All right. Good morning, everyone. Welcome to day one of the Cantor Global Healthcare Conference. For session one of today, we have Corbus Pharmaceuticals, and for representing Corbus, we have Yuval Cohen, CEO. Yuval, thank you for joining us.
Thank you for having me over.
Lots going on with Corbus right now with the CB-1 and Nectin-4 ADC, but maybe we'll start off with CB-1, given the interest in the space. What makes CB-1 an attractive target in obesity, and why such a renewed interest in the space?
I think it's a number of things. It is a highly clinically validated mechanism of action. Every single CB-1 back in the day that made it to the clinic showed weight loss, and pretty similar weight loss. And so, we know a lot about what it does and how it works. It is, has milder GI adverse events. CB-1 is not gut-centric like the incretin GPCRs and has little to no effect on gastric motility. So there are some GI side effects, they're just not as pronounced as with the incretin analogs or agonists. And it's a small synthetic oral molecule, which these days is obviously very attractive as we think about obesity as really a global pandemic affecting probably a billion people. It's important to remember, this was a very, very exciting drug class twenty years ago, almost twenty years ago.
Many, many big pharma had late-stage programs, and then there was this really extraordinary, almost traumatic setback with Rimonabant that made the entire class untouchable. What has changed 17 years later is Novo. That's really what has changed. And so Novo Nordisk, a year ago or so, bought a one of the only two companies that have a CB-1 , called Inversago. The other one is us. That's about it. And have executed a phase II study, a phase IIa study, and have announced that the data is imminent. And I think that what's making it exciting is that they have endorsed this hypothesis that we've solved for the problems of the first class by focusing on peripherally restricted CB-1 s, which is what we have and which is what they have.
Got it. On the safety side, maybe if you can talk about Rimonabant psych adverse events.
Mm-hmm.
You have also published some data around the data that was assessed by the FDA. What's your take on that?
Sure. So it's an interesting case of what we remember versus what actually happened, and that's... What we remember is very powerful, but it's not necessarily what actually happened. We published a paper in the July issue of Obesity. I think it's turned out to be a very provocative, very popular paper. I think lots of people have now read it. And what we did is we basically went back to the publicly available data, and there's a lot of publicly available safety data on Rimonabant. And what we've demonstrated, I think convincingly, is that the concerns around suicidality, which is really what doomed Rimonabant, and by proxy, the entire class, were misadjudicated. And so we have two things going in our advantage as far as this new class is concerned.
The one is we are peripherally restricted, as is Monlunabant, so that's certainly reassuring. But the other one is the notion that perhaps the underlying concern may not have quite been what we all remember it to have been.
Got it. And so you talked about peripheral restriction. How does CB-1 induce weight loss, and how much of the effect is peripheral versus central?
It's a great question, and nobody knows for sure. These are experiments that effectively are impossible to do in human beings, but here's what we do know. So CB-1 has an effect both in the brain and in the metabolic organs, so in the muscles, in the liver, in the pancreas, in adipocytes. They're all rich in CB-1. You have beautiful, beautiful data out of NIH and academia, where selective organs are knocked out for CB-1, and good things always happen when you do that. The most spectacular effect, though, is a whole animal knockout. And so our hypothesis, and I suspect Novo would agree with that, is you wanna find the optimal point to be in, which is markedly reduce the amount in the brain. Monlunabant, which is a Novo drug, is peripherally restricted.
We are markedly even more peripherally restricted than they are, and yet we're both equipotent preclinically to Rimonabant, and in the clinic, they've shown data that's very similar to Rimonabant. And we believe that's because of two things. You still need to engage the brain somewhat. You can just afford to engage it far, far, far less. So to give you an idea, in a mouse, where that's the only model you can do these experiments, we are at about 3% of the brain levels of Rimonabant, and yet we are equipotent to Rimonabant. But there's the flip side to it, which is you also want to markedly engage the periphery. So again, in the mouse, where these are much easier to do, we're at about 5X of the exposure in the periphery than Rimonabant, and that's the balance you want to find.
Withdraw from the brain, flood the periphery. We are not enthusiasts of a brain-excluded approach. It's not that we don't think it will work with a CB-1 , it's just that we think that you'll take a hit on the efficacy with a CB-1 that does not interact with the brain at all.
... Got it. So some impact on the brain is required. Completely shutting it down may not induce the level of weight loss that you would see?
We know on the other hand, the flip side of it is, you can do pair feed testing with Rimonabant, and we know from that that the weight loss is not entirely due to caloric reduction. You can pair feed the other mice with the same amount of calories that the mice on Rimonabant are, and they actually do lose more weight. You want to engage the brain. That's how anti-obesity medication works primarily, is around these hedonic circuits that you wanna just turn off the obsession around food.
Right. And your lead molecule, 913, is an ?
Correct.
Maybe explain the differences between versus neutral antagonist.
Sure. So the easiest way to think of it is, so GPCRs, remember, are evolution's favorite receptor. I think 30% of all medications that we have in a pharmacy target a GPCR. Think of the GPCR as a very, very beautiful gearbox, for those who still remember what those look like. And so, CB-1 is a particularly magnificent GPCR. So you can do a bunch of things to it. You can put it in drive, and that is to agonize it. It binds, small molecule binds to the receptor and moves it forward, makes A become B downstream. You can neutral antagonize it. A small molecule will bind to it and will silence the receptor. It's all dark underneath that. You can allosterically agonize it. You can bind far away and turn it on.
But one of the remarkable things about CB-1 as a GPCR is you can inverse agonize it. Small molecules can bind to it and actually put it in reverse. B becomes A, and that's what the first and second generation drugs do. We are not neutral antagonists. We are s, and it's that inverse agonism that is so clinically validated.
Got it. Wanted to move into some of the similarities and differences with Monlunabant, which is a key focus for investors.
Mm-hmm.
So maybe talk about the brain concentration. You touched on it briefly. How does it... If you can quantify that, how does it vary between 913 and Novo's Monlunabant-
Sure
- in terms of their preclinical models?
So, Monlunabant is absolutely peripherally restricted. There is meaningfully less of it than Rimonabant at the same dose, at the same time point, in the same mouse breed. We just happen to be markedly more peripherally restricted than Monlunabant. A reminder to everyone, these are both compounds that come from two libraries that were created pretty much at the same time and pretty much from the same origin, and they both use ibipinabant as a backbone. So Corbus ended up with a very large ibipinabant backbone library of CB-1 s through an acquisition that we did in 2018 from a company called Jenrin. And Novo ended up with a large ibipinabant-based library through the acquisition of Inversago, and that originates directly from NIH.
While our patents are different and our functional groups are different, we are very similar in terms of binding to CB-1, in terms of lack of binding to CB2. Preclinically, all of us are equipotent to rimonabant. This new class is not more potent. It's just that you can use more of it without the concern that you had from the previous class. And our only real big difference is just the extent of peripheral restriction.
Got it. So, on peripheral concentration-
Mm-hmm
... is that similar or lower than Monlunabant?
They have not published. I'm trying to remember. I don't believe they've published their peripheral concentrations pharmacokinetics. I don't know what the AUC is, but we'll need to look at that, but it would certainly be, we would expect it to be more than rimonabant simply because there's less in the brain. The moment you have less in the brain, the drug has to be somewhere else.
Right. Got it. And what about the potency? How does that compare for 913 and Monlunabant?
They seem to be very similar. Again, we both bind to CB-1, single-digit nanomolar. We both have the beta-arrestin bias. I know there was some interest around that, and so we both have-- we're both biased for beta-arrestin. Remember, we both are drawn from the same ancestor, from ibipinabant.
Got it, and so if you can expand on that beta-arrestin.
Mm-hmm
... signaling in the context of CB-1, because, the paper you mentioned, it highlighted that there is potent activity of Monlunabant-
Mm-hmm
... on Beta-arrestin that Rimonabant lacks, and that may be driving the anxiogenic effects. So what's your take on that?
I'm skeptical. So in other words, look, I can be agnostic and say that we do the same thing, so that's easy. Do I think that's what made the difference? Not particularly. And in Liu et al., in that paper, they used preclinical behavior models in mice, which I have publicly stated many times I'm very, very skeptical about. But more than that, in that paper, they do not compare Monlunabant or INV-202 to Rimonabant in those, in some of those. And so I think it's a little bit speculative, but if that is the case, why not? We'd sign up for it with pleasure.
Got it. And so the 913 activity on Beta-arrestin, how does that compare with Monlunabant?
Similar. We're biased towards it.
Okay, got it. All right, so you are in the process of filing an IND.
Mm-hmm.
Maybe talk about the timing of that and what are the gating steps.
Should be done by end of year. It is a perfectly vanilla pre-IND package. We do not need to do NHP, so that's nice, so it's a rodent and a canine, 28-day tox 'cause we wanna start as quickly as possible. Importantly, there's nothing exotic we need to do preclinically and, for that matter, we haven't identified anything that was requested of us exotic-wise in the clinic either. It's your classic SAD, MAD. Like all obesity drugs, regardless of what their mechanism of action, we have to look at CSSRS and PHQ-9. The incretin analogs do the same, and obesity is the one place where I think you kind of do not wanna be creative. You wanna follow what everyone else does because, A, that's the path of least resistance, and B, it allows you to compare yourself to other molecules.
Got it. And so what could your phase I look like, when the IND is cleared?
We would be keen on copying, to the extent that's possible, what Novo have done or are doing and what others have done. Now, there's one, for example, peculiarity that has to do with Inversago's phase Ib, and I think that it was purely because of a financial constraint. So they did a 28-day phase Ib study, which is normal, a very, very vanilla, very appropriate, but they only explored one dose, which was 25 mg a day. Normally, you do a dose response in a 28-day study or in a short study. I think it was just because they, at the time, they didn't have capital. We are more than financially capable of doing so, and so we would wanna do that dose response on the early side.
Novo have had to catch up, as it were, and do the dose response in a sixteen-week study, which is fine. It's just that you kind of want to find out the answer earlier than later.
Got it. And have you talked about the doses that you may explore in phase I, and how high would you like to go?
We have not, but your expectations are... For example, look at what both Inversago and Novo did. These are doses that are fractions or multiples of the rimonabant dose. So rimonabant in the clinic explored five milligrams a day and 20 milligrams a day. Inversago's first study was 25 milligrams a day. Novo's current study is 10, 20, which is the same as rimonabant, and 50, and so those increments are very sensible, and that's really what you wanna explore.
Got it. And the trial will be in the U.S.?
It will be under an IND. We are very, very keen on conducting obesity studies, especially early on in the US. I think all of us have seen some peculiarities in studies that have come from elsewhere, and again, we wanna keep this as drama-free as possible. So I think doing it in the US is really perfectly doable.
Got it. And will it be an inpatient or an outpatient trial?
Your SAD and your MAD have to be inpatient by definition. Afterwards, you do not wanna have it inpatient because that, again, creates, can create distortions. If you keep people, for example, in a clinic for a month, unnatural things can happen to their eating habits, so.
Right. Got it. And, Novo has publicly said that, they expect at least 15% efficacy with Monlunabant longer term. What's your take on that, and what may be driving Novo's confidence there?
Thank you for mentioning longer term. I really wanna emphasize this, everyone. It's not the sixteen-week time point that they're looking at 15%. That would be, that would be extraordinary for any mechanism of action, and frankly, kind of worrisome for any mechanism of action. If that were the case, by the end of the year, they would have lost 40% of their weight, and that can't be healthy. So they're looking. They're talking about long term, and for me, long term is a year. That's what you typically peg obesity studies at. I think it's very sensible, especially the 50 mg. So the 50 mg a day of Monlunabant, the easiest way we think about it is it's the equivalent of two and a half pills of Rimonabant.
Rimonabant at 20 mgs a day got you between 8, 9, 10% weight loss, seventeen years ago. If you do a back-of-the-envelope, then yes, you would get to mid- to high teens if you just two and a half times it.
Got it. And so what, what do you think is driving Novo's confidence around that specific number?
It's a great question to ask Novo. There is an open-label extension to the study. They pride themselves also on their ability to model. Something I think has happened, remember with Novo, it's interesting to observe it, they were very, very, very quiet about this program, even when they acquired it, and starting from around April this year, they have been very chatty about this program.
Mm-hmm.
We know the study ended in June, and I wonder, and this is purely speculation, if they got to a point where they had a critical mass of safety, because this will be driven by safety. I don't think anyone is losing sleep wondering whether these drugs cause weight loss. That's pretty much a given. It's really a question of the psychiatric adverse events, and if I were in their shoes, which I'm not, at some stage, I would see a critical mass of safety data that would cause me to feel very comfortable that the bet was the accurate bet, was the correct bet to take.
Got it. And so they have the 16-week readout coming.
Mm-hmm.
What would you like to see on the efficacy, specifically weight loss?
And by the way, I don't know if it's just a sixteen-week readout, right?
Okay.
The study has an OLE. We don't know what they're going to present, so speculating, I would think of it in three bullet points. The first one, I'm very, I'm confident will be there, and I'm pretty sure I know what the answer will be. I think I know what the answer is. The first one will be what CSSRS scores look like. Those are the prospectively measured suicidality, and that's really at the core of what doomed the first generation. I'll remind everyone, the first generation was not measured prospectively. It died by a measurement, which is post hoc, which is no longer used, so I'm very confident, comfortable and confident. Of course, it's not my data that CSSRS will be negative. I think the second bullet point is: what about other psychiatric adverse events, especially depression and anxiety?
Now, I should point out these were not a concern for FDA, and they don't seem to be a concern for practitioners either, but they're clearly of interest, and so it'll be interesting to see if they see imbalances, what do they look like, are they meaningful, et cetera. And then the third one is: how much weight did they lose, and at what time point, and we'll see.
Got it. Anything, anything specific on the weight loss magnitude that you would like to see?
It's very hard for us to model from the outside because no one ever tested 50 milligrams of Rimonabant.
Right. And so on, on that dosing front, what's your view on the 20 and the 50 milligram, and how would the brain concentration vary? Because they are going two and a half times the dose of rimonabant-
Mm-hmm
... which was dosed as high as twenty milligrams.
Mm-hmm.
So how does that compare on the brain concentration, and what does it mean for the safety side?
You're right, there should be more in the brain at fifty. You'll never find out how much, unless you can convince someone to donate their brain, which is gonna be kind of difficult. Having said that, again, if you peel the layers of this onion and go underneath that, that's where our review comes in, which is: are you really worried about it? Is there actually an issue, underlying issue at the core of this or not?
Mm-hmm.
And our hypothesis is we don't think so. We think it was a misadjudication. And so if the 50 is clean, that's a pretty strong argument in favor of our hypothesis, I think.
Got it. And longer term, where do you see CB-1 best positioned, as a monotherapy or combinations?
It depends on the efficacy. So we kept thinking about this as a class where you're not gonna break through the 10% weight loss, because we kept thinking about it as Rimonabant at 20 mg a day. At that weight loss, then, yes, you can combine it with an incretin agonist, for example, because we've shown in our paper from last year that they are additive. You could use it as monotherapy for populations where it's inappropriate to use an incretin pathway drug. CB-1 doesn't trigger sarcopenia, for example, and it's milder in terms of tolerability, or probably the most interesting thing is in induction maintenance, you lose weight very quickly, very aggressively, very well on an injectable incretin analog.
But then for the rest of your life, you just switch to a CB-1 , which is easy to tolerate, and you take it once a day as a pill. But everything is maybe upside down now because if they are gonna see mid- to high-teens efficacy with a CB-1 , then you kind of don't need to do any of that.
Right. I guess if it's a monotherapy option, GLP-1s obviously have a broad outcomes benefit.
Mm-hmm.
So how will CB-1s differentiate relative to GLP-1 monotherapy, specifically around the outcomes, like do you have?
I suspect it's gonna be very similar. I think a lot of those outcomes are driven by the incredible things that happen when you lose weight. I don't think there's any sort of magical thing. There was a very large cardiovascular study launched with Rimonabant. It was just terminated prematurely. We know there was a study in Type 2 diabetes that looked quite lovely. There were studies in smoking cessation and alcohol abuse, again, really nice signals. So I don't think there's something magical about the pathways. What's magical about them is what happens when you lose weight.
Got it. And on the diabetes and the cardio outcomes trial that was stopped, could you remind us if there was glycemic benefits with CB-1?
On the diabetes there was. I suspect somebody provocatively told me the other day that had they gone for a Type 2 diabetes label, they may have actually gotten squeaked by the FDA, but that's neither here nor there. Cardio, it was stopped at less than 3% of patients were enrolled when they-
Okay
... pulled the plug on it.
Got it. Last question on CB-1, what's the IP situation with CRB-913?
Sure. So it is a derivative library of a library we brought in, that originates in 2012, as does the Rimonabant library. 913, because of the derivative library, the patent, the composition of matter only was published last October, so it is a very, very fresh patent. Novo, by the way, are doing the same thing. There's nothing wrong that we can tell with Rimonabant other than it's a little bit old. It's from 2012, hence INV-202, which is a more novel patent and is a life cycle type of program, per their own admission.
Got it. Moving on to their next set of programs, Nectin-4 ADC-
Mm-hmm
... CRB-701. Maybe talk about the differentiation there.
Sure. So it's easier to talk about it now than last year 'cause now we have more clinical data. Last year was really just talking about preclinical data. Conceptually, it's really simple. This is a approach that's very similar to Padcev, is a nectin-4 targeting ADC using MMAE as a warhead. It just uses the novel technology of ADC, so it has site-specific conjugation rather than randomized conjugation, so the DAR is very precise. It has a stable cleavable linker rather than an unstable linker, and it has a DAR of precisely two, rather than Padcev DAR, which is on average three point eight. Remember, Padcev, because it's randomized, the DAR is kind of all over the place.
And so if you think about it that way, the idea is hang on to that MMAE for longer and have less free-floating MMAE that causes off-target tox. And so what's emerging is the following: We seem to have a differentiated safety profile than Padcev, so peripheral neuropathy, skin, et cetera. And despite the fact that we're now dosing patients for longer than the time points where you would usually see those adverse events emerging with Padcev... Then the other thing about it is we seem to have an emerging very clear signal in bladder, so that's our positive control. And an emerging signal in cervical, which was never tested with Padcev, probably for licensing reasons rather than scientific reasons.
We are now bridging it to a study in the West, the U.S. and the U.K., and the purpose of that study, and this is obviously a topic that with Summit and everybody is so topical. We want to show that there's nothing odd about the Chinese dataset in terms of the patients. So what we want to see in Western patients is that the PK is the same, and the safety is the same, and that bladder patients are responding, et cetera. If that works out, and we should be done by end of year, then one of the most interesting things to think about is, as we move forward here in the West, what is the value of the Chinese dataset. That's always going to be bigger and always more advanced.
And I think that's a question that all of us, ADC companies that have licensed from China, are actually, frankly, really excited about. And so we see how that evolves.
Got it. And you got that this asset from your partner, CSPC?
Correct.
So maybe just talk about the data more quantitatively-
Mm-hmm.
on the efficacy that you saw in bladder and some of the other indications on the safety profile.
Sure.
And what specifically in this Chinese population on the baselines or the lines of therapies that these patients were on could be different?
Very similar population, very similar lines of pre-treatment, big academic cancer hospitals in big urban areas, and again, I see. I think we're seeing that in some of these other programs that are emerging, and that to us at least, and certainly in our study, we're not seeing meaningful differences other than the fact that obviously in China, all the patients are Chinese. The other big difference is, for bladder cancer, all the patients are Padcev naive in China. There is no Padcev in China, so we in the West will have Padcev pre-treated patients in bladder. Interesting enough, we're also going to have Padcev-naive patients in bladder, which is really an unexpected surprise for us.
In terms of the efficacy, what we saw was, again, a small dataset, early days, but in bladder, which for us is a positive control, we see a similar response as to what you'd get with Padcev, so that's encouraging, and then cervical. Cervical is the first of, I suspect, a few non-bladder Nectin-4-enriched tumors that we will focus on. The way we do it is we first want to get a signal, and then we'll tell you about it, rather than the other way around, so cervical is the first one where we got a signal, so we're delighted, and we look forward to seeing what other tumor types are responding. For us as a standalone company, that's where our interests lie. In the hands of a strategic, the calculus is very different. In the hands of a strategic, bladder is achievable.
But there's 29 of us, and we love being 29 people, and we do not want, as a standalone company, to try and go head-to-head with Pfizer. We think that's not logical, but we will continue to use bladder as our positive control. It's within bladder that we can compare ourselves to Padcev, apples to apples.
Right. So you'll have bladder patients in the upcoming update, but that's not the indication that you plan to go after long term?
Correct. But imagine how many... There are a number of audiences for whom that patient group can will continue to be of interest, whether it's people sitting in this audience, whether it's strategics, and whether obviously, the regulatory authorities.
Got it. And on the safety side, maybe if you can talk about the peripheral neuropathy-
Mm-hmm.
Dry eye. You know, typically, you get lower event rates in a smaller sample, but it tends to regress to the mean. Why shouldn't that be the case?
Because of how the drug actually works and what causes peripheral neuropathy and skin to begin with. And so what we're seeing, and it's interesting that we're seeing this not just with 701 , but with other ADCs from the same platform and even other ADCs from different platforms. So we had a single case of peripheral neuropathy, which was resolved very quickly, which was numb hands as a result of hypokalemia. And then we just had three cases of skin. The most severe one was just a grade three that had involved itching, and that was resolved within eight days. So again, we'll take it.
But there's no such thing as a free lunch, and so what we've seen in China and what others have seen with these novel, more stable ADCs is, you get more ocular tox, specifically dry eye, conjunctival issues. But they are, unlike peripheral neuropathy, they are both preventable and treatable, typically eye drops. And it seems to be a hallmark of these stable ADCs, regardless of the target, and interestingly enough, regardless of what the warhead is either. And so the one difference what we've seen, Prakhar, between the Chinese population and Western population is, at baseline, half of our Chinese patients had those issues walking into the clinic.
Got it.
It's a combination of both pollution, adult smoking population, et cetera. In the West, that's obviously not the case.
Got it. And why wasn't Padcev tested in cervical or developed in cervical?
So this is speculative, but it sounds like we, we're pretty sure this is accurate. Seagen and now Pfizer also own Tivdak, which does target cervical as a label. Tivdak actually originated with Genmab, and so I think it is entirely likely that in the commercial relationship between Genmab and Seagen, which Pfizer has now inherited, there is a do not compete clause that says, "Please don't go into cervical with any of your other assets." And the reason we feel that's probably accurate is because Padcev was not tested. It's conspicuously absent from the 202 cohort. And neither Seagen nor Pfizer have since attempted to go into it.
Got it. And for your next update-
Mm-hmm.
Maybe what should we expect first from your side?
So from our side, we are wrapping up the dose escalation. This is our Western bridging study. We look forward to showcasing that in the first quarter of next year. That's not too far away. And then the next data set from China, which will be the dose expansion cohorts, that should be an update happening next year. I think traditionally we've seen Chinese companies really have strong preference for ASCO, very sensibly, and so that will be an interesting conference to attend next year.
Got it.
For many reasons.
Anything from your update in 1Q in terms of the patient population, number of patients, that investors should be paying attention to?
So it is a three plus three, four cohorts, but it is a BOIN design. In other words, we can go past three patients, so please don't expect just twelve patients, but we're not gonna get fifty or a hundred patients either. It's going to be more than twelve. We're delighted with the how the recruitment is going. And it is a basket, remember. We are gonna have a variety of known to be nectin-4 enriched tumors. One really important point that I wanna emphasize, 'cause we've seen this with another company, where I think that perhaps wasn't quite as emphasized. Chinese pharmas can do something that in the West we cannot do. They can bring in a CDX in the dose escalation already and exclude patients that aren't positive for the CDX.
You're not allowed to do that in the West. So in the West, we have to justify why we will be excluding patients based on a biomarker. And so in our dose escalation, we bring in tumors that are known to be Nectin-4 enriched, and then post-hoc, we will determine their Nectin-4 levels. Whereas in China, they could arbitrarily start with a level and then exclude patients based on that. And there's one of two outcomes for us. Either the CDX actually makes a difference, great, in which case, moving forward, we'll know who not to bring in, or we may end up with a case like Padcev ended up with bladder cancer, where there was no point in developing the CDX because pretty much everybody responded. That's fine by us as well.
Got it. And in the last minute, just remind us of the cash runway and the key milestones ahead for the company.
Sure. So the last time we disclosed, I think we had $148 million in cash, which is enough for just over three years of operations. There's only 29 of us, so we are pretty thrifty as an operation, and we like that. And then the milestones, a reminder, with CB-1, our milestone would be we'll have some more preclinical data towards the end of the year. There's a certain conference that happens in obesity that's probably worthwhile going this year. And we'll have our IND by year's end, and then start the SAD MAD early next year. And on 701, first quarter for our data and probably middle of the year for the dose expansion from China.
Got it. We're out of time, but-
Thank you.
Thank you so much for joining, and thank you.