Well, thank you and good morning. Thank you for joining the Guggenheim Securities Healthcare team at our sixth annual conference. I am Debjit, and joining us from Centessa Pharmaceuticals is Saurabh Saha, President and CEO. Thank you for your time, Saurabh.
Thank you, Debjit.
Well, from a company perspective, given your asset-centric business model, I think it'll be helpful for the audience just to get a quick overview of where you are.
Yeah, absolutely. So thank you for the invitation this year to the conference. Always a pleasure. So 2023 was a fantastic year for Centessa. As many of you know, we nominated our candidate for narcolepsy, orexin agonist, ORX750, and we've been moving forward with our SerpinPC program for hemophilia, having dosed our first patients in both pivotal studies, PRESent-2 and PRESent-3. And we moved our LockBody molecule, the first from the platform, LB101, into the clinic. So 2023 was a great year for execution, and it set ourselves up for a very catalyst-rich year in 2024.
So, I'm sorry, just wanted to let that quieten down. All right. Let's start with the orexin program, the ORX750. We've seen a bunch of issues with the competitive programs, especially if you think about liver safety, cardiac, and sleep disturbances, ocular issues. So how has Centessa designed a 750 to circumvent these issues?
Yeah. So we're very excited about ORX750. The team has done a fantastic job in selecting a molecule that we believe is best-in-class. And when we say best-in-class, there are a number of pharmaceutical properties that we've optimized against, and one, and first and foremost, is PK and minimizing the Cmax to trough ratio, ensuring that we have good bioavailability, great brain penetration, and a metabolite profile that hopefully portends to having a potentially safe profile. So I think it's important to remind folks that we generated the chemistry for ORX750 based on very strong structural biology.
Our collaboration with Sosei Heptares has allowed us to work on a proprietary crystal structure of the orexin-2 receptor, a structure that was mutated in its specific residues, which allowed us to have an altered conformation to be able to screen small molecule agonists of the orexin receptor in a very novel way, which yielded very interesting chemistry for us to optimize from a med chem standpoint, to create what we believe are very, very compelling orexin agonists.
So based on the totality of how we generated the chemistry, how we went through a very rigid screening funnel to arrive at ORX750, we feel we have a compound that can be very competitive. And even beyond that, we understood that we needed to have a very potentially safe compound, for potential indication expansion beyond just narcolepsy, but into common disorders with excessive daytime sleepiness as a component. So we tested our molecule, in a number of different safety panels, like SafetyScreen44.
We screened against a number of GPCRs, 168 or so, across 60 different families, ensuring that we have exquisite selectivity, of our molecule. And importantly, our team was able to achieve a goal that we had set out at the very beginning, is to generate a small molecule orexin agonist that phenocopies the orexin-A peptide that's lost in narcolepsy patients, and to do so, having a potency that is nearly equivalent to that orexin-A peptide. So our potency of the molecule is 110 picomolar, and the peptide itself is 30 picomolar. So the team did a very good job of hitting that benchmark.
Awesome. Can we talk a little bit about the insomnia? Is that an on-target effect, and can that sort of be navigated either by PK or it doesn't really matter for the long term?
Yeah, so insomnia could be, and this is purely, our guess at this point, is insomnia could be due to one of two factors. One is having, the PK not optimized perfectly at this point in terms of having a duration that's lasting longer than probably you'd like to have, into the night, which, which leads to insomnia. So I think a better job needs to be done, on PK, and we've paid very close attention to that, as we, are going to be developing ORX750. The second possibility is that the data that's been reported by others, in narcolepsy type 1 or NT2 patients, is that they haven't seen an orexin peptide for decades.
Now with the small molecule agonist, that pathway, that GPCR, that orexin receptor 2 is being triggered for the very first time, and there could be a hypersensitivity, if you will, to orexin signaling, which over time I suspect is gonna equilibrate, and perhaps that may lead to attenuation of the insomnia effects. Now, given both of those, what we've done at risk with ORX750 is to make sure that we're ready for all those possibilities.
And with regards to that, we've been working on extensive formulation work so that if we need different formulations to make sure we have that kinda ideal PK profile, which we hope is once daily to keep patients awake during the day, but also ensure that they have adequate nighttime sleepiness and reduced disturbed nighttime sleepiness, if you will. That is, that really is the goal, and we've done everything we can at this point to ensure that we have all those possibilities, all those levers at our disposal.
Great. Biology is very well established in this disease. Could you sort of talk to correlation between the animal models and translation into healthy volunteers and how that could impact NT1 patients?
Yeah, it's an excellent question, and it's one that I think we're fortunate in the orexin field, the agonist field, to benefit from the high degree of translatability from in vitro potency to animal models, whether it's in narcolepsy Type 1 models in mice, all the way to monkeys, and doing EEG studies in monkeys, and then seeing that correlate into human data, which has been shown in publications by our competitors in really beautiful translational work. And it's one of the rare spaces where the data just you can connect the dots from in vitro, in vivo, into human data.
What we've shown at World Sleep last year is that data, basically showing that we have a very potent molecule, a molecule that's very brain penetrant, basically checking all the boxes from an in vitro standpoint, but then showing in the narcolepsy mouse models, two different NT1- like models, where we were able to, in a dose-responsive manner, show that we're able to keep the mice awake, reduce cataplexy.
And then in our we like to call our healthy volunteer or wild-type mice, we were able to keep them awake as well, demonstrating that even in a background of normal orexin levels, our molecule works in a dose-responsive way in keeping those mice awake as well. Then, as we've alluded to last year, in our monkey studies, we've seen very good properties of our molecule, and basically connecting the dots, as we would expect. So, very exciting to be working in a space where the animal models are highly predictive of what you might expect to see in patients.
Perfect. When you think about the market opportunity here, this is a question we very often get: How many players can it sustain, or does one have to be really clean to win, take it all kind of thing?
Yeah, I think that's a really good question. The way we like to look at it is, this is a very, very, potentially very large opportunity and space. I'd like to, probably the analogy that makes the most sense is, just like the GLP started in diabetes and now ending up in obesity, I think the orexin agonists will perhaps start in narcolepsy and then end up probably treating excessive daytime sleepiness across a number of indications, which obviously affect a large portion of the population.
What's important to note here is that we don't look at where we are necessarily in terms of relative to the competitors, in terms of order of entry to market. What we look at is, do we have what we believe today is the best-in-class molecule, based on everything that's been seen publicly, from our competitors? And are we well-positioned to bring that molecule into the clinic, one, with a creative development plan, so that we can accelerate faster, perhaps, than others
have in the past in developing this molecule, and getting it to patients and ultimately to market as quickly as possible, let's say, for example, in narcolepsy, indications, whether it's NT1, with or without cataplexy. And then second, it's important to note that there's, as you mentioned, a large space out there of potential indication opportunity, beyond narcolepsy, excessive daytime sleepiness, and common disorders, which allow, us to potentially be first, to get a label for.
So we're excited because there's a lot of opportunity in this space to play. There's very few competitors at this point. The chemistry barrier to being able to play in this space with an orexin agonist is extremely hard. It's taken companies decades to be able to crack this chemistry. We're fortunate to be one of those companies. And I do think that there's, given the market size, there's plenty of opportunity for multiple players to win here in different indications and even within the same indication space.
Got it. And we expect proof of concept for ORX750 back end of this year. Could you sort of frame the expectations from the healthy volunteer data?
Yeah. So, what we've stated is that we anticipate healthy volunteer proof-of-concept data with ORX750 in 2024. And it's important to note why we say specifically healthy volunteer. As many of you probably know, being able to show that an orexin agonist works in the background of normal orexin peptide levels in healthy volunteers is the high bar. It's the bar in which you need the most exposure of an orexin agonist to be able to keep those subjects awake.
And if you're able to achieve that proof of concept, then it portends very well that in a narcolepsy population, where they have much reduced or absent orexin levels, that you'll be able to keep them awake as well, at possibly 2- to threefold lower doses or exposures. So if you hit that high bar in healthy volunteers, you basically are potentially clearing the way for not only the narcolepsy indication space but also those who have excessive daytime sleepiness as a comorbidity with many of the common disorders that folks suffer from.
Awesome. So let's move to SerpinPC. That's your most advanced asset in the clinic. Lots of differentiators here. Where do you think SerpinPC lands?... let's say, in two years from now?
Yeah. So SerpinPC, as time goes on, is really looking more and more attractive, from internally from our just our molecule and the novel, novel pharmacology of this potential first-in-class biologic brings, but also from the external environment, the competitive space. There's a lot of tailwinds that are helping us. And so, just to remind folks that SerpinPC is a subcutaneously delivered biologic that has already demonstrated in phase 2 studies to reduce all bleeds in hemophilia patients, both A and B, actually, by 96%, our latest data that we showed at ASH.
But more importantly, in this space, safety is paramount to treating physicians and the patients themselves. And so far, in 60+ patient years, we've shown no evidence of thrombosis. But even more, you know, precision around safety comes with our D-dimer data, where we haven't seen any chronic elevations that are unexplained in these patients that have exceeded the upper limit of normal. I think that's in sharp contrast to the other agents that are delivered subcutaneously in this field, where I think they're trying to thread a very difficult needle here of achieving a therapeutic index.
We feel very confident with our therapeutic index, given our D-dimer data to date. Even with increasing exposure levels, we don't see any sustained elevations in D-dimer. And as you probably know, in our dose justification part of part one of our PRESent-2 study, we're actually going even twice as high than we've ever gone in our phase two studies with dose. So that should give folks an idea of how we view safety as something that's intrinsically built into the mechanism of this molecule.
If I'm not mistaken, you're also enrolling Hem A patients, although you're specifically seeking approval in Hem B?
That's right. That's right. We anticipate half or over half of our subjects that we're treating in PRESent-2 to be Hem A patients. I think that's really by design to give us enough patients for an adequate and robust safety database because that population is more prevalent.
But what it also does is it gives the regulatory bodies, as well as treating physicians, an opportunity to look at SerpinPC as not just a hemophilia B drug, but also potential pan- hemophilia drug that is not only designed to treat Hem B, but Hem A as well, and also rare bleeding disorders like von Willebrand disease and others, which we haven't spoken much about. But this is a mechanism that really has the opportunity to have broad indication potential across bleeding disorders.
Got it. And, for SerpinPC to become, let's say, the Hemlibra of Hem B, what do you need to demonstrate?
Yeah, I think we've already, you know, we're on that journey and have demonstrated three things in our phase two studies now, over three years in dosing patients, is that, one, this could be a convenient therapy, like the Hemlibra of Hem B, if you will, not our line, but others have mentioned this, because it is delivered subcutaneously, which is a game changer, in a space where patients get IV factor. And so we know from Hemlibra's precedents that bringing a sub-Q therapy really is something that patients really desire, to change from their standard of care that's given IV. So o ne, sub-Q is going to be very important.
Two, safety is paramount in a space where, in the past, this is a very risk-averse population, understandably so, as many of those patients have gotten hep C or HIV over the years through transfusions. They really are sensitive to safety, and we've already shown in 60+ patient years, as I've mentioned, no evidence of thrombosis, which is in sharp contrast to our competitors who have sub-Q therapies that are in development.
And then third, we've, you know, so far hit that efficacy bar with a 96% reduction, not only in all-bleed rate, but also if you look at our target Joints, those are joints that bleed more than three times in a six-month period. Really, the most difficult to treat joints, because they are damaged, and they're often called as maybe false bleeds.
If you're able to reduce those, then you really are showing the power, the robustness of this mechanism of this drug, and we reduce target Joints in this population that we've been treating now for 30 years by 94%. I think that's a testament that shows how powerful this mechanism can be, all the while achieving a therapeutic index, which frankly, for hemophilia drugs, is very difficult to achieve.
What's the internal threshold or bar to advance SerpinPC into Hem A registration studies? Or is that something you would wait for some sort of a licensing or partnership?
Yeah. So while we don't comment on our regulatory discussions, our plan is, and how we've designed the PRESent-2 study, is very comparable to how others have designed their pan-hemophilia studies to get a broader label. So we're roughly about 120 patients, both Hem A and Hem B, for our PRESent-2 study, which is in line with what others have done in the space. So our intention is to take a subset of that data, perhaps this year, talk to the regulatory agencies and understand what do we need, if any, to seek a broader label.
Got it. Do you think Serpin... sorry, PRESent-3 could read out before PRESent-2, and that could trigger, BLA submission?
Yeah. So you know, I have our teams, it's like a competition here. Whichever one reads out first is gonna be exciting. But PRESent-3, just to remind folks, is Hem B with inhibitors, which is a very rare population of patients. If we're able to enroll those 12 or so faster than we enroll the 120 in PRESent-2, then PRESent-3 will win out or vice versa. So we'll see as time progresses.
... Given that patients with inhibitors cost a lot more money than patients without, how are you thinking about pricing? Because it's unique to have one single drug working across not just hemophilias, but also with patients with inhibitors.
Yeah, so we don't comment on, on pricing, but what I can say is that what's unusual as we've learned more and more about the hemophilia commercial space, both ex-US and US, is that, you know, pricing is, is very robust, both in the US and surprisingly, ex-US. So co-pricing is very, you know, competitive across the world, in a way that where reimbursement is, is pretty high, for, for drugs like this, as we've seen, with Hemlibra and others. So, we feel pretty good that we're gonna be able to price, at, at a very competitive in a very competitive way. But right now, we, we haven't commented on pricing between the two.
Got it. So now we've gone from narcolepsy to hemophilia, and maybe we can spend the last few minutes talking about oncology. LockBody, how did that come about? And maybe just a quick explanation of what LockBody really is.
Yeah. So LockBody, basically, it's a very simple technology platform where we have a traditional IgG1 monoclonal antibody, where you have your Fab and your Fc, just like any other antibody. And all we're doing is linking another pair of Fabs to the Fabs of that monoclonal antibody. The way we link those Fabs together is through a linker that's not engineered. It's a linker that's found naturally between the Fab and the Fc of a typical IgG1 that's circulating within our system.
And what we've learned over decades of empirical and experimental evidence is that that linker between a Fab and an Fc circulating in our IgG1 antibodies that we all have is highly labile in terms of breaking apart in areas of inflammation, whether it's in tumors or inflamed joints, and we see this with antibodies that are dosed for those indications, where they do fall apart and are rendered ineffective. So we borrowed a bit of well-trodden kind of experimental empirical data.
We've taken that linker, we've put that linker between two Fabs. The Fabs, the outer Fabs are tumor targeting. So in our case, it's PD-L1. So it's a PD-L1 antibody, essentially, that we're dosing. It's sterically blocking the Fabs that they're linked to, with this hinge that I'm referring to. And then that second set of Fab is a highly potent immune effector.
In the case of LB101, which is in the clinic, that's CD47, which is a immune effector that you don't want circulating exposed in the periphery, because it has demonstrated to cause anemia and thrombocytopenia, but it's sterically blocked by our PD-L1. We can also put in molecules like CD3 or CD28 that are linked to a PD-L1 and keep that effectively silent in the periphery until they get to the tumors.
When they get to the tumors, that hinge is naturally degraded as we've seen the normal Fab Fc hinge being degraded with incumbent antibodies that are given to patients for various diseases. We've taken advantage of the Fab being tumor targeting, concentrating the antibody in the tumor, and then the hinge breaking apart and releasing that immune effector at high concentrations in the tumor versus in the periphery, giving us potentially a wide therapeutic index. That's essentially the premise of LockBody.
Got it. If I understand it correctly, there are four hot tumors that you're looking at, along with two relatively cold tumors?
That's correct. Yes.
In the post PD-1 progression or PD-L1 progression setting, what's your threshold to move the program forward?
Yeah. So, for the non-immunogenic tumors, like ovarian and colon, for example, I think seeing, you know, double-digit response rates, low double-digit response rates, would be very exciting because I think traditionally, as many of you know, PD-1, PD-L1, do not really work in that tumor setting. And if we achieve the low double-digit type of response rate, in a monotherapy setting, I think that could be very exciting.
In the more immunogenic tumors, like head and neck and non-small cell lung cancer, there, we anticipate, you know, trying to see some more meaningful double-digit, response rates. We know that there's reinduction with PD-1 and PD-L1, from past experience, and we have to, overcome that hurdle, if you will, in terms of rate. And then, you know, if we see a meaningful response beyond that, we'll be very excited to move forward.
The reinduction rate's roughly in the single digits or?
Yeah, I think, you know, it's... There isn't any real hard evidence out there, but what the number that we like to use is somewhere between 5%-10%.
Got it. And, from an enrollment perspective, how is the study tracking, and, do you think you'll have proof of concept in 2024?
Yeah. So enrollment is going well. Again, proud of the team for doing a fantastic job. Among the 5,000 or so oncology trials out there, we've been progressing very fast and have gotten a lot of excitement from our PIs that are part of the study. So our plan is to announce a data set when we have, it's largely event-driven, when we have enough responses in the numerator and enough patients treated in the denominator, so that the world can basically make a very solid assessment whether or not the drug is effective and safe.
Got it. And, let's finish with the next program or the next molecule that's likely to go in the clinic, the CD3.
Yeah. So, LB206 is a PD-L1 CD3, so the tumor targeting Fab is PD-L1, and it's linked to a CD3 Fab. And many of you, you know this, that CD3 is probably the most potent T-cell activator and immune effector that exists out there today. If you were to dose CD3 that's not sterically blocked or in the periphery, you would essentially get widespread CRS. So, what we're excited about is leveraging the LockBody technology to dose CD3 in large quantities, being able to hopefully keep that silent in the periphery, but then activated in the tumor, along with PD-L1 activity.
So I think for the first time, what we're really excited to be able to do is test that hypothesis: What happens if you take PD-L1 in large quantities, CD3 in large quantities, be able to bring those both into a tumor simultaneously, basically activate T-cells to their maximal potential, without causing widespread CRS? I think that, you know, for a cancer guy like myself, I'd love to see how that molecule ultimately plays out in very difficult-to-treat cancers.
Awesome. Any closing thoughts here, given it's gonna be a pretty interesting year, across-
Every week, every month. Every week, every month this year is what you kind of wake up every day and you live for. This is 2024 Centessa's year.
Thank you, Saurabh.
Thank you.
Pleasure to host you.
Thank you.
Thank you so much.