All right, hello everyone. Welcome to another year of JP Morgan Healthcare Conference. My name is Cindy Xu. I'm an associate at the JP Morgan Healthcare Investment Banking team. Without further ado, it's my great pleasure to introduce you, Carl Hansen, the CEO of AbCellera. Thank you.
Thank you so much, Cindy, and thank everyone for coming today. It's a pleasure to be here and to reconnect with colleagues at JP Morgan. I was giving a little thought as to how I might start this today, and I came up with a metaphor, and I'm gonna give it a try. We'll see if it lands. So the metaphor is that building a company is a lot like raising a child. So you, you found a company, it's small, it's helpless, it needs a lot of support. It takes about 20 years to raise that child to maturity, and if you do it right, if you lay the foundation, you have high hopes that child can go off into the world and have a big impact. I think that's a lot like building a company in biotech.
It really does take a long time, and when you think about how far a company can go, it matters a lot how you think about the early years. So AbCellera was founded back in 2012, so we're roughly 12 years old. If you like my metaphor, we are entering adolescence, and we're really excited about the platform we've built and about the future that we have over the next decade to turn AbCellera into what we believe can be, one of the world's leading biotech companies. So today I'll tell you, in a few slides, how we think about this industry, where we have come since founding the company, and, what the future will be over the next year and beyond.
Then I have only a few slides, and I'm hoping we can have quite a lively question and answer session. Standard disclaimers, forward-looking statements. First, when we started AbCellera, we had to think very, very deeply about how we wanted to build the company. In the end, in this industry, all the companies, pretty much, that are presenting here, will win or lose based on their ability to move the needle on drug development. We're not after making incremental therapies. We want to bring science and technology together to make medicines that ultimately really make a difference for patients in need. If we do that, we will be a successful company, and we can create tremendous value.
There's been a lot of talk about the productivity lags in biopharma, and I wanted to offer that there's really only two ways to increase productivity in biopharma. The first is to get some fundamental new insights into biology, so that's science, new targets. If you have a really great target that you're able to prosecute, you can make a big difference in a short amount of time. That's a relatively difficult thing to scale because biology is complex, and we don't really understand how disease works, but we're getting better all the time. The second is that you go after the known biology.
If I was to ask everyone in this room, can you name a target that if you could drug it, you would have a candidate for clinical development that would have much better than your average chance of succeeding and would meet some very big, large medical need? If I went around the room, I'm sure everyone could name at least one. And, you know, the problem has been not that it's impossible, it's that there hasn't been sufficient investment in technology to bring those things into reach. So that has been our pursuit of increased productivity in drug development, to invest in capabilities and technology that allow us to make the impossible possible. Not an easy path.
It requires a long investment, but something that's well worth the investment, when you get to the finish line. Along that vein, since founding the company back in 2012, the main basis for the company has been investing in making an engine, a combination of technology, of people, and of infrastructure, that can reproducibly generate first-in-class and best-in-class antibody medicines. So we see investments in technology, in teams, and infrastructure as the way that we have built and will continue to build a competitive advantage in this field, and the way in which we, as an organization, can help to increase productivity or the success rate and ultimately value creation in biotech. If you do that, then ultimately you have to prove it.
So it's not enough for me to stand up here and tell you about the technologies that we have in our shop, about why they're differentiated, how we use artificial intelligence, how we use screening technologies. In the end, we have to actually deliver molecules that move the needle, that get you past barriers that were in place before, and in the long run, that deliver medicines that succeed in the clinic. Over the past 11 years, we have been hard at work at this, and today we're at an exciting time in the company because we are approaching the home stretch in building our platform.
So for about 10 years, we have invested well over $500 million in building a suite of capabilities that allow you to go from concept, so from a target and a specification for a program, right through to the beginning of clinical development. Today, based on that investment, we believe that we have the industry's most powerful engine for antibody therapeutics at the front end, and over the next two years, we'll be completing this engine with investments in regulatory capabilities, and most importantly, in CMC and manufacturing capabilities that will be online by 2025.
Since the very beginning, we also recognized that if you are spending your time and your money, particularly in the early years, mostly on capability development, then you need to use that capability a lot, both to generate value and to make sure that you are focused on the right problems, and that you're practicing and getting better at the game of making therapeutic antibodies. So we have had a robust partnership business. Since 2015, we have worked on well over 100 therapeutic programs with a variety of different partners, which include large biopharma and biotech, and we have, for many years, been focused on putting in place technologies that allow us to address the hardest challenges in the industry.
A very large effort on being able to solve problems related to complex membrane protein targets, ion channels, and GPCRs, and bispecifics specifically in T-cell engagers for autoimmune conditions and for oncology. I'll say more about that in a little bit. A few more stats. We have roughly about 600 employees. The bulk of that is in Vancouver, British Columbia, around 500. We also have a significant presence in Sydney, Australia, and a smaller group in Boston. Our work with partners has generated 10 molecules that have made it to the clinic. Two of those got a lot of attention. They were molecules done with our partner, Eli Lilly, in response to the COVID-19 pandemic.
In addition to proving that the technology worked in a way that was very accelerated, given the regulatory environment of the pandemic response, that effort also proved out the business model and brought in roughly $1 billion in revenue. So that billion dollars, along with our capital raises, means that we have now well over $1 billion of liquidity, about $800 million in cash and about $200 million in accessible government grants and long-term debt financing under very reasonable terms to continue our strategy and to push forward in making therapeutics. The partnership business over the last few years has begun to evolve. So in the early years, we had capabilities at the very front of the engine.
We entered into partnerships that I would characterize as doing the early work and handing it off and making sure that we keep a small piece of every therapeutic program, typically in the form of low to mid-single-digit royalty positions. So essentially, using technology to build up a portfolio of royalties and becoming a royalty aggregator. Over the last two years, we've been consistently communicating and executing on moving that business to fewer and higher quality partnerships, so strategic partnerships, looking for opportunities with firms to have synergy in technology. A good example of that being our recent deal with Prelude Therapeutics, where together with their small molecule expertise, we are making a next-generation antibody drug conjugate. Working with firms that bring new biology, a great example of that being a co-development alliance with Empirico.
Working on company creation, as with Abdera, which is a company that we formed together with Versant Ventures, and that's a leader in the radioisotope delivery space. And finally, continuing to work with large and enabled players, such as the Eli Lillys, the AbbVies, Regenerons, and Novartises, which we see as an opportunity to build relationships and ultimately, to have partners that we may out-license or work with in developing our own internal, internal portfolio. On the right here are some bullets that really highlight the things we're looking for in partnerships, expanding relationships, looking for opportunities to build partnerships on our T-cell engager platform, which I'll talk about in a minute, co-developments, 50/50 partnerships, and finally, company creation, as I mentioned before.
Probably one of the most recent developments out of AbCellera, at least in terms of what we've disclosed, is the advancement of our internal portfolio. So we have, for the last five years, been working on solving the problems to be able to access classes of targets that have been traditionally very difficult for the industry, but for which there are a large number of well-validated targets that if you could find antibodies with the right property, would be very attractive for development. That work is now coming to fruition. The first of those from that effort is a molecule called ABCL635, which I'll talk about in a minute. It is an antibody antagonist against a GPCR or an ion channel, and one for which we have not yet disclosed the target.
The second internal program, that we have disclosed is our program, ABCL575, which is against OX40 ligand. ABCL575 is a program that was not initiated by AbCellera, but rather came from a co-development agreement that we had with EQRx. It was one of three programs we were working on in the space of fast followers, in inflammation. EQRx, as many of you know, was acquired by Revolution Medicines, and when they were, we regained full control of that program. So ABCL575 is a fast follower. We believe it has the properties that would allow it to be a first—sorry, a best-in-class and a second-in-class behind amlitelimab, which is being developed by Sanofi. Our excitement for this molecule is primarily related to what we see as the immense potential of the class for OX40 ligand.
So it's a non-depleting antibody, and recent data shows that it is effective at suppressing autoimmune conditions and has potential to elicit an immune reset through its interaction with T-regs. It has potential to be developed not just in atopic dermatitis, which would be the first indication, but also in related conditions such as asthma, alopecia, HS, and others. Our plan right now for this is to move it as quickly as possible into clinical development and to get POC. At this point, given the broad range of applications, we anticipate entering into discussions to find partners to help for the more substantial clinical development to see the full potential of this molecule, which we do believe has the possibility to become a blockbuster. The second one that I mentioned is ABCL635.
This is one that is really on strategy for AbCellera. So we are focused on building our internal portfolio in first-in-class molecules by solving difficult technology problems. This is an antibody, as I mentioned, that's an antagonist against either a GPCR or an ion channel target. It's in the area of endocrine or metabolic disorders. And our excitement for this molecule is based on the excellent validation of the target pathway, compelling proof of concept data in primates, and what we believe to be a very large market opportunity that we would estimate to be well over $2 billion.
This molecule also would present a development opportunity that we believe we could do ourselves quite far, and so we would anticipate taking it to proof of concept and perhaps beyond, depending on the results that follow. I should also mention that this is the first and, you know, very excitingly, the first molecule to come out of our efforts on difficult targets, and we do expect to continue to add additional candidates to our portfolio over the coming years. Our goal for 2024 is to have at least one and perhaps two new potential first-in-class molecules against difficult targets announced as moving into the pipeline, and we will disclose targets if we can.
In this case, we are not, because we deem it as not strategic, given that it's a target that probably many people are not working on right now. Just to sum up, I'd also like to give an update on another area of technology development and effort that we've had going for two years, which is in the space of T cell engagers. So many of you will be familiar with this space. T cell engagers are typically multi-specific antibodies, normally bispecific antibodies, that can simultaneously bind on a target cell, perhaps a cell related to autoimmune conditions, and more commonly, a cancer cell, and recruit to it a T cell and simultaneously activate that T cell to elicit killing of that cell.
T-cell engagers have had tremendous success thus far in liquid tumors, and there's a huge opportunity for those that can figure out how to translate that success into solid tumors. There are a variety of challenges to that. One is the suppressive environment of solid tumors, and there are a variety of approaches to tackle that using co-stimulatory molecules, perhaps, perhaps checkpoint inhibitors. But there are also two other problems, and one is that in many cases you have dose-limiting toxicity that is related to cytokine release. So activating the T cells to kill the cancer and having them activated too much to result in toxicity through cytokine release. The second is that because T-cell engagers can be very potent, you want to be careful that you have a target that is very clean to the cancer.
So finding targets that are specific are very important. We have been working on both of those problems. Our approach has been first to lay the foundation by finding a very large panel of anti-CD3 antibodies. To date, we have discovered hundreds of completely unique CD3 antibodies that are highly differentiated from SP32, which is the one that is used repeatedly in the industry, has very little overlap. These CD3 antibodies can be combined with our OrthoMab platform to tune the cell killing and cytokine release, and to provide strong killing and an optimal profile of less cytokine release.
Secondly, we have been working on being able to generate antibodies against MHC peptide complexes, which expand the universe of potential targets to those not just that are naturally presented on the outside of the cell, but also to peptides derived from tumor-specific antigens that are expressed inside the cell. I've added a couple slides of data just to highlight some of the recent results on that. On this slide, we show a result of a CD3 TAA bispecific antibody that was discovered by screening a panel of hundreds of CD3 antibodies, actually hundreds of combinations of CD3 and tumor-specific antibodies.
And what's shown here in purple is the AbCellera molecule, which on the left, I guess, you see a dose response in tumor killing, and you have complete tumor killing, as you do with two clinical benchmarks, that are shown in green and in red. And on the right, you have the cytokine release that's associated with that. So as you can see, the AbCellera molecule in purple has very low cytokine release, and we've been able to show this on multiple tumor targets, getting between two and three logs of separation between maximal killing, and significant cytokine release.
So we believe that this, at least pre-clinically, provides compelling evidence that there is a solution that could be brought forward to try to increase the therapeutic window and reduce some of the toxicities that are associated with some tumors. The second data slide is one that we're also extremely excited about, and this has to do with being able to generate TCR mimetics or antibodies that are highly specific to MHC peptide complexes. This particular example is on an MHC peptide from MAGE-A4 with the HLA-A02 type. What I want to highlight here is that this was a program that was run over the course of about 12 months.
We have put in place the biochemistry required to work with these classes of targets, and our workflow allows for immediate screening for specificity right out of the gate. So we have a single-cell screening platform through which we can look across millions of cells and pull from many thousands of binders the very small number that appear to have specificity to a MHC peptide and do not bind to the same MHC molecule bound to a different peptide. With that, we were able to get a panel of MHC peptide-specific molecules that were then paired with our CD3 panel. We show on the top row very potent killing.
On the left, what you see is structures of these molecules showing that there are multiple solutions, and these are very different antibodies that approach in different angles. So this is an example of going from start to a panel of highly specific MHC peptides we've benchmarked against clinical comparables. And we are now leaning into this because we believe that this has the potential to be a tremendous advance in the cancer space, because it opens up many, many more possible targets, both for TCEs and potentially also for other modalities. So we're excited about that, and you'll hear more about that over the coming year. We intend to present updates at ACR and SITC over the next few months.
And just to sum up, so, we have, you know, at this point, we find ourselves in a position where much of the platform building has been done. Over the next two years, you will see us continue to allocate, capital and resources to completing what has been a big project, which is putting in place a GMP manufacturing facility. We expect that we'll have the first pilot runs of that going, this year, and the first GMP run slated towards the end of next year, round about. After that, we'll be shifting our resources from the building of the platform, which is something we have been doing now for 12 years, excuse me, and moving it more to the application of the platform, both in strategic partnerships and advancing our portfolio.
Second, in 2024, as mentioned, we do expect to announce at least one, perhaps two more programs that will enter into our internal portfolio, derived from our work on GPCRs and ion channels. And in early 2025, we expect our first two programs, ABCL635 and ABCL575, to be INDs. And lastly, on the partnering side, we continue to focus on strategic partnerships. The TCE platform we see as a very important basis for partnerships, and we are looking at opportunities both in oncology and in autoimmune disease. And with that, I'll thank you for your time, and we'd be happy to take some questions. Not everybody at once. Yes.
Great presentation and progress, Carl. It sounds like vertical integration is a core part of your strategy. But could you maybe share more about why you think that'll give you a competitive edge?
It's a great question and probably something I should have highlighted more. So, on the platform build, you know, when we think about technology, we sort of think about it functionally, right? So very often people get reductionist, and they say, "I've got this, you know, particular step in the process, and it makes such a big difference," whether that's an algorithm or a device, or the rest. But that's not really how drug development works. Drug development is, you know, hundreds, maybe thousands of steps that need to come together. If you don't have that capability internally, and if you need to outsource all of that, a couple things happen is, one is you cannot access the state-of-the-art technologies, 'cause typically, you know, that's not how outsourcing works.
Secondly, there's a huge friction, you know, both transactional and in time, in trying to piece these things together. So if you can bring technologies together and have workflows that, that are coordinated, you get tremendous increases in flexibility and in speed. You save time, and probably as importantly, you can, you can do steps at risk because you control the, the assets yourself, and that's particularly true on the manufacturing. So if you have your own manufacturing facility, you can shave, you know, 6, 9 months off the time to clinical, to a clinical start without even doing any innovation because you have the ability to bring those forward, in a way that you would not if you have to outsource.
So that's been the theme, and I would say it's definitely played out, and it creates a technology and competitive moat that is not easily displaced. If any one of the pieces, you know, is challenged by something in the outside universe, it doesn't impact the ultimate function, which is not just to solve a step, but to deliver drugs. Yeah.
Great. Thanks.
Thank you.
You mentioned you started developing your own therapeutics, obviously, and it looks pretty good, how are you determining what targets you wanna focus on or go after? Like, how do you prioritize them?
Yeah. Great question. So the bulk of the work that's underway has followed the strategy that was basically on my first slide. So we look and say, "Where are there targets that are extremely well-validated?" There's genetics. You know, people have done the animal models. Maybe there's a small molecule that has worked, but it's had some tox issues or some delivery issues, and where antibodies would have a fundamental advantage. And when you look at those, and then you counter select against the ones that have already been prosecuted, what you have is a technology problem, not a biology problem. So in the space of ion channels and GPCRs, we could list, you know, 20, 30 programs like that. So these are... You could go Google it.
You'll find there are, you know, public targets that have been difficult to work on. That has been the main focus in, in that sphere. Those targets end up being in indications, I&I , a little bit of oncology, pain, are sort of the main themes, moving forward. The other, the other angle, of course, is the TCE work, where, there's a, a common technological base you need to build, and then the question is, you know, what exactly is the tumor target, and what is the appropriate combination and profile and molecule that you want for development? So on the TCE side, you lay a foundation, and once that foundation's in place, and you gain experience, you can quite quickly generate candidates that, that can be brought forward into development.
In fact, the range of targets is so large there that that can only be done through strategic partnering, and that's one of the reasons we're bringing that forward for partnering. Thanks for that question. Yes.
Yeah, hi. Will there be any more continued efforts around infectious disease targets at all?
Yeah. So we obviously had a big presence in infectious disease historically through programs with DARPA and even with the Gates Foundation. And Covid, I think, was the, you know, was the holy grail of that that we executed on. Right now, you know, if there were another pandemic, we are ready to mobilize again, and I believe our partner, Lilly, would also join us in that. From a commercial perspective, we think it's much less attractive than some of the other things that we can work on. So it's not a priority now, but the capability is there, and we think it is important to have that capability in case we end up in a twenty twenty again, which is gonna happen at some point.
Hopefully, hopefully, it's a decade and not two years, but we'll be ready. Yes?
Very impressive. You got many project. I think you've got many data, but you not mentioned about the AI platform, road plan. Any insight? Especially for those ion channel on GPCR, those, those target, I think, probably you may use some AI models, especially with this big, big model. And in the future, how do you look at de novo design?
Great question. So the question is, you know, what's the, what's the contribution of AI, and what's my perspective on de novo design going forward? So, you know, I, I think I can be both a bull and a bear on this. So if you ask the question: how important is AI in drug development in the long run? Very important. And I think it's likely that, it will touch many, many parts of the drug development process. Today, we are using AI extensively. The way that we use it is primarily to allow us to, scale up experiments and do much faster and much more precise analytics to extract information from very high-throughput experimentation, so imaging, sequencing, some of the work in high-throughput expression, and developability, okay? These AI modules should be thought of as narrow, right?
They're custom-built for a task, and they are an important, but not a sufficient, piece of technology to get that particular step done. And that's where I think most of AI is today. There's obviously a lot of talk about AI as though it's going to be... People invoke it to solve the problems that we can't solve without really digging into how that's going to happen, and I'm a little skeptical of some of the things that I hear. So I wanna see - I wanna see AI do things that you could not do without it, and that bar is pretty high. You know, we have a lot of great technologies that solve a lot of problems.
You asked about the, you know, are you gonna be able to do de novo synthesis of antibodies? I think that is coming. You know, over the last year, there's definitely been, you know, some progress towards that. I have not yet seen progress where, someone found an antibody that solved a problem that actually existed. So the problems that I've seen have been toy models thus far. It doesn't mean it's not gonna happen. We think it's important. We're keeping our eye on it. We have the data and the experimental capabilities to allow us to keep pushing on that front and to make sure that we're there when it happens.
But for now, for now, it's like a let's wait and see, and let's make sure that we're actually working on moving molecules into the clinic today while not missing the boat for what's happening in the future. Thanks.
Sorry, just to follow up on that, 'cause that was-
Yeah.
An interesting question. Do you have any thoughts on... You know, there's been a few companies now that are extremely well-funded that have all started, you know, relatively recently, mostly in the small molecule space, right? With the NVIDIA platform. But there's been a couple in the biologic space and, you know, maybe one or two in the ADC, that have a lot of funding. They're talking about how they're gonna use, obviously, AI to transform this feature. But is there some, you know, thought process, either concern or, like, something you guys are gonna start integrating in that you've seen from some of these other companies coming out?
Yeah. So, you know, we are investing in these approaches, and it's really hard to benchmark where you are relative to others. It hasn't been... Like, what we are not doing is staking the entire company on it and talking only about that. And, like I said, I think that, you know, AI is real, and some of these problems you will be able to solve. But I do believe people very, very badly underestimate the timelines and the efforts that are required to bring those to the point where they're useful. So if you'd indulge me a second, like, an example is, you know, AlphaFold is an example. So AlphaFold, tremendous breakthrough, and as soon as that happened, you know, there's all this inbound: "Okay, now you can generate proteins. You know, everything is obsolete.
AI has arrived. AI is gonna predict outcomes in the clinic." And there, there's no logical connection between these things. So, you know, AlphaFold works because there's 100,000 structures. There are 100,000 structures because, you know, over 50 years, people did the experimental work to put that in, and that particular problem, the data sets there are, are perfectly clean by the nature of what it is. Here's the sequence, and here's the coordinates. So everyone can use that. So people see that result, and they, they infer AI has arrived, and it will hit every field. But I think the right conclusion is, wow, it took a heck of a lot of data in order to get that breakthrough in AI.
In my particular problem, if I want that to happen, I need to make sure I have the data, and it needs to be clean, it needs to be put well together. In some places, you know, maybe we're getting there, but in a lot of the things that people are talking about, I don't see it. And in some of the things, it's also, it's not at all clear how you close the loop. You get a ton of data, but you don't know... Like, how are we gonna predict, clinical outcomes? We need to run clinical trials. So if I build a model, I can't tell you if it works for 10 years, and, you know, $100 billion. So that's not gonna come anytime soon. Yeah.
Thank you.
... Switching to a near-term, a few near-term items, are there any key milestones that we should look out for in 2024? And how are you thinking about capital allocations over the next few years?
Great question. So, as I mentioned, you know, some of the big capital allocation is in completing building the engine, specifically on the manufacturing capabilities that are in place. You'll see that happen mostly in 2024, a little bit into 2025. We are starting to put some real capital to work on the internal program. So as programs move into IND-enabling studies, we'll be doing manufacturing, and starting to do some of the work that gets to be more expensive, but that doesn't really get expensive until you get into later stage clinical development. In terms of, you know, catalysts, if you would, you know, last year, we expected to do a validating deal on the TCE platform. That didn't happen.
We think we missed the timing but are more bullish than ever on the potential of that platform to generate some real value in partnerships. So if we are able to, you know, bring those together this year, I think those could be those could be significant events for the company. Looking forward, you know, the biggest value inflection is going to come likely from the internal programs. As I mentioned, the first of those are still early. They won't be IND until early 2025, so I'd say that's when it really starts to get interesting.
And the last piece, if we bring more candidates forward and we're able to give more scientific disclosure, depending on how those go and exactly which targets they are, there is, you know, the potential that we could show an exciting potential first-in-class target that would both have obvious commercial potential, but also would highlight the progress we've made on the engine in solving these tough problems. So that's how I would think about it. Yeah. Anything else? One more, yeah.
Well, you just mentioned Uni-Fold. You just mentioned AlphaFold, those kind of very outstanding techniques. But how do you think about the next big work that could be as outstanding as the AlphaFold, especially in the antibody discovery?
I'm sorry, I didn't get the question.
Oh, you mean, you know, we have seen that after the AlphaFold, there are many, many AI companies bumped out.
Yeah.
But, what the next one you think could be as much outstanding as AlphaFold?
You know, the question came up before, "Hey, do you think, Like, are we going to be able to, to do generative design, like, right from scratch of proteins that do what we want?" I think that's in the future. I don't think that's today. So, that will come over time, and, you know, that could be a really big deal. That'll be a big deal in places where, you know, technologies have not already proven up to the task of making great therapeutics, and, so it's important that you point that at the right problem. Maybe backing up, like, where will AI really make a difference? I think you have to ask yourself: What's gonna be more important?
Is it generating, you know, new drugs, or is it better understanding biology to find better targets and have better insight into what we should be going after? I think the latter will be more important. Like, most failures are because you picked the wrong targets. I think the challenge is gonna be it's easy to come up with a hypothesis, using, you know, computational methods and genetics. It's very hard and a lot of work to work up those hypotheses. So there's an experimental bottleneck that we're gonna have to cross if we really wanna get the juice out of that.
Thanks.
All right. Thank you, everyone. I really appreciate the conversation and the questions.