Alkermes plc (ALKS) Earnings Call Transcript & Summary

Great. Good morning, everyone. Thanks so much for joining us. I'm really pleased to be joined by Richard Pops, Chairman and CEO of Alkermes. Thanks so much, Rich, for joining us.

Thanks for having us.

Yes. Maybe before we jump into your commercial business and your pipeline. I can start with a big picture question here. Alkermes has been undergoing this evolution over the last couple of years. Maybe just level set for us, what spurred this change? Where do you see the company today? And where do you hope to see it in the next 5 years?

Well, first of all, it's great to be here with you. It's great to meet you in person for the first time, and thanks for doing the work on the company. This company is a super interesting company because it's never really been the same for any long period of time. It's evolved from an early, early focus on very narrow aspects of neuroscience into a broader drug delivery company with a lot of deals with big pharmaceutical companies that drove a royalty business, and then taking those royalties and segueing into proprietary products business. So today, with $1 billion top line, all driven by our own medicines that we've made and marketed ourselves. A few years ago, we just decided that we wanted to keep moving higher and higher up the value chain. And if you think about it from the early days, when you're taking other people's drugs and improving them, that drives a certain amount of value creation. But the ultimate zenith of value creation in our business is making brand-new molecules, doing new things for patients that haven't been done before. And we got interested in this area of neuroscience associated with the sleep/wake cycle and circadian rhythm a number of years ago. And we had a hunch that would bear fruit. But of course, you don't know until you do. And the interesting thing scientifically has been the segue from our making technologies that would take other drugs and improve their PK or PD through formulation to everything we do now is based on our own chemistry and has been for a number of years. So one of our great technical strengths is in our chemistry. And that turns out to be an essential prerequisite to going after this orexin space. The molecular design is quite complicated. So it's been a continuous evolution. And coincidentally, and I don't think any of us could have predicted the certain macro environment we're in right now. For big pharma and small pharma with the premium now for new molecules that can command a higher price point without payer control for long periods of time, they've never been more valuable. So we're really pleased with where we're positioned right now because the commercial business drives substantial and sustained profitability that funds the pipeline, and we see the pipeline growing and expanding over time.

Great. Well, maybe to your point, $1 billion in top line revenue, yet, I think the focus very much this year is on your pipeline on the orexin portfolio that you have here. And you've referred to the space before the new white space for the field. Maybe speak a little bit more about what in particular makes you so interested? What makes you convicted in the orexin opportunity?

The short answer is data because the promise of the orexins became manifest from a neuroscience perspective, fairly recently. I mean the first papers were published in 1999, 2000, identifying the circuitry in the brain for the first time that actually drove wakefulness. There have been all kinds of development work, as you know, on the sleep side in service of having more wakefulness during the day. But now having discovered this bundle of neurons that drives wakefulness in the brain across multiple regions of the brain, the question was, could you replicate the natural ligand, which is a neuropeptide with a small molecule and drive that wakefulness rather than just helping people sleep better, so they're more wakeful. So it's a much more direct intervention with the circuitry that's relevant. So we and I think Takeda were really the leaders in this, demonstrating that indeed the small molecule agents could recapitulate what the neuropeptide does. And now where the program stands, I think between our data and others data is that it's very clear that these small molecule agents can drive significant wakefulness in NT1 patients. And the open question right now that I think will make a long way to answering in our Phase II data will be what's the tolerability like of doing so over time in these patients. And so far, the data look very good. So your conviction grows as you get more and more data sets under your belt.

Maybe to that point, before we jump into the Phase II and the expectations ahead of those studies, just remind us what you did see in the Phase I study here. What can we expect this week at the SLEEP meeting? And maybe help us put that into context compared to these other assets that are out there.

So the secret -- you can take different approaches to your early clinical work. Ours was informed by the fact that we made the decision that healthy volunteers who are sleep deprived really didn't serve much predictive purpose at all. And we knew that because we ran a Phase 0 study where we just ran through the methodology of how we would test an orexin agonist. And in that, we interrogated whether we kept patients up until 4:00 in the morning, whether that would drive sleep pressure and now like as to what you would see in NT1 patients. And we determined no. So what we did in our program was we went into Phase I in healthy volunteers in a single ascending dose format, which is typical and then moved very quickly into a multiple ascending dose format, which allowed us to credential multiple doses to allow us to go right to patients. So our Phase Ib study was conducted in patients with NT1, NT2 and IH. And in a very data-intensive crossover design, each patient received multiple doses, and we try to get a sense of the dose response that we see using the hallmark efficacy test, which is this maintenance of wakefulness study. And the data were quite clear, and that's what was so exciting about it. It wasn't ambiguous. In a dose-dependent fashion, we could see increases in the maintenance of wakefulness test across NT1, NT2 and IH, albeit at slightly higher doses for the NT2 and IH than the NT1. But we showed in that study that ALKS 2680 was highly potent once-a-day dosing in a dose-dependent fashion. That was all the information then we fed into our modeling to design Phase II, and Phase II was underway now. So at SLEEP this week in Seattle, you won't see any brand-new data from the Ib study that's been widely -- you'll see some new quantitative EEG qualifying that as a biomarker, you're looking at it in the brains of patients. And I think you'll see some panel discussions and just more education about the orexin pathway.

Great. So your Phase II study, Vibrance-1 expected to read out early 3Q, so coming soon. Maybe you can help frame expectations for what you would like to see from that study as it relates to MWT, cataplexy rates. The tolerability, as you mentioned, is going to be top of mind there. I would love to hear your thoughts.

I just think it's such an exciting moment now because this is why we develop new drugs, is to get new data sets like this. So when I think about it in the aggregate between Vibrance-1, 2 and 3, we'll have circa 300 patients with new information about orexin-2 receptor agonism in a whole range of patients and making a contribution no one's ever made before, which is this whole range of doses rather than just a single dose that might be efficacious. So for Vibrance-1, Vibrance-1 is the NT1 patient population. It's a much more monodisperse patient population. So we expect data very similar to what we saw in the IB, that is at the doses we've chosen, which are 4, 6 and 8 milligrams, we expect to essentially bracket the doses that we might end up using in Phase III. Of course, we won't know until we look at the data, but the expectation is that we've created a range of doses and that gives us the ability to model then for Phase III, what the more precise doses are. So the goal is at the lower dose to have activity, but really, really excellent tolerability. At the highest dose, maybe we have maxed out what we see and have more side effects and define what the edges of those parameters look like to model for Phase III. Vibrance-2, I think, will be even more interesting in a way because it's a much more variable disease. And there's no predecessor data of efficacy of our orexin-2 receptor agonist in NT2 because the Takeda study was not successful. So we have a higher range of doses, 10, 14 and 18 in that patient population. And when we get those data, then we'll get a sense of how important the variable doses are because the disease is much more variable in that case as well as with IH. So I think in the aggregate, this will be the first real major contribution of once-a-day dosing, a range of doses and a range of diagnoses, looking at a 6- to 8-week double-blind period, looking at really fully elaborating tolerability profile and I think incrementing the field in a significant way.

And then when you think about your NT1, for Vibrance-1, when you think about MWT, I mean, you're seeing with Takeda's asset, you can get pretty close to 40 minutes on the MWT. Do you think that's necessary? What is the appropriate maybe threshold that you'd like to see across your doses?

For me, the most important thing is not the absolute number, it's dose response. Because if you have dose response, you can essentially dial up the duration and different patients will have different clinical presentations and also different desires. So in a 40-minute MWT, I think the important thing to understand about the test itself, it is actually a little more complicated than people often realize. The MWT is run over 8 hours. So it's actually a 40-minute test conducted every 2 hours over 8 hours. So the number that you see is an average of that delta from baseline across 4 different tests over an 8-hour period. So it doesn't tell you much about the shape of the curve. It doesn't tell you about -- a lot of stuff is buried in that number. The biggest drugs in the field right now move an MWT latency for NT1 patients, which might be under 5 minutes and move it into the low teens at best. And as you said, the orexins can almost max out the test, whether you can max it out early for the first few tests or all the way and keep people up all night. So I think that what I want to see in our Phase II is what we saw in the Ib, which is dose proportionality, pushing the MWT to limits that no one's seen before with oxybate. And we know from patients that this is not just the only measure. The MWT has the virtue of being a number. You can do stats on it, it's objective. But it tells you nothing about the quality of the wakefulness of that patient. And I think in the fullness of time, this will be the virtue of this category because what we hear from patients in the Takeda study and our study is the quality of the wakefulness compared to being on a stimulant or other ways of driving wakefulness per se. So if this is indeed the natural circuitry in the brain that drives overall wakefulness, that applies to cognition and to vigilance and to mood and all kinds of other domains that we're actually going to learn more about in these studies as well.

Got it. Maybe one point there. Just remind us what measures you are looking at within these Vibrance studies to be able to detect the quality of wakefulness or these other domains, cognition, mood.

So in the first tier that you'll see at the primary analysis of the primary endpoint at 6 weeks. So the first top line you'll see, that will be the classic MWT and then the Epworth Sleepiness Scale and the cataplexy rates as captured by the patients on diaries. That triad is the classic. Beneath that and the FDA recognizes those. FDA does not recognize the PROs, and we and Takeda, I think, are going to do a lot of work to try to validate some of these in the context of this disease. So we'll be looking at a whole range of PROs you'll see in a secondary set of analysis later in the year. But we'll be looking at -- nothing is bespoke. These are off-the-shelf measures that have been used in other studies.

And then as you think about the AE profile that you would be comfortable with and that physicians have spoken to you about being comfortable with, maybe talk to us about that, particularly in the context of what you did see in the Phase Ib.

What's remarkable about the class as a whole so far [indiscernible] is that given how potent these molecules are and how profoundly they're affecting key circuitry in the brain, I think categorically, we can say that the side effect profile has been mild-to-moderate and transient and quite advantageous at this point. We know from our data and others that the hallmark AEs that appear are insomnia and polyuria, which is urinary urgency. Those, based on the Takeda 8-week Phase II data, appear primarily in the first week and then diminish as patients get used to being on this new therapy, tend to be dose-dependent and transient, mild to moderate. Our data set will provide even more information because we'll be able to push the doses higher and see what we run into as we push the dose higher. It may just be the same things with more frequency and severity and new things may emerge. We'll see when we unblind. But we have a data safety monitoring board that's looking at unblinded data through all our studies. And so obviously, we keep rolling with a green light on those. So I'm hopeful that the way that the class will emerge will there will be -- all drugs have side effects, but they will be as described, mild to moderate, transient and not leading to patient discontinuations and counterbalanced by the quality of the wakefulness that patients are experiencing, which is perhaps different than they've ever had before.

So just to confirm, no discontinuations in your Phase Ib from these AEs.

No.

And if they stayed mild, moderate, transient, do you anticipate...

[indiscernible] that I'm going to look at Sandy and Blair. Thumbs up. [indiscernible].

Do you anticipate that would impact uptake in a commercial setting, if it was truly just mild, moderate and transient.

Well, think about what patients are doing right now with stimulants and oxybates and not getting the efficacy that they want. So absolutely, I think that's -- I fully expect these drugs to have side effects, I fully expect them to be dose dependent, and I fully expect them to have efficacy that has not been seen before. But that's why you do the Phase II to make sure you can confirm that. But there's so much excitement in the community. The reason I'm going to leave here this evening to fly to Seattle to go to SLEEP conference is will meet with all the patient groups, narcolepsy and IH as well as the KOLs. And if you compare it to a year ago in Houston and 1.5 years ago to Brazil, the field has moved from talking a lot about oxybate and sleep apnea machines to orexin. And it's because more and more sites are using them in clinical trials even on a blinded basis and getting hands-on experience with their patients and what they're doing.

How concerning are the visual disturbances?

Well, right now, there's not enough data to make a [indiscernible] statement because we had one occurrence in our NT2 population and one occurrence in our IH population at doses we're not testing in Phase II. But that doesn't mean we're sanguine about it. And the point is, there's a very logical case as to why some type of pupillary or ocular thing could be an on-target effect of orexin-2 receptor agonist as you raise the dose. So in Phase II, what we're doing is in contrast to 1B, where we had no baseline ophthalmic understanding people could have had glaucoma or intraocular pressure changes, we didn't know. But in Ib, all patients have a baseline ophthalmic exam. And to the extent they have any AE, they'll be worked up in a correct way to try to understand if there's any basis for it. But we're blinded, so we won't know until we unblind, but I think it's a really important question.

And then maybe just quickly touching on NT2 here and IH, you've spoken about the patient population is much more heterogeneous than NT1 gives you the confidence that you'll be able to actually detect out a signal from these populations since the patient baseline characteristics may be so different.

The primary reason we believe that is the Ib study. We've done it already. And in that study, it's important to note patients in any way. I suppose there's a way of running that Ib study where you could have hand selected certain NT2 or IH patients that look more like NT1 patients, i.e., with small low sleep latencies. We didn't. We took all comers and we saw a reliable dose response at these higher doses. So our modeling turned out to be quite correct and where we thought the doses would have to go. So in the NT2 Group, the MWT, recognizing there's a 40-minute test. To make the numbers simple, let's say that in NT1, your average baseline is 5 minutes. You have a 35-minute dynamic range then to play with. Let's say, in your NT2 population, let's say the baseline average was 25 minutes. Then of course, you only have 15 minutes to do it. And the 25 minutes probably doesn't represent a central tendency. It's probably just an amalgam of somebody's got 5, somebody's got 30. So it's a much more heterogeneous population. So that's why we love the Phase II design. We'll just look at the data and see what it tells us. But because we can drive -- if this circuitry in the brain is driving wakefulness, NT1 think of as a knockout model. You knock out all those neurons, what's the phenotype. Basically, you're fighting in your sleep all the time. But the QEEG data from healthy volunteers as well as the data from NT2 and IH has suggested that, call it, even supraphysiologic doses that administer hit this receptor and drive the circuitry, they'll drive more wakefulness. And so I think our belief is if you see a strong signal in NT1, you should see a signal in NT2 and you should see a signal in IH and go beyond that. Maybe then you should see signals in patients with other psychiatric and neurological disorders.

Great. Maybe one last question here for your Phase II studies. They all embed kind of a post dosing or I guess, post -- there is a 6- or 8-week dosing regimen where the patient can modify their dose. Maybe speak to us about the importance of that, what you hope to learn from it and how it will help inform your Phase III?

This is also a new knowledge that predecessors haven't been able to do because without a range of doses to play with, you can't ask this question. So what are we doing? We have a double-blind period, 6 week in the NT1, 8 week in the NT2 and IH, where patients stay in their dosing lane. So they're randomized to 1 of 3 doses or placebo. They stay in that lane for that fixed period of time, double blind. The primary assessment is at the end of that double-blind period. Then all patients go, if they want to continue, they move to the middle dose. That would be 6 milligrams in NT1, 14 in the NT2. And they stay on that dose and they know they're on that dose for a 2-week period. And then for the subsequent 5 weeks, they get to choose, do they want to go up or down. And that will help answer the question about whether or not maxing out the MWT is always the greatest thing to do or do people have different individual preferences. That will be additional information to help us select doses for Phase III.

And then maybe if we can switch to the commercial opportunity or how you see this market evolving? I think the company has put out numbers in terms of the prevalence for NT1, NT2, IH patients, and it's quite large, yet a fraction of those patients, only a fraction are treated. So how and where do you see an orexin agonist fitting into this landscape across all 3 indications?

This answer is informed by more and more conversations we're having with physicians, and we'll do more of it this week at the SLEEP meeting. So I think the common response we're hearing is that there will be a very deep penetration in NT1 because it's a disease-modifying therapy. You're replacing the deficient neurotransmitter effectively. And the quality so far of the data that we've seen is sufficiently advanced over what perceived and the tolerability profile of the things we've talked about and because the clinical presentation of the NT1 patient is pretty consistent. NT2, you hear ranges 30% to 50% of patients might be targets for orexin. Why is that? It's not that it's necessary that you're not driving efficacy in those patients, but these will be expensive drugs. And for many NT2 patients with fairly modest disease, they might do well just with generic modafinil or something like that. So they won't need all things that an orexin-2 receptor agonist might bring with it. But I think that will really depend on the data we need to see for NT2 and IH. I know originally, you may recall, we weren't going to develop 2680 in IH, keeping it only in the narcolepsy indication. But after we got the Ib data, we met with a lot of the thought leaders in IH, which had overwhelming response. There's a huge unmet need. There's a paucity of drugs. These patients need something. So we started -- it really makes sense to extend to that as well. So I think it will be driven by the quality of the data, but I think that it starts with NT1 and then ripples out from there.

When you think about these markets, is there a scenario where orexin, to your point, NT1, you see deep penetration there. Is that cannibalization of the existing drugs? Or are you essentially growing the market because so many of these patients are not being treated right now?

Yes, that's a critical question. I think a lot of investors haven't looked at it carefully enough, someone say what fraction of the oxybate market will you take? It's important to understand that oxybates are not the modal way of treating this disease. Of the 80,000 patients who are treated for narcolepsy in the country, about 16,000 get oxybates. And a lot more of them have tried oxybates and decided not to use oxybates. Oxybates are very difficult to be on for a sustained period of time. But for the patients who found their way to them and use them, they like them, and they're going to continue to use them. I think many of them will. So most of the market is underserved. And I think that those 16,000 patients drive a couple of billion dollar market. So I think about the inverse, you've got 64,000 patients who aren't getting everything that they need. And I think that's why I think the numbers get so exciting from a financial point of view, if you think about having a disease-modifying therapy in this category.

Disease-modifying. So do you see scope for combination use with other agents? Or is this truly somebody will opt for an orexin agonist as monotherapy?

To be determined. I think that in NT1, it should become frontline monotherapy. But then the question whether the oxybates. And I think that remains to be seen. In our Phase II study, we are going to explore using polysomnography in the double-blind period, what the quality of sleep architecture looks like for somebody who's on an efficacious dose of an orexin-2 receptor agonist. That's new information. We don't know that. One school of thought says, if we drive meaningful wakefulness for 12 hours in a day, you should sleep better, should consolidate sleep better. That's hypothesis. The other hypothesis is that for the patients who really like what the oxybates do, i.e., really guarantee you go to sleep at night and defragment your sleep in a way, there'll still be a place for those. And I think both are possible.

And then as you think about the commercial opportunities, you mentioned IH has relatively fewer options compared to NT1 or NT2. Would it make sense to focus your marketing efforts towards specifically IH? I mean you've done this with VIVITROL as you think about alcohol dependence versus opioid dependence. Is that a strategy you would look at?

I don't think you need to. I think the core of the bulls eye is going to be NT1. It's the most explicit demonstration of the disease and a huge unmet need. And I think as clinicians see that this agent can be used in their NT1 patients, it will naturally flow into these other diseases of hypersomnolence. I think it's so important to try to get a label that accommodates all 3 because the differential diagnosis in the real world is quite complicated. People don't run MWTs in the real world. And in fact, they don't put people in the sleep lab in the real world. The reimbursement just -- people are treated based on the clinical symptoms as they present. And the clinical symptoms between NT2, IH and even NT1 sometimes can be almost indistinguishable. The hallmark distinction between NT2 and NT1 is cataplexy. But what we've learned is that cataplexy can be explicit when you lose muscle tone with high emotion to be very subtle and not even detected necessarily by the patient. So that NT1 patient may look just like an NT2 patient and vice versa sometimes because we have NT2 patients with sleep latencies under 5 minutes. And then IH, the differential between NT2 and IH, we're finding in our clinical trial, sites will say we have NT2 patients who enroll, they come in to qualify the patient, patient has IH and vice versa because the distinction on that differential diagnosis is the presence or absence of a certain number of these sleep onset REM periods, which can only be detected via EEG in a sleep lab. So that isn't happening [indiscernible]. People are showing up with a clinical presentation, excessive daytime sleepiness and the doctors have to make the call. So if you have a drug that's only indicated for one, where the payer says, we have to prove that it has, it's much easier if we can just sort of run the gamut and say NT1, NT2, IH range of doses go for it.

Got it. Maybe following up on that, as you think about what the potential label could look for 2680, you've now touched on the potential to be used across these 3 different indications, multiple doses, once daily. What are the puts and takes to be potentially second to market behind Takeda?

Well, I think Takeda deserves an enormous amount of credit for being the first mover in the field and doing it responsibly and publishing the data and just doing it correctly. Their drug is, I think, a proof of the pharmacology, but it's an incomplete product because there's -- it's given twice a day in an interesting regimen because it's given once in the morning and then another dose 3 hours later, I believe, 3 to 4 hours later. So it's not a classic BID dose. It's given in that way. And it can be indicated only for one of these diagnoses with NT1. And so I think it proves the point, but it opens up a huge vulnerability from a commercial perspective for a drug that could be across all the indications with a range of doses given once a day. And hopefully, that's what we have. The admonishment always is you have to wait for the data. You have to see what we have, and there's always risk in developing small molecule drugs, particularly for the brain. But so far, so good and what's so exhilarating is that -- this is why these companies -- within the company, we've been waiting years to get this data set to see what it looks like. And I think it's going to answer a lot of questions.

Maybe in the last couple of minutes, we can touch on your commercial portfolio. But before we do, just quickly, as you think about 2680 and commercializing this, recognizing that you have commercialized your products exclusively in the U.S., is this a strategy you would also look to employ with 2680? Or would you actually take on global commercialization?

Here's what you don't do. You don't launch your drugs at a lower price than in other countries. 2680 for us, from a commercial perspective, is a breadth of fresh air. In our existing portfolio is largely Medicaid and Medicare, where the payers put up enormous barriers to access, gross to nets are very high, restricted formularies, multiple competing drugs. This is a completely different space. So it's a space where they come in high prices, disease-modifying therapies, more analogous to some of the biologics than traditional CNS small molecule drugs. And we can absolutely develop it for Europe ourselves and Asia. I can see in Japan partnering for a number of reasons. But I think we're thinking about this drug on a more global basis. By the way, the reason we don't commercialize VIVITROL or LYBALVI or ARISTADA in Europe was because of the price differential because Europe was just unwilling to pay a price that was either directly relevant to the U.S. price at all, which is a fair U.S. price. These are not super expensive drugs. And it just didn't feel like it was either economic or fair to do that. And I think that decision was a good decision.

And just quickly, the Medicaid Medicare exposure for NT1, NT2 and IH patients.

Low. This is largely commercial [indiscernible]. I think we modeled it. I'm looking to Blair, somewhere around 80% commercial price...

And maybe in the last 1.5 minutes that we have here on your commercial portfolio, anything you'd like to share here as you think about the growth trajectory moving forward for these 3 products? And then what are the puts and takes to achieving your EBITDA guidance for this year?

I'm really pleased with where we are on the commercial portfolio and the way we've sort of aligned with the Street on where it is and where it will go. So it's taken away a lot of volatility quarter-to-quarter as people -- we've guided clearly to it. We're in that lane. And there's just -- in this type of environment to be able to rely on a multiproduct $1 billion top line that's profitable as a source of capital to fund our business. It's rare. And so I think that the -- I often say that we live in the postapocalyptic world at Alkermes because we deal with government price controls all the time. We deal with restricted formularies. We deal with high rebates to goverment system. And we can thrive in those systems. It's just a different business model than selling very high priced medicines to very small patient populations, which is a lot of biotech. So I think that for LYBALVI, it's still in a really nice growth phase. ARISTADA is a long-acting injectable atypical antipsychotic, benefiting from the fact that we've expanded our sales force this year. And VIVITROL continues to surprise after a long time in the market, it continues to grow largely in the alcohol indication.

Great. Well, with that, Rich, thank you so much for joining us. Thank you, everyone.

Thank you.