Good afternoon, everyone, and welcome to Morgan Stanley's Global Healthcare Conference. I'm Sean Lahman, Head of U.S. Midcap Biotech Equity Research. For important disclosures, please see the Morgan Stanley Research Disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative. For this session, we have Absci with Founder, CEO, and Director Sean McClain, and CFO and Chief Business Officer Zach Jonasson. Welcome.
Yeah, thanks for having me.
Gentlemen, maybe just to kick off proceedings, Sean, contextualize the discussion and give us a bit of a view on Absci.
Yeah, definitely. We're a generative design company really leveraging AI to be able to go after the hardest and most challenging problems that still exist. Being able to go after hard targets such as ion channels and GPCRs, these undruggable targets, we think of AI as being able to decrease timelines and decrease overall costs. It also opens up this fourth dimension, which is being able to unlock new novel biology where other technologies are unable to address. You see that in our partnerships with Merck, AstraZeneca, and Almirall, but also in our own internal pipeline as well. We have a portfolio that's focused on INI. The first asset is in the clinic. We'll have a Phase 1A readout in the coming months. Additionally, we'll have another asset in the clinic, either end of this year or beginning of next.
This is ABS-201, and this is for androgenetic alopecia targeting the prolactin receptor. Essentially, think of hair regrowth and re-pigmentation. We're excited to have a Phase 2 readout on that the second half of next year. I think it's a really exciting technology paired with some exciting readouts that we have coming over the next 12 months.
Wonderful. Thank you, Sean. Just a macro question. With the tremendous rise in biotech innovation, how are you thinking about Absci's competitive position here? Will this influence your R&D and BD strategy?
Yeah, it definitely has influenced how we think about our own strategy, moving forward. You have to give China credit. What they've been able to show that they can do over the last five years is really remarkable, being able to develop high-quality assets in a very cost-effective way. They're going after targets that are known, really kind of taking this fast-forward approach. It's really made us rethink how we're wanting to approach our own internal pipeline and how we're approaching it with large pharma, how we stay differentiated, and how we continue to stay relevant and innovative. That's really where we've focused on these hard-to-drug targets, such as being able to create agonists towards GPCRs or blocking ion channels. Where others have struggled, that's where we want to focus in on. We're focusing on biology that's been known, but again, hard to drug.
That's really been a driver for us over the last couple of years and kind of shifted our strategy based on what we're seeing in China. I think that pressure, the pressure to innovate, I actually think is really important for the U.S. biotech community as a whole. I think we're actually going to see a lot of benefit come from that.
Wonderful. Now, I am going to ask you this next macro question, but we're asking these questions of all companies. It is a little circular to ask an AI biotech company this AI question. As an AI tech-enabled biotechnology company, can you describe the key ways your platform is leveraging AI and thinking about AI's future disruption potential?
Yeah, AI is truly transforming how we're doing drug discovery. It's really shortening the timeline. If you look at our TL1A asset, we are able to get that in the clinic in roughly 24 months from starting on that program. Normally, it takes five and a half years. We're seeing a decrease in the overall cost. Cost upwards of $30 to $50 million, traditionally even upwards of $100 million to get an asset into the clinic. We are able to get our TL1A asset in the clinic for roughly $15 million. You are seeing a decrease in both cost and time. As I mentioned, one of the things that we're focused in on and we see as a big differentiator is this fourth dimension, being able to unlock new novel biology with AI.
I think this is where you're going to see a lot of value creation over the next few years. An area that I'm really excited about, while us at Absci aren't necessarily focused on it, is how can you use AI on the clinical development side as well, in order to make sure that you're getting the right patient recruitment, you're looking at the right biomarkers, and helping make sure that you're getting the right enrollment into your trials. You're already seeing, I think, some of this early transformation, but we are in the early innings. One of the things that I can't emphasize enough is the importance of data. At Absci, we've always been a data-first company. We've generated data to train our models and validate it. This has really created this learning loop that has allowed us to rapidly innovate and ultimately generate the best models.
Whether you're generating it in-house or getting a consortium of data, what we've seen for ultimately building the best models is making sure you have access to the right data.
I'd make one point to that, which is I think a lot of people maybe don't appreciate AI as something that's constantly evolving. It's not a static platform at all. What's exciting for us is we look at how our platform has expanded in capability over the last 12 months, even the last six months. The examples Sean was pointing out about these difficult targets that we've been able to address, those were our last generation of models. We're already on to the next generation. For me, what's really exciting is you think about the future. That pace of capability, as you mentioned, it's not linear. It's non-linear. We are definitely on that curve. The ingredients for that are the data that Sean mentioned, as well as compute.
Exciting, exciting. What has been the most impactful on Absci on the regulatory side? It's FDA, MFN, tariffs.
Yeah, you know, we really haven't been impacted by any of the MFN or tariffs. We're not a manufacturing company, and we've been doing our Phase 1 clinical work in Australia where it's cheaper and faster. We haven't been impacted by any of those initiatives. The other set of initiatives where the administration's been promoting or advancing more rapid development through FDA, so obviating animal testing, we think that fits really nicely into the kind of things we're doing. I mean, we work on polyspecificity models and other models that address safety. We think some of those initiatives are right in line with our objectives.
Wonderful. I guess onto the platform, you know, you build a proprietary AI drug discovery development platform. What differentiates your architecture from other AI drug discovery and development systems?
Yeah, it's really this wet-lab-in-the-loop, this ability to generate protein-protein interactions, essentially antibody functionality data that we can use to train our models, but also using that same technology to inform how accurate the models are as well. This has really allowed us to rapidly iterate on what model architectures are best suited for the problems we're looking to solve, what hyperparameters, and then particularly how you curate your data and what data is necessary. We have the six-week cycle time, and we're constantly trying to push that down because the faster you can create that learning loop, the faster the actual innovation progresses. I think that's been just a very key differentiating feature. I look at where we're at with AI and with how the semiconductor industry has gone. Every 18 to 24 months, you have a new chip coming out.
I think that's very similar to what you're seeing at the AI drug discovery space. It's kind of what Zach was touching on. It's not static at all. It's ever evolving. Again, the faster that you can iterate, the faster you can stay ahead of the competition. Ultimately, at the end of the day, it's about the product. How can you create differentiated products that ultimately are going to be first-in-class and best-in-class?
Sure. You know, being at the beginning of that sort of innovation, which is not linear, and the pace of discovery accelerates, and you've got wet-lab-in-the-loop. Do you get to a point where you've got enough data points to train a model that you no longer need the wet lab?
Yeah, absolutely. I do believe that you're going to get to a point where you have a model that is static for solving a particular problem, and you're really just relying on the inference, and you're on to solving the next problem. I think the way we look at it at Absci is you could get to the point where you're able to predict an antibody to bind to an epitope of interest. The next step then is, okay, how do you predict the functionality, not just the epitope? You're always kind of going on to the next problem. That's where it goes back to this learning loop and the wet-lab-in-the-loop and why it's so important. You will ultimately get to the point where you have static models, but then you're on to the next problem.
Sure. What are the biggest challenges in designing antibodies which have historically been undruggable or challenging to drug targets, you know, like ion channels and GPCRs?
Yeah, absolutely. It's really the surface exposure. There's not much surface exposure of these ion channels and these GPCRs, which creates a, you know, a difficult time for, let's say, the immune system in an immunization campaign, struggles with being able to hit these antibodies or to hit these targets because there's just not much surface exposure. With an AI model, it doesn't really matter how much is surface exposed. You just need some, and if you have an epitope that is surface exposed, you can then generate an antibody towards that. Being able to block an ion pore has always been a struggle. You can now, with our models, be able to design deep CDR or long CDR loops that combine deep within the crevice to block it.
In the case of a GPCR, being able to generate an antibody that agonizes the GPCR very similar to how the ligand would. These are some of the problems we're trying to solve and why I think some of the traditional approaches have failed.
Sure. Thank you. What are some of the smart features you've engineered into antibodies so far, such as pH-dependent binding or agonism/antagonism?
Yeah, absolutely. We've been able to show that we can really engineer in a lot of different features. We were able to show, on a particular oncology target, where we could essentially get differential binding in the tumor microenvironment where it's more acidic, but it doesn't bind in the healthy tissue at neutral pH. That opens up a lot more targets that you can start to go after in oncology when you can have that differential binding. Additionally, if you look at some of these GPCRs that are coming up in metabolism, a lot of these are peptides or the original ligand is a peptide. You can essentially mimic how that peptide is binding to the target to agonize it, or if you want to block it, you can develop an antagonist towards it.
These are some of the ways that we're using de novo design to start to engineer in the properties that we want versus this trial and error process that drug discovery has always been.
Okay, thank you. Maybe just to move away from the platform for a second and start running down the pipeline. On ABS-101, what are the key objectives for the upcoming Phase 1A mid-term healthy volunteer data with the readouts expected, I believe, in 2025 for your TL1A asset?
Yeah, we're just right around the corner. We're really looking forward to the readout. The number one thing will be safety, obviously, but we'll also be looking for the PK profile. We're expecting to be able to do at least once quarterly. We'll also be looking for PD in that target engagement. This will be in healthy volunteers. We'll be looking finally to see if we get a good signal around a low ADA rate, which is what we expect given the epitope we selected.
Sure. Thank you. I guess still on ABS-101, compare it to other TL1A antibodies in terms of potency, durability, and patient convenience, or what are you hoping to achieve there?
Yeah, I mean, certainly, we've done head-to-head comparisons against the first-gen molecules. I would say we see advantages in potency, half-life for sure. Most of those are once every two-week dosing. We also see some advantages in the ADA profile, which we think is going to be important for the final drug product here. I would say one thing that is maybe a little bit overlooked is we see really great bioavailability and tissue penetration. We do believe long-term that could be something that translates to additional efficacy. We also have monomer trimer binding, which also could get at additional efficacy in certain patient types as well. We think the molecule is very well set up to compete against the first-gen, and we believe it's in a good position to compete against the next-gen molecules as well.
That's true. Thank you. Can you describe your progress in partnership discussions and what you believe might be the value inflection points of the program? Where do you foresee partnering to be optimal?
Yeah, we've got a lot of engagement with large pharma as well as tier two pharma companies that are interested in the TL1A program. We do believe with the breadth of indications that are now under investigation for that mechanism, there's a large partnership or buyer audience. We've engaged with them. There are a couple different points across the timeline where we think it would make sense to potentially move into a transaction. One of those is after this interim readout later this year. Another one will be when that Phase 1A closes. We have the final readout there, which would coincide roughly when we'll have a bispecific package at TL1A with a novel arm as the second arm, which we understand from several of our discussions is of high interest to pharma, particularly having a first-in-class bispecific.
Thirdly, we're fully prepared, capitalized, and in our forecast is to take this program through a two-way study as well in patients.
Moving on to our one in androgenetic alopecia. What makes that ABS-201, you know, a compelling molecule for the treatment of that disease?
Yeah, absolutely. If you just look at the standard of care, you have minoxidil and finasteride, and what you see there is patients being frustrated with having to take this either orally or topically daily. Convenience is a big factor. Patients at the end of the day, a lot of them aren't seeing the overall efficacy that they would like to see. Minoxidil is only efficacious in a certain patient population. Women really don't have a great option either because they can't take finasteride, and minoxidil gives them hair in unwanted areas. There really hasn't been a lot of innovation within hair regrowth in the last 20 years.
What we're seeing with the mechanism of ABS-201, going after the prolactin receptor, is that if you block this receptor, what we're seeing in both mice as well as non-human primates is that you end up shunting the follicle back into the active growth phase, and you actually start to get hair regrowth. We've seen this hair regrowth in stumptail macaques and in mice. The stumptail macaques is really quite exciting. Essentially, these are monkeys that naturally go bald, and when you block the prolactin receptor, they go from their bald, gray hair to a full head of hair, and it being jet black, showing that you can essentially re-stimulate the follicle growth as well as achieving pigmentation as well. We actually have seen this translate into the clinic as well.
Our Chief Innovation Officer actually discovered this mechanism when he was CSO at Bayer, and they were actually looking at this particular mechanism in endometriosis. It was a serendipitous find that the prolactin receptor was involved in hair regrowth. Essentially, the way they discovered this was the mice that had the drug regrew their hair faster than the control arm, which led them to believe that, led them to discover that this is indeed involved in hair regrowth. They ended up out-licensing this molecule to a Chinese company, Hope Biomedicine, and they took it into a Phase 1B. They were able to show that they could get 14 hairs per square centimeter. Talking with the PI on that study, Dr. Rod Sinclair, he had mentioned to us that they had severely under-dosed in that study, ultimately getting receptor occupancy well below 90%.
That was also confirmed in our non-human primates study that we did with that molecule. We've engineered the molecule to essentially be able to achieve greater than 90% receptor occupancy and only have to dose two to three times. We do believe that this mechanism has been validated both on the preclinical side as well as the clinical side. There's really a massive opportunity for this particular market in general. We're really excited about this. I think one of the great things about this upcoming trial is that we will be in the clinic at the end of this year, beginning of next, and we're going to be doing a combined Phase 1, 2A study. In the second half of next year, we will have a 12-week interim efficacy readout looking at the hair regrowth.
We're roughly 12 months away from a really exciting, pivotal Phase 2 readout in this program.
With the molecule, you've got greater receptor occupancy and less frequent dosing. With your platform, why couldn't you have done that with a human designing the molecule? What has your platform enabled you to do to come up with that molecule?
Absolutely. I would say humans actually tried to design the first molecule, and the first molecule had lower receptor occupancy. In order to get full receptor occupancy, you'd actually have to dose 24 times, versus the two to three times that we've engineered. Essentially, what we were able to do was increase the overall affinity and potency of this overall molecule. We also engineered in an extended half-life. The molecule that was originally designed had a half-life of, you know, roughly two weeks. We were able to engineer ours to, at least what it looks like, will translate from NHP to humans. I think we're going to roughly have a once quarterly dosing, so two to three doses over a six-month period. This was all done with our AI platform.
Sean, one other thing I'll mention too, and this is true of all of our programs, is redesigning the developability so you can formulate these molecules quite easily. The first-gen molecule that Bayer developed has got severe formulation conditions. It's not a very stable molecule. We think that's roughly capped out at about 60 mg/mL, which is very low. We should be at a 200 mg/mL formulation, and we're on track for that for this trial.
Amazing. Maybe just to talk about the market opportunity, you know, we wrote something on this very recently. What do you think the market opportunity is?
We've done some market research. We're doing a little bit more now, and we've also had, I think we've been really fortunate to have some great advisors, including the former Head of Commercial at Allergan, as well as the former CEO. When we've looked at this, we've done patient surveys, KOL surveys. We think this is an enormous opportunity on the order of $10 billion, putting aside pigmentation. If we restore pigmentation, then I think that that market would go up significantly from there. This would be a cash-pay market. When we look at what the consumer needs here, what they really want is they want something that works. In practice or out in the field, we're seeing that minoxidil, for example, works in maybe 5% to 10% of patients, and the results are pretty mediocre.
If we can deliver a robust efficacy with durability, that's what patients are looking for. They don't want daily use. They want something that they could do like once or two or two to three injections is perfectly feasible. It looks to us like the price point could be quite significant, enabling gross margins that would be north of 90%.
Wonderful. Onto many additional pipeline programs. Could you provide an update on earlier-stage programs like ABS-301 and ABS-501? You know, what are the next steps of these programs?
Yeah, absolutely. 301, we showed some really nice target validation data earlier this year. We are now transitioning into in vivo efficacy studies. Both with 301 as well as 501, we are looking to partner those. We do not want to take these into the clinic ourselves. We think that these are best in the hands of large pharma, or pharma that focus in on oncology. We do not plan to invest the capital further to take these into the clinic. We do plan to out-license them. We have a whole host of other leads that are focused in on INI and metabolism-based targets that are earlier in the pipeline, which we haven't disclosed yet. We do plan to nominate and disclose a new drug candidate either later this year or beginning of next.
Again, that is definitely kind of focused on the INI, you know, similar to 101 and 201.
Wonderful. Is it the way to view your business over the longer term that you've got some proprietary pipeline programs ongoing? For you, it's more about validating molecules to get into the clinic, and then maybe the preferences over the long term, that's where it stops for you, and then it's milestones and royalty from that point. What you see your machine as is something that you essentially out-license to big pharma where they'll approach you to optimize molecules for them.
Yeah. I would say in general, yes, with the exception of ABS-201. We do think that we can take ABS-201 deep into the clinic, and even submitting a BLA on our own. This is an indication that is actually pretty cost-effective to develop, and we have the domain expertise in-house to be able to run the necessary trials. We do think by taking this further, we retain more optionality and ultimately create more shareholder value by keeping it and taking it deeper into the clinic. It's something that we have very high conviction in.
Thank you. Can you manage the status of your partnership with Amgen, and what are the next steps for the biospecific ion channel program?
Yeah, as you probably saw, in the end of July, we announced that Almirall had selected a second target to work on with us, which is a bispecific program. That was really based on our success on the first program, which was successfully designing a highly specific antibody to an ion channel. That program, the bispecific, the new program, I should say, is underway now, and the ion channel program is currently in optimization. It's moved to the next phases of development.
Cool. Thank you.
I think this just goes to show, once you have success on one challenging target with the large pharma, they want to move on to the next. I think that really demonstrates the value of the AI platform and being able to go after these hard, challenging targets.
Sure. You've mentioned a potential large pharma partnership this year. What are the key attributes you look for in a partner?
Yeah, I think ultimately we want a partner that complements what we're doing. If you look at a lot of the partnerships we've had to date, they've focused in areas that we're not focused in on. This partnership is going to be focused on oncology. Being able to leverage their therapeutic area expertise, their disease biology expertise in oncology, and then leveraging our platform to design first-in-class, best-in-class assets, we see as a great strategy for diversifying our portfolio outside of our therapeutic area focus.
How do you think about the sovereignty of data? If you're partnering with a big pharma and you generate some insights to get a molecule into the clinic, how do you separate the ownership of the data?
Yeah, I mean, just to take a step back, some investors have asked us recently, why do you even do partnerships? Why don't you just do your own programs? There are a couple of reasons why we do them. One of those is the data. One of the reasons is for diversification, because as Sean mentioned, we're typically working with partners that have expertise in indications that we're not focused on.
Sure.
There is non-dilutive capital, which is great. The third reason is in each of these programs, we generate data in-house. We're able to keep that data, and we keep the results of that data in terms of how we implement it to improve our models. You can think of it this way: all these partnerships are our partners who are underwriting the development of the platform. There is good reason to keep looking to do partnerships in parallel with our own internal programs.
Sure. Wonderful. With recent capitalizing from $1 million to 2028, how you allocate, how are you allocating resources across the various programs?
I can go ahead and take it. We go through this on a regular basis. Obviously, we are very committed to moving ABS-201 through that Phase 1A two-way. We think that that readout is going to be very pivotal, and we're committed to bringing ABS-101 forward to a place where we get a transaction in the range of what we're looking for. Those are two fundamental core tenets of what we're doing here. On top of that, we look at our platform where we make investments in continuing to build the capabilities. As we look at where those capabilities are, we then apply them to create new assets. As Sean mentioned, if you look at the early programs in our pipeline now, which we haven't announced, they're all these really difficult targets. That's reflective of the advancements we've made in our models over the last year.
Thank you. How do you assess capacity for additional partnerships while scaling internal development?
Yeah, I mean, this is the thing about AI that is also exciting, right? I mentioned that we're on this non-linear advancement in capabilities. It's also the case that AI opens up additional capacity as it gets better. We're getting more efficient. As an example of that, you know, a year ago, when we do a campaign internally, you might generate half a million or even more designs that we would then test in the wet lab. The models have gotten more and more efficient and precise. Today, we're generally generating 100,000, so almost a factor of five reduction in terms of what we need to generate in the wet lab. We look at our expanding capacity, and then we decide how to allocate that. That comes down to allocating some of it to partnerships and discovery programs, and then the rest to our own internal programs. Cool.
I guess, generally, biotech's been a tough place this year because of the uncertainty on the regulatory and, you know, uncertainty on rates, uncertainty on the M&A environment, which is just the trifecta. Not fantastic, but I'm wondering if that might have led to some hesitation of potential partners to form partnerships with you. Once these things are lifted, we get more certainty on rates, some M&A activity, and it seems that the regulatory picture gets clearer and clearer. Would it be fair to assume that you might expect a greater pace or cadence of new bound?
Yeah, I agree with everything you said. I think in this environment, there's a lot of cautiousness and a lot of deep diligence before deciding to embark. I think about our future, and we're certainly going to continue to do partnerships, but we're going to be much more selective about the partners because we do want to allocate more and more capacity to building our own programs. Those programs could be programs that we only take to a DC and then do an out-license. We just see this enormous opportunity in these hard-to-drug targets where they're not worried as much about any competition from China. We truly have a competitive advantage there, and we really want to press the gas pedals all the way.
Fantastic. We're almost out of time, but with that said, is there any question I didn't ask that I should have or any message that you'd like to leave investors with?
Yeah, I guess the one piece that I just really want to drive home is that we do have runway into the first half of 2028. That allows us to ultimately get through the Phase 2 readout of ABS-201 and have a year of runway past that. It allows us to ultimately get through our Phase 1B, 2A, and TL1A and get to a transaction there. Additionally, that runway doesn't take into account the large pharma partnership that we've guided to this year, which we do think will bring in upfront to further extend that runway. We feel that we are in a really good position to ultimately execute on the current pipeline, getting us through really important key value inflection points.
Additionally, ABS-101 has always, quote unquote, been our lead, but I would say that it's essentially a co-lead with ABS-201, especially since ABS-201 is going to have a Phase 2 efficacy readout before ABS-101. I think kind of thinking of ABS-101 and ABS-201 almost as co-leads is a really important piece given that we will have a Phase 2 readout on that next year.
Wonderful. We're perfectly out of time. Thank you, Sean. Thank you, Zach. I appreciate you attending today.
Yeah, absolutely. Thank you, Sean.
Okay. Welcome.