Welcome to the UBS call. Mr. Trung Nguyen, you may begin.
Good morning, all. Welcome back from our short break. My name is Trung Nguyen. I'm the pharma analyst here at UBS. It's my pleasure to welcome Dr. Robert Plenge, who's the Head of Research at Bristol Myers Squibb. Robert, thank you so much for joining us today.
Yes, thanks for having me.
Before I start, I do have a quick disclosure to read out. I am a research analyst. I am required to provide certain disclosures relating to the nature of my own relationship and that of UBS with any of the companies which we express a view on the call today. These disclosures are available at ubs.com/disclosures. Alternatively, please reach out to myself, and I can provide you them after the call. Also, for investors who have a question, please use the Raise Your Hand feature. We do have a moderator in place, so we can try to get your questions asked. With that said, let's get into it. So we are delighted to have Bristol to talk on this subject. You are one of the leading companies in the protein degradation space.
I think the best place to start is, Robert, if you can give us a quick background into yourself, you know, your role at Bristol, how you got into that role, and then perhaps summarize yours and Bristol's enthusiasm for the protein degradation space.
Yep, certainly. So, I've been at Celgene/BMS for seven years. So I joined Celgene in 2017, and then BMS by way of the acquisition. So, I've been involved in targeted protein degradation during that duration. I lead the research organization. We work very closely with the development organization and other parts of the company to deliver on our pipeline. And that piece is gonna become very important as we talk about targeted protein degradation, because I think a competitive advantage for us is how all of these pieces really fit together: knowledge of targets, knowledge of molecules, how to do clinical development, how to manufacture, and ultimately, how to get drugs to patients.
One other just kind of disclaimer on my side. I will use 'we' a lot in this discussion, but needless to say, there are many, many people who contribute to our knowledge and excitement around targeted protein degradation, so it's really a collective team effort. So to your question, I think, you know, why are we excited about targeted protein degradation, and why do we believe we're one of the leading companies? You know, I think this is a field which has really matured a lot over the last several decades, and we're really proud that many of us, including our scientists at BMS, have contributed to seminal discoveries.
So how thalidomide works, for example, binding to cereblon, the mechanisms of action of two approved medicines, Revlimid and, and Pomalyst. But I think what really makes us excited, and really speaks to our leadership, is simply the number of clinical programs that we have ongoing. So we have eleven programs, that are in clinical development, including four that are either in registrational trials today or will be in registrational trials, in the next, in the next year. These include three distinct modalities, which we can actually talk more about. So molecular glues, CELMoDs, for example, as well as, a ligand-directed degraders, LDDs, some call these PROTACs, and then finally, degrader antibody conjugates or, or DACs.
And then, you know, finally, I think just our commitment to the, to the broader space of targeted protein degradation with regards to, to next-generation mechanisms. So we can talk about all of these in more detail, and I'm sure that we'll get to each of them.
Excellent. So yeah, let's talk about, and perhaps a bit more generally, your protein degradation platform. As you say, you and Celgene, you're certainly considered one of the pioneers of the space. So perhaps can you give us a brief journey about how you guys have built out these platforms, and specifically, how Bristol's platform may differentiate from others that are in the space today?
So let me start with just the evolution from Revlimid and Pomalyst, our IMiDs, to what we could now call our CELMoDs. And so initially, the IMiDs were actually discovered more empirically. And since that time, we've continued to learn about how to make these molecules. And so now they're far more rationally and intentionally discovered. In some cases, we use phenotypic screens. In other cases, we use binding to individual targets. So this has been a very long journey, and we've learned a lot along the way. Let me just give a couple of specific examples around our platform. The first is how we've actually expanded our CELMoD library over time.
So initially we had, you know, several thousand CELMoD molecules in our library. And based upon all of our understanding, we've really been able to expand that library manyfold. We know that these molecules bind to cereblon, and then they recruit neo-substrates, so transcription factors, Aiolos and Ikaros, for example, that get ubiquitinated and tagged for degradation. With our library expansion, we've been able to profile representative molecules across the entire human proteome. And what we begin to see are novel proteins that did get degraded. And with these novel, we call these neo-substrates, with these novel neo-substrates, we're beginning to learn kinda new rules about how these ternary complexes are formed, how cereblon recruits these neo-substrates in the presence of these CELMoDs.
We're beginning to discover entirely novel targets and novel motifs, and these are now entering clinical development. So I think, you know, the first point to remember is just how we've expanded our CELMoD library based upon knowledge of the molecular mechanism of action. The second component of this is how we've extended our knowledge of CELMoDs into two other approaches to targeted protein degradation. Ligand-directed degraders, again, we call these LDDs. Others will refer to these as PROTACs, as well as degrader- antibody conjugates, and specifically for us, CELMoD degrader antibody conjugates.
And the reason that this is part of our platform, and I think that we believe that we're differentiating, is that we can really understand aspects of cereblon molecular glues and how to hijack this ternary complex mechanism to design these LDDs or heterobifunctional molecules. We know how to think about optimal linker length, chemical flexibility of those linkers, for example, different features of neo-substrate degradation, the kinetics of neo-substrate degradation. So there's really a lot that goes in to designing these LDDs. A third component, and I think this is probably underappreciated, is just our extensive collaboration network, not just within academia, but also within industry. And this allows us to think about novel targets, think about novel molecular glues.
We have partnerships to use artificial intelligence and machine learning, to really help with our capabilities. And so we're able to put all of this together, our knowledge of the clinical programs, manufacturing, knowledge of CELMoDs, LDDs, degraders, antibody conjugates, and then our partner network to really, I think, be, I think, leaders in the field of targeted protein degradation.
Awesome. That's a great summary, and that leads me nicely onto the next question. So you've just talked about your work with the molecular glues, the, you know, heterobifunctional degraders, or the LDDs, the ADCs. Perhaps can you talk about some of the programs under each of those platforms? Which ones are you most excited for, and where should we have the next set of updates in the next sort of 12-18 months?
Yep. I'll start with the CELMoD program. So we have two myeloma programs that are in registrational studies today, iberdomide and mezigdomide, and then a third program, CELMoD program in registrational studies for lymphoma, and that's golcadomide. Just to touch briefly on iberdomide and mezigdomide, we're really thinking about how to position these in the broader landscape of multiple myeloma. And by the way, I think this is another competitive advantage for us because we know not just about the biology of these individual molecules, but we understand the broader landscape of many of these diseases, including multiple myeloma for mechanisms beyond protein degradation.
But for iberdomide and mezigdomide, iberdomide we're positioning in earlier lines of therapy in multiple myeloma, as well as in the maintenance setting. And this is important because tolerability in this subset of patients is very important. Mezigdomide is really interesting because it's our most potent CELMoD, and it was very intentionally designed to selectively and potently degrade two transcription factors, Ikaros and Aiolos. And this degradation was optimized even in cell lines that are resistant to our IMiDs, so Revlimid and Pomalyst. Mezigdomide also combines very well with other proteasome inhibitors and other mechanisms, and so we're positioning mezigdomide in more advanced and harder to treat patient segments.
Just quickly on mezigdomide, just to say a few words about it, and then I'll pause and be happy to kind of take your direction, Trung, on where to go from here. But, you know, mezigdomide was very interesting because one of the things that we learned from our IMiDs is about how cereblon actually targets and then degrades these target proteins like Ikaros and Aiolos. And what we learned from mezigdomide is that mezigdomide will actually lock cereblon in the closed state, which is the active state. And this does this much more effectively than the IMiDs, and it's one of the features that allows us to have far more potent and selective degradation of Aiolos and Ikaros.
I think the phase I, II data that we published last year in The New England Journal of Medicine, both for the doublets and the triplets, look very promising in terms of overall response rates, and duration of disease and progression-free survival. We expect data on both iberdomide and mezigdomide in 2026, so in the next couple of years, for both of those registrational studies.
... Excellent. And then within that toolkit of the other types of protein degradation that you have on board, just do you think that people need to have a broad portfolio? So do you need to have a full toolkit here to see the greatest potential for each of these platforms?
I mean, we believe it, it helps tremendously. I mean, there are so many examples where we learn across our, our various programs. This includes knowledge around kind of translational data sets, knowledge about how to manage some of the side effects of, of these, these drugs. Neutropenia is, is one example of that, how to combine them. And I think, also about how to manufacture and make them, how to have the right, formulation, and then how all of this actually feeds back onto our preclinical programs to design, the next generation of molecules. So, you know, we believe that having this full clinical development portfolio and a, and a very, I think, robust, toolkit is, is a very, very big competitive advantage for us.
Okay, excellent. And then, finally, just more on the more broader questions before we dig into some of the CELMoD stuff and, some of your other areas. If there was a novel target that you thought would be a great, it would be great to suit a degrader target or a drug, you know, how do you go about making a decision on, you know, how would you use this target, whether it's a, a CELMoD , an LDD, an ADC degrader? Like, what goes into that decision-making?
Yeah. So let me just provide a very broad framework for how we think about programs even beyond targeted protein degradation. So, you know, we start with a deep understanding of causal human biology around the target of interest. So in the case of targeted protein degradation, they're very high-value targets, like the androgen receptor, BCL-6, a number of others as well, and we can talk about those. But it's really more than just identifying a target. It's about how we will design a molecule to therapeutically perturb that target. So we call this matching a modality to a molecular mechanism of action, or matching modality to mechanism for short.
And then finally, about how to actually bridge from research into development to actually show a clinical proof of concept, because in early development, there are obviously key inflection points which tell you your drug is working and behaving the way that you think that it should. So that's a broad framework for how we think about all of our programs. And then in the case of a target that's suitable for targeted protein degradation, you know, in some cases, there are functions of proteins that go beyond enzymatic activity. So some proteins may have scaffolding function, for example, and you want to degrade both the enzymatic activity or get rid of the enzymatic activity in addition to the scaffolding function. There are other targets that just you can't modulate with traditional enzymatic inhibitors.
So for example, transcription factors are a great class because, you know, these are proteins that bind to DNA and big complexes, and that is a very difficult process to interrupt. But if you can degrade those targets, we believe that that will be a beneficial approach for patients. So BCL-6 LDD is an example of that. And then finally, there are certain targets where there may be enzymatic inhibitors, but resistance occurs over time and we can actually think about how to make molecules against different parts of the protein to overcome those patterns of resistance.
So, the broad framework is causing biology, matching modality to mechanism, path to clinical proof of concept, and in particular, for the matching modality to mechanism segment of that framework, we think about ways in which targeted protein degradation is the ideal method of perturbing a target of interest.
Okay, excellent. So let's touch upon some of the leading assets you have in development, the CELMoDs. So, iberdomide and mezigdomide, you touched on the mechanism of binding against cereblon and the degradation of, E3 ubiquitin ligase. Revlimid also acts on the same sort of pathway, but it modulates this activity. These two drugs, degrade the protein versus this modulation. So could you talk to how this differentiation could improve the efficacy beyond Revlimid here, or, you know, overcome that resistance which you just touched upon?
Yep. So maybe just to clarify, so the way that all these CELMoDs work is they're molecular glues. They bind to cereblon, they then recruit neo-substrates. The specificity and the efficiency and depth of degradation of those neo-substrates is really what delivers a clinical activity. So, Revlimid, Pomalyst, iberdomide, mezigdomide, golcadomide, all work in that same broad kind of molecular glue fashion. But the differentiation really comes from a few different features. In the case of iberdomide and mezigdomide, it's the ability to lock cereblon into this closed conformation, and that's the active conformation. And so Revlimid, for example, locks cereblon in this active conformation in about 20% of molecules.
In contrast, mezigdomide and golcadomide are closer to 100%, and that allows much deeper and more efficient degradation of neo-substrates like Ikaros and Aiolos. I think there are other features that go beyond kind of the IMiD class for CELMoDs, and that include other features around, you know, how a drug can actually stay and penetrate... how it can penetrate into a tumor, how it can stay into a tumor, and it can have a prolonged PD effect. I think golcadomide is a good example of that in the lymphoma. And then I think a third feature is how we can actually dial up or down selectivity. So there are some molecules which we think could actually add to clinical benefit.
Some neo-substrates that are the targets of degradation that can actually add to clinical benefit, and then others can actually add to toxicity. We can empirically dial these up or down based upon our knowledge of the mechanism of action. Again, this kind of comes back to this broader theme of because we have so many programs in clinical development, and we can learn not only from our translational studies, but also in research across a variety of animal models, that I think we have a pretty good handle on how to dial up and down some of the target degradation against these neo-substrates in a way that again creates a real competitive advantage for us.
... And, can you perhaps talk about some of the data that we've seen so far, some of this early data in efficacy and also safety, for both of these products? And then how translatable do you think that can be in a more registrational trial?
Yeah. So, I, you know, I think for iberdomide and mezigdomide, again, for positioning these for earlier lines and or in multiple myeloma. Iberdomide we're positioning for earlier lines and in the maintenance setting, and then mezigdomide for later lines and more resistant refractory patient populations. You know, one of the things we can actually see from our own translational data is not just the effect on degradation, but the other component of these drugs, which is not just direct tumor cytotoxic antitumor activity, but also the immunomodulatory features of these medicines. So, the effect on T-cell exhaustion, for example, or the effect on NK cell mediated killing.
So those are a few examples about how, from a translational perspective, we see, you know, differential profiles. In terms of the clinical activity, and I think the real key for us is our ongoing registrational studies, where we'll see more information. But the earlier data that we have from phase I and phase II, in terms of both efficacy and a toxicity profile, is very favorable for multiple myeloma for both iberdomide and mezigdomide.
And you touched perhaps on this a bit earlier, when you were talking about combinations, but how do you think about adding dex to these compounds? Is dex enhancing of the benefit and tolerability of these studies? Because you've done some studies with mezi, without dex, if that's right, and we haven't seen that data yet.
Yeah. In general, I think the addition of dexamethasone, a steroid, to IMiDs and CELMoDs, is seen as clinically beneficial. This includes not just on the efficacy side, but also on the tolerability side. In terms of efficacy, there is improved cell-autonomous killing when you add dexamethasone to CELMoDs. We have preclinical data that we've actually published on that shows upregulation of various apoptotic mediators of cell killing. It adds to the efficacy of CELMoDs. From a tolerability perspective, you know, we think that dexamethasone can also help with one particular aspect, neutropenia.
Although we don't fully know all the details of the mechanism of action, there's probably a component of it which has to do with demargination, but there are other features as well. We have done some exploratory studies on the neutropenia side with some of our reagents, including the human cereblon and mouse model, and we're working through some of the specifics in terms of the mechanism. But in general, dexamethasone is thought to be beneficial for both killing and for tolerability of these agents.
Yeah, it does seem that neutropenia is on target tox for these agents. If we can move on to your other CELMoD, golcadomide. You presented some data at ASH last year. Just how differentiated is this CELMoD? What's interesting to us here is it's in the treatment of non-Hodgkin lymphoma, so perhaps what's your thesis in developing it for something like DLBCL, specifically?
Well, so, you know, you asked at the beginning, you know, "What are, you know, some of the favorite, favorite programs?" And I always get in trouble if I pick just one, because depending upon who you ask, you know, everyone has their favorite program. And I know that there are some members of our team where golcadomide is absolutely, the favorite. One interesting feature of this is how it was optimized specifically for non-Hodgkin lymphoma. So again, kind of building on this intentional and rational design on the next generation of molecules.
Here, what was a little bit different about golcadomide was to focus on other properties beyond just the degradation of Ikaros and Aiolos, and specifically how these medicines penetrate into bulky tumors, such as lymphoma, and how to get distributed, how they stay in the tumor and are active for a long period of time. That's one of the features of golcadomide that I think is particularly unique, and we're beginning to see that play out in clinical development. We did present some data at EHA just a few weeks ago that provides some updated information on the program.
And this includes rationalization for golcadomide from our preclinical models, but I think more importantly, how we're beginning to see this drug work in high-risk segments of DLBCL and follicular lymphoma. And this is really a target for us, is to go after first-line DLBCL in high-risk patients, and it's about, you know, 30%-40% of patients fall into this high-risk category. What we presented was evidence that showed a dose response between the two doses of golcadomide in this patient population. We showed data that demonstrates improvement across cell of origin in DLBCL, and also we provided we showed data around the persistence and durability of MRD negativity.
So I think multiple lines of evidence are making us very excited about golcadomide in first-line DLBCL, in particular, in high-risk patients. And as I mentioned, our phase III study is ongoing, and we expect to see data in the next few years.
... Okay, awesome. And, just into the last five minutes here, I'd love to talk about your ligand-directed degraders, so the LDDs, and also, the CELMoD ADCs. If we start with the LDDs, can you compare and contrast how those, how CELMoDs work versus LDDs, and then your excitement for the BCL-6 LDD program in lymphoma?
Yep. So, CELMoDs work as molecular glue, so it's a single molecule, tends to be smaller, binds to cereblon, and then that recruits these neo-substrates like Ikaros and Aiolos to this, this ternary complex. LDDs, again, people refer to these as PROTACs. They're heterobifunctional molecules, so one end will bind to cereblon or another E3 ligase. There's a short linker, and then another part of the molecule, the, the, this is the heterobifunctional piece of it, will bind to a target of, of interest. And so, we can actually bind targets like BCL-6, I'll talk briefly about that, but also targets such as the androgen receptor.
I do want to make sure I emphasize that, we have, I think, very encouraging data for our androgen receptor LDD, that we expect to advance into phase III registrational trials within the next year. But specifically for BCL-6, this is a target, which is an incredibly high-value target in lymphoma. It regulates germinal center dynamics and B cells, both for activation and maturation. But as a transcription factor, as I mentioned earlier, it's really, really hard to drug. So what we've done is we've made this heterobifunctional molecule. One end binds to cereblon, the other end binds to BCL-6, and we can actually show in our preclinical models that we can actually degrade BCL-6 and also have a clinical benefit in these preclinical models.
But what's really exciting is that we now have an IND that has cleared. We are now in human dose escalation studies, and we should have data to report on this in the near-very near future to establish this relationship between BCL-6 degradation, and the impact on clinical activity.
Okay, that's useful to know. And then finally, just the last question on the CELMoD ADCs or DACs. We are a few ways away from the clinic here, I think. You summarized the concepts at the start, but, you know, can you perhaps talk about your excitement here, and then how difficult is this manufacturing process of a CELMoD ADC versus just, you know, normal ADCs? Is this difficult to manufacture?
Yep, yep, yep. So just clarification, we're actually in the clinic today with our lead program, which is a CD33 GSPT1 antibody drug conjugate or degrader antibody conjugate. And I think what makes us really excited about this particular molecule is that we know that each component of this is clinically active in patients with AML, but each individual component has limitations when given systemically. So GSPT1. There's a CELMoD that degrades GSPT1. It's clinically active. We know that there's benefit, but there are side effects such as hypocalcemia when it goes throughout the body. Similarly, CD33, the other component of this, clinically active. There was a drug called Mylotarg that was approved way back in 2000.
But there are also side effects, including veno-occlusive disease , which was thought to be caused by the drug that was conjugated to it. So we then asked ourselves, you know, "Can we combine both of these together to actually have the clinical benefit of each while limiting some of the adverse side effects?" And that's the concept of a degrader antibody conjugate. And as I mentioned, we are in clinical development with this program and phase I dose escalation now. So we're really excited about showing data on that, you know, in the coming year or two.
Excellent. So we're heading up to the half-hour mark here. Ellie has Triana Biomedicines on at 11:30. Dr. Plenge, that was timed to perfection. Thank you very much for your time and taking us through Bristol's developments here in the space. For those that listened in, we hope that was interesting for you, and we hope you enjoy the rest of the day. Have a wonderful day, everyone.
Thank you.
Thanks, Dr. Plenge. Thank you for joining today's-