Joining this session with Caribou Biosciences. My name is Alec Stranahan. I cover SMID biotech at Bank of America. I'm pleased to welcome Rachel Haurwitz, President and Chief Executive Officer of Caribou. Thanks for being here, Rachel.
Thanks so much for having us today, Alec.
Yeah, great. Looking forward to the discussion. Just for the benefit of the audience, those that have either not looked at Caribou for a little while or haven't seen the latest updates, you can give a high level overview of what the company's doing and the recent data that we've seen?
Absolutely. Thank you. Caribou is advancing two off-the-shelf CAR-T cell therapies. vispa-cel targets CD19 for lymphoma, CB-011 targets BCMA for multiple myeloma. Each of these programs we shared actually quite significant clinical updates towards the end of last year, in both cases, we are seeing phase I data that give us the confidence that each of these programs has best-in-class potential. What really unlocks unique outcomes for these patients is how we use our CRISPR technology platform. We have our own proprietary next generation CRISPR technology. We call it the chRDNA technology, its unique attribute is that it drives editing outcomes that are orders of magnitude more specific than first generation CRISPR-Cas9.
We've leveraged that to edit, to armor our cell therapies in ways that we believe are unique in the competitive landscape and that we think are really critical to unlocking the kinds of high response rates and durable outcomes that we've seen with both programs. We're really excited that actually both programs have abstracts that were accepted at EHA for oral presentations next month, which will give us a chance for both of them to provide really, we think, important data updates.
Great. We can definitely unpack what we're looking forward to at EHA. At a higher level, you mentioned the chRDNA platform. You've also got some kind of unique edits that you're incorporating, PD-1 knockout, for example. You can just talk through kind of like the rationale for that approach and how it's armoring or I guess camouflaging.
Yeah, if you wind the clock back a few years, we were certainly not alone in our vision that off the shelf has to be the strategy to bring cell therapy to much larger patient populations. We were really differentiated in how we set about doing that. Most of our peers moved forward with roughly the same construct, healthy donor T-cell. You get rid of the T-cell receptor so you don't cause graft versus host disease, and you get the cells to express a CAR, and they call that a product. We looked at that and just scratched our heads. We couldn't quite square that circle because we know these healthy donor cells are foreign to the patient's immune system, and they're therefore going to be more rapidly rejected than auto CAR T.
You have to do something to bridge that biology. With CB-011, our myeloma program, we're actually directly trying to address that. We use an immune cloaking strategy. It's a multi-edit approach to try to slow down that immune mediated rejection to buy additional time for additional anti-tumor activity. We've seen the PK for that product, the persistence for that product is about twice that of uncloaked cell therapies, we think that's working, and we think that's really a key part of the differentiated clinical outcomes that we've seen so far.
With vispa-cel in lymphoma, we've taken an orthogonal strategy and said, "We don't necessarily need to tweak how long the CAR T cells are there if we can drive better anti-tumor activity while present." That really drove us to the PD-1 knockout strategy. The intent, of course, is to take the brakes off the CAR T-cells so that during that window of time, about 30 days while they're present, they pack as much anti-tumor punch as possible. We think it's that PD-1 knockout that is helping vispa-cel drive outcomes that are on par with autologous CAR T-cell therapies.
It feels like we have the first gen, which is the ex vivo CAR Ts. You've got allogeneic, like what you guys are doing. Now we have in vivo CAR Ts, which are kind of unproven, earlier stage. It's kind of each one has its own lane it feels like, but allogeneic is the sweet spot. You can just talk from a technological perspective, the differentiation and the benefits that allogeneic affords.
Yeah. Absolutely, I'll build with what do we see in the commercial landscape today, which is obviously exclusively auto CAR Ts. In spite of the extraordinary clinical and real world data that have cemented auto CAR Ts as the gold standard for second line large B-cell lymphoma patients, only about 25% of patients get auto CAR Ts. We're actually really excited that we've achieved alignment with the FDA on a phase III pivotal design to directly hit that head on by developing vispa-cel for patients who are not getting auto CAR T and are not getting auto transplant. We see significant opportunity to address that unmet need and build a really compelling future commercial case for vispa-cel by doing so. The key differentiator for vispa-cel, again, with the backdrop of today's commercial reality, is twofold.
The majority of those 75% of patients who don't get auto CAR Ts don't get them for one of two reasons. Either their disease has already progressed too far or is progressing too rapidly to wait, right? The reality of these N-of-1 therapies is it takes 8 or 12 weeks from the idea of getting an auto CAR to actually getting your dose, and a huge fraction of patients simply cannot wait. The other key driver is geography. Most of these patients are cared for in the community, but it's really only the top-tier academic sites that offer auto CAR Ts, and there are a lot of real world reasons why many of these patients cannot pick up and move to go, let's say, to MD Anderson to get their dose.
We believe vispa-cel is really well situated to address each of these issues, specifically as we think about the unmet need, that is exactly the patient population we're solving for in our phase III study. As we think about geography, because of the efficacy and the safety profile we've seen with vispa-cel so far, we know it can be delivered in the community. Actually our phase III study intends to leverage both sophisticated community hospitals alongside, of course, top-tier academic institutes. Your point is well taken. The field and the technology space continues to evolve, what role might in vivo CAR T play in the future? I'll say probably first and foremost, it's early days. We've seen some safety signals emerge already in that space.
There's probably a decent amount of engineering that it's gonna take to get either the LNP, mRNA or lentiviral platforms to function and function safely for a large enough patient population. What I'll highlight is whether it's auto CAR or in vivo CAR, in both of those settings, it's the patient's own T cells that need to do the heavy lifting. Whereas with allo CAR Ts, we don't ask anything of the patient's T cells. We are leveraging healthy donor T cells that have been armored, be it through PD-1 knockout or otherwise to really boost their anti-tumor potential. We think in both of these landscapes, allo CAR Ts have a really important role to play for these patients.
Right. We can definitely talk about the manufacturing advances that you guys have made, but just double-clicking on the phase III design since this was a recent update. You've aligned with the FDA, 2nd line LBCL patients who are ineligible for auto CAR T and transplant. I guess, can you talk us through the design regarding the endpoints, the comparators, the patient groups crossover, et cetera?
Yeah. We are really excited to have achieved alignment with the FDA on this study, and believe that the study design gives us incredibly high likelihood of technical success.
As I mentioned a moment ago, we're really focused on what I call the have-nots, the more than 50% of patients who can't get auto CAR T and are not getting auto transplant. By focusing on this patient population and running a randomized controlled trial, the control arm regimens therefore are chemo immunotherapy regimens that don't have curative intent. We think that setup gives us incredibly high likelihood for technical success. We are planning for a 250 patient population, 1-to-1 randomization. In the control arm, physicians will have a choice amongst four different regimens. Two of them contain polatuzumab, Pola-BR and Pola-R-GemOx. Two of them don't because we know some patients will see pola in the frontline setting, R-GemOx or tafasitamab.
As you look at all of those regimens, either clinical trial data or real-world data, all of them have PFS curves that just kind of drift back down to baseline. The only question is exactly how fast they do that. Looking across all of them, the median PFS is like 4.5 months. Now compare that to the phase I data from vispa-cel, where at our update at the end of last year, we saw this incredible plateau in the PFS curve, just like you see with the auto CAR T, where somewhere between 40% and 50% of patients are achieving these very durable outcomes. We had not yet reached median PFS or 12-month progression-free survival was about 53%. We really like that setup.
We also know that potentially accessing a cell therapy for patients who otherwise by definition have no access to cell therapy is a really important part for recruiting patients to this study. We plan to have crossover so that if a patient experiences a progression event on the control arm, they can cross over and get a dose of Vispa-cel. The primary endpoint will be progression-free survival. Now it's event-driven, not a landmark analysis, again, zooming out, we really like the setup and how high the likelihood of technical success is.
Okay. That's a really great overview. Could you just going one level deeper around, what drove the 250 number in terms of powering, how is the study blinded given that there's a crossover component?
Yeah. It is not possible to blind this because it's pretty obvious to everyone involved whether you're getting a few days of lymphodepletion and a single dose of a cell therapy or repeat cycles of, say, Pola-BR or one of these other regimens. It is not possible to blind a study like this. In terms of the size of the study, we've looked at a number of different scenarios as we think about the performance of vispa-cel and the performance of the control arm. We've modeled endlessly, as you might imagine at this point, and have high confidence that a 250 patient population gives us quite high power in a huge fraction of these scenarios to detect success.
Okay. Great. We can talk about the November update just as kind of setting the stage for what we expect in the pivotal. You showed efficacy and durability that was actually on par with autologous CAR T. Some of the conversations I was having after was like, at this level of activity, why don't we just compare it against the auto CARs? I appreciate the unmet need in kind of the driver, and it might actually be faster to read out given the study or the setting you chose. You can just talk to the efficacy that you were seeing there and how that should hopefully translate to the pivotal.
Yeah. Your headline is exactly correct, copy paste the auto CAR Ts, whether you're looking at overall response rate, complete response rate, or how durable these responses are. Really interestingly, what we saw in our phase I data set is that if a patient achieves a response by the month 3 efficacy analysis, the likelihood that they remain in response is incredibly high. There were very few patients who relapsed post three months, we're really excited next month at EHA to share an update that will be many additional months of follow-up looking at these patients to get an even better picture of how durable some of these outcomes can be.
Okay. Obviously, the longest duration patient in the study, is out to three years, and they're still in response.
As of that update, that's correct. Actually, the very first patient we ever treated who wasn't a second-line large B-cell lymphoma patient, so he doesn't fit in that cohort that we've been focused on, he's been in complete response for more than four and a half years at this point.
Wow. That's obviously auto CAR or better types of activity.
Exactly.
Yeah.
You could elaborate a little bit more on the EHA presentation. You've got an oral there, that's in two weeks. What do you hope to gain from, in terms of incremental on the follow-up? How do you think you'll be able to position that data set to help recruit investigators for the pivotal?
Yeah, it's a great question, look, keep in mind, the way we've been thinking about developing vispa-cel is we want to see auto CAR T like outcomes, now we're planning for a pivotal study where the control arm is chemo immunotherapy regimens with no curative potential, pretty rapid progression-free survival kinetics. We love the setup for this study, as I mentioned a few minutes ago, what we expect to see based on our historic data is what I'll call more of the same. Continued evidence of how durable the responses are with vispa-cel following just a single dose.
We think, whether it's last year's PFS curve or what we might be able to put out later this year, any of those data sets are going to be incredibly attractive to this patient population because they don't have access to anything with curative potential. Part of one of your earlier questions, Alec, was getting at why not think about a head-to-head study against auto CAR T. As you might imagine, we thought about a lot of different development paths for vispa-cel and actually discussed many of them with the agency.
There is a potential path forward to run a head-to-head study. You wouldn't need to show superiority, you would simply need to demonstrate non-inferiority. That's also fairly large and therefore fairly expensive and long study to run, and we saw such an opportunity to serve the lion's share of the second line setting, who is not getting auto CAR T today with our study design, in a way that is addressing unmet need right on the nose.
Is more capital efficient and time efficient. I should say based on the payer research that we've done, and the KOL interactions that we've had to date, we expect on the other side of an approval that vispa-cel will be more broadly used than just the patients who would most appropriately fit this pivotal study. Payers, for example, have shared with us an expectation that it'd be a fairly simple physician attestation to say patient's not gonna get auto CAR T, patient's not gonna get auto transplant, and they're off to the races with vispa-cel.
That could be as simple as the patient can't or won't travel. They can't wait for the auto to be produced. Just simple things like that.
Exactly.
When we think about logistics for enrolling the pivotal, is patient selection based on the criterias we mentioned easy to implement? What training have you had to or are you putting in place for your investigators?
Yeah, that's a great question. We certainly want to collect the data on why any individual patient is being enrolled in this study. In particular, is it because they are deemed medically ineligible? If so, why? Versus are they access challenged? The real world constraints, be they financial, geographic, or insurance-based. I'll gloss over the medically ineligible criteria by saying they all sum up to can't wait, right?
There's something about their disease biology that means it is not appropriate for them to sit around for 8 to 12 weeks to wait for a dose of CAR T. They need something now. We are getting tremendous enthusiasm from sites both here in the U.S. and ex-U.S. We expect this to be a global study to drive rapid enrollment. People are super excited about having something that looks and smells like an auto CAR T, is readily available off the shelf.
Okay. We'll talk about the manufacturing next, but one more question on kind of the site, the complexion of the sites that you expect will open on the clinical study. Are these larger academic centers that have already established their auto CAR infrastructure, or is it the vast majority of the country where these capabilities haven't been built out, but there's a strong desire to have a cell therapy option?
Yes and yes. We're aiming for, let's call it a 50/50 mix here in the U.S. between top-tier academic institutes that have ready access to auto CAR Ts, as well as these sophisticated community hospitals who, for a variety of reasons, are unable to stand up auto CAR T practices but are using bispecifics today. I highlight that because the care approach necessary to serve patients with a bispecific is exactly what you need for an allo CAR T, right? It's the same AEs, it's the same monitoring, it's very similar SOPs.
In many ways, it's actually easier for these sites to deliver an allo CAR because it's a single dose compared with the bispecifics where it's repeat dosing, and therefore you have these very prolonged periods of time where those bispecific patients need to be monitored for AEs, and in particular address the infection challenges that they face over a long period of time.
Okay. That makes sense. Talking about the ease of manufacturing, this is something you guys have obviously been optimizing over the past several years. COGS seem very low, fraction of the price of an auto obviously, which is quite expensive and laborious to make. It's more like a drug than a cell therapy, like your analogy of the bispecifics. I guess, how does the cost structure change the commercial case? If you could just add a point on kind of how you're positioning the antigen makeup within the batches for patients.
Yeah, absolutely. Look, the supply strategy for Vispa-cel makes an incredibly compelling commercial case for this program. This is not the very expensive N-of-1 bespoke strategy that people are used to when they think of cell therapy. To your point, this truly looks like a drug that anyone here at the Bank of America Conference would recognize in their businesses. We anticipate cost of goods sold at launch to be 96% lower than the auto CAR T's. It creates just tremendous opportunity as we think about building that business. You've highlighted kind of a key piece of our biological strategy here, which is based on a learning that we've come to over the past few years, that very modest HLA matching between patient and donor helps drive these auto CAR T-like outcomes.
Very quickly, there are two different classes of HLA, 12 different alleles total, and we found that matching any two out of the 12 helps drive those auto CAR T-like outcomes. We can solve for this on the supply side without restricting patient access to Vispa-cel. For example, here in the U.S., based on our modeling, we only need about 10 different lots from 10 different donors to be able to serve 99% of lymphoma patients with a two-plus matched product. Actually to get that long tail into the 2-plus category, our average match in the U.S. will be something like seven alleles.
That's great. Definitely scalable.
Right.
Way more scalable than auto.
Yes. Tremendously so.
Yeah. I wanna shift gears and talk about CB-011. This is your anti-BCMA allogeneic CAR T for multiple myeloma. Slightly different, complexion in terms of what you would be comparing yourself against more bispecific heavy in multiple myeloma. You'll also have an oral presentation for this asset at EHA, but you could just speak to what you're seeing in the dose escalation and what we'll be expecting at the conference.
Yeah, absolutely. I'll start with the benchmark for success, which is exactly as you highlighted the bispecifics, right? For years, as we talked to myeloma physicians, they uniformly point us to the bispecifics and say they'd love to see an allo CAR T that could at least drive responses on par with the bispecifics because then they'd pick that one and done every day of the week over the repeat treatment burden and the long-term infection challenge that bispecific patients face. That was really the goalpost that we had in mind as we've been developing this program for a few years. At the end of last year, as you well know, Alec, we shared our first ever update, which was a pretty meaty update.
It was all 48 patients enrolled in dose escalation. One of the many decisions we made based on those data was the recommended dose for expansion, which is a single dose of 450 million CAR T cells. We had already enrolled 12 BCMA naive patients in that cohort, in the recommended dose for expansion cohort. The response rates that we saw there far exceed those of the bispecifics and are actually approaching auto CAR T-like outcomes.
We're really excited to leverage some of the podium time at EHA next month to share an update on those patients. We'll be really focused on the fact that many months have passed by. This will be much longer follow-up on those patients because, of course, not only is response rate and depth of response important, and by the way, 91% of those patients achieved MRD negativity, that's really exciting, but also the duration of response. We know this data set will allow us to tell that story.
We're also busy with that program. I'll say, we're actively enrolling patients in dose expansion. We've committed to sharing dose expansion data by the end of this year. Though our update at the end of last year really focused on that BCMA naive cohort, we're now really trying to think about both sides of the BCMA equation. We're continuing to enroll BCMA naive patients to continue building that data set. We're also enrolling post-BCMA patients. We know that's a large and rapidly growing patient population here in the U.S., and we're eager to understand what benefit CB-011 can provide for those patients.
Great. Looking past EHA to the second half update from the dose expansion, I guess what would you want to see here to give you conviction that this is truly best in class? At what point would you approach the FDA to talk about a further study or even a path to pivotal?
Yeah, it's a great question. By building data sets in both BCMA naive and BCMA exposed patients, we're really maximizing optionality as we think about the future development path for CB-011. These dose expansion data will be a key piece of the puzzle for making decisions about how to prioritize our resources as we continue developing the program. We were so excited that just recently the FDA granted RMAT designation to this program. It actually gives us the opportunity to engage earlier and more proactively with the agency. We expect to engage with them later this year. Ideally we'd be in a position where we're not only providing that dose expansion data update, but also providing some guidance that we're hearing from the FDA about how to leverage these data learnings for next steps for the development of the program.
Okay. Great. Is there an expectation from the dose escalation or a bar in terms of activity that you're striving towards, and what's the regulatory precedence for whatever registration endpoint would look like?
Yeah, great question. Starting with that dose escalation data set, what KOLs have told us for years is look like a bispecific. What does that mean? The bispecifics have overall response rates in the kind of 60%-70% range. We're well north of that with what we've seen so far. How about PFS? What we've largely heard from KOLs is they'd love to see median PFS somewhere in that 12-15 month range. That's an important benchmark for those BCMA naive patients. Probably stay tuned as we think about how best to articulate the goalpost for post BCMA patients. Those are obviously, even higher unmet need and higher risk patients.
We think there's a lot of opportunity for a healthy donor T cell product for those patients. In terms of the pivotal path, like I said, there's a lot of optionality as we think about the development strategy. I almost worry we're a little bit kids in a candy store right now with where you could go with CB-011. Kind of on one end of the spectrum would be an initial strategy into a fairly large patient population running a traditional randomized controlled trial and thinking about how best to sequence that versus there might be some pretty straightforward capital efficient faster paths that even could be a single arm trial. For example, in some of these post BCMA patients, or other well-defined patient populations where we think CB-011 could have a lot of benefit.
Okay. Just thinking about the cadence of enrollment for the BCMA naive versus the BCMA exposed. Are you seeing any difference in terms of the enrollment rate from either of the populations? Just thinking about the information that we'll gain from the expansion.
Yeah, that's a great question. I'd say our expectation is to be able to read out a bigger data set on BCMA naive than where we left the end of last year with our dose escalation, and to have enough patients in the post BCMA cohort to say something meaningful about that too, right? We'll be leveraging those data to help drive those decisions. Will it be the exact same numbers between those two cohorts? Time will tell. Enough patients to drive these learnings and decisions is our objective, and we're on path to do that.
Okay. Just thinking about what you want to keep for yourselves versus potentially partner out, especially thinking about larger like a head-to-head study against bispecifics and multiple myeloma, for example. Do you have like a preference within the pipeline for what you want to kind of or what you feel like is reasonable for you to push forward on your own versus what think you can drive the most value through a partnership?
Yeah. I'll tell you what's on my whiteboard, I'll readily acknowledge it takes two to tango, right? There might be very different opportunities that present themselves in the future to us. Our objective right now is to develop vispa-cel on our own. We believe we can be the company that pushes vispa-cel across the finish line from a regulatory perspective and could launch vispa-cel here in the U.S. To your point, as we think about multiple myeloma, we feel that is a disease area that is really ripe for partnership, as we think about simply how large the unmet need is and how significant the commercial opportunity is.
Okay. Your guided cash runway, it is second half 2027. Which carries you through the start of the, of the pivotal for CB-010. How do you see that runway or the optionality within the pipeline to extend that?
Yeah. Our balance sheet today fully funds the CB-011 dose expansion work that we were just talking about, really excited to get to those data later this year. It funds our ability to kind of get started, do some of the important startup work for the ANTLER phase III pivotal trial. It does not fully fund the company through the readout for that study, we will need to raise additional capital to do that work.
Okay. Definitely having the RMAT for CB-011, like even having that feedback from the regulator in hand, like I imagine that's valuable to your partnership discussion.
Absolutely.
As well. Just in the last minute or so, just so we've got a nice checklist for what to be looking for from Caribou, over the next 12 months, if you could run down like the updated catalyst calendar for you guys.
Yeah.
That'd be great.
Yeah, absolutely. Look, come and see us in Stockholm. We're excited to see both vispa-cel and CB-011 on the podium at EHA next month. With respect to CB-011, as we've been talking about, we've also guided to dose expansion data by the end of this year. A rich series of data updates for both programs.
Okay. Do you think we'll be getting updates on manufacturing too or like non-clinical updates to be looking out for?
Yeah, that's a great question. We've also taken some time already this year to tell the translational story behind both of our programs. We think there's a lot that we've learned that continues to really support why we've seen such differentiated clinical outcomes, driven by kind of the unique edits and the biology of our healthy donor allogeneic CAR T cells. From a manufacturing perspective, we are fully scaled. We are at a commercial scale for vispa-cel already. Really excited that we're there, checkbox on that one, and leveraging that for manufacturing for our pivotal study as well.
Okay, great. With that, we're out of time, so we'll have to leave it there. Thank you, Rachel, for the great conversation, and thanks everyone for attending.
Thanks, Alec.
Thank you.