Precision BioSciences, Inc. (DTIL) Earnings Call Transcript & Summary

All right. Good afternoon, and thanks for tuning in to the JPMorgan Health Care Conference. I'm Eric Joseph, senior biotech analyst. Our next presenting company is Precision BioSciences. And here to talk to us a little bit about the company is CEO, Matt Kane. Before I hand the presentation over to him, let me just remind folks that there is a Q&A after the presentation. [Operator Instructions] With that, Matt, thanks for sharing some of your time with us this afternoon.

Yes. No, thanks, Eric, and it's great to be with you again, and everybody that's on the line today. Since we're in this virtually, I'm going to try to guide us through the deck. So let's turn to Slide 2. Before we begin, I've got to remind everyone that the statements we make may be considered forward-looking. And so please refer to our most recent 10-Q for more information about our business and various risk factors. So let's turn to Slide 3 and get this going. At Precision BioSciences, we are intensely focused on delivering on the promise of therapeutic genome editing to overcome devastating diseases. And our foundation is a wholly proprietary therapeutic-grade genome editing platform called ARCUS, and it's something that we developed right here at Precision BioSciences. The ARCUS genome editing platform offers numerous technological advantages for the development of potentially curative human therapeutics. And we'll talk a lot more about ARCUS later today. But by leveraging these advantages, we've developed one of the industry's leading allogeneic or donor-derived CAR T platforms. And we've also quietly built what is arguably the leading in vivo gene correction platform focused on curing genetic and infectious diseases. My co-founders, Jeff Smith & Derek Jantz, are two of the true pioneers in the field of genome editing and together have been developing gene editing technologies for over 20 years. Giving us some unique and very deep insights into how to translate these powerful editing tools into greatly needed therapeutics. We've also invested at a very early stage in GMP manufacturing and now have a facility capable of supporting both our CAR T and in vivo programs. And finally, we, today, have a strong balance sheet that gives us a cash runway into 2023. Let's turn to Slide 4. So with ARCUS, we have a very special proprietary genome editing platform that we are leveraging to develop 2 cutting-edge therapeutic approaches: in vivo gene editing to permanently correct genetic diseases and allogeneic CAR T therapies to overcome cancer. And as a company, we are highly focused on unlocking the full value of ARCUS to address serious diseases and we think made a tremendous amount of progress in both areas during 2020. Let's turn to Slide 5 to quickly recap the year. So in 2020, we formed what can only be described as a transformational partnership with Lilly in the field of in vivo gene editing that included up to 6 targets, and we named one, Duchenne muscular dystrophy that we're really excited to move forward today. This came with an upfront payment of $135 million and importantly, extended our runway into 2023, giving us a really strong financial position from which to invest in our most promising programs. And last month, we also announced interim results from our CD19 targeting allo CAR T PBCAR0191. And in this update, we were pleased to report an 83% ORR in patients with NHL and ALL and this program was combined with an enhanced lymphodepletion regimen. We'll have quite a bit more on that in just a bit. We also expanded our CAR T collaboration with Servier with 2 new targets aimed at solid tumors and 2 additional targets aimed at hematological cancers. And I'm like, look, really proud of our team for their work to initiate 2 new clinical studies in our allo portfolio, PBCAR20A targeting CD20 and PBCAR269A targeting BCMA. And finally, we delivered a very important milestone late in the year with the IND filing for PBCAR19B. And this is our next-generation or stealth cell CD19-targeting CAR T, which is designed to extend cell persistence of the CARs. Let's turn to Slide 6. And just briefly, I wanted to touch on some of the important achievements we made with our in-house GMP manufacturing facility that we call MCAT. First, we completed a tech transfer from our CMO to MCAT for both the CD19 and CD20 targeting programs and also manufacture the clinical materials for our BCMA program. More recently, we were also able to produce the first batch of clinical trial material for the upcoming stealth cell program. So we continue to view this as a very important strategic asset for the company. Let's turn to Slide 7. And as a result of the progress we've seen, we now have a deep allogeneic CAR T pipeline, including 3 products in the clinic and a fourth expected to be in the clinic in 2021. And beyond the named programs, we've also added 4 new undisclosed programs to the pipeline and plan to study our BCMA program in combination with a gamma secretase inhibitor in multiple myeloma in the first half of the year. So turning to Slide 8. Let's discuss our CAR T platform. Our CAR T cells originate from healthy donors, and through a patented single step process, we knock out the T cell receptor by knocking in the chimeric androgen receptor, or CAR, into the TRAC locus, giving us consistent expression of the CAR and helping us to avoid GvHD and do so in a single step. The CAR also contains our proprietary N6 co-stimulatory domain, which is designed to control the expansion of our CAR T cells. We have a 10-day manufacturing process that is shown to produce consistent batches of cells, composed primarily of naive T cells in an approximate 1:1 ratio of CD4s to CD8s. And the data that I'll review in just a bit will include cells from 5 different batches across 4 different donors. Let's turn to Slide 9 and in an overview of the clinical trial design for our lead CD19 program. So in this study, we've enrolled separate cohorts of NHL and ALL patients using a standard lymphodepletion regimen containing fludarabine and cyclophosphamide, and we previously shared data across several dose levels that demonstrated clear signs of anti-tumor activity and a well-tolerated safety profile. We've also recently disclosed data from 6 patients that received an enhanced lymphodepletion regimen, which consisted of increased doses of both fludarabine and cyclophosphamide. Now if you turn to Slide 10. You can see that this enhanced LD arm was correlated with quite impressive antitumor activity. In the NHL patients, we demonstrated a 95-fold increase in peak cell expansion. And importantly, more than a 90% decrease in tumor burden across all these patients. So look, these patients are achieving very deep responses with the enhanced LD strategy. And look, while early, our investigators have found this approach to be quite compelling. Let's turn to Slide 11 and the response rates we've seen in NHL patients. Now we enrolled into a pretty tough patient population. All the patients had aggressive disease, and we did allow patients with prior auto CAR T and prior stem cell transplants. And yet, we achieved a 75% complete response rate in NHL at day 28 with the enhanced LD strategy. And to our knowledge, this is the highest complete response rate seen to date in NHL patients treated with allo CAR T cells. And across all the patients we've treated, regardless of the lymphodepletion, nearly 40% also achieved a CR, and look, between these response rates and the greater than 90% reduction in tumor burden with enhanced LD, it's clear that PBCAR0191 has very potent and very rapid antitumor activity. Let's now turn to the safety profile on Slide 12. So across all doses of PBCAR0191, we are seeing an acceptable tolerability and safety profile with no DLTs, GvHD and no Grade 3 or higher CRS or neurotox. Now we did report one death from sepsis, and that occurred in a patient who had 9 prior lines of therapy, prolonged and persistent cytopenias and an infection just prior to enrollment. So as a result, we have modified our eligibility criteria so that these prolonged cytopenias or serious infections within the past 30 days will not be allowed to enroll in our study. So let's turn to Slide 13 and talk about the next steps with this program. So if we step back, look, while our end goal is, of course, to be able to safely deliver deep and durable responses in a high percentage of patients, one of the guiding paths from the outset of this study has been to rapidly optimize the area under the curve, or AUC. And as we show on this slide, and we're also looking to draw from insights that we've garnered from the autologous CAR T experience. And we think we have 3 ways to potentially make this happen. First, we can look to pull the peak of the curve up. And we've already shown that we can effectively do that with the enhanced LD strategy, and we remain focused on enrolling more patients using the enhanced LD. And because of the safety profile we've established, we are also going to be able to increase the total dose as another means to increase this peak, and we are doing that now as well. Second, we can look to stretch out the curve, overextend the persistence of the cells, and we're exploring some novel approaches to lymphodepletion to see if this might be possible. And finally, because this is a readily available allogeneic CAR T, we can also increase the total AUC by scheduling repeat doses, in effect, creating multiple curves, and we're doing this now as well. So 4 distinct paths to rapidly optimize the AUC and our quest to safely achieve durable responses for this program. Let's turn now to Slide 14. Now another way for us to potentially increase the persistence of our allo CAR Ts, which, remember, are foreign cells is to engineer them so that they cannot easily be detected and eliminated when the patient's immune system reconstitutes. We call this our stealth cell strategy, and this is designed to avoid rejection by both T cells and NKs. So building on top of our base CAR-T platform, the stealth cell incorporates a short hairpin RNA that suppresses beta 2 microglobulin to reduce MHC Class I expression. And we believe that if we can reduce Class I, we should also reduce the ability of the patient's T cells to recognize and kill the CAR Ts. But importantly, we don't knock out B2M, we just suppress it to about 10% of normal levels. And we do this because that we found that doing so in vitro leaves just enough Class I on the surface to avoid triggering a response by the NKs. And to further protect ourselves from NKs, we've also incorporated an HLA-E decoy. So very importantly, though, we do all of this still using that single step engineering process in a manufacturing process that is effectively identical to the one that we developed for our base CAR T programs. We're anticipating being able to test this stealth strategy in the clinic in the first half of 2021. So a very, very exciting program we're looking forward to learning more about. Let's turn to Slide 15 and our CD20 and BCMA programs. So this year, we are expecting to share interim data from both of these studies. They were both started in the first half of 2020, and we're expecting to get into our third dose level for each program this quarter. So very solid progress thus far with both of these programs. And in addition, as I mentioned earlier, we're also expecting to begin a cohort in the first half of this year that includes a GSI in partnership with SpringWorks for our multiple myeloma program. So that's CAR T. Transitioning next. I'd like to turn the presentation over to our Co-Founder and Chief Scientific Officer, Derek Jantz, who will pick up on Slide 16 to discuss ARCUS and our in vivo gene editing programs. Derek?

Thank you, Matt, and hello, everyone. So ARCUS isn't CRISPR. It's actually based on a naturally occurring gene editing enzyme called I-CreI that comes from Chlamydomonas reinhardtii, which is an algae. And it's a very rare example of a gene editing enzyme that actually evolved for function in a eukaryotic cell with a very large genome. And because of that, it has a number of pretty unique attributes that we think make it a much better starting point for the development of therapeutic gene editing enzymes. Chief among those is it has a really exquisite degree of specificity that prevents it from cutting off-target sites in the genome. And this is in large part because I-CreI has the very unusual ability to turn itself off after it has made its intended gene edits. So we can actually express the enzyme for a really extended period of time in a cell or animal without having to worry about it sort of wandering off and finding off-target sites in the genome to edit. And we really -- we view this as the primary safety challenge that any gene editing technology is going to need to overcome, particularly for in vivo editing applications. A second important attribute of I-CreI is the type of cut that it makes. When it cuts the genome in the cell, it leaves behind a pair of 3 prime sticky ends. And those are important because they act as a sort of signal to the cell that, that particular DNA break needs to be repaired through a high fidelity repair process called homology-directed repair, or HDR. And really the practical significance of that is that the enzyme is very good at inserting DNA into the genome. It's not just a tool for knocking genes out. And as Matt mentioned, in fact, our entire CAR T platform is based on the ability of ARCUS to insert the gene encoding the CAR with high-efficiency into the T cell genome. Lastly, the enzyme is very small, which makes it very easy to deliver with any of the common gene therapy vectors or particles. And this is actually turning out to be a very significant practical advantage for in vivo gene editing in which delivering ARCUS to the right tissues and to the right cells within those tissues is actually half the battle. ARCUS was developed at Precision, and we own and control the IP, which puts us in a pretty unique position in the gene editing field from an IP perspective. So that's what's great about ARCUS. If you want to turn to Slide 17, here's the challenge, and this is the reason that everybody isn't working with I-CreI. It turns out it's actually very difficult to reengineer I-CreI to make it recognize a new DNA sequence for a new therapeutic project. It's not nearly as easy as just changing out the guide RNA like it is with a CRISPR. The protein engineering process that we use in-house is iterative and it involves multiple cycles of testing and optimization of the nuclease to get to what we consider to be a clinical candidate, which is a nuclease that has a high rate of on-target editing but really no significant detectable off-target editing. And that iterative process can take 6 months or longer, but that upfront investment in time that it takes to make the nuclease, in fact, results in a much, much higher quality nuclease at the end. Then there's also so much IP and know-how that goes into this process that it represents a very high barrier to entry for anybody else. With the next few slides, I wanted to give some examples of ARCUS in action to really give a sense of the power of the technology, particularly for in vivo editing. And that starts with Slide 18. These are results from a study that we've conducted with Jim Wilson's group at the University of Pennsylvania starting in 2017 with the goal of knocking out the PCSK9 gene in nonhuman primates. We made an ARCUS enzyme for PCSK9. And Jim's group put that nuclease into an AAV vector and delivered it to NHPs and the results look like this. On the lower left of this slide are the serum PCSK9 levels. And what you can see is that immediately following delivery of the vector, PCSK9 levels drop by over 90%. And importantly, once they go down, the PCSK9 levels stay down now out past 3 years, and we're continuing to follow these animals. And then on the right, you can see a corresponding drop in LDL-cholesterol by more than half. So literally, what we're looking at here is a one-time treatment, one-time infusion that cut their cholesterol in half. And at this point, we really have no reason to think that this isn't permanent. The vector is more or less gone by now, the ARCUS is gone, but the gene edit is stable and is being inherited by subsequent generations of hepatocytes. And this is really the fundamental difference between editing and RNAi or conventional AAV gene therapy in that we're able to achieve what we think could actually be permanent cures by changing the genomic DNA. On Slide 19, we have a similar example. In this case, from our TTR project, showing knockout of the transthyretin gene responsible for TTR amyloidosis. So it's a different gene, but really the same story as we just saw for PCSK9. We see a very rapid reduction in serum TTR levels following administration of the vector. In this case, actually greater than 95%. And once the TTR levels go down, they stay down over time. On Slide 20, I wanted to provide a quick example from the eye. In this case, our gene target is a mutant form of the rhodopsin protein that is the primary cause of autosomal dominant retinitis pigmentosa. And we've demonstrated in a pig model of AdRP that we can restore rod function by delivering ARCUs via a subretinal injection. And we measured this either by electroretinogram, shown on the left or a visual acuity test shown on the right in which we actually got to make the pigs run through a maze, which was very entertaining by all accounts of the people that were there to do that. And I really just wanted to show these to demonstrate the versatility of the technology for in vivo editing and demonstrate that we can achieve a very broad spectrum of edits in a variety of different tissues. Our gene editing pipeline is -- in vivo gene editing pipeline is shown on Slide 21. We recently, as Matt mentioned, announced an agreement with the Eli Lilly that includes 3 targets, one of which is Duchenne muscular dystrophy and then our lead internal program is primary hyperoxaluria type 1. All of the other projects on the slide that are shown in teal are research projects that have achieved proof-of-concept in a large animal and are available for us to either partner or to use to backfill our internal pipeline. On the next slide, Slide 22, is a quick overview of our Eli Lilly partnership. There are 3 programs in our active pipeline. And then Lilly has the ability to select up to 3 additional gene targets down the road. The economics are on the right. And that all works out to about $2.7 billion in total deal value. Under the agreement, Precision is responsible for everything up to clinical, and then we hand it off to our friends at Lilly. If you want to jump ahead a couple of slides to Slide 24, I wanted to provide a quick overview of the approach that we're taking with the DMD project because it's really pretty unique. What we're doing is we're using a pair of ARCUS enzymes to cut simultaneously and delete a 0.5 million base pair segment of the dystrophin gene that encodes exons 45 through 55 of dystrophin. We picked that particular region because about 20 -- about 60% of DMD-causing mutations fall somewhere in this region. So by just deleting all of it, we can potentially treat up to 60% of patients. When we do this, the 2 sticky ends that are created by the ARCUS enzymes find one another inside the cell and just ligate back together perfectly to create a new gene that is an in-frame fusion between exons 44 and 56, and this results in the expression of a dystrophin protein that is almost full length, minus those 10 exons. And we know that those 10 exons are actually largely dispensable for function because this deletion mutation of exons 45 through 55 actually exists in the human population and has a very mild phenotype. On Slide 25, just some data showing that this actually works, at least in cell culture. When we treat myoblast from a DMD patient with increasing doses of the 2 ARCUS enzymes, what we see is that up to 30% of them have the desired gene edit in the DNA and that corrected DNA then leads to corrected mRNA, which then leads to expression of dystrophin, shown on the right-hand side of this slide, albeit a slightly truncated form of dystrophin. So essentially, what we've been able to do here is reactivate the latent dystrophin gene in these patient cells, which I think is just really cool. On Slide 26, we wanted to add this figure to sort of differentiate the approach that we're taking from the approaches that the microdystrophin and ASO groups are using in the clinic because we do get asked about this all the time. Aside from the fact that we're actually producing a dystrophin protein that is pretty close to full-length and fully functional, the biggest difference is that we're targeting satellite cells, which are the stem cells that live on the surface of the muscle fiber and give rise to new myofibers by correcting the satellite cells, we expect, really, a significant amplification of our effect because a single corrected satellite cell can give rise to a really large number of permanently corrected nuclease in the myofiber. So effectively, the delivery bar for us could be a lot lower because we need to target a much smaller population of cells due to this amplification effect. And then the correction should be a lot more durable because once the cells are edited, they will stay edited. I do want to touch very quickly on some new preclinical data from our primary hyperoxaluria program, so if you want to jump to Slide 28. The PH1 disease is caused by mutations in the glycine biosynthesis pathway that result in the accumulation of a toxic metabolite called oxalate that causes extremely severe and potentially fatal kidney stone accumulation. We're interested in addressing this by knocking out a gene called HAO1 that acts at an upstream point in the pathway and will lead to the accumulation of glycolate instead. A glycolate is a nontoxic soluble metabolite that is easily excreted. On Slide 29 is our most recent nonhuman primate data demonstrating that a single administration of an AAV vector encoding our HAO1-specific ARCUS results in really dramatic reductions in HAO1 mRNA as well as the resulting protein, which is called glycolate oxidase, or GO. And on the right-hand side, you can see the corresponding metabolic shift in those animals as we see high levels of glycolate start to appear in the serum. We included these data because we think it's actually pretty easy to benchmark this against the published nonhuman primate data with siRNA products that are in late-stage clinical development and we really think that this demonstrates that we get significantly greater therapeutic effect with a single administration of ARCUS than the siRNAs are able to achieve even with ongoing repeat administration, again, really highlighting the potential advantages of editing over other approaches. And with that, I will turn it back over to Matt to bring us home.

All right. Thank you, Derek, nicely done. Let's turn now to Slide 30. So look, as we enter 2021, we do expect to achieve many important milestones for the company, which should advance our story and serve as the key opportunities to unlock some of the enormous potential value we think we have with ARCUS-based gene editing. On the in vivo editing front, just last week, we fully closed the Lilly partnership, and we'll seek to rapidly move those programs forward. We also plan to give an update on our PH1 program in the first half with additional preclinical data and plans for the program. We plan to have even more visibility on the CAR T side and expect to initiate our clinical program for the CD19 stealth cell program in the first half and also expect to begin dosing our BCMA program in combination with a gamma secretase inhibitor also in the first half. In addition, we plan to share additional data on PBCAR0191, our lead CD19-targeting CAR T around the midpoint of 2021 and anticipate having data on over 40 patients, including double-digit numbers of enhanced LD patients and, of course, longer follow-up. We also plan to update on our learnings with repeat dosing, higher doses and lymphodepletion strategies. So we expect this to be a very important update for our CAR T platform. On the corporate side, we also expect to complete the full spin out of our food business, Elo Life Systems, to really allow its ongoing operations to be funded with external capital and independently grow into its full potential. And finally, we plan to provide an early look from data from both the CD20 and BCMA programs. Turning now to our final slide. To conclude, I want to thank everyone for joining us today to learn more about Precision BioSciences, our proprietary ARCUS genome editing platform and the significant advancements we're making in our quest to overcome cancer and cure genetic disease. So thanks, again, everyone. And Eric, I'll turn this back over to you.

Great. Thanks, Matt. Thanks, Derek, for that presentation. I thought we'd start the Q&A with the in vivo gene editing programs. We saw an update from another company that was using another editing technology to down-regulate PCSK9 with some data, some long-term follow-up data and NHPs. I guess how would you sort of contrast those data with what you previously described in your work with Jim Wilson in terms of adding any efficiency and, to the extent possible, off-target editing?

Yes. Well, first of all, it's great to see some other groups having some success in translating gene editing tools into large animal models because we think large animal validation is absolutely going to be critical to see gene editing used broadly in the in vivo setting. So it's been -- we've been saying that for a long time, and it's great to see others placing heavy focus on that as well now. But I'll let our CSO, Derek, take the rest of that question.

Yes. I think what I really appreciate about our PCSK9 data, and the reason I like to share it is the fact that we have so much follow-up from that initial administration of the vector. There really is a lot of data that has been generated by us and by Jim Wilson's group as part of this study to really help us understand how ARCUS behaves in the animal and different ways to deliver it and different ways of thinking about things like immunogenicity, safety of the approach in addition to, obviously, the efficacy of on-target editing efficiency. So it really is a -- it's been a great collaboration with Penn that has generated a very, very nice large data set that I think is allowing us to draw a lot of general conclusions about using ARCUS for in vivo editing.

Got it. And just picking up on this new NHP data. And forgive me if I missed it, but can you just talk about some of the delivery modality that you're using there, whether it's -- I believe in the PCSK9, you're using AAV. And previously, you've also talked about LNP as an optionality. What's being applied here and sort of what modalities do you anticipate taking into the clinic?

Yes. That's right, Eric. We've successfully used both AAV and LNPs with ARCUS in the past, but I'll let Derek comment a little more on exactly what we did with this PH1 study.

Yes. The PH1 data that we've shared was generated using AAV, although as Matt said, we have used AAV and LNPs somewhat interchangeably, each of the technologies has its own advantages and its own disadvantages. As far as which of the 2 approaches we intend to use for the PH1 program, in particular, that will be part of an update that we'll give on PH1 later this year.

Okay. Okay. And on looking at HAO1 knockdown, pretty impressive knockdown levels. And I guess it's early, kind of, days right now in terms of looking at follow-up but do you have a sense of -- I guess, how long have you tracked these out so far? And do you have a sense of the durability of -- the stability of knockdown? And then secondly, where would you want to be in terms of normalized serum glycolate levels? Are you sort of in that normalizing range right now, at least in the models?

Yes. We are still at a relatively early stage with the PH1 study, but this is where the nearly 4 years of follow-up data we have with PCSK9 gives us so much confidence. We're seeing the exact same sort of trends already. But let me let Derek comment further on the reduction.

Yes. In terms of -- you're right. In terms of stability of the edit for PH1, we have no reason to think that it isn't going to track exactly like the PCSK9 project, exactly like the TTR data that I showed because, again, once the gene has changed, it's changed, and that's permanent. In terms of magnitude of the effect, in fact, with the data that we showed, we overshot where we need to be. And we -- one of the things that's so nice about the PH1 project is the fact that we do have Alnylam out in front, really blazing that trail and Dicerna out in front blazing that trail. So we do have a really good sense of kind of where we need to be to have significant therapeutic benefit. And in our most recent study, we actually went well beyond where we need to in terms of the efficiency of knockout, which is terrific because that tells us that we have a lot of room to work with in optimizing our dose. We can go down presumably significantly from where we were in that study. And still achieve a significant therapeutic benefit.

Okay. And I guess for the DMD program, it's early since the announcement of that partnership. But a similar sort of delivery -- choice of delivery modality question there, right? So difference is instead of targeting the liver, you're trying to reach skeletal muscle. This choice of delivery modality -- just some of the early thinking there is, is it most likely going to be an AAV-mediated approach? Is there opportunity for LNPs in that indication?

Yes. Well, I can say, first of all, we're really happy that Lilly picked DMD for numerous reasons. One, there's just a huge, huge unmet need, which we'd love to be a part of helping to solve, but also because the editing challenges really speak to the strengths of ARCUS. So we talk about delivery, this is where the small size comes in, the fact that we can put multiple ARCUS nucleases onto a single AAV vector. This is where those 4 based 3 prime overhangs really matter. It doesn't work without that. But let me let Derek talk a little bit more about how we're thinking about delivery.

Yes. AAV is the obvious choice in this case because we can leverage the learnings from the microdystrophin groups. We know what works, we know what doesn't work. And as Matt mentioned, we can very easily get both of the ARCUS enzymes into a single AAV vector, which in this case would be absolutely critical, given the number of cells that we have to target, presumably, in order to have therapeutic benefit. If we had to split the nucleases between two vectors and expect them both to get to the sales at the same time, I think that is probably too high of a hill to climb. So I would say likely AAV, again, with the caveat that we are targeting primarily -- or interested in targeting primarily the satellite cells, so it's not necessarily an apples-to-apples comparison with the microdystrophin AAVs.

Okay. Okay. So with the cell candidate there, I noticed -- well, I'm just curious to sort of get a sense of what else we sort of might learn from the stealth candidate program, I guess, on the way to the clinic. But if I'm remembering time lines correctly, you could be in the clinic this year. Remind me again, I guess, where you are with that program and sort of how to think about additional data updates from the stealth program.

Yes. No, I'm glad you asked because we're really excited to see what the changes we've made to create the stealth cell are we're going to do and how it's going to perform. So you're right, we filed the IND for the stealth cell program late last year and are optimistic that we'll be able to start dosing patients into this program in the first half of this year. And maybe I can turn this over to our Chief Medical Officer, Chris Heery, to talk a little bit more about what we think we might see and what we think we'll be able to learn from this program.

Yes. Thanks, Matt. The beauty of having worked on a particular target like CD19 for a couple of years is we can narrow in on where the opportunities might be with the particular limits and strengths of a given technology. What we think stealth will do is unlock some of those limits that we've seen. So we presented data in December where we showed that the enhanced lymphodepletion regimen clearly drives high peak CAR T expansion and is tied directly to deep responses, at least one of which was durable. And so as we think about how do we get even further, the next step might be to allow those cells to persist a bit longer. And using these techniques to avoid immune rejection within the cells allows us to do all of that without having to modify lymphodepletion significantly. So we still believe there may be -- it's still possible that there may be a viable path for our first-generation of cells. But if we find that, that is not sufficient to get where we want to go, we think stealth will get us over the edge there. So just as a reminder of what it is and just the time lines, yes, we filed the IND at the end of last year. We expect to be in clinic in the first half of this year. We knock down beta 2 microglobulin line to prevent T cell rejection, if we don't knock it down to 0, we don't knock it out to prevent NK-mediated rejection. And additionally, to prevent NK rejection, we knock in HLA-E. And at least in vitro, that looks very impressive on preventing both T cell and NK rejection. We will know a head-to-head dose level comparison within this year about what the impact of stealth is.

Okay. Okay. All right. All right. I think we'll have to leave it there for time, guys, but I really appreciate you joining us this afternoon. And thanks, everybody, for tuning into the webcast. Have a great afternoon.

Thank you, Eric and everyone. Appreciate it.