Good afternoon, everyone. Thank you for attending Jeffrey's Global Healthcare Conference. My name is Kelly Xu, one of the senior analysts on the biotech team. For this first chat session, we are very pleased to have Mr. Adrian Dimond, Chief Executive Officer from Foghorn Therapeutics. Welcome.
Thank you.
Maybe first, we could recap the potential in addressing the Chromatin Regulatory System and the BAF Complex, leveraging your unique platform. Also maybe discuss the current pipeline and what has been keeping you busy for the past 12 months.
Sure. As you know, Kelly, we're focused—and thanks for having me here today—we're focused on the Chromatin Regulatory System, which is responsible for gene expression. The simplifying analogy I like to use is one of architect and blueprint, where the genes are the blueprints and the architect here is the regulatory system, determining which genes are turned on and off and which sequence, et cetera. In cancer, obviously, that goes awry and you get undifferentiated growth. The reason we've been interested in this area and initially started with the BAF Complex is, over the course of the 2010s through the last decade or so, every time you looked at a tumor being sequenced, there were a whole bunch of genes or proteins that were mutated in this particular remodeling complex.
In fact, about 20% of all cancers have a—pardon me—yeah, 20% of cancers have a mutation in this BAF Complex. When you extend that to looking at transcription factors and helicases and other chromatin-binding proteins, about 50% of cancer has some sort of relationship back to this biology. We have built the company with the expressed intent of being able to study and interrogate this biology in a way that others have not before. We have built a platform that allows us to create a whole suite of biochemical, biophysical, functional assays to look at these proteins and complexes of proteins. We have layered on top of that small molecule capabilities, specifically both enzymatic inhibitors, but also over the last five to seven years, protein degradation, which we have become very efficient on. We are really interested in finding novel targets.
We're not a me-too company trying to go after the same target that 20 or 30 other people are doing. We're really interested where we can apply the platform on very difficult proteins that require exquisite selectivity. I think that's been proven so far in our pipeline when you think about SMARCA2, CBP, EP300, ARID, some of the other things that we're doing. We're currently at clinical stage. We're in the clinic with a phase one that's being operationalized by Eli Lilly with our selective SMARCA2 inhibitor. That's our FHD-909 program. I think the next 12-18 months are going to be a very exciting timeframe for the company as we get further into the dose escalation and have potential data coming on that.
Potential second target that's yet to be disclosed with Lilly that may get disclosed over that timeframe, progress on CBP and EP300 and ARID1B, which I'm sure we'll talk a bit about today. An exciting time for the company.
Fantastic. Maybe also another question built on that is, I think based on the biology you just described for the chromatin system, so it's not so hard to think about this could be a very good drug target. What has been the challenges to actually develop a drug for chromatin system? Is it because they're all like super complex or maybe it's a very, very big protein as of the subunit? It's very hard to find the binding site, the docking sites. What has been the challenges behind, given that this is we know the biology for a long time, but we haven't seen much of success for drug development by targeting chromatin?
Yeah. So I'll give you a few examples of that. If you think, let's take SMARCA2 as an example, right? The SMARCA2-SMARCA4 synthetic lethal relationship has actually been well established. It came out of Project Drive, which was work that Novartis originally did. It's been confirmed through the Achilles CRISPR knockouts that the Broad Institute did. The challenge being, in a lot of the cases of these proteins, they have paralogs or very similar related proteins. SMARCA2 and SMARCA4, just to use that as an example, if you look at the primary amino acid sequence, right, it's about 90+% similar. If you look at the ATPase binding pocket, it's almost identical.
If you go and you screen SMARCA2 and SMARCA4 in a large diverse set of, through high throughput screening, you're going to get chemical matter that by and large is going to hit both proteins. The same is true if you think about CBP and EP300. You have the added complexity in the case of the SMARCA2/4 proteins that they are actually part of a bigger remodeling complex that's got other proteins. What's interesting, you alluded to sort of the challenge in finding binding pockets, is these proteins undergo conformational changes when they incorporate into some additional complex. That can reveal different chemical matter as you're going about trying to find something that's unique. It's actually all the things you said. There's an issue of purification sometimes of these proteins. They're very intrinsically disordered in some cases.
That's the case for ARID1A, 1B, which is one of our synthetic lethal pairs that we're interested in. It does require, which I think we've become very expert at, really great and significant protein production and protein biochemistry capabilities so that you actually have the starting reagent, if you will, when you start to set out to screen these proteins and find selective chemical matter. We've done this now on SMARCA2, on CBP, on EP300, on ARID1B, some other stuff that's in the collaboration with Lilly. It's all of these things combined that I think have made it very difficult to find appropriate selective chemical matter. Most of the time, in the case of SMARCA2/4, it's dual inhibitors. In the case of CBP, EP300, it's dual inhibitors or dual degraders. Selectivity is really the big challenge with these proteins.
All making sense. Maybe for the phase one trial running by your partner, is it like a 100% Non-Small Cell Lung Cancer or it's more like a primary Non-Small Cell Lung Cancer where we might be able to see some other tumor types? Also, in terms of patient enrollment, do you actually select patients with certain mutation types?
Yeah. Just to take a step back, let's say phase one dose escalation, Eli Lilly is responsible for all the day-to-day operational considerations. The collaboration has sites open now in the United States as well as Japan. There are about five sites open in Japan. I think we're close to having 15 sites or so open in the United States. To your question, the dose escalation portion before we backfill, and I'll distinguish that from expansion, which will come after all of this, is taking all histologies, so non-small cell, but in addition, any tumor type, as long as it has a mutation in SMARCA4. The key inclusion criteria has to be you have to have a SMARCA4 mutation or you have to have lost the SMARCA4 proteins. You're going to have an outright deletion of the gene.
The goal as we dose escalate, once we get into whatever we deem to be the therapeutic range, ostensibly we'll know that through both our PK and the exposure levers we're trying to get to an IC90, but also empirically through seeing tumor shrinkage responses, whatever may be, we'll elect then to have these backfill cohorts, which will be homogeneous for non-small cell lung with the goal of having more clear SMARCA4 mutations that are actually loss of function. Not all SMARCA4 mutations, not surprisingly, are equal. Again, we're looking for this synthetic lethal setup where the cancer has lost the function of SMARCA4, making it highly dependent for its survival on SMARCA2. The intent is to have those backfill cohorts, maybe one backfill cohort or maybe two, somewhere between 10-20 patients in those cohorts. That will be part of the dose escalation package.
What is the percentage of Non-Small Cell Lung Cancer patients bear all types of SMARCA4 mutations? And out of that, what percentage actually have loss of function mutations?
Right. So roughly 10% of all Non-Small Cell Lung Cancer patients have a SMARCA4 mutation, any mutation, truncation, missense, deletion, et cetera. If you think about, let's say, 10 Non-Small Cell Lung Cancer patients with a SMARCA4 mutation, somewhere between 6–7 out of 10 patients would actually have a loss of function mutation. You could have, and again, it varies based on the literature. We'll obviously find out more about that as we get further into clinical development and ideally commercially at some point. There are mutations. Obviously, if you have a pure deletion of SMARCA4, then you've lost the protein. Truncating mutations can do it.
There are also missense mutations that, if you have these hotspot missense mutations either in the ATPase domain where it renders the catalytic activity of the protein inert or in the helicase domain where it can no longer interact with the nucleosome and double-stranded DNA, that is also going to create loss of function. There is a preponderance of these patients who will have some sort of loss of function. Now, one other interesting point to make on the mutation, we will back up a second, which is the patients who have these SMARCA4 mutations specifically in Non-Small Cell Lung Cancer have a worse overall survival when treated in the frontline setting and specifically are more refractory and less likely to respond to pembrolizumab, to a checkpoint inhibitor. What becomes really interesting is when you overlap concomitant KRAS mutations.
For every 10 patients that have a SMARCA4 mutation, roughly three of those 10 patients will have a concomitant KRAS mutation as well. The data, at least that's available in the literature, suggests that those patients are going to do far, far worse. Probably less than a 10% response rate, the data show actually zero in the study and median PFS of probably a couple of months. Those patients are really unfortunately in a bad way from a prognostic perspective.
I see. So can I make an assumption here? It will be not very straightforward to set your bar on efficacy given the actual mixed background of different mutation types for this phase one, dose escalation, and also dose expansion data. So with the patients with poor prognosis, for example, comutation with KRAS, the efficacy bar could be lower than other SMARCA4 background.
That's right. I think in terms, and that's why it's important, obviously Lilly will be the ultimate decision maker in terms of movement to expansion, which would include monotherapy cohorts and likely combination, although we haven't disclosed exactly what we'll do. We've put data out at AACR showing combo with chemo and KRAS inhibitors as well as pembrolizumab, all which in our view were pretty compelling, showed meaningful regressions in the preclinical setting. I think when it comes to looking at the clinical data, that's why this backfill cohort where it's a homogeneous population for non-small cell histology and then obviously the comutations and loss of function of SMARCA4 we'll have to assess, this goes back to what's the denominator. What is likely going to be a patient population just by definition, they have to have exhausted all available approved care.
When you think about what's the right bar, if you will, I think is the question you're asking. I think if you're seeing anything in the 10-20%+ range from a response rate with duration past two to three months, we're going to be pretty excited by that because just as an example, dose of Taxol in a second line setting, the most recent study I looked at with Dato was about a 13% response rate, median PFS of 3.7 months. Third line, it's less than 10% and median PFS is about two to three months. And SMARCA4 patients, by the way, do worse than that average pool data that I just mentioned to you. So I think if we see efficacy in these refractory patients, at least I'll be excited. I'll leave it to Lilly.
Obviously, they've got to make the decision ultimately once the data are in.
Super helpful. How established is SMARCA4 mutation screening at your study centers?
It's part of Foundation Medicine panel, Tempus, the Illumina. I always forget the Illumina specific brand name, but it's part of these panels already. I think it's something that the physicians historically have been looking at because it's a marker of poor prognosis. They haven't had obviously therapeutic intervention. Knock on wood, that's going to change with what we're doing. It's a pretty established part of the sequencing for these tumors.
Great. And one peer, SMARCA2 degrader, has showed the initial efficacy signal in the SMARCA4 mutant background. What are the key learnings there?
Yeah. I think you're referring probably to our colleagues at Prelude who I think have done a good job showing target validation of the SMARCA2 pathway, albeit at least in our eyes with a molecule that's probably suboptimal to the task. I think there's several learnings that we've taken away. Some of this has come from studying this biology over the better part of seven to nine years, which again, in our hands, whether it's an inhibitor or a degrader, and again, 909 just to be clear is an inhibitor. Regardless of that, you have to cover this target and you have to keep it suppressed. What does that mean? For an inhibitor, that means hitting this at an IC90 level and keeping it sustainably inhibited over the course of administering the drug.
For a degrader, it means achieving a DC90, meaning 90% of the protein is degraded and keeping it at that level, not letting that protein rebound. That is one of the key learnings. I think the second learning, which based on how they've communicated their data is, again, you want to try and get to this homogeneous patient population. The challenge with these genetic lesions, when you start looking across a lot of different histologies, you may see dose responses at low levels, high levels. In fact, these tumors actually are behaving differently based on how reliant they are on, in this case, SMARCA2. I think that can lend itself to confusion. I think the third thing, which we've always been, or I've been consistent on as long as I've been at Foghorn, is we have no intention of putting data out on an interim basis.
I think it's very hard for investors to know where they are on the signal to noise. I always think of these things as sort of sinusoidal. You don't know where you are until you're actually done. I think we want to be really thoughtful about when we communicate data, we want to be thoughtful about what the denominator is and the homogeneity of the patient population. Again, we've purposely tried to ensure that our molecules are going to cover these targets in a very sustainable way.
Great. Since SMARCA2 is actually a compensator protein for chromosomes, to target SMARCA2, what kind of toxicities do you anticipate?
Yeah. The best way I can answer that, obviously you run the clinical trial to find out in the setting. I'd make sort of two comments on this, which is historically we've had actually a dual inhibitor, which was our 286 program we're no longer allocating capital to. This was an equipotent inhibitor of both SMARCA2 and SMARCA4. What we saw in that setting was what I would call more of a sort of chemotherapeutic sort of tolerability profile. Muscle weakness, fatigue, some nausea, some GI stuff, some skin rashes, things of that nature, which just made it just generally not tolerated. I think the good news based on, again, what we've seen from Prelude, at least to date, has been, at least for a phase one oncology agent, sort of a relatively typical safety profile there.
Nothing that stood out, at least based on what we've seen publicly. I guess the third point I make is in the preclinical setting, as we've tested 909 in the animal models we've shown, we've been able to achieve a pretty wide therapeutic window. If we applied our dual inhibitor to 2 and 4, we would never be able to get the regressions that we've shown because it just wouldn't be as well tolerated in the animal model. Obviously we'll have to demonstrate the safety profile in the clinic, which we're doing. I think there's certainly in our mind a wider therapeutic window and we'll ideally avoid some of the tolerability liabilities I just mentioned with dual inhibition.
Okay. Based on your commentaries, it's pretty obvious that this molecule has a potential to be combined with KRAS inhibitors like other targeted therapies and also Non-Small Cell Lung Cancer. We already have IO established their clinical benefit. That's also another obvious combination therapy. When do we expect those potential new clinical trials to explore the opportunity beyond the single agent? Maybe what kind of, or in another way, what kind of data could trigger and the size of the dataset could actually trigger a decision for the next step for this program?
Sure. Maybe just set context. At AACR here back in April, the collaboration put out data in combination with KRAS, combination with chemotherapy, and combination with pembrolizumab, all of which showed very nice regressions. In the case of chemotherapy, it was more additive. In the case of KRAS and pembrolizumab, it was, at least as one measures it quantitatively, certainly synergistic. The collaboration has not yet communicated exact timing or exactly what sort of studies, but I think it is reasonable to infer that these are obviously of interest. It is as simple as just thinking about the treatment paradigm for non-small cell lung, which is when you get, if you are in that first line metastatic setting, you usually, unless you are an EGFR positive, you are going to get an immune checkpoint inhibitor plus or minus chemotherapy.
Second line, if you're KRAS positive for mutation, you're going to get a KRAS inhibitor, potentially down the road a pan-KRAS. Chemotherapy is obviously reserved if none of those are options. I don't want to speak for Lilly at this point, but from my perspective, we'll want to see monotherapy efficacy to want to proceed into combination studies. I'm not a fan of praying for synergy in the absence of monotherapy data. I think that's not to say it will never work, but I think that's a higher risk proposition. I think it's reasonable to expect that as we get into efficacious range and, knock on wood, I see responses that will be encouraged as a collaboration to move into those combination studies. I think it's reasonable to expect that probably in the 2026 timeframe, but it's a dose escalation study.
We'll just have to see how fast things move or how many doses we have to get through and how much backfilling. I think it's going to be predicated obviously on seeing some monotherapy activity such that we're convinced that 909 is adding a reasonable amount to that combined efficacy study.
Super helpful. Maybe we can touch upon CBP and EP300 programs and what potential indications would you pursue for these two degraders and where are we on the clinical development progress?
Sure. To set context here, CBP, EP300 related proteins, they're histone acetyltransferases. Moving acetyl groups on various different proteins as it regulates chromatin. You know that probably better than I given your background in this space. I'll talk about CBP first. We've got a selective degrader, CBP. It's about 1,000-fold selective over the EP300 program. I think there's two potential paths here in terms of direction for clinical development and not mutually exclusive. One is the synthetic lethal angle where we would look for patients who have EP300 mutations, which now render the cancers dependent on CBP. Those are tumors such as gastric or bladder or lung. Different percentages. There's about eight or nine different large histologies with varying degrees of percentages of mutations, similar commentary about looking for loss of function. These strong loss.
What's been interesting, we put out data on this at ACR is that we've noticed combination potential with the CDK4/6. We had some in vitro data that we put out showing that actually CBP degradation actually looks relatively powerful in combination with the CDK4/6. Now we're in the midst of looking at that in in vivo animal model studies. We have a plan over the course of the second half of this year to really dig into that. That could really be an interesting way to develop CBP. I think it also would be potentially very interesting to various large pharmas who have CDK4/6 Inhibitors who are looking to expand that franchise. That would be in positive breast cancer. Right now, in terms of status of that program, we basically have what I consider to be our Pre-DC.
The properties are, it's a once a week subcutaneous injection. It has the potential to actually be every other week, but we want to go into our safety studies in phase one with the once a week so we can manage, if at all, any safety issues. Right now, it looks like a very well tolerated molecule. We'll start what I consider to be the IND enabling sequence, which will be our dose range finding in two species. That will start this summer. It will lead, assuming that goes well, into GLP tox studies later this year, beginning early of next, which will then track to an IND in the first half of next year. I think this could be an exciting molecule potentially for us, but also potentially in partnership depending on how the positive breast cancer stuff plays out.
EP300, I think about that molecule as somewhere between six to nine months behind where CBP is. Similar story here, again, selective degrader, about a thousand fold selectivity over its paralog. Again, this is one I'm actually particularly excited about. That's because there's a private company, our colleagues at Cellcentric, they actually have a dual CBP EP300 inhibitor. They had compelling, very compelling data, in my opinion, at ASH last year in multiple myeloma in what was, I believe, fourth or fifth line patients who were refractory to a bunch of the IMIDs and BCMA therapy, et cetera. In their highest doses, I think at about a 70% or so response rate, again, proving that this is a very interesting mechanism. This is frank dependency, by the way, on EP300. It's not a synthetic lethal setup.
I think the challenge, similar story to SMARCA2, the challenge with dual inhibition of CBP EP300 is you run into an impact on the heme compartment. Specifically in the myeloid lineages, you run into issues with thrombocytopenia, neutropenia, anemia, which if it's single agent, maybe you can manage through that. As you know, in multiple myeloma and a lot of these other blood cancers, you're doing a lot of combination work. We're excited by that program. We've got a lot of work ongoing to move that forward. That could potentially be oral. It could be potentially parenteral. If it's parenteral, similar profile, we're on track to have a subcutaneous, less frequently injected molecule there. That's one that I think over time probably will benefit again from a strategic partnership, whether it's initially as we go in the clinic or a little bit later down the road.
I think a lot of these cancers are going to be large tumor types where a strategic partner, similar to Lilly, will be beneficial. Really excited by both those programs, making a lot of progress. We'll have updates on both of those. I think probably as we get into the October timeframe around the targeted protein degradation conference, where we usually highlight some of the progress we've made.
Yeah. Very interesting story. Maybe lastly, can we quickly restate the key milestones and catalysts for the next 12 to 18 months?
Yeah. I think it's going to be an exciting time for the company. I don't have any official collaboration guidance for 909, but I think given that we started dosing the 909 study as a collaboration back in October of last year, I think the bookends could be towards the end of the year through the first half of 2026. I think, again, that's going to depend on how many cohorts and how much backfilling. I think it's reasonable that we'd have a sense of the efficacy there. Again, that will lead into decisions, as I mentioned, by the collaboration for expansion and combination. That's something to be paying attention to. We've got a second target with Lilly that is yet to be disclosed. There's a potential that could be disclosed if Lilly decides to take that forward. That's obviously their decision.
That's something that we're really proud of. I can't name the target, but it's something that we believe we've potentially got a best in class molecule for. CBP, as we've talked about, be on the lookout as we progress that. We update people as we're progressing that through towards IND, the positive breast cancer. I'd look for something on that towards the end of the year in terms of in vivo data. EP300 update likely in the October timeframe. We didn't talk much about ARID1B, which is another synthetic lethal, really difficult to drug. I think we'd also look for data in sort of towards the end of the year, as we get in the second half. Next year, we have the potential for multiple INDs. So potentially one with another one with Lilly, CBP, potentially EP300.
SMARCA2 degrader, won't comment too much on that, that's also with Lilly, but potentially some further progress there to disclose. A lot going on within the company.
Yes. So much to anticipate and looking forward to it. We're going to wrap up here. Thanks again, Andrian, for a fantastic discussion. Thanks everyone for staying with us.
Thank you, Kelly.