Building A New Antibody Discovery Platform With Infinimmune's Wyatt McDonnell, Ph.D.

On this week's episode of the Business of Biotech, Wyatt McDonnell, Ph.D., cofounder and CEO at Infinimmune, talks about how single-cell biology and computational rigor led to the formation of Infinimmune, and a new way to discover human-derived antibody drugs. Wyatt explains Infinimmune's 'discovered in humans' antibody discovery platform, how the company was launched and initially funded, and how Infinimmune landed a Merck collaboration worth up to $838 million. 

Access this and hundreds of episodes of the Business of Biotech videocast under the Business of Biotech tab at lifescienceleader.com.  

Subscribe to our monthly Business of Biotech newsletter.

Get in touch with guest and topic suggestions: ben.comer@lifescienceleader.com

Find Ben Comer on LinkedIn: https://www.linkedin.com/in/bencomer/

Transcript

Welcome And Guest Overview

Ben Comer 0:04

Welcome back to the Business of Biotech. I'm your host, Ben Comer, Chief Editor at Life Science Leader, and today I'm speaking with Wyatt McDonnell, Ph.D., co-founder and CEO at Infinimmune, a company using a database of human blood and tissue samples, plus artificial intelligence to identify, engineer, and develop new antibody therapeutic candidates. We'll talk about why and how Wyatt helped launch the company four years ago, what's different about Infinimmune's approach to discovery and development, how the company landed an antibody discovery deal with Merck in March, worth up to $838 million. And what's coming next. Wyatt, thank you so much for being here.

Wyatt McDonnell, Ph.D. 0:48

No, thanks very much for that introduction, Ben. Great to be here and happy to share our answers to some of the questions people have with uh with the audience. And you know, as a listener, thanks for the opportunity.

From Wet Lab To Computation

Ben Comer 1:00

We're uh we're pleased to have you, Wyatt. I think it's an uh an interesting company. I want to start off with uh some background about you. And my first question was on computational biology, how you got interested in that field initially.

Wyatt McDonnell, Ph.D. 1:16

Yeah, so you know, that's a story that starts uh right out of college for me. Um so I went straight to grad school after I did my uh my bachelor's degree. I had spent uh you know three out of my four years in undergrad doing hands-on lab work, um, you know, got really interested in bacteriology, microbiology, immunology. Uh, went and joined a grad program that was very concentrated in all of those three areas at Vanderbilt University. And then, you know, I think you know, one of the things that really pulled me to Vanderbilt at the time, uh, when when you walk on campus, so you know, you fly in, you stay at this uh this little Hilton that's like right by the baseball field, you get a shuttle bus in the morning, you're all suited and booted and fresh-eyed, ready for your interviews. You get off this little traffic circle, and there's three buildings you're looking at. So there's a research building on your left, there's a hospital in the middle, and then there's one of the other big research core buildings on the right. So you take a walk 200 feet in any direction, and you're in and out of the hospital and out of the research buildings. And Vanderbilt had spent a ton of time investing in all of this infrastructure and you know computational architecture, really, where you know they had recruited at the time uh you know some really prominent geneticists. It's now where the All of Us Human Genome Research Program is operated nationally across all of the inbound samples. So it was really just this incredibly talent-dense environment where you could see this intersection of bench to bedside in a very unusual way, where the university and the medical center had made some very steep and clinically impactful investments in computation and in wet lab development and and basic biology. So that was sort of a pretty incredible environment to grow up in as a scientist. And then right around the first year, first two years of my of my PhD, um, later when I when I came to 10x genomics, I found out that one of my future co-founders had been doing all of this work at 10x that I was also doing in a in a different way at Vanderbilt and in the labs that that I trained in and that I worked in. So I'd spent this time building single cell assays using expensive automated robots to try and make that high throughput, working with some incredible molecular biologists to try and miniaturize some of the reagents involved, try to bring the cost of those down. And my PhD supervisor, uh, one of one of them, uh, he ran the genetics core and the genomics core, the vantage core at Vanderbilt. So basically, every time somebody went to that core with a library, you had an opportunity to ask them, hey, what are you working on? What are you doing? So Vanderbilt at the time was one of a very small number of institutions that had a 10X Chromium controller, a Fluidime C1, and a BioRad DDC controller all in the same NGS core facility. So I basically watched one of those customer lines grow while the other two lines did not grow. And turns out there was this whole intersection of yet another new angle of computational biology. It's just like, wow, we're not, you know, looking at you know a fruit smoothie thrown into a blender and coming out. We're now looking at what you can read from an individual cell. And that just was such a no-brainer, like, oh my gosh, you have to go work on these problems because we didn't know anything about it. And this was also right around the time when the you know, even the National Cancer Institute started telling investigators, hey, for a new R01 you put in, we're sunsetting the use of bulk sequencing technologies. Like we expect you to have a single cell technology component. So my grad cohort full of super smart interdisciplinary people, but very few of them were working on computation, and then all of a sudden started having problems and data sets that clearly needed a computational angle. It's like, well, I don't know, seems like there's an intersection there that would be pretty fruitful to work on and pretty interesting, get you a lot of exposure to a lot of different problems. And that turned out to be a pretty reasonable instinct and assumption, and uh, a lot of other people uh figure that out too. And that, you know, was part of how 10x built this incredibly productive research and development group was they just started asking around the community, like, what are you working on? And do you see how that could plug into what we're developing from a product perspective? So yeah, that sort of like 2015 to 2020 era was really, I think, a magical time to like go to DNA sequencing 2.0, get into single cell resolution, and you can see the you know, the academic and the the clinical impact of that technology just a decade later, which is also pretty impressive. Yeah.

Single Cell Data And AI Parallels

Ben Comer 6:25

Yeah, you meant NGS uh next generation sequencing for anyone listening who wasn't sure about that acronym, but it it sounds like the Vanderbilt obviously had some incredible technology, but it was it also sounds like it was a period of of evolution and change just in terms of the next generation, next generation sequencing technology. And I'm wondering if if what you witnessed there is similar maybe to how AI, you know, was starting to change, and maybe this happened first and AI was second, or maybe they were happening at at the same time. We're gonna talk about how you're using AI in a few minutes, but I'm just curious if you know that experience of seeing how the technology was changing in next generation sequencing uh is similar or mirrors in some way what's been happening with with AI.

Wyatt McDonnell, Ph.D. 7:13

Yeah, I I think it is. Um, and I and I think there's some there's some parallels there that are that are pretty sensible from a from a scientific perspective. So I think there's some early examples in you know this what can you do with single-cell transcriptomics set of papers where first you have the establishment of single-cell transcriptomes, then you lead into this natural set of questions. How do those transcriptomes come to pass from an epigenetics perspective? How do those transcriptomes move forward into a protein perspective? And you see the evolution in all of those little paths. And now we have AI that's asking the same set of questions from a, well, we have all of this data for one set of modalities. What else can we add to have a set of models that have a better understanding of how all of this works in the first place? So we want models that understand text, but we also want models that understand things that are a little harder and more important, like abstraction and communication. And those are things that require how do you take text and tell a story? How do you take text and turn that into a picture that communicates what you're thinking to a lay audience, to an investor, to somebody you're trying to recruit? And I think it's really fascinating to see these parallels develop where there's always some set of early ideas that open up a couple of roads. And the answer always seems to be you're going to need to come back and collect more data, even when you think you have one data type figured out. And if you can figure out how to make a couple of data types work together, you can do some pretty astonishing science and you can learn some pretty useful and interesting things about biology and translate that forward into solutions for patients.

Ben Comer 9:08

So you encountered uh 10x genomics while you were the company, while you were at Vanderbilt, and then you went to work uh for them after you finished uh school. Is that right?

Wyatt McDonnell, Ph.D. 9:19

Yeah, that's that's about right. Um, yeah, so I I started in 2015 at Vanderbilt and I finished my PhD late 2018, early 2019. I actually went directly onto the faculty at Vanderbilt, so completely skipped the postdoc phase, which you know, I would say if you're thinking about that and listening as a grad student, I think that's a repeatable decision. And also if you're thinking about that as a grad student and are thinking about doing a postdoc, that's fine too. Um, like there's there's a couple of paths there. But part of how I made that transition onto the faculty was you know, I had some incredible mentors at Vanderbilt, uh, Simon Millal, Evelyn Georgiev, John Kurda, uh Dave Aronoff in the Department of Medicine who had seen my contributions to a lot of different research groups and said, you know, it would make sense for you to join in a more full-time integrated setting here rather than you know, sort of doing this traditional walk up the postdoc to researcher progression, you see. Um, and unfortunately, 10x had a really interesting uh job offering at the time. And uh my future manager at 10x uh had put some code on GitHub several years prior. I had raised a couple of issues and proposed a couple of fixes to some things that, you know, in fairness to Mike, I think he had published this really interesting set of nature biotechnology papers. And then he went to work for 10x. And it's like, man, have you ever tried to work a full-time job for a company like 10x Genomics and tried to maintain code you wrote while you were in the academy? It's a really hard balance to split, right? So I was maybe, maybe a month, not more than two months into my faculty job at Vanderbilt when an offer came in from 10x and I've been talking to them for a little bit and I was like, well, boy, that's an opportunity you can't pass up. And it's like you can always go back into the academy. And I had seen some incredible researchers at Vanderbilt with industry experience, you know, some like some of the folks who invented you know, things like, you know, small molecule fragment synthesis and coming back in to train whole generations of grad students on like, how do you deploy that technology into new problems? Like, so if that's possible, maybe I should take a little excursion out, test my, you know, test the waters a little bit, cut my own teeth, see what it's like. And I I don't think I could have possibly made a better decision. Um, and and 10x was recruiting really well and really aggressively at that point in time. And you know, that was part of how they came to become this this powerhouse of technology that they still are today was like, we want people who understand how we can build new products to look at new biology that we couldn't query previously. Yeah.

Ben Comer 12:11

Well, tell me a little bit about that uh cutting your teeth process. You were at 10 uh 10x genomics, I think, for three years. Tell me, you know, a little bit about what you learned and then when it it dawned on you that you know maybe you could take an additional step and start your own company.

Wyatt McDonnell, Ph.D. 12:27

Yeah, happy, happy to talk about that. Um, yeah, so broadly there are not a lot of life science tools companies that can do what 10x does. And what I what I mean by that, uh one, when you have an organization that's full of PhDs, even outside of the academy, you know, this is why people joke about professors arguing with each other at universities. It turns out that dynamic still exists in industry, right? So TenX sort of did this incredible selection process for people with PhDs that were able to articulate where their knowledge ran out and where they were capable of asking questions to colleagues who knew things they didn't. And they were capable of building that organization, you know, Serge, Ben, Kevin, uh, Mike, their CTO, they took that talent establishing mentality, not just in early research and development, but through every department of the company. So that's still actually true, even in the commercial organization. It's true in the field support community at 10X, where you have technically savvy people who are really good listeners, who are good communicators, and are good at identifying problems and a set of solutions to that problem early on. So I joined the computational biology department, uh, moved out to Pleasanton. Uh, you know, you walk into the Stone Ridge Mall complex, which they've expanded several buildings now. Uh, go up to the fourth floor where computational biology is seated, and you encounter this very, you know, this is all pre-language models, pre-like Silicon Valley is so back, right? But you're literally sitting back to back in an open office cubicle configuration desking system that is so innovative, you want to go to work in the morning, and you don't really want to leave work very badly, if at all, by the time 5, 6, 7 p.m. comes around. So 10x started building all of these proficiencies, not just in computational biology, but in microfluidics, in physics, in chemistry, in high-speed imaging, and all of these groups came together and started building all of these projects in this really impressive start small with a small team and then scale that team up as you get closer to product launch, and then scale that team out across the organization as you're supporting a product on the market. So I started and, you know, initially some work that I had done in grad school, I think my my first month at the company as somebody who didn't know anything about how companies worked, I emailed a couple of the senior leadership team members and said, Hey, I'm Wyatt, nice to meet you. I'd like to talk to you about some stuff, which is a pretty crazy thing to do as a first timer, but it was even more crazy, you know, how how generous senior leadership at 10X were, because at the time, and like I didn't know this, they were doing their IPO roadshow and they still took time to sit down with everybody who would email them and like hear new ideas as they were coming up. So I told several folks, hey, I think this set of technologies, the Chromium platform, could do some incredible things if pointed at antibody biology in particular. It's like, okay, that's that's interesting. Like, let's flesh that out a little further, let's think through that over the course of a couple of months, see what could come to pass. So this is 2019 going into 2020, right? Well, a little thing came along not long after called COVID, and all of a sudden, antibodies. Everybody in the country knows what an antibody is because everybody is thinking about how the hell do I get an antibody? And how do I know my antibody status? And what kind of antibody therapy could I receive? Like, how do we get antibodies with vaccines, right? So it sort of provided this really natural segue into the signature product that I worked on. Uh so I worked on barcode-enabled antigen mapping, and that was a technology that let 10x customers pair antibody sequences with the antibodies and antigens that interact together in a single experimental workflow. So you might recognize that sounds something like what we do today at Infinimune, and it is. We do this in a pretty different way, technologically speaking, and we do it at a much larger scale, hundreds, uh actually coming up on thousands of antigens at a time versus maybe five to ten that you know even some pharma companies are still doing basically the same way. But I think the things you learn from that at the end of that several year journey, one, a really good interdisciplinary team can do things that other teams cannot. So if you can get chemists, engineers, biologists, and actual software developers all in the same room, uh unbelievable what you can do with a team like that. Two, if you look at the gap between what the technology enabled and what was actually being pursued with it, being able to understand how to close that gap is what builds companies like 10x Genomics. Being understand, you know, being able to understand how you operate in that gap can also build incredible therapeutics companies. And that's what we now do at Infinimune. And really, you know, I think you know, if you were to go and read some of Serge's interviews from 10X, he could not be more right. This is the century of biology. We now have a way to read human immune systems in unprecedented detail. So for me and for our founding team, the obvious next step to make therapeutic antibodies wasn't really being chased. And you know, I spent a bunch of time touring different pharma customer sites, doing webinars like here's how you could think about using this technology, here's how you could use it for antibody discovery. And pharma are so great and focused and resourced. And unfortunately, it's really hard to take all of that focus and resource and narrow it and point it really, really intensely. So it was sort of a sense of like, well, there isn't a total addressable market case to do what we do at Infinimune. It is obvious that there's so much incredible biology that is completely untapped and not understood. Uh the NIH funding environment, uh, that's not really a way of developing any of this. And it's certainly not a sustainable, repeatable, low-risk way of engaging. And I think over the past few years, we've had a pretty good run of it so far. And like largely that that turned out to be a pretty good set of experiences and sort of foundational to uh to what we work on. I think the other thing that is like ideally, so if you're somebody like me, idiot coming out of grad school, doesn't know what you're doing, doesn't know anything about any of these different things, there's a lot of value to just getting a haircut and getting punched in the nose, and it's like, oh my gosh, I have written some of the worst garbage code on the planet, proudly published it on GitHub, described myself as a computational biologist, and then you go work on a team like that at 10X, and you're like, oh, there are real software people from Google here, and I don't know what I'm talking about at all. There are people who have trained full-time exclusively in computational biology for decades of their life. I don't know any of how they do, you know, what they do. Like, I remember leaving my interview from 10X and going, wow, so that's what it's like to fail an interview process. And because, like, you know, you're given like proper software engineering questions in computational biology that don't come up in grad school, or it's like, I need you to write me three algorithms on the whiteboard right now, unaided, no chat GPT. That didn't exist at the time, like no lead code interview questions, those didn't exist at the time. Tell me how to compare two Kmers of lengths eight, 10, and 12 and find the two that are closest in the set. And it's like, well, that didn't come up in my qualifying exam, and that didn't come up when I was writing my K99 or R01 applications, right? So it's just really good to know what you don't know and just have that enforced on you so strongly. I think it makes you a better scientist and it makes you more comfortable with uncertainty. And that is really at the core of what we do in biotech as a sector. We have to get really good at those two things high-quality science and operating in uncertainty.

Ben Comer 21:21

Absolutely. And uh you mentioned the conversations you had with executives at 10x Genomics. At that stage, were you already asking them questions about company formation, you know, like how to fund a startup or anything like that?

Wyatt McDonnell, Ph.D. 21:36

No, not not at all. Because, you know, part of part of that learning experience is like, how do you build consent in an organization full of incredibly smart technical people? Like basically, if you want to convince people to do something meaningful, doing that in a positive fashion and not in a critical fashion is the only way you get anything done because all scientists. Largely are smart and finicky. And it's a lot better to approach people and say, like, hey, I have an idea. What do you think? rather than I think we should do this. Do you think this is a stupid idea? Right. So there's a there's a way to do that right that 10x built an incredible culture around. So I think in the in the earliest, you know iterations of what I worked on at 10x that turned into products, a lot of that was why should we do this? Who is this going to help? Who are the researchers? Who are the customers that are going to benefit from being able to have a new generation of antibody and T cell receptor sequencing? From that, who are the customers that are going to benefit from having antigen specificity on top of that? What are the solutions that they're trying to bring into their labs? What is working well? What is not working well? So a lot of it was really more how do you know when you have a promising product opportunity? And coming out the other side of that, you know, you get exposure to all of these different uh inbound opportunities where people are building on top of 10x technology or have compatible 10x technology. And it's like, okay, you're called in for 15 to 30 minutes to share an opinion or a perspective on something that has come up from top down. And it's like, oh wow. Yeah, there's a lot of other smart people who are still out there and still don't work at 10x. Like, I need to be thinking about what we, you know, what we work on. And so at the, you know, at the think at the time, you know, it's it's also hard if you're a company like 10x to it it's it's challenging, right? Because on the one hand, you have all of these people who can do all of these incredible things, and it I think is really tempting to want to let them run and do all of these different directions all at once. And I actually think one of the best things that happened during my time there was the discipline that Serge and Ben and Mike had top down. It's like, no, we're really focusing in these lanes so that we establish early technologies in these new areas of biology, and we're not going into therapeutics or other things like diagnostics right now, and we're not talking about that because it wouldn't make sense. Like we're we're executing really well. We know we have traction with all these customers, we keep hearing about all this biology, they still can't ask it. Like we have to keep our eye on the ball. And I think they made the right decision. And um, the other end of that for me is like, well, I've wrapped up a couple of projects here. I've met some of the smartest people I've ever met in my life. I see an opportunity to be one of the customers of 10x, and we are a 10x customer at Infinimune. It's like there's all of this biology that I think I have to go lead a team to go explore. And and part of part of that was, you know, coming out of COVID. Um, you can read a set of papers that uh we published uh out of 10x and a set of patents as well, where we identified basically in a week, several thousand incredibly broad neutralizing antibodies directly out of humans that still are bulletproof from a variant perspective today. We published some of that work in the cell family of papers, and it's like, man, once you see that kind of workflow, you're like, ah, research and development of products is great, but that's therapeutics. And like, I really want to work on therapeutics now, right? So it's sort of that siren song of if that is true for infectious targets, man, I think the textbooks that I learned from, the textbooks I taught from, are wrong about human immunology. And I would like to go see if that's true. So that's what we do. Uh it's a pretty, pretty compelling thing to chase. And that's the basis of our product platform today with the drugs we're working on. Yeah.

Founding Infinimune And Raising Seed

Ben Comer 26:02

And so you had, you know, you had this kind of forge this scientific directive. You knew how what you wanted to pursue, how you wanted to pursue it. Can you talk a little bit about just the the I guess the business side, you know, how you started up the company, um, how you secured that initial 12 million uh in seed funding, and and what was that like? We're just working through science, scientific side aside, what what was it, what was it like on the company founding and and startup piece?

Wyatt McDonnell, Ph.D. 26:32

Yeah, definitely. So there's a re as it turns out, there's a reason that most companies in this space start in a very small number of cities, right? And you're looking at Boston, you're looking at the Bay Area, you're looking at San Diego, and really not a whole lot else in between in the United States at least. Um those are the big ones for sure. Those are the big ones in the US, right? Um so I think all founders should do this exercise. Sometime if you look at the companies you really admire, and you look at companies that have been funded in the past five years, it's good to sort of build out these how did that company get to where they got? So you can usually trace this back to who was the first check or who wrote the first checks plural into companies that you as a founder really respect, you know, either from a technology they've built, products they've built, you know, trajectories they pursued in a meaningful, aggressive timeline. And for us, you know, that basically turned into a small list of early funds in the Bay Area that had invested in several companies that our founding team really, really admired. So uh one of the first people I called was uh Josh Elkington at Axial VC. And Josh uh ended up being, you know, I think really the first ardent early supporter of the company. Uh he said, yeah, you know, let's let's go walk by my office in Berkeley, let's meet up for half an hour. And that turned into like a four-hour hike around most of the Berkeley campus, several surrounding city blocks. Uh, we sat down and and caught a meal together, and he said, Yeah, this is the bones of something very good. I think you should make some slides, and I think you and I should chat again over the weekend, and then let's see who else in my network would be interested in in talking to you. So basically, one lunch meeting turned into our seed fundraise, and so you know, that was just after we'd incorporated uh in July of 2022. And he said, Okay, here's 10 to 15 funds you should go meet. Go pitch them, go figure it out. So you're walking around with all of these horrible, ugly slides trying to convince people we have this idea, we have this plan, it's coming together, we will figure it out. And then you have more meetings like the one we had with Josh. So I have similar setup. I went into Playgrounds' offices down in Palo Alto, this beautiful cannery that they've turned into an incredible venture studio. Same thing. This is supposed to be a lunch meeting, a little 30-minute meet and greet. And it turns into like an hour and a half, maybe two-hour whiteboard session with Ben and Jory, who then let our seed round and was like, wow, look at all of these antibodies that exist in humans that you can't find with modern DNA sequencing technology. You can't even find it if you buy a kit from 10X, which is arguably the only way you would find them in the first place. So if the textbooks are wrong about what a human antibody looks like, what else are they wrong about? Probably they're wrong about how human antibodies drug human targets inside the human body. So basically, we I think second week of July had incorporated the company and we closed the seed round just three months later and brought in 12 million uh for the seed. And we just said, we're gonna get to work, we're gonna get our platform established, we think we can develop some pharma partnerships before our series A. And we think we can do this in a very capital efficient and build a portfolio of assets. So weirdly, about 95% of what we said we were gonna do at the seed, we ended up doing as a company. And that was very satisfying to see, and it's been helpful in telling our story to you know, future investors, future partners, current partners as well. Um, you know, it turns out if you have a little focus and discipline and do what you say you're gonna do, that's the highest signal you can give somebody outside of the company that they should keep talking to you, keep learning about what you work on, and keep coming back to the conversation.

Ben Comer 30:52

You excuse me, you you close this initial seed round. Um, what could you say, Wyatt, about I guess hiring initial the the initial executive team and and how you approach that? And you know, I I assume that you it sounds like you learned something really positive uh from the way that 10x genomics uh hired. Um, but uh and maybe you used a similar strategy, but um what what would you say about that? That you know, bringing the money in and you know, yeah, you have a company now, you need to staff it, you need to, you need executives that you count on. Maybe they were already in place. I don't know.

Wyatt McDonnell, Ph.D. 31:27

Yeah, I I I think that's I think that's right, Ben. So uh, you know, I weirdly we actually took the same approach even to our seed round, where it's like, you know, the people we wanted on our cap table were people that could change our minds about how we thought. We wanted investors who could tell us, I don't think you're thinking about that quite the way you should, or you thought about it this way. And that's in our experience as a as a team been a pretty good ethos to adopt. So uh, you know, I I think uh there's a few things I care about. You know, again, interview processes I think there's ways of doing them wrong, and I think those ways outnumber the right ways of doing it. Uh, and it's also a thing that's very specific to the team you work on. So I think approaching those conversations as you're trying to get to know somebody. It's really hard to get to know somebody in an interview process, and it's really easy to ask people canned questions that they will give you a canned answer for, and then they haven't learned anything about you, you haven't learned anything about them, and that's a pretty good recipe for things going poorly shortly long term, right? So I I think hiring for judgment under uncertainty and finding people who are comfortable not knowing the answer is really, really important. So getting to an honest, I don't know is an excellent signal in somebody as it's like, you know, if we're hiring you as a subject matter expert, whether that's in business development, preclinical development, clinical development, we as your teammates need to know when you have run over your skis. Like we need to know when you're out of your depth and your expertise has reached its limits. Totally fine to do that, but we can't work well in a team where people are saying something confidently and they're wrong. Like that's very destructive in our in our team in particular. So finding people who've worked in environments where the right move wasn't obvious, wasn't on the menu, wasn't just like sitting in front of them, that's a really valuable thing to, you know, screen for. And I think, you know, honestly, my approach over the past several years has really evolved mostly towards behavioral interviews, because I think that's where you get evidence of how somebody thinks, and you know, evidence of how they actually operate versus their their pitch about how they've operated. And you want to pick people who have done the job before, but still want to be in the trenches, rolling up their sleeves, doing the work. You know, there's a window in some folks' career where they have the experience and they still have the bug for it, or they have the experience, they still have the bug for it, and they went on to management. And they're like, man, this whole hierarchical organizational thing, I just really wish I was still in the weeds doing the work myself. And that's really, I think, where I like to hire. And then lastly, you want people who are curious and people who are kind. Um, you know, if you ask somebody, you know, tell me about a time, you know, where you know you, you know, you know, harmed your team or like team's progress was hindered by something you did personally. Like, how did that go? Who did it affect? What'd you do differently? You want people who have self-awareness and can answer that question and understand what they could have done differently, what they should do differently. And that's really at the core of a very, very functional team. Smart people who are very hands-on, who care about not disappointing their teammates and are honest about what they don't know and can be creative about fixing that gap and filling that gap in.

Human Derived Antibodies At Scale

Ben Comer 35:23

Yeah, that's uh that's excellent. I uh I want to ask you next, Wyatt, about um uh Infinimune's uh discovery and development platform. Infinimune, is that suggesting an infinite number of antibodies or or what's is there anything behind the name there that you want to mention?

Wyatt McDonnell, Ph.D. 35:40

Yeah, you know, behind the name, uh we thought it would be good to have immune in the name. We thought it would be good to remind people of the very successful companies that have had a name plus another syllable in front of it, like Metimune, for instance. And we wanted people to be aware of you know what we felt the potential looked like, which is it seems like there's a almost infinite amount of antibody biology that can be translated directly from human to human, and also is relatively unknown and unexplored. So that's that's where the name came from in part. Um it also helped that we had a completely clean SEO profile. So at the time, if you googled Infinimune, zero results, and that is great when you need to be high index and searchable, right? So yeah.

Ben Comer 36:32

Absolutely. Well, you uh you know, you said that um perhaps the textbooks or maybe they are definitively wrong, the textbooks, about human antibodies. Uh, I noticed on the website, you know, it it mentions discovered in humans uh in terms of your process for drug discovery. Could you just maybe talk a little bit about that and what that means and maybe why it's different from, you know, the way that antibodies have been surfaced in the past?

Wyatt McDonnell, Ph.D. 36:59

Yeah, definitely. So uh the platform is something that we invented at Infinimune. It's built on top of single-cell technology that our founding team helped develop at 10x Genomics. The basic idea is that instead of making antibodies in mice or selecting them from synthetic libraries, we can read antibodies that the human immune system has already made, evaluated, refined, and curated, right? So that the human immune system spends a lifetime curating its own answers to disease. And since you know, Miller, sorry, uh, Kohler and Milstein back in the early 70s first described monoclonal technology, we spent like 50 years as a sector trying to reinvent what the human immune system already does every single day. So in Finimune, what we do, we sample human memory B cells at scale, we sequence and characterize them at single cell resolution, and we identify antibodies that are already, you know, good or more narrowly functional. So antibodies that are already affinity matured, antibodies that are already highly developable, antibodies that have already been validated by a person's own biology. And it turns out that when you do this carefully across many different donors, you find convergent solutions in a couple of different forms. We published this paper in Nature in 2022 that shows somewhat surprisingly that antibody generation in you and I is semi-deterministic. So basically, if you pick an antibody that you and I have in common and you look at its heavy chain, there's an 80% chance or higher that we also have a light chain that is paired with that antibody. And that gets to how do you have a stable molecule that's antigen-specific? Native pairing of a heavy chain and light chain, we now know to be super important in the development and selection of high-quality antibodies. Separately, if you start seeing targets surface over and over and over and over again in all the different humans you screen, you start to get a feel for what targets the immune system has nominated as functionally important to regulate with autoantibodies. And that's the source of most of our pipeline targets today. And it's a strong signal that you can pick up way before phase one, right? Like if you want to look at target biology and how it translates to a human, ideally you'd have inhuman signal before you ever dose a patient. So that's really at the core of what we do today. And and most therapeutic antibodies that are in the clinic today were not discovered in humans, right? They were made in transgenic mice or they were selected from yeast or phage display libraries or generated as hyperdomas, you know, all these things, which again are useful, but they're substitutes for the real thing. So discovered in humans just means what it sounds like. We find drugs in the people that we're trying to treat or in people that are protected from a disease that we're trying to treat. So, you know, we think the human-derived antibodies have been through the best and most stringent quality filter that exists, which is whether or not that antibody could survive and function in a human body. And we've also learned along the way that they tend to have better developability profiles, better tolerability profiles, and even hit epitopes that other other antibody discovery technologies might miss entirely. So it's a different starting point. You get different drugs out and you get higher quality drugs out early.

Ben Comer 40:36

And then the AI is helping, I guess, reduce the noise to signal between the um the antibodies and targets. Is that right? Is that generally right?

Wyatt McDonnell, Ph.D. 40:47

Yeah, so we so we get this question a lot. And uh, so basically the first week on platform, we're screening physically in the lab. So there's no prediction, there's really no AI involved in that early discovery process. We're screening millions of memory B cells against hundreds of targets in parallel all at once, and we're reading that using next generation sequencing. AI in discovery at Infinimune, we use in a couple of different ways. So it's genuinely useful for protein engineering, you know, sequenced property prediction, uh, optimization, filtering of large candidate sets, uh, engineering of initial discovery variants. This is where we use it. Every discovery program we run has a wet lab component that you can't AI your way out of, but you can use AI to build basically molecule-specific engineering models. And we found that to be incredibly productive in our own pipelines. So basically, the biology ends up being the constraint, uh, not the compute. Uh separately, we also use AI for you know practical things, you know, things like this the Zoom call, for instance. You know, we use it for meeting summaries, you know, we use it for getting up to speed on the literature, we use it for internal research tools. Um, so like we, I guess we lean towards underclaiming rather than overclaiming. You know, our platform is a biology platform with computational components and a protein engineering platform where AI has some incredibly useful plugins and has made some awesome advancements in our own pipeline.

IFX 101 Timeline And Manufacturing

Ben Comer 42:25

I want to uh talk about the deal with Merck in just a second. Um, but before I do, I wanted to ask you about your your lead internal clinical candidate, IFX 101. And just give me a sense, I and I that's it's not yet in human clinical trials. Can you give me a maybe a ballpark idea of when that might enter clinical trials? And then also what can you say about the manufacturing process for these therapies? Are you using a same kind of bioreactor that you know the monoclonal therapy monoclonal antibody therapies have used in the past? Uh yeah, get just give me a sense of of uh that that I think is your lead uh development candidate, IFX 101 and and yeah, the manufacturing.

Wyatt McDonnell, Ph.D. 43:09

Yeah. So uh we expect IFX 101 to enter the clinic in early 2027. Uh, we're developing it for uh antibody disorders where the antibody biology matches cytokine biology or cytokine receptor biology. So we're developing it in atopic dermatitis. We also see some other opportunities in immunodermatology where this barrier biology has not really been treated with. There's a lot of programs in immunoderm that deal with T cell inflammation, but less so barrier inflammation specifically. So we're on track to enter the clinic in early 2027 with IFX 101. From a manufacturing perspective, so IFX 101 is several things. First, it is a half life extended human antibody. Second, it's manufactured in Chow cells in a standard process at KBI Biopharma, our C DMO partner. So it is built here in the US, in uh North Carolina, and in Durham. And as a molecule, it's been engineered for a high concentration subcutaneous formulation. So over 200 milligrams per milliliter, which allows us to pack a lot of drug into a standard small injection volume. So the idea is that patients will be able to, at the end of this, dose at home with a small injection volume rather than an infusion. So that combination of a really long half-life and a developable high concentration formulation is what enables us to think about quarterly or less frequent dosing in the clinic. And we work with KBO, KBI biopharma for a couple of reasons. But first and foremost, they are one, an excellent partner. Two, they're handling cell line development using this incredibly high quality cell line they acquired several years ago called Selexis. So super high yield, very stable, uh stable not just at hot temps, but also cold, uh really productive molecule. And separately, just a sterling reputation for that cell line being portable to many, many different manufacturing environments, which is really important when you're talking to you know, not just pharma partners or investors. But if you think about one day you're going to have to manufacture this molecule around the world, you want to have a really high quality cell line that can actually transfer around the world successfully to lots of different teams. And at their core, KBI began as a formulation shop. So they have a lot of expertise in building these high-quality, high concentration formulations, which are what we think the future of patient care looks like for antibody biologics in particular.

Ben Comer 46:00

Let's talk about the Merck partnership. Um there are a lot of you know, young companies, startups that have an innovative platform that would absolutely love to sign a deal with a you know blue chip partner uh like Merck. Um what can you say, I guess, about the initial approach and engagement and you know how how Merck, and I don't, I'm not gonna ask you to read Merck's mind, Wyatt, but how do you how do you think, you know, uh from your perspective, uh this partnership came together?

Wyatt McDonnell, Ph.D. 46:30

Yeah, so I mean I think it's it's fair to say that Merck has one of the most sophisticated antibody groups in all of the industry, and they're discerning about what they engage engage with you know when it comes to small companies like us. I think two things really resonated. One, our platform finds drugs that other approaches don't find, antibodies against difficult targets with developability profiles that tend to come out of human-derived material, with immunogenicity profiles that tend to come from material directly matched to a human immune system. And then, two, you know, the team. Merck does a lot of diligence on the people, not just the technology. And we have a group that has deep experience in single cell biology and antibody discovery. We have built our own pipeline. We weren't waiting for a partner to tell us, you know, we think you should build this molecule against this target. So that you know, the fact that this is the first antibody partnership they've disclosed economics on in quite some time tells you something about how often they decide to do this. And really it's matched to Infinimune addresses antibody biology in a way that is very complementary to what Merck can do, complementary and you know, compatible with internal discovery of Merck and matching their quality standards. And you know, do they think we can actually produce molecules that will match clinical quality in a partnership like this? And being able to show them progress on our own pipeline over the course of that conversation was incredibly uh important. So and then you'd asked uh like how did we make the first connection with Merck as well? Yeah. Questions that they have. Okay. Yeah. So so the first conversation was with their search and evaluation team for platforms uh in December of 2024 through an introduction from one of our advisors. So their early questions were the right ones and also pretty common ones. So, how does a platform actually work? What differentiates it from what's already out there? What's the evidence that the antibodies you find at Infinimune are better than antibodies found other ways? So we ended up meeting in person shortly thereafter at JPM in January 2025. We met again at Bio in June, and you know, throughout, you have several different teleconferences with a sort of expanding group of research and development, platform SE, business development, legal, as you figure out, okay, here's what we believe, here's what we believe we can turn this into from a partnership perspective, over what timeline do we have something that's a fit to both parties? And the answer is yes. So I think it took roughly 15 months from first conversation to announcement, and you know, much shorter than that in terms of actually pulling the deal together and getting through diligence, because you know, there's diligence along the way was substantive, basically deep scientific review of the platform, technical diligence on our underlying methods, including our language models, our AI component, like what kind of predictions are you making? Is this generalizable? And then ongoing engagements as we showed them, you know, updates on platform performance in the pipeline. Like our internal programs, as we move them to KBI, like, well, is it a molecule you can make? You've said it's gonna be great. Did that actually come to be true? You've said that it's gonna have an incredible half-life. Did that come to be true? And the answer is yes. So like now we're you know, coming up on very early into the clinic next year, and Merck got to see that story develop, and that's how they build conviction around us as a as a partner.

Ben Comer 50:22

Yeah, I think that's really useful to point out, and I'm glad you did, Wyatt, is that not only are they you know looking at the work you're doing for the the targets that they're interested in, they're watching to see if your technology that you're developing internally is is gonna pan out. And I I I think the targets have not been released that that Merck is interested in, but are is is Merck getting access to your platform to do some experimentation, or are you um you know, develop, I guess, identifying antibodies that you think might be interesting to Merck given their parameters and presenting to them. How does that work?

Wyatt McDonnell, Ph.D. 50:58

Yeah, so one type of deal you sometimes see in the space is you know, we'll see companies like Adamab or Alloy Therapeutics describe, which is, you know, we've completed transfer of a technology platform to a partner, and we've completed, you know, execution of that platform through to a clinical candidate, through to development. Ours is a slightly different story there. So we're still retaining our platform, operating our platform, not transferring it out, not licensing it out. We onboard targets of interest to Merck, and then we operate our platform. And Merck owns the outputs, the candidates of the platform for those targets. So basically, we're doing all of the high-speed, high-quality, early discovery of candidates that Merck then advances through non-clinical studies, clinical studies, and ultimately commercialization. And that's why it's a multi-year, multi-target uh partnership. And then these are targets that are not currently in our pipeline. So from a program perspective, it's additive, not subtractive to what we're currently working on.

Ben Comer 52:07

Well, I uh I hate to cut this conversation short. We're running out of time, Wyatt, but I did want to ask, you know, just in terms of uh potential uh partnerships, additional partnerships, do you have the capacity? Are you looking for potentially additional partnerships like this one? Are you focused, you know, primarily on Merck and your internal development candidates at this point?

Wyatt McDonnell, Ph.D. 52:30

Yeah. So our top priority as a company right now is delivering IFX 101 and IFX 201 to First and Human safely and on schedule and continuing to build our platform behind it. Uh, and we believe that you know the Merck partnership is, we think, a pretty credible signal that our platform works. And the next signal for us as a company is going to be our own clinical data. And the clock on that will start in in early 27. Uh, in terms of additional partnerships, uh, we have one other discovery partnership with Immunome, which is uh part of the former Seattle genetics management team alongside the Merck Collaboration. We are selective about partnerships. You know, our criteria are you know, do we learn something we couldn't learn alone? Does the partner bring something we don't have from a targets and you know knowledge perspective? And does it not get in the way of building our own pipeline? So if a conversation meets those, we'll have it. And if it doesn't, uh, we won't.

Ben Comer 53:26

Got it. Well, um, Wyatt, thank you so much for being here. I I enjoyed speaking with you.

Wyatt McDonnell, Ph.D. 53:31

Yeah, thanks very much, Ben. Uh, have a great rest of your day and looking forward to catching up again down the road.

Ben Comer 53:35

Likewise, we've been speaking with Wyatt McDonald, co-founder and CEO at Infanimune. I'm Ben Comer, and you've just listened to the Business of Biotech. Find us and subscribe anywhere you listen to podcasts, and be sure to check out our weekly video cast of these conversations every Monday under the Business of Biotech tab at life science leader.com. We'll see you next week, and thanks as always for listening.