Good morning, everyone. Thanks for joining us here at the Leerink Partners Global Healthcare Conference. My name's Tom Smith. I'm one of the senior biotech analysts here at Leerink, and it's my pleasure to welcome our next company to the stage, Alumis, where we're joined by CEO Martin Babler and CFO John Schroer. Gentlemen, thanks for joining us.
Thank you. Thanks for having us.
Of course. Alumis is one of our top picks for 2026. We see this as like a huge year of execution for the company. Obviously very impressive top phase III psoriasis data for your TYK2 inhibitor and envudeucitinib presented in January, we're looking forward to seeing details from that study here, perhaps at AAD towards the end of March. Perhaps.
Yeah. It's on the AAD website.
It's on the AAD website.
There you go.
... we'll definitely be tuned into AAD. A year of additional important data points in for you, phase II-B readout in lupus, some other external potentially validating data points coming for the TYK2 class. A lot of data-
Yeah.
... coming this year. With that, maybe Martin, maybe if you could just recap, I guess kind of the last 12 months and what you're looking forward to most over the next 12 months, and then we can go into some Q&A.
Excellent. Thank you. Alumis is a precision immunology company. We were actually founded really on the concept of doing a lot of genomic work and then translating that into the clinic. As one of the key targets that we discovered early on as being really meaningful in immunology, is TYK2. We pursued that target with envudeucitinib. We created a second molecule with A-005, which is our brain-penetrant TYK2. Over the last 12 months, we've really advanced both of those molecules and created a pipeline behind it. We also last year did a merger with Acelyrin. That gave us additional cash, really with the idea that it would help us read out some key data sets. The first one was the psoriasis data that we read out in January.
The second one is the lupus data. With the Acelyrin merger, we really were in a position that we could read out both of those and still have a cash buffer after that because we felt it was meaningful to really re-establish the TYK2 class in a different way. The reason why we did that is because from very early on on the genomic side, where we saw that people that are homozygous carriers of a TYK2 mutation that down-regulates the kinase function, they actually had a way stronger protective effect than the ones with the heterozygous carriers with one allele only.
To the pre-clinical data, to the phase I data, to the phase II data, we could always show that if you really inhibit TYK2 very well, you should get a very, very different outcome than we've seen from the first-generation TYK2 inhibitors. That's exactly what we demonstrated early in January when we released the top-line data for our TYK2 inhibitor, envudeucitinib, in psoriasis. The key metric there really that we shared is the 24-week data for PASI 90, PASI 100, which is really what the clinicians look for to understand the efficacy of a drug in psoriasis today. We shared that our PASI 90 values are about on average 50% 65%, sorry, and that our PASI 100 values were above 40% if you average out the two studies.
Which really means that from a PASI score, that is at the top end of oral therapy today. We have also shared that we have really interesting data on the quality of life and the itch components of the study. We haven't shared that data yet. That will be shared at AAD. The AAD has made public, as I just mentioned on their website, that there is going to be a late breaker involving envudeucitinib, so I can share that today this way. There will be a press release about that, but that's still embargoed at this point. The other piece that we shared is that we actually have a very fast onset of action. We saw separation of PASI 90 and PASI 75 already at week four from placebo.
We do believe that really demonstrates that this is a class to be reckoned with, because it has not just good skin clearance, but has other features that might not be seen with other classes in oral therapy. Overall, we were very pleased with the safety. We reported that it's very consistent with our phase II program, where we had the phase II 12-week data set and then the long-term extension, given that this is also six months of data and not just three months of data. Of course, the numbers are slightly higher with the than just a 12-week data set, but very consistent safety between the two trials that we've conducted. Very pleased with the profile that we see for envudeucitinib.
That's great. Yeah, let's dive into some of the psoriasis details and you mentioned, yeah, we think this is truly top tier efficacy among the orals, but maybe you could just level set that and contextualize it versus some of the others. The Takeda next-gen TYK2, the J&J Protagonist oral IL-23, and I guess what, you know, how we expect to compete against those two agents while acknowledging we haven't seen all the details, certainly for the Takeda compound.
I can't really comment on how we stack up against the Takeda compound because I can only speculate at this point because there's just not enough data. For us available. Against ico we would say that there is an interesting aspect that we basically have an IL-23 plus inhibitor. There is clearly the effect that we see with an oral IL-23, where we're very competitive, maybe on certain aspects, even slightly ahead. When you look at the overall package, we do believe that once you look at additional aspects, whether it's quality of life, whether it's special area, whether it's basically the aspects of itch, we do believe that there actually is going to be a very competitive profile with our molecule as well.
We do believe that for envudeucitinib, we have a profile that stacks up very well, and ultimately, though, it's important to understand that we also believe, like our competitors, that the oral market will increase quite dramatically. There's room for more than one molecule in that market. The reason why we believe it will evolve quite dramatically is even today, actually, there is more patients on oral drugs in psoriasis than there are actually on biologics. The majority of patients are on topicals. What we believe you will see is that patients will move from a topic to an oral faster than they have in the past because fundamentally, physicians know that IL that psoriasis is a systemic disease.
What really is going to happen is that you will have a systemic therapy for a systemic disease. At the same time, what's happening today is that most patients that are on oral therapy are only on it for about six months and then move on to a biologic. That will also change dramatically because these drugs work really well. We have, from our phase II, still the majority of patients that were in our phase II actually on our oral drug three years later. Patients will stay on oral drugs a lot longer before they move to a biologic, and that will also basically squeeze the biologic market.
The third one is that there will be patients that we believe based on our market research, and I think it's confirmed by our competitors' market research as well, that will switch from an injectable to an oral drug as well. Overall, we believe actually the most important thing to understand is that a large growth of the oral sector of the market and there really opportunity for more than one drug, and where we believe it's mostly going to be distinct segments and preferences for physicians on which one they will use. Within that segment, we will have a very competitive profile.
That makes sense. Mechanistically, I mean, the story from the beginning was basically like greater target inhibition, more target coverage, longer target coverage. It seems like that really played out in the top-line efficacy data that we've seen. May you just talk about, I guess, like your perspective of how that played out on efficacy. Did it play out at all on safety? Is there any safety tolerability things associated with that thesis? Then how do you expect this to manifest in the long-term 52-week data we'll get later this year?
Yeah. A couple of aspects. The first one is if you actually look at the development of this drug, we can literally draw a line from all the way to genomics, to the in vitro preclinical, to the animal preclinical, to the phase I, to the phase II, and now to phase III, where more inhibition leads to better outcome. We've always made that point, and I think there's very few cases where it's so clear across everything. Interestingly, to your point, the safety aspect of that is very interesting because we know from the genomics that there's no phenotype for people that have this mutation that downregulates their kinase function. We do believe that that safety is not necessarily associated with that, but these are all kinase inhibitors, and so the question is, how selective are they?
One of the fascinating things about TYK2 is that there was a time a couple of years ago where we were so obsessed about the TYK-JAK relationship and selectivity that we've all forgot to look at the whole kinome. What we see now is that basically, the selectivity across the kinome probably plays also into the safety. Despite the fact that we have more inhibition of TYK2 as a target, we actually have not seen a worse safety profile because just inhibiting TYK2, we know from the genomics at least, that should not really lead to a lot more safety signals.
I think that's when you look at our profile right now. Remember, we were the only TYK2 that did not have to dose reduce from phase I to phase II phase III because we did not see safety issues with our molecule. I think that has borne out at least all the way phase III right now. The other aspect really is how does this ultimately translate into long-term data? We don't know yet, but the one thing that we're excited about is that ultimately, and we've shared this publicly, that when you look at our lines, especially for PASI 75, we expected, like in every case, for effective therapies in psoriasis, that between week 20 and 24, you probably see the peak.
For PASI 90 and PASI 100, you've seen in our phase II data, and we see that phase III as well, is the lines have not yet flattened out. We hope that we will see a continued evolution of the numbers, and we'll see that when we see our long-term data. We shared that we will issue top-line data in the second half of this year for our 48-week data set, basically.
Yep. long-term data in the second half of this year. You've also guided to NDA filing in the second half of this year. can you just remind us what that filing package will include? You had the long-term extension experience from the phase II-B. Do you have to wait for the long-term onward data to submit? just remind us, like any other gating factors to submission.
As you might remember, we have a little bit more of a creative way that we designed our phase III, mostly because we wanted to make sure that we have a chance to really get a chance to catch up with some of our competitors, at least from one data set. The reality is we have to still do the withdrawal study because part of the label, we want to make sure that we can show durability and we can show maintenance. While our competitors have the complete data set by the end of last year, our complete data set will be available to us at the end of the second quarter, early third quarter of this year. We basically will submit in the second half of this year.
The package will include, basically the phase III data, including the withdrawal portion, and the safety. We also have, fortunately, an extension of our phase II. At the time of launch, we will actually have three-year safety data, which is something that nobody's had before us, and hopefully we'll get that in the label as well.
Great. You've alluded to this. We talked a little bit about how you see the segmentation between orals and biologics and maybe within the orals a little bit of the segmentation. Is there one sort of like key driver, decision-making driver you think there will be for clinicians among the orals? Like, should we be looking at higher levels of response, PASI 90, 100? Should we be looking at durability, dosing convenience? Like, how do you think about that?
This is going to be a very interesting dynamic because we do believe that there is an understanding that physicians currently have about psoriasis drugs and what oral drugs can do and there is an opportunity here as well. Fundamentally, the skin clearance is going to be a key driver. When we talk to thought leaders and dermatologists today, they will tell us that the way they evaluate drugs today is they look at the PASI 90, the PASI 100 at week 24. That's the key standard metric that they use. However, there are also other aspects that we hear clearly from clinicians that matter. One of them is special areas.
The other one, when you actually ask physicians what they're looking at the first three criteria physicians and patients both mention is efficacy, safety, and then itch. We already know from the head-to-head of the IL-23 versus the TYK2, that their IL-23 beat TYK2 on many aspects. In one area that actually IL-23 was not as strong was around itch, and that's the number one symptom patients want relief on. We know that the IL-23 plus mechanism that TYK2 has actually, really one of these aspects is that itch component, where the itch actually gets resolved a lot faster with a TYK2 than with an IL-23. We will see.
Ultimately, we'll share that data at AAD, but we believe that that will play a role because it feeds into the quality-of-life element as well, and that's where we've also seen a differentiation, basically. I would say special areas, the itch, the quality of life component, and then, ultimately, the PASI scores are all elements and the safety. Given the safety of these drugs, compared to broad immunosuppression, we do believe the safety will play a role, but probably a subordinate role compared to some of these other aspects.
Great. Looking ahead to AAD, you're orienting us to the itch data, that will, I think, obviously be very, very interesting. Maybe just help frame some of the other datasets, detailed datasets we can expect to see. I guess I would expect, kinetics of response, responder curves, detailed safety data. Anything else that you would kind of highlight and orient us to be tuned into as we get that late breaker?
Yeah. Fundamentally, I think it's your basic, you know, PASI scores. We'll talk about quality of life. I think the kinetics of response, especially the speed of onset is an important one. Also, as you alluded earlier, how do these curves look? Do they still look like they're pointing upwards versus they flattening out? Those are probably important. Then there's a couple of other key elements. The reality is these late breakers are 10 minutes in length, so you try to cram as much into it. We're working very hard to also have a publication out relatively soon, as that has a lot more information in it. Then we'll continue to educate.
The areas that we are not able to incorporate in the AAD presentation to the detail that we want is, because it's just not yet done completely is some of the special areas and then the biomarker work. As you know, we started out by saying we're a precision immunology company, so we've done a lot of biomarker work. We've also identified actually a biomarker specifically related to TYK2. When we looked at that in the phase II, there were just not enough non-responders. We have the same situation here a little bit, but we'll certainly look at, is there a correlation between that biomarker and efficacy ultimately? There's some other things that we're doing. That will come later, but you should get a really clear picture of what this molecule looks like at AAD.
Interesting. Just on the biomarker potentially predictive of response, who's going to respond to a TYK2? How should we think about that?
This is a really interesting biomarker that we are working through whether we can ultimately predict response. It's more likely that this will actually be interesting in the lupus dataset than in the psoriasis dataset, because to figure that out, you have to have non-responders. I think one of the things that we emphasize a little bit on the phase II and we are now going to work through a phase III, we think about responders in psoriasis as PASI 75 or higher. The reality, though, for a patient that has not had a lot of response to many other drugs and finally gets something, and it's even just PASI 50, that actually is meaningful.
When you look at the waterfall plot from our phase II at the 40mg BID dose, actually, there was not a single patient that did not respond. The challenge with the responder analysis in psoriasis really is that if you have no non-responders, then you can't really define it. We're still doing that work, and that will be work that we'll share later in the year or next year.
Interesting. Okay. We'll stay tuned for that. I wanna ask about the-- You're working on a once-daily formulation.
Yep.
How important is it for you to get to once-daily? I guess, what's your level of confidence that you will get there, and what's the latest on that front?
Yeah. The first thing is that I think it's important to understand that the once-daily certainly is preferred by patients, but it's not a 1-0 equation. In fact, what we saw that when we asked patients, "Would you rather take a twice-a-day drug with no food effect versus a once-a-day that actually has a fasting requirement?" 2/3 of patients tell us that they will take it twice a day any day before they take the other one. So it's meaningful enough that we have a big effort going on there. The one piece I wanna make sure that people understand is that we have now shown that basically the trough level of a TYK2 inhibition is so critical for efficacy. Normally, when you do a once-a-day formulation, you actually want the Cmax and the AUC optimized.
We have to optimize Cmax, Cmin, and AUC. We actually have succeeded at this, and we do have a once-a-day formulation, but there's one aspect of the profile that we did not like, and that's why basically we're refining it. To your question, we will have a once-a-day formulation. This will be a once-a-day drug. The question is how fast can we get it, and that actually depends on the engineering solution that we need to basically get that one kink out.
Once we know what we're using to get that one kink out, I can give you an exact timeline. It is not gonna be at the time of launch. It's basically a lifecycle extension, which will also help us get additional patent life. Ultimately, it will be once a day. We do believe with the profile that we have, we still have a very competitive molecule, even with a twice-a-day formulation.
That makes sense. Let's talk about, I guess, potential launch timing and potential launch plans and how we're thinking about potentially scaling like a commercial organization to be able to launch in psoriasis versus partnership or other business development activities. Like, how are we thinking about monetizing the incredible data set that we have in psoriasis?
Let me first say that we have a lot of internal capabilities and capacity for launch preparation of what needs to happen right now. We are doing all the work in preparation for the commercial launch. The key question for us is twofold. The first one is, when do you scale up, and how do you scale up? Do you scale up internally or externally? In that context, I think the more important question is not psoriasis. The more important question is we have a huge opportunity here by having a TYK2 franchise with two really great molecules, and there's about 20 indications that we've identified. The likelihood that we, as a single company, are going to develop and launch on a global level about 20 indications for TYK2 is extremely slim.
For us, the key question is more: when do you partner? How do you partner? How do you identify a partner? Does the partner have the same vision for this franchise as we have? For us, it's more about, do we wait for the lupus data to have those conversations, or do we have those conversations right now? Then do we ramp internally, and how much do we ramp internally versus how much do we ramp externally?
That makes sense. Let's talk about lupus. We've seen some initial proof of concept for Bristol's TYK2, so TYK2 in lupus. Maybe just if you could apply, I guess, a similar sort of like mechanistic rationale, what drives your confidence of success for envu and lupus, and maybe talk about trial design and I guess how we've optimized around maybe some of the deucravacitinib experience with respect to trial design.
Yeah. We actually are really excited about lupus for our envudeucitinib program. The reason is that there's really three scientific rationales and validations. The first one is the genomics for lupus look very similar to the genomics of psoriasis, and we know now that our genomic analysis in psoriasis has very much panned out. The second one is, as you mentioned, there is data from deucravacitinib, another TYK2, where they actually, in every single dose, saw a clinically meaningful and statistically significant response in lupus patients. The third one is that the mechanism that we're using in lupus is really the interferon mechanism, and we know from anifrolumab that that is a very effective mechanism in lupus, one of the only two drugs approved in the disease.
Those three points really validate that TYK2 should work in lupus. In terms of design, we have specifically looked at a couple of elements that we believe are important. The first one is the biggest issue in lupus trials is always the placebo effect because it's a waxing and waning disease. How do you optimize for that? The second one is what is the patient population is more likely to benefit? That one, we know from the anifrolumab data that actually patients that have skin involvement have responded better to anifrolumab and the interferon pathway inhibition than others.
We actually are enriching our patient or have enriched our patient population for patients that have SLE but have a skin component. The third one is really how do you measure success and how do you pick the right endpoint? That's where we actually believe that picking BICLA as the endpoint is a better choice in this population, because in BICLA, while it has more domains and more points to measure, actually it has a better and more discreet way to measure skin manifestations and the resolution of them. Instead of just having a one zero, you have it or you don't, it actually is a gradual identification. We believe all of these things matter.
We have done two other things to really make sure that we have as good an active effect and as a smaller placebo effect as possible. The 1st one is we basically made sure that patients really have active disease, so we have an adjudication panel that looks at every patient at the time of the randomization and at the time of the first dose, that they have active disease and hopefully are on the waxing phase and not on the waning phase of the disease. We also made sure that we reduce especially steroids as a concomitant med as much as we can. There's a four-steroid taper, and if you don't match that taper, you're basically a non-responder. The third one is really making sure that the physicians and the investigators understand the tools that we're using really well.
One of the biggest issues in lupus actually is that the tools that we're using are very, very crude. We're doing that in two ways. One is we did a lot of education around those tools with the people actually who invented the tools in the first place. The second one is that we're crawling over the data all the time with a crawler that we built and basically look at inconsistencies in the dataset. If there's any inconsistency in the dataset, we have two things that we can do. We can immediately go back to the site when it's still fresh in their mind of what actually happened and why is there inconsistent data here.
That turned out actually some of the big issues in previous trials, where people waited at the very end to adjudicate and didn't realize, why is it that, you know, the way this was measured versus that was measured was not consistent. The second thing that we do, is basically we go back and re-educate folks if we feel like that's necessary.
Got it. Okay. Essentially more real-time adjudication.
Very much so, yes.
some of the data. Okay, excellent. Just help frame for us how you think about sort of bar for success? Like I guess among investors, we kind of debate, you know, does envu need to look better than dukor? Obviously, the primary endpoints are different, but on a somewhat apples-to-apples basis, as much as you can in lupus, like, how do you think about what a good dataset looks like?
The way we actually think about it is that as long as the dataset is as good as the anifrolumab data, but an oral drug, you already have a drug in lupus. That's the lowest end of the bar. Ultimately, the question with dukor is a slightly different one because we know that dukor actually doesn't have linear kinetics and has a lot of variability. One of the issues with dukor is you push the dose, but you're not necessarily getting more exposure, but you're getting more side effects. The key question there is ultimately what the efficacy and the safety and therefore the patient benefit looks like.
That's a little bit hard to predict given that these are doses that have not previously been studied in large populations, and so we'll have to see what happens there. Our prediction still is that more inhibition of the target will lead to a better outcome, and that's the piece that we'll just have to see once we see the data, because the one harder piece to predict is that we're not using the 6mg QD or BMS is not using the 6mg QD dose in lupus, but it uses the 3mg BID dose. So that will lead to a slightly different outcome.
Yep, that makes sense. You called out, you know, evaluation of roughly 20 indications where we could potentially take envu. We're gonna see some, I think, very interesting, hopefully informative data towards the end of the year from Takeda as phase II readouts in IBD. Maybe just walk us through sort of like your framework for indication evaluation and kind of like next steps for envu.
Yeah. We're going through a very intense process right now to really look through all these indications that we've identified and really do a prioritized list. I think it's really important to understand in two ways. One is, there is the IL-23 axis and the interferon axis. For the IL-23 axis, we have now shown wherever an IL-23 works, we should actually work as well. That's certainly a place we could go. We still want to better understand a little bit what the exact profile from zasocitinib looks like to understand how competitive we are and what that does that mean for us.
The other aspect is the interferon pathway. If we have really, or when we have positive lupus data, then I think there is a whole slew of indications here that open up that are really more driven by interferon. That for us is a very interesting axis. It's those two pieces. The other aspect that is really important for us is we are in a very fortunate situation that we actually have a lot more degrees of freedom with our TYK2 franchise because we have two molecules. If you think about pricing, if you think about IRA, if you think about patent life, et cetera, et cetera, we can now really choose which one we use for what. I think that is a big shift on how we communicated about A-005 and envu before.
Now that we have phase III data, we know what safety profile we have in psoriasis, we know what efficacy profile we have, we have to really think about this as a franchise. What are the indications that we're using A-005 for? What are the indications that we're using ENVU for? For us, actually the most important thing now is to really not think about these as two distinct molecules for two areas, because with A-005, we have shown that we get the same level of inhibition as we get with ENVU in the periphery, but we also on top of that get really great inhibition in the CNS. Think about those now as two members of a franchise that you can optimize across two molecules price-wise, patent-wise, IRA-wise, et cetera, et cetera.
That's great. Just in maybe the last 30 seconds, just on A-005, we've got into putting this into a phase II MS study in the first half of the year. What's the latest on that? Are we on track for that? Just maybe a 30-second soundbite on A-005.
Yeah. A five has so far shown the best brain penetration of any TYK2 inhibitor. It's basically a one-to-one ratio from the blood to the brain. Of course, the best way to optimize that would be to go into an MS study. We are considering basically looking at an imaging study with A five. That's an imaging study that very similar to what we did with tolebrutinib at the time, a three-month study, where you hopefully would see at about an 80%+ lesion reduction, and ultimately then would basically with that study, really understand what is the best dose to use in CNS disease. The idea there really is just the fastest way to go there.
There are some other really interesting indications, Alzheimer's, Parkinson's, ALS, et cetera, where you have basically a very strong neuroinflammatory component in a neurodegenerative disease, and there's a lot of interest there. For us, that is one of the focus areas. As I said, we also have to look at A-005 in the overall picture as well, and not just at CNS diseases.
Yep, that makes sense. All right. Perfect. A lot to look forward to here over the next 12 months. We'll stay tuned. Thank you, Alumis team, for joining us and sharing the insights.
Thank you. Thanks very much for having us.