So good afternoon. Welcome to TD Cowen's 44th Annual Healthcare Conference. Thanks to everyone listening in to our fireside chat here with Centessa Pharmaceuticals. I'm Stacy Ku, one of the biotech analysts here at TD Cowen, and we'd like to welcome Saurabh Saha, CEO, Gregory Weinhoff, CFO, Mario Accardi, President of the Orexin Agonist Program, and Kristen Sheppard, Head of IR, for being here with us today. Thank you so much for being here. While Centessa is not a covered company here at Cowen, they're going after very interesting indications and large opportunities, and importantly, should have a very interesting catalyst, catalyst-rich year. With that, we're gonna go right into Q&A. Saurabh, if we could start with ORX750, a lot of investor interest and focus on your orexin receptor two agonist program.
So just really some quick background on the asset and just provide some background.
Absolutely. Well, thank you, Stacy, for the invitation for Centessa to speak to you today. So we are super excited about our orexin program. We have Mario here, who's the President of the Orexin Program at Centessa, and what we've tasked Mario and his team to do is to try to achieve the best possible molecule you can achieve in terms of developing an orexin agonist for narcolepsy and other sleep-wake disorders. So it's worth taking a step back. What does that exactly mean? So the team was really tasked to seek out a molecule that had the optimal PK profile that would be desirable for both narcolepsy and non-narcolepsy indications.
To have a dosing regimen that would be once daily, ideally, a drug that's delivered at the lowest dose possible, so most potent molecule you can achieve, but to do all of that very safely. So it's worth taking a step back to understand, how do we achieve this? How do we achieve this? So what Mario and the team was able to do is exquisite, exquisite medicinal chemistry work to be able to generate the chemical matter that's resulted in ORX750. And that goes back to first principles in that understanding the biophysical properties of the orexin-2 receptor.
So what made the chemistry possible is by having a proprietary crystal structure of the Orexin-2 receptor, one that can't be downloaded off the internet, but one that's proprietary through our collaboration with Sosei Heptares, where the structure was mutated to open up the pocket where the Orexin A peptide lodges, and be able to then design small molecules that can agonize the receptor and have the desired effect. So that resulted in ORX750, as well as, a bounty of other follow-on molecules that Centessa is now pursuing in this space. So we're very excited to get ORX750 in the clinic as soon as possible and then have healthy volunteer POC data readout in 2024.
Oh, wonderful. So one piece that you've alluded to or highlighted in the past is therapeutic index. So help us understand kind of what that means. Obviously, we've seen a de-risked kind of pathway forward for your agonist, but how should we think about therapeutic index and how you might differentiate?
Yeah. So when we look at an orexin agonist, we just don't see only narcolepsy as a potential indication or NT1. We see beyond NT1, we see NT2, we see idiopathic hypersomnia. We see basically any disorder where excessive daytime sleepiness is a morbidity associated with that disease. So that could be patients who have normal orexin levels. And to achieve that, you need much higher exposures of an orexin agonist to be effective in the latter cases, where excessive daytime sleepiness is a comorbidity of various diseases. So if you work backwards from those principles, you need to have a molecule that has a very wide therapeutic index. So the team was tasked to find such a molecule and to do the chemistry necessary to achieve what we think is a best-in-class profile. And what does that mean?
That means that in our screening funnel, when we came up with ORX750, we had to go through a number of different screening filters, if you will. So we profiled our molecule through the most wanted list of enzymes and receptors that cause potential toxicity, like the SafetyScreen47 panel. We screened against a GPCRMax panel, which has 168 receptors, that we made sure that we were only selective for the orexin-2 receptor, and also significantly more selective, 9800-fold more than the orexin-1 receptor. That was just part of our funnel, if you will, to achieve the safest molecule possible, at least based on non-clinical studies.
But then also make sure that we have a metabolite profile that predicts, at least in humans, non-clinically, that we would actually have a molecule that behaves the way we want to when it's metabolized.
You talked about safety. Obviously, that's an interesting aspect of the orexin agonist space, as the efficacy is not really called into question, but safety is. So just wanna talk about the different potential adverse events that have been kind of seen in the space, and your views on what's important, what's you know less, let's say, prohibitive of adoption within narcolepsy.
Yeah. So, there are safety, you know, AEs that you may have seen with other molecules in the orexin space that are, you know, associated with on-target effects of the molecule, like polyuria, like an increased, slight increase in blood pressure or heart rate. But those are all on-target expected AEs, because basically triggering the Orexin-2 receptor increases arousal, and when you increase arousal, you get those kind of effects. Now, what we've seen, and this is based on published data, is that these are often transient or they're manageable in the case that most of the patients who are on orexin agonist stay on the agonist, and, and they prefer to do so.
... Do you view, let's say, visual disturbances or hallucinations as concerning as we think about kind of the benefit-risk profile for narcoleptic patients? Obviously, you talked about some of the other issues that might actually be limiting, but curious your views.
Sure. Well, first of all, in terms of visual disturbances, there is really very little data out there. First of all, I would say that, well, from what we heard, is that they're transient in nature, and they seem to be happening at the higher exposures, probably even beyond the therapeutic window that you would need for Narcolepsy Type 1. So it's worth just mentioning that. Secondly, we hypothesized that it could be a, at this stage, either on-target or off-target effect. If it is on target, it could be Cmax mediated, it could be Orexin-1 mediated. At this stage, it's difficult to know, Stacy.
I guess back to what we've done with ORX750, is that, first of all, we have a molecule that is extremely selective at orexin-2, 9,800-fold more selective than the gold standard, I would say, FLIPR calcium mobilization assay. And we also have a molecule with a predicted human PK that is really an optimal peak-to-trough profile, again, potentially minimizing those Cmax-mediated adverse event. But I guess it's early days. One thing that I would stress is that the effect size of this mechanism, orexin agonist, is just so so large that maybe you don't even need to potentially to go and dose to such and reach such high exposures to drive a therapeutic a therapeutic profile that one is really looking for in order to be better than standard of care.
One last question on kind of the safety profile. As you look through kind of the signal to noise of a lot of these different issues that, let's say, concern investors, to you, what do you think are the most important ones to consider for adoption? So, for instance, you know, cardiovascular-
Of course
... issues versus, let's say, liver tox versus something like, you know, insomnia or anxiety.
Absolutely. Well, first of all, I think one of the beautiful thing about this mechanism is that the biology is so well understood, right? One of the main objectives here is to be able to reach a certain wakefulness threshold, a certain brain exposure, in order to promote wakefulness and suppress cataplexy. That is really the objective. In doing so, you want to be able to have an optimal PK profile in terms of having that diurnal exposure and then have that PK profile then come back down at night and allow the individual to go back to sleep. Of course, DILI is an incredibly important focus area.
We've front-loaded a lot of MetID and advanced assays in order to de-risk that as much as reasonably possible pre-clinically. As Saurabh mentioned earlier, we're also focused on a highly potent molecule in order to bring down those predicted human doses. In doing so, we also brought down the risk for DDI, which is important not only potentially for the rare insomnias, but most importantly, maybe even for the common disorders, where you need to maybe co-administer orexin agonist with other treatments.
Wonderful. So just curious, as you look to the competitive landscape, obviously efficacy is not in question, let's say, broadly speaking, among the class. But what different aspects, in terms of the profile, efficacy profile, do you think is most interesting, among other, let's say, the other, lead programs that are already in clinical?
Well, from an efficacy perspective, again, what you're looking to do here is, on the registration endpoint, is get as close as possible to what a human, a healthy individual would experience in terms of Maintenance of Wakefulness Test, which is really around your 35+ minutes on that scale. So that's one of the things that, of course, we'll be looking at in our clinical, in our POC study, in 2024. In terms of other efficacy parameters in the—I would say in the narcolepsy population, well, as you know, treating EDS and suppressing cataplexy by bringing that weekly cataplexy rate to as low as possible in Narcolepsy Type 1 individual is highly important. And I would say just treating the broader EDS across the indication is really what is...
What we know the target can be extremely powerful at doing, and we've seen that with clinical precedence. There's also some I would say some other clinical features that for which an orexin agonist could potentially play a role. I mean, maybe even fatigue, maybe even, maybe even disturbed nighttime sleep. We've heard that orexin agonist can also potentially, at least at the preclinical level, consolidate sleep architecture. So it could be really interesting to evaluate whether an orexin agonist could indeed be a monotherapy for this indication for the very, very first time.
Perspective, and it's interesting. When we talk to clinicians, again, they talk about kind of the unmet need and the tolerance from narcolepsy patients to see some of these AEs. But also, when we talk about kind of all these different endpoints, obviously, maintenance of wakefulness, Epworth Sleepiness Scale, we talk. When the clinicians look at it, they're like, "Okay, when is this patient driving?" Right? So it really becomes such a simple task that the patient needs to overcome. So we think it's very interesting.
That's right. Exactly.
So, again, another question on the competitive landscape. So they've recently disclosed, Takeda has disclosed that they're going into phase 3 for NT1, but not NT2. So just curious, you've obviously talked about a lot of the thoughtfulness that's gone into your kind of development of the product, but where else do you think you could kind of differentiate as we think about the broader opportunity?
... Well, first of all, yes, we recently heard some of the news about progressing, Takeda progressing to a phase 3 study in NT1. I would say at the fundamental, one of the goals that Saurabh tasked us with was making sure that our ORX 750 and the whole candidate nomination process would allow us to go into indication expansion. That's very much how we've nominated 750. And at World Sleep last year in October, we showed a comprehensive set of preclinical data in different animal species, including Narcolepsy Type 1 animal models, but also showing wake promotion in wild-type mice, which essentially are proof of concept of that indication expansion opportunity. So, there is potential to go much broader than narcolepsy.
As you may know, Stacy, we also have one of the great things about Centessa is that we actually have a treasure trove of orexin agonist, all discovered through structure-based and rational-based drug design. We do have follow-up molecules that we're progressing on. The whole common disorder space, EDS and MDD, EDS and Parkinson's disease, residual EDS and OSA are all really, really exciting opportunities for which an orexin agonist could have a significant role.
Remind us to go after NT2 or idiopathic hypersomnia, what kind of dosing are we looking at for, let's say, a normal functioning orexin system?
Well, again, based on clinical precedent, on our own preclinical experiments, we've seen that in non-orexin loss indications, you need probably around 3-4-fold higher doses for a given efficacious profile. I think the important thing to consider about NT2 and IH is also these are very much diurnal clinical features, right? In IH, the individual experiences sleep inertia amongst other symptoms such as excessive daytime sleepiness. But Narcolepsy Type 1 really is probably the most different of all because you have that disturbed nighttime sleep. So you can also start to think how you would probably modify dosing regimens, dosing in order to treat these different pathologies.
You would need to push the dose?
Okay. Yes.
So obviously back to your product, 750, what disclosures should we expect this year? Initiation of the phase 1 study, what should we expect? What kind of studies are you gonna run? Are we seeing single ascending dose, multiple ascending dose, and, are we gonna get data this year as well?
Yeah. So, we're super excited, as you heard from Mario, about ORX750, and the guidance that we've given for 2024 is that we anticipate having healthy volunteer proof of concept data in 2024. And I just wanna make a point on why we say healthy volunteer proof of concept data, it's because that is the highest bar. If we're able to achieve good MWT scores or minutes awake in that patient population, or that population, I should say, that portends well, as Mario just mentioned, into all the other indications where orexin levels are normal in patients.
So we believe if we hit that proof of concept, and we have that data this year, that portends well, not only for NT1, where the doses are, you know, 3-4-fold less, but also for broadening our scope of ambitions, if you will, for indications outside Narcolepsy Type 1.
Okay. And, we'd note that you probably also would get a very good idea of safety after phase 1 study, in addition to level of-
That is correct.
Okay. I'm curious, you, Mario, you talked about the preclinical investigation for 750. Are you able to look at safety at all, kind of in a preclinical setting? So what level of conviction do you have around at least some of the AEs, maybe visual hallucinations, a little bit tougher to look at, but anything else?
I just wanna comment that, as you can imagine, it's a competitive space now, right? So the bar internally at Centessa that Saurabh set for myself and the team was extremely high to nominate an orexin agonist. So we feel very confident about the preclinical profile of ORX750. And that's why, amongst other things, we focused on potency, we focused on CNS penetration, we focused on a differentiated, very different metabolic profile in order to give us confidence and de-risk as possible all of these features that you've just described. So, yeah.
To specify, would you be able to at least look at liver toxicity or any type of cardiovascular AE or blood pressure or anything like that in a preclinical setting?
Well, we performed the standard package-
Yeah
... that one would anticipate when you file for opening a study. So not only on top of that, but we went beyond, as I mentioned before, doing the GPCRMax panel, the SafetyScreen47 panel, but as well as Mario mentioned, metabolite ID, we made sure that in our screening cascade, all these steps were achieved.
Wonderful. So, if we could transition to SerpinPC. We're in a registrational program right now.
That's right
... for hemophilia B. So, again, just some real quick background on the target, and the most recent data that's been presented so far.
Yeah. So SerpinPC is in two pivotal studies. We've already dosed patients in our pivotal studies. The studies are ongoing. They're enrolling quite well. We're very excited about this particular mechanism. It's very unique. It's unique pharmacology. It's unlike, I think, any that's been seen in hemophilia, in the hemophilia space before. And I say that because, one of the underlying challenges for developing a hemophilia drug is that you're trying to stop bleeding by creating a clot, but at the same time, you're trying to make sure that you don't overclot and create pathologic thrombus from occurring, resulting in strokes or, or MIs.
So with SerpinPC, we were able to achieve what, really, I don't think has been achieved to date, which is actually have a meaningful wide therapeutic index compared to the agents that are in development today. So, there are three anticoagulant pools that you could target, and they are TFPI, the antithrombins, that you can deplete, or protein C. So the competitors that are in development today target TFPI, and they target the antithrombin pool. And that basically leads to unchecked, you know, procoagulative state. Protein C is not what SerpinPC touches. It's such as it blocks activated protein C, which is a subset of protein C.
So we are very careful in this mechanism, being able to generate enough clots to prevent bleeding, but then not create over clotting to occur, because of the mechanism of action of a drug only targeting the activated Protein C and not the pool of Protein C. So that we are able to achieve what I think is very difficult in a hemophilia drug is a wide therapeutic index. And when we tested this drug non-clinically, we were able to see at least a 100-fold window between exposures necessary to achieve clotting versus actually seeing any kind of micro, even microthrombi, basically at 100X the exposure. So we're super excited about this mechanism. There's a...
The KOL community is very excited, and it certainly is gonna address an unmet need, where there now could be a very effective subcutaneous option for hem B, and but also safe at the same time, which is really important for this community.
Wonderful. And then let's walk through the regulatory pathway. Clearly, right now, you're only pursuing hem B, and some of the other, let's say, antithrombin, thrombin assets are going after both hem A and B. So just walk through the clinical trial design. It's a little bit unique, so kind of, I'm curious, you know, obviously, you spoke to the FDA and got agreement there, so just walk through the design and implications.
Yes. So there are two pivotal studies, PRESento-2 and PRESento-3. PRESento-2 is intended to generate data in hemophilia B without inhibitors. But what's important to note is we're also dosing hemophilia A patients with and without inhibitors, and that's a 120 patient study. There are two parts to that. The first part is what we're calling a dose justification part, where we have three different dosing regimens: once weekly, once every two weeks, and once every month.
From there, we'll pick a dose and a regimen, and then we'll move forward into part 2, where we have a minimum number of hem B patients without inhibitors, that were on-demand or on prophylaxis treatment, and then they'll go on SerpinPC prophylaxis treatment for 24 weeks, and that'll serve as the primary endpoint for this study, and we can go out to longer to 48 weeks, as an endpoint as well. In part, in PRESento-3 of this, of our pivotal study is hemophilia B with inhibitors. So that's a smaller population, where we will treat approximately 12 patients, at a predefined dosing regimen of 1.2 mg per kg every 2 weeks, and assessing their ABRs, annualized bleed rate at week 24. So those two pivotal studies are meant to serve as registration studies.
Okay. And I believe we're gonna get updates from the registration study this, this year. So, what's gonna be this interim part one, disclosure that we're gonna get?
Yeah. So our interim analysis is on part one of the study, and there we'll have basically a limited interim analysis that's intended to disclose what dosing regimen that we ultimately pick. We'll have a discussion with the FDA, and per to your earlier question, we'll address the question of whether or not we can actually go into potentially hem A as a possible population, because we are enrolling them anyway to be able to go beyond hem B and get a pan-hemophilia label. But that's a discussion we have yet to have.
Okay. And as we think about, the primary endpoint, that would be for patients that are not on routine prophylaxis-
Yeah, that's right.
To be on-demand only.
Exactly. That's an important point to note, is that the patients, which serve as the primary endpoint for the PRESento-2 study, are patients with hemophilia B without inhibitors who were on-demand. So they're bleeding quite a lot, on an annual basis, and they will then go on SerpinPC as prophylaxis therapy, and their endpoint will be addressed, at week 24. Now, what's interesting is that we've already done essentially that study now 3 years. We've been dosing patients with both hem A and hem B who have been on-demand, and just to remind folks that we were able to achieve a 96% reduction in ABR, an ABR of 1, median ABR of 1, which we disclosed at both ASH and EAHAD, and most recently, an oral presentation at EAHAD earlier this year.
Wonderful. So curious, as we look to, let's say, the hem B market more broadly. Talked about the safety, obviously, that's very differentiated. Who do you think is gonna be kind of the target market? Who's gonna switch? Is it gonna be the younger patients, newly diagnosed patients, maybe those that are not severe enough for gene therapy? Just curious your thoughts.
Well, I'm ambitious, and I'd like to think everyone would want to switch to SerpinPC, given the profile that we've been seeing to date now over three years, where we haven't had a single episode of thrombosis. If you look at the AE table that's attributable to SerpinPC, it's zero, except for one injection site reaction very early on. And that's actually quite encouraging. And I think in a population of treaters and patients, who have had a very difficult past, where they've gotten hep C or HIV due to transfusions, factor transfusions, this is a population that's highly sensitized to safety being at the front of their minds. And as they're seeing and as they hopefully will see continued safety with SerpinPC, they'll be-...
I think encouraged that this makes sense as an option for them. So we hope to get younger patients, we hope to get patients who are going to switch from on-demand from prophylaxis. So I think the entire space is open for us.
Okay, wonderful. And then last question on SerpinPC. You've previously talked about the ex-US opportunity, so just remind us what you're thinking there?
Yeah. So what's interesting about the hemophilia space as a rare disease is that, we anticipate, you know, half the revenues actually come from ex-US, 'cause it is actually highly reimbursed, well reimbursed outside of the US as it is in the US, which is unusual for a rare disease, but, that's an exception that exists today. And we hope that the opportunity is just as large in the US as it is in the ex-US.
Wonderful.
I'm excited about that.
In the final moments, we'll talk about LB101. Is that the right way?
Yes.
So talk about your LockBody, the technology around it. Obviously, it sounds like you're in dose escalation, phase 1.
We're in dose escalation.
Background, and then, Yep.
With LB101. The LockBody technology, in brief, is essentially a monoclonal antibody with a set of Fabs that you typically have in an antibody, but we simply attach another set of Fabs to those Fabs that exist in your typical IgG1. But what's key is we attach those Fabs using a linker. It's an, it's an all-natural type of linker. It's the same linker that's found between the Fabs and the Fc of your typical IgG1 antibody. And we know, through empirical evidence, over decades, really, that that linker is highly labile in areas of high inflammation, where protease cleavage occurs, and various mechanisms, which causes the Fab and the Fc to decouple when that linker is degraded. So that's why we chose that linker to put between two Fabs.
In LB101, the Fab that's exposed, if you will, is a PD-L1. So it's essentially a PD-L1 antibody. But the Fab it's attached to is a CD47. So it's sterically blocked by this linker. The PD-L1 is blocking the CD47 activity. So when you dose these patients, they're essentially seeing a PD-L1 antibody, but when this molecule, LB101, goes into the tumor, you're seeing that not only are you getting PD-L1 blockade, which is great, you're now seeing the CD47 open up because the linker is being degraded in the tumor microenvironment. So you get the added benefit now of both immunomodulators helping you fight the tumor. And this is an active Fc, which should be noted.
As an active Fc, you're essentially bringing a third modality into play here to try to combat these tumors. What's really encouraging is that we're making great progress, obviously, on the clinical study, but the non-clinical data in cynos and NHPs showed that at 50 mg per kg, we were able to dose up to that level without seeing any anemia or thrombocytopenia over 28 days, which was remarkable. That's what's given us a lot of confidence that this platform, this technology platform, is behaving and hopefully will behave in humans as we expected non-clinically.
Okay. And obviously, hard to talk about the timeline, but when can we, let's say, broadly expect next updates?
Yeah, so it's event-driven. So when we have a meaningful number of patients, obviously, dose and a meaningful number of responses, and enough of a safety database, if you will, from the phase 1, that's when we'll share the data.
Okay. So share data from multiple indications, kind of multiple patients in each indication?
That's right. Six different tumor types for the study protocol.
Okay, wonderful. In the last few moments, we'd like to open it up for Q&A.
According to the program, what kind of patients do you have?
So per protocol, we treat four immunogenic cancers, so non-small cell carcinoma, lung cancer. We have head and neck cancer, squamous cell carcinoma, and, and we have colon and ovarian cancer as our non-immunogenic cancers. So those are the ones patients who haven't previously received PD-1, but for the other indications, they may have received PD-1 or PD-L1.
How many patients?
We haven't disclosed that at this point. Standard dose escalation per protocol.
Just a quick one on the orexin program. Do you have to achieve maximum tolerated dose inside of that for in-human patient studies, or does the FDA, you know, be more flexible with that, than maybe, you know, higher, higher dose than you would use in typical doses?
Yeah, it's a great question, and for competitive reasons, we haven't disclosed what our clinical study design is at this point.
Okay. In the LB's?
In our phase 1 design.
Take one more.
When you think of the potential theoretical, use risk of the orexin drug class, not necessarily involved.
Yeah. So, Mario, why don't you go ahead?
A couple of things. First of all, we've seen on the, at the preclinical level that, Orexin-1 seems to be the receptor that is, that is mostly related to, reward-seeking behavior. We haven't seen that, at least based on preclinical experiments with orexin, with Orexin-2. I would say that, we've also carried out a broad, abuse liability panel, internally, which we feel pretty confident about. So, yeah, at this stage, it's, of course, early data. It's a, it's a new mechanism, but, as of now, especially with a selective Orexin-2 agonist, there is nothing that points into the direction of, of, of reward-seeking or abuse liability. But of course, this would need to be confirmed in the clinic.
Yeah, it's important to note that, as part of the criteria for selecting ORX750, we wanted to design a molecule that mimics as close to possible the Orexin A peptide. So not too potent, not under potent, but potent, akin to the peptide itself, which is 30 picomolar, and our molecule is 110 picomolar. So if you're able to phenocopy the Orexin A peptide, the loss of that through a small molecule replacement, if you will, then you essentially are bringing as natural of reconstitution of that pathway in these patients as possible. So, kind of an indirect way of answering your question is that that should mitigate potentially from any type of abuse liabilities.
Anything else? Okay, looking forward to the next updates.
Great. Thank you.
Thank you so much for the time.
Thanks.
Thanks for listening in.