All right. Good afternoon, everyone. I'm Andy Berens, Senior Biotech Analyst at Leerink Partners, day three of our healthcare conference in Miami. We're very excited to have with us Boundless Bio's CEO of the company, Zach Hornby. Zach, thank you for joining us.
Thanks, Andy. It's good to be here. Thank you to you and Leerink for hosting.
Great. Why don't we start with an overview of Boundless, for those that may not be familiar with your company?
Yes. Boundless Bio is a company that was established about six years ago to address one of the largest unmet needs in oncology, which is patients who have oncogene amplification-driven tumors. This is about 25% of all cancer patients. Unfortunately, these patients generally do not respond to any standard of care, such as targeted therapies or immunotherapies. Our scientific founders have spent the past decade-plus trying to understand what's the underlying pathophysiology of oncogene amplifications and why do they not benefit from standard classes of drugs. One of the learnings that they made through their research is that oncogene amplifications, unlike other forms of genetic alteration, frequently do not occur on chromosomal DNA, but instead occur on distinct circular units of cancer-specific DNA, which is called extrachromosomal DNA or ecDNA. This is novel biology that only cancer cells rely upon, not healthy cells.
Once one has the understanding that there is this differential biology in cancer, it opens up an avenue for inquiry, which is, are there unique vulnerabilities that accompany this dependence on this ecDNA? Can we, as an industry, exploit these vulnerabilities in the form of new drug targets and new drugs? That was the premise upon which Boundless Bio was founded.
OK. Yeah, and it certainly was the premise that attracted me to your company, what you were doing. Very difficult. A lot of drugs have failed in this indication. Also the hypothesis about the ecDNA and the role it could play. I guess, why don't we start with that? How important do you think is that ecDNA hypothesis to your story and the investment thesis?
I think the ecDNA hypothesis was critical to the formation of the company and kind of the raison d'être of the company, the differentiation of the company, in that it gave us a unique angle on cancer biology kind of to look where others are not. ecDNA is the lamppost under which we are looking. Also, it is possible that when we find things under that lamppost, they may have application only in ecDNA, or they may have application more broadly. While our company is formed around and our platform is based upon vulnerabilities associated with ecDNA, we are not overly dogmatic in that we will follow the signal wherever the science takes us. It is possible that some of the targets we discover or some of the drugs we're working on will have application both in ecDNA but also in other settings.
OK. What is it about amplifications that make it difficult for traditional interventions to work?
Yeah. One of the core premises of these amplifications that occur on ecDNA is that because ecDNA is not on a chromosome, it does not adhere to the classical Mendelian genetics that we all learned about in high school. Normally, every cell in the body has the identical genome. You do not want your liver cells to have a different genome than your heart cells. EcDNA, these little circles, adhere to a different set of rules in that they can constantly change from one cell division to another. You can get more copies of ecDNA, less copies. You can get ecDNA that encodes certain genes but not other genes. What they do is they afford a genomic malleability to the cancer cells that allows the cancer cells to evolve real time and therefore find new clones that have better fitness advantage, particularly when under therapeutic pressure.
The implication is that you have this population of cancer cells. You try to treat it with a drug, whether that's a chemo or an immunotherapy or targeted therapy. These cells with ecDNA are actually changing their genome to find versions that are more resistant to that therapeutic pressure. Those more resistant cells start to perpetuate and become the dominant clone. Next thing you know, that's the new bulk tumor population that's completely resistant to the therapeutic pressure. To your question as to why do these tumors not do so well, it really comes down to the fact that they can evolve. They can change their genome and evolve and become resistant in real time.
Right. OK. I think it's important because I think some of the classes of drugs that you're leveraging, it's helpful to understand kind of this whole hypothesis around the ecDNA. Is it fair to say that what the ecDNA provides the cancer cell is that the cancer cell's addicted to some specific oncogenic pathway? When you block that pathway, that's occurring in the regular chromosomes, the cancer cell can then switch to this ecDNA repository. It's kind of like a backup to produce the proteins necessary to continue the process.
That is a good way to think about it. Let's give some examples to make it kind of tangible. For instance, we know that ALK inhibitors work very well with patients with ALK fusions. Typically, when an ALK fusion occurs, that's on the chromosome. The cancer cell is addicted to that ALK signaling pathway. Therefore, when you potently inhibit it, the cancer cell dies because its previous addiction has now been shut down, and it can't rely on some new genetic pathway absent some sort of point mutation. That's why these ALK inhibitors are so durably potent. In the case of these amplifications that we're talking about on ecDNA, if the cancer is addicted, let's say, to an EGFR amplification that's on ecDNA, when you inhibit the EGFR, the cancer cell can quickly downregulate its reliance on the EGFR because it's not a chromosomally locked aberration.
It's on these ecDNA that are malleable. They can quickly decrease their copy number of EGFR, and instead, they can replace that with a new dependency, let's say, on FGFR or on MET. The cancer cell can quickly switch out its growth dependency to something new that's not under pharmacologic pressure. That's how it's able to evade the prior therapy.
Interesting. OK. I think an important part of the investment thesis validation is identifying this ecDNA and then leveraging that to show that there's a therapeutic advantage to enriching. What can you tell us about your diagnostic tool? Where are you currently with that? How is it being integrated into the development program?
We've developed a diagnostic tool called ECHO, which stands for ecDNA Harboring Oncogenes. This is a clinical trial assay that's in use right now. Importantly, it is a bioinformatic algorithm in the form of a software, meaning it does not require a new sequencing platform or a new approach to sequencing or new biopsies. Instead, what it does is it ingests the standard data coming off a standard sequencer, like an Illumina sequencer. It uses that data as input, and then it recalculates those data in a proprietary way to detect, is there a circular DNA sequence present? If so, which genes are encoded on that circular sequence, and how many copies? Ultimately, it's reporting out presence of ecDNA, what genes are on the ecDNA, and how many copies of those genes on the ecDNA.
How sensitive and specific is that test?
In certainly all the validation work and modeling, it was quite sensitive and specific. It had a greater than 90% accuracy when comparing against orthogonal methods of ecDNA detection, like FISH. Ultimately, we will have to associate sensitivity and specificity with clinical activity. That is part of the clinical experiments that are underway right now.
OK. How are you incorporating ECHO into your programs now?
Right. The way that ECHO is being incorporated right now is as a retrospective analysis for patient segmenting. So what I mean by that is in our current phase 1/2 study called Potentiate of our lead asset, BBI-355, the qualification for patients to enroll into the study is that they must have an oncogene amplification. That can be assessed by any local assay, meaning like a Foundation Medicine or Guardant or Caris or Tempus or like Memorial Sloan Kettering Impact Assay. Any of those, so long as they detect an amplification, are sufficient for enrollment. Once the patient's on study, we then have their tissue biopsy sent to our central lab, where we perform the ECHO analysis, and then we determine whether that amplification is ecDNA-based or not. That becomes later part of our statistical analysis of the anti-tumor activity that we observe.
OK. When will we start to see data coming from that approach?
What we have disclosed is that we anticipate sharing preliminary proof of concept data from this phase 1/2 study by end of this year. We would disclose any ECHO findings in association with those data.
OK. What do you hope to see there? I mean, obviously, focusing on ecDNA would help validate the hypothesis that formed the company. It could also make the market size of the opportunity smaller. What's your hope for this data that you find?
Yeah. Ultimately, just follow the science. In terms of putting some ring fence around that, for those who are unfamiliar, the design of the study has both a single agent component as well as some combination components. In the single agent setting, we will be exploring some gynecologic cancers that have high replication stress. Those include tumor types like ovarian cancer and endometrial cancer. The reason for that is that this target, CHK1, was identified as an ecDNA synthetic lethal target because when amplifications are present on ecDNA, they lead to high replication stress. If replication stress is unmitigated, it will lead to mitotic catastrophe and cell death. What cancer cells do that have high replication stress is they self-regulate by invoking the DNA damage response pathway to pause the cell cycle, resolve the replication stress, and then move on.
CHK1 is the master regulator of the DNA damage response pathway. These cells are reliant upon CHK1 to self-regulate and manage through the replication stress. Therefore, if we inhibit CHK1, it's synthetic lethal in tumors with replication stress. That is part of the rationale for going to these gynecologic tumors that are kind of known to have replication stress and have previously shown sensitivity to other DNA damage response pathway inhibitors ranging from CHK1 inhibitors, WEE1 inhibitors, PKMYT1 inhibitors, ATR inhibitors. The other component of the study is the combination modules, where we're taking patients with specific oncogene amplifications, and we're combining with a targeted therapy directed to that amplification. This is now in a tissue-agnostic manner of basket design.
Two of the examples that are ongoing right now, patients with EGFR amplifications in any tumor type, we co-administer with an EGFR inhibitor. Similarly, patients with FGFR amplifications, we co-administer with an FGFR inhibitor. The idea here is that those single agent targeted therapies really do not deliver much activity in the amplification setting. We're hoping that this novel combo approach will deliver meaningful activity and durability. Now, to bring all this back to your question is we identified CHK1 as a target for ecDNA because ecDNA gives rise to high replication stress. However, ecDNA is not the only thing that gives rise to replication stress. It is possible that our CHK1 inhibitor will work when there's replication stress that is not caused by ecDNA.
When I started my answer by saying we will be data driven, I can't tell you for sure whether the presence of ecDNA will be required for a CHK1 inhibitor to work. If it is, then it's great because we've got a diagnostic and proprietary way of finding these patients. If replication stress more broadly, like any amplification, confers sensitivity to a CHK1 inhibitor, then to your point, that's actually a larger market opportunity that doesn't require an ecDNA specific diagnostic. I think either outcome is OK. We will simply follow where the data takes us.
OK. A question I've had personally is just, is the ecDNA and I'm not sure it's actually it's not necessary to answer the question because it could be applicable the other way. Is the ecDNA actually driving the oncogenic process? Or is ecDNA a byproduct of rapidly proliferating cells that spits out this ecDNA? I'd just like to hear your thoughts on that and whether it really matters. I mean, maybe it doesn't matter.
In certain tumors, we think it's unquestionable that ecDNA is the driver because there is often only a single molecular alteration present, and it's an amplification, and it's on ecDNA. There is no other putative driver in many settings. However, there are other settings where, and this is an area I know you're fond of, where the patient starts with a KRAS point mutation on the chromosome. They get treated with a KRAS mutant inhibitor, like an adagrasib or sotorasib. They have an initial response, and then they become resistant. Once we profile the resistant tumor, it turns out it now has a new amplification on ecDNA. In that case, ecDNA was not the original driver of the tumor, but it was a primary driver of the resistance and therefore still meritorious of trying to therapeutically intervene.
Back to your question, I think ultimately for us, we would like to be able to address populations where it's either the initial driver or it's a secondary driver of resistance. Either way, we'd want to try to benefit those patients.
OK. That goes back to this whole concept that ecDNA can be a repository for an addicted cell when you block it with one of these other agents, which then gives support for your hypothesis and your approach of combination therapy.
That's right. Exactly.
OK. All right. Remind us of the, because I think it's taken a while to get to this point to validate the diagnostic test validated. I think people are very interested in what you guys are doing. I think they're waiting for the data that kind of show that you guys are doing something different than what everybody else is doing. What's going to be the flow of information specifically from the programs?
Yeah. For the lead program 355, I'll reiterate that we've guided to preliminary proof of concept, efficacy, and safety data by end of this year. As I walk through the design, there's the single agent setting ongoing in the gynecologic tumors. There's the basket study design in the patients with EGFR amplifications, FGFR amplifications. The trial design also includes another combination cohort, which will be co-administration with a CDK4/6 inhibitor in patients with CDK4 or CDK6 amplifications, though that arm starts later than the other one. Probably will not have data for that other arm this year. I think in terms of what people can expect by the end of the year is data ideally from multiple of those cohorts that I just mentioned. We anticipate it will be a meaningful end and with some durability of follow-up.
OK. The diagnostic, you'll have data showing retrospectively on.
Yeah. I mean, the diagnostic will simply classify for anything we see or any patients enrolled, was ecDNA detected or not. Then we can draw any conclusions based on that. Think of it like a segmenting analysis.
Right. Do you guys have data showing that when you block, when you administer a targeted agent to a tumor with the relevant oncogene, that the ecDNA increases? Or it sounds like you have a way of measuring that that oncogene shifts to the ecDNA. Are there data to show that?
Yeah. We've published some of this in the past at conferences. I think at prior AACRs, we've published particularly this MAP kinase example that I just alluded to, which is that the patients or the tumor models start with either a KRAS G12C point mutation or a BRAF V600E point mutation. They're treated with a standard targeted therapy, either as a single agent or often in combination with like an EGFR inhibitor. When the models become resistant, they manifest these amplifications that we have shown to be on ecDNA. We've not only shown this in preclinical models. We've shown it in actual patient specimens of patients who are treated on the clinical trials of adagrasib and sotorasib.
Even just as an example, Roche had a nice publication in the New England Journal of Medicine for their initial kind of pan-tumor proof of concept of divarasib, where they showed activity in multiple tumor types throughout the body. They also characterized the resistance that ultimately emerged. If you look in one of the tables, more than 60% of the patients where they were able to molecularly characterize the nature of the resistance, it was amplification based. Based on our own experience, that's very likely on ecDNA.
OK. I think it's worth discussing, obviously, the CHK1 is the lead program. And there's another company developing CHK1, Acrivon. Maybe we can talk a little bit about what they've shown. They have their own diagnostic. And it's a different approach than what you guys are doing, but it is obviously the same class. So maybe you can give us a brief summary of what they've done.
Yeah. Let me start with a little bit of the history, which is most pharma companies over the past decade plus tried to take CHK1 or CHK1/2 inhibitors into the clinic because it is a validated cancer target. Most of those programs did not advance past a certain state. The only compound that did deliver some single agent activity, including durable responses, was a molecule that had been discovered by Array and Lilly called prexasertib. In the initial phase one study, it actually showed pretty encouraging activity, particularly in ovarian cancer among other tumor types. Lilly advanced it into phase two, and then the signal attenuated, and they stopped advancing it.
A few years ago, Acrivon in-licensed this compound prexasertib from Lilly with the thesis that it has some single agent activity and with a prior or with a thoughtful biomarker, you could really enrich that response rate to a point where it could be approval. That compound is a very potent CHK1 inhibitor. Acrivon also refers to it as a CHK1/2 inhibitor. It's IV administered. They have shared some preliminary data over the past year that with their proprietary biomarker, and by the way, they haven't disclosed the specifics of the exact biomarkers, they are showing an encouraging anti-tumor activity both in ovarian and endometrial cancer. I think it's still early data. The N are still relatively small. The response rates and durability they've shown thus far are definitely encouraging.
If they hold up, I would think it would be on a path to registration doubling. From our point of view, it's great news. We have a different compound. We like our profile. It's oral. It is CHK1 selective. We're using a different biomarker. We're generally going into different patient populations. I'm encouraged that thus far they've shown that CHK1 can be a druggable target under the right conditions.
Is hitting CHK2 a liability, or is it a benefit?
In our experience, through our target discovery models, CHK2 has never come up. Like, it's never revealed itself as important to what we're trying to do. There are some hypotheses that CHK2 inhibition can lead to certain types of GI tox, as an example. From our point of view, we don't think it contributes anything. Just based on first principles, like, we want to have compounds that are selective for the target we're trying to hit. We intentionally made our compound as CHK1 selective as possible.
OK. Their diagnostic, you alluded to it, it's not measuring ecDNA. We believe it's a measure of replication stress.
My understanding from their disclosure is it's some measure of pathway activation. I believe what they use is like a digital proteomics looking at three phosphoproteomic markers that they refer to as pathway activation. My guess, without knowing for sure, is that they are detecting CHK1 pathway activation in the context of replication stress. I think there's probably some overlap. We are very specifically looking for amplification, or ecDNA as a root cause of replication stress. They might be looking at replication stress that results from those same things as well as other things. It may be that there's like an overlapping Venn diagram of the biology that they're looking at and we're looking at, but I don't know for certain.
OK. That test has some limitations commercially, right? Their test.
My understanding is it requires shipment of tissue to a central site administered by Akoya, which I believe has been recently acquired. Beyond that, I'm not too close to their approach. What I like about our approach is what I mentioned earlier, is that we're using standard sequencing data, and it's just a bioinformatic software. There is no requirement of any new diagnostic platform or new diagnostic workup. I'm really fond of our approach because I think it fits right into how patients are currently treated and where biopsies are currently analyzed in today's standard of care.
OK. I know we only have a few minutes left, but maybe you have a second program that's starting, the Kinesin program. Can you give us an overview of that approach?
Yeah. Thanks for asking. I'm really excited about this program. It's kind of the epitome of the vision of what Boundless Bio could be in that using our platform, which is called Spyglass, we have been searching for targets that are synthetic lethal in the context of ecDNA or amplifications more broadly. We discovered a novel kinesin target that as we searched the scientific literature and patent literature, really there's very little known about this target. In the patent literature, there was zero, meaning that to our knowledge, nobody is actively working on drugging this target with any modality. As we started to unpack the biology, it's just very exciting and differentiated. It comes down to kind of an esoteric feature of ecDNA, which is that chromosomes have centromeres, and that's critical during mitosis. Again, back to high school biology.
The chromosomes all align at the mid pole. They duplicate, and then they get pulled evenly into daughter cells. That's how genomic fidelity is maintained in every cell in a multicellular organism. EcDNA lacks centromeres, so they don't align at the mid pole the way chromosomes do, and they get asymmetrically distributed during mitosis. They do still have some structure. It turns out that this kinesin we discovered is essential for kind of guiding the ecDNA to the proper location during mitosis. Therefore, what we discovered is that initially with genetic knockout or knockdown, but now with pharmacologic inhibition, if you disrupt the functionality of this kinesin, the ecDNA kind of mis-segregate. Instead, they actually aggregate, and that ultimately becomes cytotoxic to the cancer cells. This is a novel target. We've been working on it for a few years. There were no assays.
There was no crystal structure. Like there was nothing. We had to start from ground zero. It has been a real labor of love for our drug discovery group. They have made tremendous progress, particularly over the past year, ultimately to the point where we now have some very attractive oral degrader candidates. We have disclosed that we believe we are on track to declaring a development candidate by mid this year and therefore filing an IND for half of next year. We are super excited to bring this forward into the clinic. We have not disclosed the specific target yet, just for competitive reasons, because we have put so much effort into it, and we do not want to immediately have someone dropped on our coattails. We do look forward to talking more about this target as we get to the clinic.
Is there a liability associated with targeting Kinesin in normal cells?
We believe no. Based on DepMap, it is a non-essential target. There are even human genetic conditions of knockout of this target that are viable. We believe it's a very attractive target based on what we know so far.
OK. What we like about this program relative to the CHK1, I guess, is the CHK1 is, in essence, I guess, identifying cells that are susceptible to CHK1 inhibition.
Yeah.
It's not actually disrupting any process in the ecDNA pathway.
Let me modify that a little bit. It's not depleting the ecDNA, but it actually does disrupt the transcription of the ecDNA. So what we found is that when we inhibit CHK1, we talked earlier about oncogene dependencies, like an EGFR, FGFR that's amplified on ecDNA. When we inhibit CHK1, there's no longer any functional oncogenic protein. It is pretty interesting. To your point, it's not quite as direct. It's kind of like first derivative impact.
Right. Right. I mean, this one is definitely affecting a process that's critical to the ecDNA proliferation.
Yeah, absolutely.
OK. Thank you, Zach. Look forward to Boundless' progress. Lots of cool data coming. I know there are a lot of investors that are definitely listening and watching what you guys are doing. Let me just ask one last question. Is there anybody else now that you guys have kind of become public and brought this whole focus more into the public domain? Do you see other large pharma interest? Anybody else doing what you're doing?
Yeah. There's not direct competition per se, but there's one venture-backed company. There's a company backed by SR One called Eik, based partly in Boston and partly in London, that's doing they've also built a platform for ecDNA-based target discovery. They're still private and somewhat in stealth mode, but we're certainly friendly with them and rooting for their success, because I think a rising tide lifts all boats. In terms of pharma, we certainly have a sense as to who within pharma is interested in ecDNA and who has nascent discovery groups. What's in the public domain, it's obtainable through Google searches, excuse me, is that both AstraZeneca and ex-US Merck have funded postdoctoral work in ecDNA discovery at various research universities. That's at least explicit that ex-US Merck and AstraZeneca are active in this domain. There are others as well.
OK. Great. Thanks for walking us through the story. Congrats on all the progress, and look forward to the next steps.
All right. Thanks, Andy. Thanks, everyone.
Thank you.