All right. Excellent. Thank you, guys, for joining us. A pleasure to have the management team from Centessa. I know this name has become very eagerly tracked. Every little update you guys are putting out now is under the microscope. Saurabh, I'll let you kick things off.
Yeah, absolutely. Well, thanks, Umer, for the opportunity to speak to you this morning. We're obviously super excited at Centessa for our orexin franchise pipeline that we're building. As you've seen over the last few months, to your point, Umer, we've released some very significant data, data that uniquely positions us as a best-in-class molecule in the space, both in terms of the effect size that we're seeing and very difficult to keep awake healthy volunteers in the middle of the night, and the corresponding safety data that goes with it, which, frankly, has even surprised us.
We weren't expecting to see such few on-target AEs that are really no different from placebo at this point, but seeing such a large effect size with the wide therapeutic range from the very first dose that we even tested at one milligram in our sleep study all the way up to the latest doses that we've disclosed at 3.5 mg. So very excited for the field. This is viewed as a narcolepsy drug, but it's much more than that. It's a drug that keeps people awake. It's a drug for excessive daytime sleepiness. But now, as the data is evolving in the space for orexin, we're seeing benefits in cognition, memory, attention, executive function, in addition to wakefulness. So I think this is just the beginning. It's the tip of the iceberg for what's to come with the orexin agonist.
Outstanding. Well, there's several layers to unpack there, but maybe just at the outset, can you remind us the strength of the signal from Takeda? And is that really the only main clinical data point on indications beyond narcolepsy, what Takeda was able to show on executive function and memory, et cetera?
Yeah. I mean, if you look at the orexin agonist and just the pathway in general, you're releasing serotonin. You're releasing dopamine. You're releasing histamine. You're naturally going through a signaling pathway that's releasing neurotransmitters at the right levels because you're letting the body do that through signaling through the orexin pathway. And that has now translated into some incredibly powerful data, some great work by our competitors to show that this pathway is going to hopefully be effective beyond just keeping people awake, but actually improving their mood, improving their executive function. These are all aspects of diseases where, if you look at the eye chart of potential indications where an orexin agonist could play outside of the rare hypersomnias, those are all benefits that you can have on top of wake promotion, cognitive benefits in neurodegenerative diseases, as well as neuropsych indications.
Excellent. Also, maybe perhaps remind us, Saurabh, your next-gen orexin programs, what's going to be similar or different about them? And are they tailored for certain indications, or how are you thinking about that?
Yeah, it's a great question. So ORX750, which is in phase II now and being dosed in NT1, NT2, and IH, is positioned for the rare hypersomnias. And we are following closely behind with ORX142 and then with ORX489. The intent was not to fix what's not broken with 750. 750 is exactly what we'd like to see in terms of the pharmaceutical properties, in terms of how it's behaving. In fact, in my 20+ years doing this job, I've never seen a molecule, a small molecule that has behaved as well from a PK standpoint, the linearity of dose exposure response and safety, which is what ultimately you want to see in any kind of drug period for ultimately for approval purposes. I haven't seen anything like it with ORX750.
And with 142 and 489 falling closely behind, our goal was to not change something that didn't need changing. However, we did note that with 142 and 489, the potencies are even getting better and better. And now we've approached essentially the potency with 489 of the Orexin A peptide itself while maintaining selectivity. I think that's important because oftentimes you can kind of trick in med chem as you trade off selectivity and potency. You can play that game. Our chemist didn't. We retained both of those features. And what that means is lower doses. So, Umer, to your point on differentiation, with 489 and 142, one can pretend with the lower potencies that you can have even lower doses, which is great. I mean, we're already at somewhere between 1mg and 5mgs for 750.
But now with 142 and 489, this could be even lower than that, which is important because some of the diseases that we hopefully plan to go into, on top of standard of care, neurodegenerative disorders or neuropsych disorders, where EDS is a component, you want to be able to dose on top of standard of care at doses that are really, really low so you don't have potential DDI risk. And I think the lower the dose you go, it's just less mass of the drug in your body, the less burden on your liver, the less burden on your end organs, and the less chance of a drug-drug interaction, which is why I think these molecules are even further differentiated for the broader indications.
Got it. So orexin receptor 2 versus 1 selectivity, I think you guys are tracking broadly something close to 10,000x.
That's right.
In my mind, even if it's 100x, that's pretty good. I guess, how do you think about that 10,000 versus, let's say, something which is 700x, for example, the Harmony molecule?
Yeah, it's a really good question. I'm glad you're asking that. So when you have an antagonist, for example, you want to be able to slam the target 98%, 99%. So you want to have the ability to really shut down that pathway. And so there, selectivity may not be as important. You can get away with less. But if you want to agonize a pathway, especially the orexin pathway, where just a little bit of activation of that pathway can lead to signaling, selectivity becomes a bigger, much bigger issue because there, in agonist, you don't want it to tickle other pathways even a little bit because then you're starting to get activation.
So with antagonist versus selectivity, you think of it as antagonist versus agonist.
Exactly. That's exactly right.
Got it. So I guess, I mean, you guys don't necessarily have a view on the Harmony molecule per se, but you guys do think if your selectivity was not closer to 10,000, are there certain off-target side effects you would have anticipated?
Yeah, I would imagine we would, but that's hypothetical.
Sure.
I'm just imagining that. That's why in our selection criteria, our filter, if you will, for picking 750, one of the criteria to Mario and credit him and his team was to have a very, very four-digit, five-digit, almost approaching level of selectivity over the orexin 1 receptor. But not only that, he and the team profiled to Safety 47 panels, GPCR max panels, an extensive array of GPCRs, an extensive array of non-GPCR targets with incredible selectivity. The team could have progressed an orexin agonist two years before 750.
We could have put one in the clinic at least a year and a half before, but we decided to hold back until we achieved that level of selectivity, potency, and most importantly, this narrow Cmax to trough ratio, which we're seeing play out when it comes to the safety benefits of this drug and large therapeutic index is that we're not seeing these kind of as much as these on-target AEs that normally one would expect with orexin agonist because of our narrow Cmax to trough ratio.
OK, got it. Do you think visual disturbances are a class effect?
Yeah, I think there's a lot of discussion around visual disturbances. We haven't seen it, so we can't really comment on it. And so I think you'd have to ask the other companies in the space that have seen it.
But Takeda has never seen it across three molecules.
Yeah. And what's interesting is Takeda is given BID. So that's why if you'd ask me to kind of guess what's happening here, I think drugs that are given QD that have a high Cmax to trough ratio, where you've got to keep the tail high enough for folks to stay awake later parts of the night or the day, depending on what study you're doing and what population, you're going to have to dose the drug high. And if you dose the drug high enough, you're getting all this AUC and the Cmax above a certain threshold to stay awake. That's kind of wasted, if you will, because the key is to stay above that threshold just enough and keep the tail above the threshold long enough. So our goal really was to minimize that Cmax because we knew that was just kind of wasted space.
And ideally, like what you saw with the 925 study, which I think is beautiful work by Takeda, they've just done amazing work in this space, with an IV drug, you're getting a flat PK curve, essentially. And so you can define where that threshold is. And they don't see these kind of visual disturbances or other effects that you wouldn't want to see as you're approaching a high Cmax. So that's just a guess. I don't think we know. I think you'd have to ask the other people.
Got it. But just to be clear, at least in SAD, in single ascending dose, you guys went up to 5mgs , and you're not going up any further.
Yeah, we haven't disclosed if we're not going up any further, but what we've seen in our 5mgs SAD and what we've seen in our 3mg MAD, and we are dosing higher in terms of the MAD to approach the SAD, obviously, to give us that room, is that we haven't seen any appreciable differences between our drug in terms of the safety profile on-target AEs compared to placebo. If you look at our table, which we've disclosed, we haven't seen that in our SAD studies, in our MAD studies, our sleep studies. The safety is just very consistent. We've now dosed over 100 subjects. So this isn't like small n anymore, and these are healthy people where the safety will show up in healthy people.
I think our competitors have shown clearly that the signals you see in patients are the same signals you see in healthy volunteers. There isn't a new AE that shows up in patients that isn't seen in healthy volunteers if you dose high enough. I think that's a really important point for the entire space to understand clearly, and we haven't seen that.
So maybe that's a good segue then to the exact trial design because I think there's still some confusion. So maybe just to recap, in healthy volunteers, you guys did 1, 2.5, and 3.5. And we saw the wakefulness data on those three doses. Separately, in your SAD, you went from 1, 2, 2.5, 3.5, and 5. And you're saying you may or may not go higher.
Yeah. I mean, right now, what we've seen at 2.5, which we disclosed in September, I'm trying to remember, I think it was 32-minute MWT, a nice difference between placebo. And then at the 3.5, where I think 34 and 20+ , that is a great place to be if you are trying to achieve efficacy in NT1, which is a third, roughly a third of that dose. So we have a lot of room there. And with NT2 and IH, we anticipate the range of effect size we're seeing between 2.5 and 3.5 to be very compelling for an ultimate NT2 IH drug. So to your point, we've gone up to 5. We can go higher. We haven't really seen any of these on-target AEs come up at any, again, difference from placebo. So we can go higher if we need to. But the question is, why go higher?
Why then push the MWTs to the point where you're going to create insomnia in patients if we're pushing all the way up to 40? I think if you talk to the KOLs, and just I think what makes sense is that you don't want to be at 40. I think you want to be somewhere where we are now in the mid-30s or so. I think that's the sweet spot. But we retain the optionality to go higher if we need to. And we certainly have the therapeutic index. And as I mentioned, our PK is so linear that we can, which is remarkable, almost to the nanogram per milliliter per hour, see exactly what that number is in terms of threshold for people to stay awake with the drug on board.
And if they're one or two units above or below, we can predict if they're going to be awake for 40 minutes or five minutes and fall asleep. That's the level of precision you can have if you have a drug with linear PK, linear exposure, linear response, and the safety and therapeutic index to go with it. And as you go higher, you get tighter and tighter and tighter data. And the heterogeneity all but disappears. And that's where you want to be. Ultimately, this is drug development. That's the goal.
Remind me, you guys have never shown the PK curve so far, right?
We haven't.
Is that something you guys would show when the data gets presented?
Yeah, it's certainly a possibility. Right now, we haven't shown it purely for competitive reasons.
Got it. One of the questions I had after some of the initial data came out was, theoretically, if it's Cmax-related AEs on visual disturbances, et cetera, you're going higher in dose. That could happen. So clearly, you went to 5mgs on single dose, and it was all clean. Separately, what we had not seen was multi-dose data. So in the recent update, you guys had data on the 2 and 3mgs multi-dose. Could you remind us, are you guys going beyond 3mgs on multi-dose?
We are going beyond.
So, you'll go to 5?
We are going. We haven't specified which doses. But we want to create the situation where we have enough headroom in our very creative credits to the team study design in phase II. We have the opportunity to start, and we've disclosed those doses, start at very, potentially very effective doses. But then we feel we have the ability real-time to adjust those doses up or down based on the effect size that we're seeing in NT1, NT2, and IH. And having that headroom from the SAD and the MAD in the phase I gives us that optionality to be able to titrate the dose up and down, as opposed to picking the doses in phase II ahead of time, waiting a year later to see what happens.
We're going to be able in phase II, real time. Phase II is ongoing. See the data, and then be able to say, OK, let's go up 0.5mg . Let's go down 0.5mg . Let's establish a dose range to justify the dose to the regulatory agencies when we design and propose the phase III study after that data is achieved.
Got it. OK, great, and remind me again, the trial was done where, the existing trial that we're seeing data from?
U.S.
It was done in the U.S., but it wasn't on ClinicalTrials .
It was not on ClinicalTrials . That's correct.
Got it. So the next step now is, presumably, you're multi-dose through 5mg . You have that by, I would imagine, sometime in 1Q. Are you guys going to a large phase II, or what is the development look like?
Yeah, so we have a different philosophy on ClinicalTrials in general. As you've seen, we went from an open IND in May this year, actually, and into phase II already, dosing 100 subjects, getting safety data, efficacy data in a matter of four or five months. That was our phase I. That gave us a confidence, as all clinical studies should do, is you need enough confidence to then say, OK, now I'm going to go into my next study, understanding what the dose could be and what the safety profile could be. That's all these in-between studies should matter in phase I, phase II. Phase III is very different because that's when you're actually powering everything up to get your endpoints met and safety database met for approval. But everything in between to us is simply informing what we do for phase III.
How do we get to phase III working backwards in the fastest possible way? Look, you saw phase I, three and a half, four months. Phase II could be very quick because we know we have an idea where the doses are going to be effective, what dose range we need. And we've also introduced a randomized discontinuation part in our phase II, which I think is a very clever way of getting an inkling as to what potentially one of the pivotal studies in a phase three could look like, in addition to your traditional phase III type of study design. Basically, our phase II doesn't have to be hundreds and hundreds of patients.
It has to be enough subjects, six in NT1, eight in NT2, 12 in IH per dose cohort, can do multiple doses, and then have that to be able to say, OK, this is our phase III doses we're going to test, and this is our design because it's being informed by the phase II with a high probability of success.
Got it. So I guess maybe just remind us then, what are the cadence of readouts into 2025? I think there's a little bit of lack of clarity on what that is. We should expect multi-dose data. We should presumably expect some efficacy from that. And then there's like a phase IIb that's starting next year? I just wanted to ask.
Yeah, so we're separating the molecules out. ORX750, we anticipate sharing data at a conference from our phase I, the totality of the phase I.
But that's 20 25, not 2020.
That's 2025. And then with 750, we have already started phase II. So we've publicly said that we will have data on NT1, NT2, and IH in 2025. We haven't specified when in 2025.
This is also not on ClinicalTrials ?
This is. It is on ClinicalTrials , I think. It's going into ClinicalTrials soon.
It's going in.
Thank you.
This just got started.
It's started. It's been underway, and so in phase II for 750, we anticipate generating this data real time, which is what I've mentioned, is that we have the ability to see the data after these cohorts that are blinded, obviously to physicians and the patients, but be able to see by dose, by indication also. We don't have to wait for NT1, NT2, and IH to all read out and then kind of unmask or unblind it and see the data set. We can do it by indication, by dose level. That data, and it's up to our discretion as to which parts of that data or the totality of data we'll show in 2025, with the intent of getting to the phase III or the registration studies as quickly as possible after we have the confidence to, after we see that data.
Then 142, which is ORX142, is our second orexin molecule, which should be in the clinic next year. We should be able to generate clinical data as well, informing, hopefully, the broader indication space beyond the rare hypersomnias.
Got it. And what will be the first indication for that?
We haven't disclosed that yet.
OK. We'll be in clinic next year as well. So maybe in my simple mind then, can I just repeat it back to you just so I have it straight? Before we get to the, so phase II is up and running, presumably data is second half next year, I would imagine. It's probably not sooner than that. Obviously, you guys haven't guided. But separately, the phase I's multiple ascending dose, would that be data we probably get a press release on in 1Q?
The phase I data that we've already generated.
Oh, no, the multi-dose at 3 and 5mgs .
The phase I data, yes. I mean, we will release that data in 2025. We've already released a 3mg multi-dose data.
Yes, exactly.
Yeah, so any further multiple dose data, we will release in 2025.
OK, got it. And is it reasonable to assume also that you guys do have visibility on a blinded basis on safety for new studies as well? And you guys have been perfectly comfortable.
Yes. I mean, perfectly comfortable, meaning that I can't disclose what we've seen to date, but we feel very confident that we have a best-in-class molecule.
I guess my last one would be, how do you guys, well, I have to, because I wanted to ask about what happened on hemophilia, but I'll come to that in a second. How are you guys thinking about, I mean, clearly, you've been developing and going at a rapid pace. Do you guys think you need a partner, or is this a program you want to go just solo, blaze right through? How do you guys think about that?
I mean, we've certainly seen in the rare hypersomnia space a number of companies take on that. It's a $5-$6 billion market, a population of patients that you could certainly, as a small biotech, be able to build a commercial franchise to be able to go after that, or enterprise to go after that. But the way we've been positioning our orexin agonist is to be much more than just the rare hypersomnias. It's to go after all the indications out there, neurodegenerative, neuropsych, where excessive daytime sleepiness, cognition, mood, memory, these kind of things are important. That's an eye chart of indications.
Our job as a company is to generate data as fast as we can with ORX142 and ORX489 in small, very inexpensive proof of concept studies, if you will, in all those indications that we possibly can, to be able to show that, look, the orexin agonist can be very effective beyond the rare hypersomnias for a number of different disorders where there's comorbidity associated with these cognitive issues as well as sleepiness issues, and then be in a position to decide, OK, these are the ones that we may want to take forward, or it becomes a Pembro moment where can you spend, you'd have to spend billions and billions of dollars going after dozens of tumor types of different lines of therapy all the way to first line.
It's probably not dissimilar to orexin in that sense, where there will be, and there already is the thinking of that moment where, OK, how do we capture the entire space of opportunity here, and can we do this alone as a company?
So is it reasonable for me to assume that narcolepsy and the rare indications, you guys can probably go solo, but for the large, large indications, maybe a partner makes more sense, or?
We retain all that flexibility and optionality going forward. But our job is to continue to generate data, so we have that.
Excellent. And maybe just as a one-liner, hemophilia, did the biomarker pop up? Did the D-dimer levels pop up? Is that what happened?
There were no D-dimer elevations, no chronic D-dimer elevations, no thrombosis. This was a decision that was very, very difficult for us as a company because of the emerging orexin data that we were seeing, and it just became a strategic decision of, for every dollar that we spend, how much of those pennies would go toward one or the other, and it just made sense that at the end of the day, if you look at it, one additional indication on a label for one of our orexin drugs probably supersedes the entire market.
I see. So this could get out licensed and still get developed elsewhere.
Yeah, I think we'll see what the appetite is for someone to develop this drug. We think it's a good drug. And we want to make sure, and patients that are currently on the drug are getting the drug. And we want to make sure.
I spent too many slides on this drug. Outstanding. Well, listen, thank you.
Thank you.
I know this is an exciting year coming up, so congrats.