...Thank you attending Jefferies Biotech Healthcare Conference. My name is Kelly Xu, one of the biotech analysts here. We are very pleased to have Dr. Harpreet Singh, CEO and the founder from Immatics, joining us for this fireside chat session. Welcome, Harpreet, and maybe to start off, could you give us overview of Immatics unique TCR technology platform, and how does it enable pretty good pipeline?
Well, thank you very much, Kelly. Thank you for having me, and thank you very much, ladies and gentlemen, for your interest. The questions on the platform, so Immatics is a company that's been working on a mass spec-based technology platform to discover novel targets based on peptide-HLA complexes, as well as generating actually unique and very specific TCRs. So the foundation of our platform is a mass spec platform that starts with actually natural native tumor tissue, eluting millions of peptides. Our database is actually unique in the sense that we have more than 500 million peptide sequences in our database. Over time, we've looked more specifically at 2,000 of these targets and finally homed in on 200 targets of interest that are used as a starting point.
This unique database of naturally presented peptides on tumor tissue also covers lots of normal tissue, so allows a differentiation to kind of understand how tumor tissue differs from normal tissue in terms of target presentation. It has a secondary function. It also allows us to actually identify very clean T cell receptors, 'cause we can very early on recognize cross-reactivities of such T cell receptors, and at the end of this process, what we have is what we call a clean pair, attractive targets such as PRIME, which is depicted here on this slide, combined with a TCR that only recognizes peptide and does not cross-react with other peptides.
Okay, terrific. And a few weeks ago, you surprised us with the new data update of IMA203, sorry, Gen One, and PRAME targeting TCR-T in melanoma and the overall data looks very promising on both response rate and the durability. Could you tell us what led you to disclose the data? And it would also be great if you could set a stage for the next phase of a development plan.
So the initial reason to actually create the data cut is that we're in ongoing discussions with FDA in preparation of an upcoming pivotal, i.e., phase II/III trial that we are intending to conduct in melanoma. And so that was data cut for this for the updated discussions with FDA. When we saw the data, we actually thought, "Okay, it's good to put the data out," also to fuel engagement of physicians and KOLs as we are currently rapidly ramping up in terms of execution or site footprint. So we put out the data and, and this has caused actually a lot of enthusiasm among KOLs. We had tons of discussion at ASCO, where we see strong interest from melanoma KOLs, both for cutaneous and uveal melanoma KOLs for this drug.
I think an interesting side effect that we could achieve with this data is that it provided context, 'cause as you're probably aware, a peer of us also released data on a PRAME compound, a bispecific, not a cell therapy, and so allowed the market and investors like you to actually look at, the data in cell therapy versus, sort of our peers' bispecific.
What about the development plan for registrational trial, and also, are you thinking to move to, like, a front line?
So this data actually warrants now to move into the next spec. We'll talk about the data in a moment, and the next step for us, clearly, is to kind of get this now on a fast track to registration. This is ideally achieved in a kind of randomized controlled trial that also has a component or accelerated approval embedded in such randomized controlled trial and follows all the recommendations that FDA has also set up in 2023 for, in the industry guidance for accelerated approvals.
Do we have a timeline on when to start, and what are the remaining questions to answer, from Immatics side?
So the timeline that we've guided towards is still to start the trial in this year. We're expecting actually to be able to start the trial by end of this year. At this point, we are now finalizing these discussions. We had, we have a RMAT designation that's been awarded to us for IMA203 for multiple indications. Think about a RMAT designation as a kind of re-entry pass. You can go to the FDA, as often as you like and kind of have iterative discussions. This is actually very beneficial to have very constructive, dialogue with the FDA and jointly with them, develop the best and fastest, possible concept. And so based on that, we've already aligned of a number of aspects, but there's still some clinical design aspects that are outstanding, that we will solve over summer.
We hear the excitement on the clinical signal so far on 55% OR and more than 10 months of durability of response, and there is speculation that it could be even better than the commercially available TIL therapy. But the question is, why the TCR-T program can actually enable this high efficacy, efficiency, given that it only targets one tumor antigen, but the TIL actually cover more than probably thousands of tumor antigens?
Yes, so this is a function of actually the PRAME target. We see that the PRAME target is actually present on almost all tumor cells. We often find this in 90%-100% homogeneously distributed. In the tumor environment, it's presented at very high copy numbers, so high target density, typical hundreds to thousands of copies, so actually these tumor cells can be very easily engaged by our TCR, which has a very high sensitivity. So those are the hallmarks of a good target. You want a lot of that target around, and you want it on almost every cell, and if that's the case, and this is recognized, then almost every cell is killed.
This may be different to checkpoint inhibitors or TILs, where people think about really, these agents working through neo-antigens that are expressed, in, in the tumor microenvironment and TILs or checkpoint inhibitors, then sort of reactivating T cells that are recognizing such neo-antigens. And then it depends very much on the distribution of these neo-antigens in these different tumor cells, how well the efficacy actually will play out. I mean, maybe speaking about the data that we've just shown, I've put up one slide only. Actually, that's the only slide I will show today with the recent data. We've put out this data just prior ASCO. This is in actually 30 cutaneous, uveal, and mucosal melanoma patients. Majority of these are cutaneous melanoma patients. We've observed a confirmed response rate of 55% in 16 of 29 patients that were efficacy evaluable.
One patient wasn't evaluable because that was still upcoming for the second scan. A disease control rate of 90% in 27 of these 30 patients. We've observed tumor shrinkage in 87% of these patients. The median duration of response is, at the moment, 13.5 months, but very importantly, as you can see, the vast majority of these response are ongoing, so there's an opportunity to further expand that median duration of response. We've also recently shown that most of these response are actually deep responses, deep being defined as responses, sort of deeper than 50%. We've seen in our experience that such deep responses have a very good opportunity and chance to translate into long-term durable response, so that's another factor that matters. The confirmed response rate is actually similar across all melanoma types.
So we have 56% response rate in cutaneous melanoma, 54% in other melanoma types. That includes uveal melanoma and mucosal melanoma. By the way, these rarer melanomas, as you probably know, are much harder to treat, so it's also very encouraging to see, although these patient populations are smaller, that these very, very hard to treat patients actually also have an opportunity to achieve deep responses. And let me say this as a fact, and that's a point because we often have discussions: Why do you do both cell therapy and bispecifics?
We think cell therapy, although more tedious and despite the autologous nature, may have this unique opportunity to actually see deep responses in a majority of patients, and that's a huge opportunity for these patients, not just in terms of it's very important extension of survival, progression-free survival, but also think of quality of life for these patients. When these large tumors that really affect day-to-day life of these patients are significantly reduced by 50%, 60%, 70%, 80%, I mean, that's a relief of pain. That's relief of all kinds of other symptoms that these patients have when they are actually really plagued by these very large cancers.
How would the registrational trial to be designed to include uveal melanoma? Are they going to be in the separate cohorts and also with different regulatory bar?
That's actually still an ongoing point of discussion with the FDA, hence open. Clearly, the focus for us is on cutaneous melanoma. If you think about patient numbers that we want to serve, we're talking about 3,000 A2 positive, BRAF positive patients just in the U.S alone that are suffering from post-checkpoint, so checkpoint, refractory, relapsed, and BRAF/MEK inhibitor, refractory, relapsed patients. That's kind of last line patient population, 3,000 patients compared to 300 uveal melanoma patients. So focus is clearly on cutaneous melanoma. We may actually start with cutaneous melanoma to have like a cleaner, faster, easier design, also quicker discussion to get to accelerated approval. But we are very much focused also on the medical need of rarer cancers.
Now, not something that we would do on its own, but as an add-on to a large prep patient population such as cutaneous melanoma, also from commercial setup and the kind of investment that's required and the kind of site footprint in the commercial manufacturing setup, that is absolutely valuable to add these, these patients and to deliver something that these patients typically do not have. I think the best drug that's currently approved uveal melanoma has recently reported a 5% confirmed response rate, there are upcoming drugs that go up to 20%, 30%, but I'm not aware of any drug that actually can induce more than in more than half of these patients, a deep response that's durable.
Can you also compare and contrast the manufacturing process to TIL in terms of procedural steps and also the turnaround time?
Yeah, first of all, I'm very glad that actually IV cell therapy is approved. This is the first time that actually cell therapies are now made accessible to a broader population. It's an older technology, obviously, a pioneering technology that requires a pretty tedious setup. So the first element actually requires a tumor surgery. You need to access those tumor-infiltrating lymphocytes as a starting point for your manufacturing, and then it's a relatively lengthy manufacturing time of more than 3 weeks to expand these TILs to billions. And then after infusion, also TILs require post-infusion high-dose IL-2, which comes with known pretty severe toxicity. So that actually requires patients to be relatively fit before entering the therapy. So that's quite a tedious approach. It also requires coordination in the hospital between surgeons and oncologists.
In contrast, our IMA203 TCR-T is much simpler. There's no surgery required. There's a simple leukapheresis or apheresis as a starting point, to kind of start manufacturing. Manufacturing takes only 7 days, plus 7 days of release, so it's a much shorter turnaround time, which also has a significant impact on cost of goods. That also matters for commercial viability. And then we do not require any post-infusion high-dose IL-2. We actually only do low-dose IL-2 for 10 days, which can be also easily done. An outpatient setting is not associated with any additional toxicities.
... does the low-dose IL-2 administration cause any ICU stay from your phase I/II experience?
No, not at all. So the few cases of high-grade events that we have seen in patients, we're talking about less than 15% of high-grade CRS and less than 5% of high-grade ICANS, all of them grade 3, all of them resolvable with tocilizumab treatment or other agents and transient. Those are occurring basically as a direct consequence of infusion of the TCR-T product, and it's really a byproduct of the efficacy. Because imagine, I mean, these T cells, the onset to response is very fast, and we see responses almost always happening. Actually, I can show you that in the next slide. Almost always happening within the first 6 weeks. If we look at patients that have visible cutaneous lesions, we see that some of these lesions actually start disappearing within a couple of weeks.
So there's a very fast onset, which indicates that these T cells are acting very fast. They're destroying, particularly in patients with large tumor burden, they're destroying a lot of tumor cells within a very short time, and that is actually then causing inflammation, inflammatory effects, CRS, and associated ICANS.
Okay, terrific. You also have optimized the Gen 2 ongoing, and you said expectation this could be suitable for pursue other tumor indications beyond melanoma. And could you tell us why? What are the product characteristics actually make you to strategically thinking in that way?
Gen 1, IMA203 Gen 1, or let's just call it IMA203, is a product that is solely comprised of functional CD8 cells.
So while there are CD4 cells in the product, they're not functional, so think about this as a CD8 cell-only product, right? That's kind of half of the adaptive immune system. In Gen 2, we co-transduce CD8 alongside, so that's a coreceptor CD8 alongside the T cell receptor. And so now the CD4 cells are also expressing CD8. We're basically converting CD4 cells into CD4/CD8 double positive cells, and now these CD4 cells have the capability of being fully functional. So we are adding dose. The ratio of CD4 versus CD8 cells is 1:1 to 2:1, so we're doubling up to tripling the dose of available transduced T cells. That's obviously beneficial.
But we've also seen that these CD4 cells actually may have a high firepower per cell and may actually sort of then lead into more durable responses over time.
When do we expect the next data update from PRAME TCR-T program?
So we've guided towards a next update for Gen 1 in the H2, and also expect, probably at the same time, to provide an update on Gen 2. Gen 2 is currently at what we call dose level 4A. We are intending to kind of maximize the firepower of Gen 2 by further dose escalation. That's a process that has started, and we expect that dose escalation then to be completed, in the H1 of 2025. So the data set in Gen 2 this year will be still relatively early, focused on dose level 3 and 4A, and still focused very much on melanoma and sarcoma.
At this point, for Gen 2, it's more about developing Gen 2 and maximizing the firepower and showing in melanoma and sarcoma patients in direct comparison to Gen 1, how we can actually further enhance efficacy.
It is an exciting time for another PRIME program in a different modality in the bispecifics format along with the CD3 arm. And at ASCO, we have another PRIME bispecific showing the clinical update in melanoma. And I'm curious, from the molecular design and the preclinical evidence perspective, do you see differentiation from your IMA203 program?
Yes. So first of all, I mean, just to kind of highlight that we are a company that's among the biotech field, pretty unique, that is actually sort of developing both the TCR modalities. So what you see on the left side is autologous TCR-T. We also have allogeneic, but we won't talk about this today, and TCR bispecifics. And the reason we've decided to do that is because we do see, and we do envision principal differentiation between these two different modalities. Obviously, autologous cell therapies are a process that requires leukapheresis lymphodepletion, as we know. So we're making these patients jump through a few hoops, although our hoops are much easier to jump through than TILs or other cell therapies. But in return, these patients deserve a kind of better efficacy profile. Indeed, that's what we're seeing.
55% response rate is actually very encouraging and very promising. Some colleagues call it unprecedented. And this is different from what we envision for bispecifics, which obviously have the vast opportunity and to be deployed as off-the-shelf agents across classical pharma supply chains. But we envision that they may be not as potent as cell therapy, and so we've given guidance on this in the past, very consistently in the past two years, and that's we think about early decision making based on confirmed response rates at RP2D and have defined a lower threshold of 40% for our cell therapies. That's pretty high for last line solid cancers, and 20%, which is more like the industry standard for biologics.
But it's clear that confirmed response do remain a very relevant metric for us in early development. Now to your question, within our bispecifics, we actually also differentiate quite substantially from first-gen bispecifics. So not just first-gen ImmTAC, but also first-gen BiTE like developed by Amgen, in terms of a number of elements, so only utilizing high copy number targets like PRAME, like our MAGE-A4/ A8, which is a completely different one that's been used by others, a different target. Second, by deliberately utilizing a low-affinity T-cell engager against CD3 and TCR, that is actually allows the product to be dosed higher. And thirdly, by employing an extended half-life, so the first-gen BiTEs had half-lives less than 6 hours. 6-8 hours, I think, is the gp100 program Immunocore KIMMTRAK.
I believe the current programs are less than 12-24 hours. So our half-life, preclinically, we've published that, is at around 8-11 days. So think about our bispecifics more like IgG antibodies and also developed that way. And those elements allow us to dose these, to dose our bispecifics much higher and less frequently, while actually having a very strong, steady-state PK in these patients, and those are kind of hallmarks of our second-gen bispecifics.
Just to confirm, the 20% OR you just cited is also the bar you're thinking about for the second line melanoma program for your PRIME bispecific?
So yes, our bar has not changed, also not in light of recent data. We think that for early decision-making, seeing a minimum threshold of responses and confirmed responses at RP2D is absolutely relevant. Now, we will publish data in the H2 of both our bispecifics. IMA401, which is a bispecific against our MAGE-A4 and MAGE-A8 peptide, a different target that's been used by, by others. And IMA402, which is targeting PRAME, that's the same epitope that we're actually also going after with our IMA203 cell therapy. Both programs will be still in dose escalation. In IMA401, we're expecting something like 25+ patients, or at least 25 patients, with a substantial number of patients already at therapeutically interesting doses, but not necessarily already at RP2D.
In IMA402, so our PRAME compound will be 15+ patients, obviously still a substantial number of these patients still in very early doses. So think about both of these releases in conjunction, not necessarily released at the same time, but in conjunction to demonstrate proof of concept for our second-gen enhanced bispecific platform, where it's about safety, PK, but already at this early stage, dose escalation, showing first signs of clinical activity. And let me be very crystal clear: for us, obviously, long-term disease control matters. It matters for the patient. But in early-stage development, for us, the metric to decide whether we move this forward or not are confirmed RECIST responses. That's our orientation, and a 20% bar at RP2D is absolutely, still valid.
Do you plan to also move the bispecific program to frontline in IO combo?
Actually, when you look at the differentiation, that is even the default. So we have so far thought about cell therapy on the left side as something that, we would primarily develop as a monotherapy in a last line setting, like 2L+ setting in cutaneous melanoma. So in post-checkpoint patients and post-platinum-treated patients in ovarian cancer, in post-chemotherapy or other agent patients. 'Cause this is where the medical need is the highest, and such a platform can be actually established best, while also given the potentially different potency of bispecifics, this is better positioned in combination settings, and that means that there is a clear opportunity, and that's the default for moving quickly with our bispecifics into earlier or frontline settings in combination with checkpoint inhibitors and other agents.
Now, with the recent data, as we've just published that in IMA203, with 55% response rate and good durable response, actually, we have been really asked a lot of questions. Why wouldn't we also consider moving cell therapy into the frontline setting? That is now a serious consideration on top of our current consideration 2L plus, and that's something that is also commercially very attractive. I mean, think about 3,000 patients in the 2L plus setting and almost double that patient number in the early line setting.
So this is in terms of market size and, revenue potential, peak sales becomes, even just in melanoma, and remember, PRAME will go beyond melanoma, but even just in melanoma and just in cell therapy, this suddenly becomes commercially a very, very different opportunity than we thought about this a few years ago.
How will you position two PRAME programs in melanoma market? Would you say TCR-T is more for like, for more like refractory settings?
Yes. At this moment, yes, that's the default positioning. That's the label that we're kind of, we'll start with the 2L plus monotherapy label in cutaneous and potentially uveal melanoma patients that have been pretreated already with checkpoints and BRAF MEK inhibitors in the case of cutaneous melanoma patients, while bispecifics would be presumably better positioned in earlier disease in combination.
Great. And you also have another bispecific program targeting MAGE-A4 and A8, and also, first-in-human data expected in H2. This could be overshadowed by the PRAME focus by investors. Could you talk about, like, why to target both MAGE-A4 and A8, given that most of the programs target MAGE-A4? What's the advantage, and what should we expect for the data update in H2?
Yes, I understand the enthusiasm for PRAME. We share that, but MAGE-A4 A8 is a fantastic target. When I say MAGE-A4 A*08, everybody thinks about the protein, but it's not the protein we're targeting. It's a specific peptide from two proteins, MAGE-A4 and MAGE-A8, that is highly expressed or highly presented on the surface of tumor cells in very relevant indications, not just including melanoma, as we have for PRAME, but also, squamous non-small cell lung cancer. Squamous and neck cancer show even higher levels of MAGE-A4 than PRAME, so it makes it a highly exciting target. The presentation level of this of our MAGE-A4 peptide is approximately fivefold higher than the kind of classical, well-used, and well-known MAGE-A4 peptide that's been utilized by others. So there is a huge opportunity with this peptide on top of our PRAME opportunity.
As I mentioned before, there will be more patients with more follow-up on this data, so this will be most likely the first data set that we'll put out, and that will educate us very strongly on the potential of our second gen bispecific platform that we think actually can provide more potency and significantly better patient convenience. Because in contrast to the earlier first line bispecifics, this can be dosed already every other week, and we see potential doses every three and four weeks, and that is a game changer for these patients. If you're a patient, you don't want to go to the hospital every week, and that's a real point for commercial viability that's very relevant for us.
Lastly, maybe we can lay out one more time the key milestones for the next 12 months.
So this will be a very busy H2. We will kind of educate, obviously, the market on the design of our phase II/III randomized controlled trial in at least cutaneous melanoma to achieve a label for our cell therapy, IMA203 Gen 1, and start the trial this year. This is the plan. We will also have additional data in Gen 2, in dose escalation, and we'll have the first data sets on IMA401, which we indicated will be presented at a prominent medical conference in the H2, and IMA402. So our two bispecific compounds against MAGE-A4 and against PRAME. A busy H2. We're very excited. Stay tuned, and thank you very much for your interest.
Terrific. Looking forward. Thank you again.
Thank you, Kelly.