that emerge to the active site inhibitors, Iressa, Tarceva, osimertinib, do so by blocking the binding to the active site or by destroying the site to which the molecule binds directly, neither of those are particularly relevant for our molecule. The second thing is, one can always speculate broadly across any TPD platform about ways resistance might emerge, which would certainly be discrete to a tyrosine kinase inhibitor. And so I think from both of those angles, the answer would be a yes.

Great. Thanks, Adam. The next question is, would you expect any activity in exon 20 mutations?

So the short answer to this question is no. The slightly longer answer is that this molecule is built to bind to an allosteric site, which is actually created by the L858R mutation, which speaks to what sort of allows one to realize exactly why it is so specific for that activated EGFR protein. And we would not expect activity in exon 20 insertion nor exon 19 deletion.

Okay. A couple more questions specific to mutation. So have we tried to drive resistance in the L858R patients to see if and how quickly there could be mutations in the allosteric site? The comment is looks very selective on Slide 12, but the cut-off for the decrease on Slide 13 was about 50%. Were there any proteins that were decreased, but less than less than 50% in those cell lines?

So I think there are 2 questions here. Stew, I'll fill the first one, and I'll punt the second one to you. So the first question is effectively, have we studied how -- what mechanisms of resistance we think might emerge to our molecule in L858R-driven non-small cell lung cancer? The answer is no, we haven't studied that in depth at all. And so we don't have answers for that at this time. And then, Stew, the second question was really about the data we put forward on our selectivity and the degree of degradation of other off targets.

Yes. I'd say this molecule is one of the most selective ones in our entire portfolio. We do see some protein changes at less than 50%, but those align almost universally, as I recall, in fact, entirely universally, with proteins, which are downregulated upon EGFR inhibition. As to say, when we run a control lane with or experiment with osimertinib, we see the same proteins going down. So what you're seeing is the cleanest on-mechanism assessment at less than 50%. When we go lower than that, then we have some pharmacological downstream effects that we're controlling for. So the bottom line as near as we can tell, there are no additional on-mechanism off target degradation effects.

Okay. Next question is what dose was used to obtain the plasma blood tumor concentrations on Slide 16?

Yes. So I'll answer that one. We ran all of our studies, the top dose at 50 mg per kg p.o. And in fact, that's where we've measured those plasma to brain tumor concentration. So in fact, as you look through all of the efficacy data there, our maximum concentration, both for the xenograft, the allograft and [indiscernible] model, all done at 50 mg per kg p.o. b.i.d.

Thank you, Stew. Next question is, where are we in terms of IND-enabling studies, including tox coverage, et cetera? And what key steps remain? And then a second question related, given what we've seen in our initial data, what are our expectations for the human dose?

Sure. So I can speak to both of those questions. I'll take the second one first. Human dose, we're in the process of modeling as additional data comes in. At this point, there's a range of doses that we expect would be active, all of which I'm comfortable with, reasonable doses for an oncology small molecule. Beyond that, because our resolve is still improving, I'm reluctant to speculate as to exactly what the active dose would be. In terms of IND-enabling studies, I mean, I think it reads out from our time line where we intend to file an IND in the end -- in the middle of next year that there is significant work to be done. That's, let's call it, 10 to 14 months out, depending on how you count. And I think that at this point, any toxicity data we have is preliminary. So I think better to hold off now before we discuss in detail what remains to be done.

Yes. I'll just add that if you remember, this program was originally part of the Roche collaboration. It was something that they decided, for portfolio reasons, not to move forward. And so some of the normal work that would be done as part of the collaboration therein, in terms of the official sort of CMC work, we're having to start that up as well. So that kind of speaks to some of the time lines as well that results in that mid- 2022 IND submission.

Okay. Thanks, guys. The next question, if CFT8919 as an allosteric degrader is used in combo with an ATP pocket binding inhibitor, TKI. What would that mechanistically cause blunted degradation of 8919, though may enhance inhibition of EGFR phosphorylation and for that sake, would we use 8919 sequentially after a TKI or is it better to be used in combo?

I can answer this question. I think speaking from a high level, I'm excited about the idea of what some would call vertically integrating on EGFR, using 2 molecules which bind to different sites to effectively block EGFR signaling. Yes, one is a degrader, one is an inhibitor. I think that, that is advantageous. I think that in short, what one would expect would be less protein around because our molecule has degraded a lot of it, which effectively makes it easier to achieve maximal or near maximal or adequate inhibition by putting the combination. One could speculate 2 things. This would result in less EGFR signaling that is present with dosing of current drugs like single-agent osimertinib. And one could also speculate that by coming at EGFR from 2 different loci or 2 different nodes, it's the same story as a cocktail against TB or a cocktail against HIV. It would prevent the emergence of on EGFR lesions to either molecule. Was there a second part of the question that I had intended to ask you about, so I can't remember. I may have answered the question in full. Kendra, feel free to redirect me.

Yes, Adam, it was just about just sort of mechanistic rationale, right, in terms of combination studies and then you...

All right. How it will be best done clinically? I think at this point, it'd really be speculative. Yes, we're doing preclinical studies to assess how the molecules work together, how they would work together best clinically. And I think at this point, it would be just speculation. But there are different ways they could be combined to maximally increase the therapeutic index. I think it's a great question for this program and for the field more generally.

Okay. Thanks, Adam. So the next question is about us mentioning there was no activity versus SALL4 and GSPT1. And the question is, can we comment on activity with IKZF1 and 3? And how do we interpret, if at all, the slight lower rate of expression with the 25 milligram versus the 10 milligram, and that's looking at the bottom left graph on Slide 14, the green versus the gray?