All right. Welcome back to TD Cowen's Genetic Medicines and RNA Summit. I'm biotech analyst Brendan Smith here, part of the TD Cowen Biotech Group. I'm joined today by the CEO of Generation Bio, Geoffrey McDonough. I'm going to start off with just a few background slides before we dive right into some Q&A. So, Geoff, do you want to go ahead and take it away? Thanks so much, Brendan, and thank you for the kind invitation to have this fireside chat. It's always fun. Today will be no exception on a Friday afternoon. It's really a pleasure to just give a couple of slides, and I'll only show two slides to give some background on Generation Bio. This entry slide that talks about the limits of genetic medicine is all around limits in two dimensions: delivery and cargo. Delivery is about bringing LNPs in a specific way to systemic delivery beyond the liver and the spleen to new cell types and tissues like T cells and HSCs. And on the cargo side, it's about non-immunogenic, long-lasting DNA cargo free of viral delivery. So it's a delivery and a cargo set of limits that we're looking to eliminate here at the company. Just as we are in a public markets environment, a reminder of some of the forward-looking statements, considerations. To underpin that desire to remove the limits on genetic medicine, a slide each on the two platforms that we've developed here over the last five or six years. The delivery platform is something called cell-targeted lipid nanoparticle delivery. And like all lipid nanoparticles, these are re-dosable. But unlike every existing LNP that we're aware of, these are highly selective in getting to tissues beyond the liver and the spleen. And I think for the first time, opening up the potential to use LNP delivery for a wide range of genetic medicine cargoes to tissues beyond the liver and the spleen. As I mentioned earlier, we're prioritizing T cells and HSCs as well as hepatocytes. But in principle, as I'll talk about, this is a platform that can move to many tissues over time because the tissue-targeting moiety or ligand that is driving the selective delivery is changeable and highly adaptable. And so that side of the equation is cargo independent. We can deliver siRNA, mRNA, guide RNA, and our own cargo immune-quiet DNA. The cargo side of the equation is this immune-quiet DNA. And as I said earlier, it's about unlocking the idea of years-long expression per dose from the nucleus of target cells from an episomal cargo that does not integrate. And that gives us, obviously, durability, but also titratability because each dose lasts a long time. When you add a second dose, you can increase the basal level of expression to match the needs of individual patients. And of course, this is a gain-of-function platform where we're really adding something to what cells can do as opposed to knocking down. So the combination here, in the same way that our delivery platform is really independent of cargo, because we've engineered immune-quiet DNA to remove the innate immune stimulatory nature of DNA, it's really amenable to multiple delivery modalities. And of course, our anticipation is we'll use our ctLNP delivery system to bring iqDNA to hepatocytes, to T cells, and eventually down the road to HSCs. So that's one out of two slides. I'll show one more slide, which shows how we're translating these platforms into an early portfolio where we've really focused on taking novel approaches in three program areas. The first one is delivery in vivo to T cells with an initial focus on understanding the application space and autoimmunity. The advantages of a re-dosable in vivo ability to program T cells are obvious. You can avoid conditioning. It can be point of care. It's re-dosable and adjustable. And here, too, the cargo is flexible. We can choose cargo based on the duration of action that we're looking for. And in autoimmunity, I think that's quite important. The length and depth of B cell resetting is something that's not quite understood yet. And having that flexibility in our cargo feels really important. And fundamentally, this is about expanding an already large market by allowing it to be addressed in vivo without the kind of overhead and time and cost that is there for ex vivo approaches. The second program area we're focused on is sickle cell disease. Here, we're tuning that cell-targeted lipid nanoparticle delivery system to go to HSCs. This is adopting the very same kind of indel editing approach that's been used by the approved products and just resetting that to the same advantaged platform of a re-dosable point of care kind of administration, which can, again, vastly expand the number of patients who can take advantage of this approach. I think the competitive differentiation, as it is for T cells, is quite clear as we bring this forward. Then lastly, importantly, is this focus on hemophilia A for the liver. The horizon in hemophilia A, since I've been working in the field over the last 15 years, has been advancing in two dimensions. The first is duration of coverage. Now the state of the art is about a week for most patients. For some patients, as bispecifics advance, they can get to two weeks. We'd like to really expand that envelope so that patients can live free of their hemophilia for years at a time. And that's the profile that we're aiming for with iqDNA here. But to retain that ability to individualize coverage. So the trough floor has been also elevating. It's about 10%-15% today. We'd like the ability to individualize that coverage to 10%, 15%, 25%, depending on the number of doses and the patient needs, producing endogenous Factor VIII, and here, too, on the basis of very different costs and duration, expanding patient access and competing effectively to create much longer windows of a hemophilia-free living condition for patients. So three program areas: T cells, HSCs, and liver. Autoimmunity in T cells to begin with, sickle cell for HSCs, and hemophilia for the liver, each of which, of course, can lead to a deeper franchise portfolio development. All of it sitting on a platform that we've developed to create really outsized COGS advantages over all current approaches, which is sitting on a foundational belief we have that we'd like to be able to serve more of the world at a time than just the modern sort of allopathic markets have historically provided for. We've got the company set, so we have time to build and mature this portfolio over the next few years with cash runway into the second half of 2027. So maybe that's a good place to stop so I can keep my promise on 2 slides and go into the discussion. So I'll stop sharing here. All right. Great. Yeah, this is a fantastic introduction. So I think maybe we'll kind of jump and double-click into each one one at a time here. So maybe we can start with the ctLNP, so the cell-targeted lipid nanoparticle delivery system. I mean, you gave a great overview of everything that you have going on. I know at ASGCT earlier this year, you guys really presented a bevy of updates across the board, really. So maybe instead of diving into the specifics of everything at once, what are kind of the most important updates to ctLNPs that you presented at ASGCT? And maybe help us understand why those were the priority and why this is the right time to make those changes. Yeah. So I'm not sure if I can cover a full bevy of updates, but I'll give you the highlights. I think the first thing to say is that LNPs, by their nature, have 99% of every administered IV dose absorbed by the liver and the spleen. When I say it's by their nature, it's by their charge characteristics, their size characteristics. And we think of this, if you want to get to new tissues, the data we showed at ASGCT was focused on delivery to T cells. The problem with the liver and the spleen acting like a sink is that if 99% of the dose ends up there, then you have 1% or less of the dose with exposure to systemic circulation and for a very short time, so that the opportunity to get to cell types and tissues like T cells is very limited. So the reason that we made this set of disclosures at ASGCT is that we wanted to show that by engineering our cell-targeted lipid nanoparticles to specifically avoid that sink of the liver and the spleen. So to get to the data, we showed that in humanized mice, we're delivering less than 1% of the dose to the liver and the spleen, which means that 99% of the dose makes it into systemic circulation. We have an extended half-life of over 7.5 hours in monkeys compared to minutes for every other LNP that we're aware of. And that is what allowed us to set the foundation for showing highly potent and efficient and selective transduction of T cells with in vivo delivery. This, of course, is the foundation for that first column that I showed of developing in vivo CAR-T for autoimmunity. So just to take the headlines of what we showed on delivery, again, in a humanized mouse model, at 0.05 mg per kg, 10-fold lower than the highest tolerated clinically validated doses of LNPs. We're transducing close to 100% of target T cells. We are driving no off-target delivery to splenic and myeloid B cells and other myeloid cell types. And we were able to show that not only does this obtain for GFP, but for the first time, we showed that we can do it with the delivery of a functional CAR. So when we're doing it with CARs, we're transducing up to 30% of T cells with about 4,000-5,000 CAR molecules per cell. And we're able to show that that CAR is functional in tumor killing and in in vitro assay. Progressing from an idea in principle that we weren't delivering to liver to now we are delivering on target to T cells in an efficient and a selective way with a level of CAR density that is quite relevant for what we know today about ex vivo approaches. Okay. That's great, actually. Maybe we can talk a little bit more about just the T cell ctLNP. So I think we've heard for a few years now about the more liver-targeted LNPs. I think people probably are historically a little bit better oriented with those. But these T cells that you're kind of touching on here have a pretty wide potential. But I guess for Generation Bio, I guess, what are the priorities within the T cell ctLNP vertical? Are these programs that you would potentially pursue on your own as part of the Moderna collaboration, a mix of both? I guess, where do you see the T cell LNPs fitting into that larger platform? Yeah, great question. So as you just said, we began a collaboration with Moderna about 18 months ago on the basis of this ctLNP platform. And our goal together was to optimize that platform for specific delivery to T cells, which I think we've crossed a very meaningful horizon around just recently, which I think is allowing both companies to do what you say, which is to contemplate where is the first place to take this delivery system into T cells. As I said at the outset, we have a lot of flexibility around cargo. And the way to think about our relationship with Moderna here is that we have broad division between us as it relates to cargo. So Moderna has a broader and a majority of rights as it relates to messenger RNA, and we do as it relates to iqDNA. So I think as we think about parsing these programs, boy, there's a lot of people investing right now to understand how important is it to have short-acting transduction and creation of CAR-Ts, and how important is it to have a higher or larger area under the curve for T cell transduction, specifically as that relates to the ability to reboot the immune system. So I think it's probably a little too early yet to say exactly where we will land. Of course, we can't speak for Moderna in that sense. But I think for G Bio, what we'd like to know is, does IQ DNA as a cargo offer a differentiated profile in how long a CAR-T is formed for in vivo, and how prolonged and how significant can that lead to a reset of the B cells? That's the state of the art as we look at deploying the system. The way to think about our rights broadly is mRNA with Moderna and iqDNA for Generation Bio. Got it. Okay. All right. And I know you also have some HSC-targeting ctLNPs, I believe. I guess, where are those in development kind of relative to T cell targeting? And then maybe kind of a similar question, how you're thinking about relative positioning of the HSC platforms. Yeah. So our thinking was with an emerging approach that could get very selectively to extrahepatic cell types, what other circulation available cell types could be therapeutically interesting? And HSCs really jump out when you do that analysis. For the first part, because we have this really elegant validation of an indel approach for sickle cell disease and beta thal with the two most recently approved products here, which work ex vivo. The idea that you could take a very well-validated biologic approach to recapitulating that indel, but in an in vivo delivery system was really appealing to us, especially because the great majority of patients, both with sickle cell and beta thal, are unlikely going to be able to access the current therapies given their necessary price point. So that's the start. Before I get to the portfolio question underneath, I'll just say that the way our ctLNP system is engineered, there's a core base composition, which we call a stealth LNP because it stealthily goes past the liver and the spleen. The conjugation, the attachment chemistry, and the ligand are ultimately adaptable. That's why we refer to it as a platform. That very same stealth composition stays the same as it is for T cells, but we adapt the conjugation and the ligand to go for HSCs. We're about 3 or 4 months behind the T cell work. We're now in the process of identifying many, many ligands against actually a variety of receptors. Now we're testing those in vivo. The process that we experienced with T cells is that it's not about finding the ligands. It's about fine-tuning the density of the ligands, the way that they interact with the receptors, and the behavior that results from that interaction with the receptors. So we're in the process of dialing in what that will look like for HSCs. Now, to get to your portfolio question, I think sickle cell and beta thal are destinations worth the travel in their own right. And there are a whole host of haploinsufficiency diseases underneath those where one allele is inadequate in HSCs and leads to progenitor disease. Many of those are in the hematologic space. PKD is one of those. There are protoporphyrias that are like this. And many of the myelofibrosis conditions result from core HSC mutations. So I think there's a very interesting space here under HSCs. The other thing I would say, just as a reminder to people like me who met up in medical school or thought about this 25 years ago, HSCs are literally the foundation for all of our circulating hematopoietic cells. So targeting HSCs can be done at the most primitive, highest parental level. But again, there's probably not only one ctLNP for HSCs. We can target HSCs at different levels in different lineages of progenitor cells over time. So we think HSCs within wherever we start can be very interesting. And there are levels within HSCs, I think, that we can tune delivery to explore. So we see this as just a really fertile area where delivery and a range of cargo can lead to a lot of product opportunities. Got it. it. Okay. All right. That's great context, I think. So if HSCs are, I think, about 3-4 months behind from the T cell LNPs, maybe then how are you thinking time-wise about those relative to the liver-targeted ones? And then I want to definitely ask you about some of the actual targets that you're exploring here. I think we're pretty familiar with GalNAc, but I know you guys have a few others that are kind of top of mind. So maybe help us visualize liver. Yeah, absolutely. So I think the HSCs program, the next kind of milestone or steps is about showing the same profile of in vivo delivery for HSCs, as we've just discussed for T cells. For liver, I know you want to come on to the delivery for liver in a minute, so I'll just focus on the cargo for liver. Our view is that the delivery of, for example, siRNA or mRNA to the liver is not a gigantic gating problem for therapeutics today. Although I think it's worth saying that the upper limit of dosing for LNPs in the liver remains about 0.5 mg/kg. So it's still not a very efficient or potent delivery modality, even for mRNA or siRNA as LNPs stand today. But the real thing that's required to change the curve for patients, for example, with hemophilia is duration. It's about amount and duration of coverage. I think DNA is uniquely positioned to give that kind of years-long respite from disease that I think will be important for making inroads into what is the largest rare disease market today at $13 billion. iqDNA, I think, is the cargo that can get us there. That is an engineered, partially single-stranded and structured DNA that we invented to remove the innate immune stimulation that is integral to double-stranded DNA. It was the challenge that we centrally struggled with here for the first five years at Generation Bio. It's a new molecule. We invented it last year, and we're in the process of working through what's going to be required to get it ready to be a kind of industry-ready cargo, which means working through the production and the purification, the final molecular tuning, and so forth to bring it to a level of efficiency and robustness to take forward. So now I come to your question. Will Sickle go faster than Hem A? We're not sure. As soon as we kind of cross the horizon for each program, we'll disclose and guide on the timing for the clinic for each of those programs. Okay. All right. Great. That's very helpful. I think within liver, I mean, I think there's a fair bit of understanding for GalNAc as a target. I think it's fairly well established at this point. But I think you're also exploring this at scFv and then VHH, both ASGPR. So I guess, how are you thinking about those two relative to GalNAc? Why not stick with GalNAc? Have you seen early data to kind of suggest that this would kind of benefit maybe not just Hem A, but the ctLNP targeting the liver a little bit more broadly? Yeah. Yeah. So maybe to answer your question, I'm just going to start with something we learned in T cells. This is an interesting thing about how learning becomes recursive in a platform, a company. When we started with GalNAc and liver, what we noticed was that it was a relatively nonspecific interaction between GalNAc and ASGPR. In other words, there's a very strong relationship between the sugar and the receptor. And you can have nonspecific interactions between the sugar and other elements that are present in the epithelium of the liver. Nothing wrong with that. And it did prompt us, though, to ask the question, what if we would try, for example, a small-format antibody to your question, like an scFv? And we found that the performance in the liver was quite differentiated from GalNAc, both in terms of potency and in biodistribution, which actually prompted us when we invented the stealth LNP to say, well, maybe we could try small-format antibodies for other tissues, like as we went on to do for T cells and as we're currently doing for HSCs. As we worked on T cells, we started to appreciate that there's a step beyond scFvs, these VHH domains, which are much smaller, and they are very easy to work with once you work out the techniques in LNPs. And what we found over and over and over again is that if we find a really good small-format ligand, we can find better ones. There's an optimization curve and cycle that can make them more ongoing like this because they get more potent. So that kind of allowed us to come back around to the liver and, to your point, ask the question, well, if an scFv is good, maybe an antibody is better. And maybe that becomes something that we can iterate to get yet better performance in the liver. And that's what we're finding, is that the antibody approach is just more specifically tunable than a sugar approach is. And that's why we started to explore these biologic ligands in addition to GalNAc. We haven't made a decision to move away from GalNAc. It's just to say that in a world that's always progressing, we'd like to follow our best and latest approach there. Got it. Okay. All right. I want to shift gears a little bit to now actually talk about the cargo. I mean, we kind of touched on it a little bit, but at ASGCT, you had some really interesting data in mice and in non-human primates to kind of showing that this next-generation iqDNA that we've been talking about has pretty minimal interferon and interleukin responses. I guess, help us contextualize just maybe the one or two most important changes you made from the earlier generation of standard to immune-quiet DNA and why those biomarkers are the ones that make sense to look at here. Yeah, great question. So well, let's start with the fact that DNA is fundamentally the building block of life that has a nuclear kind of basis or center of gravity. And as a result, it's not surprising that higher organisms are pretty oriented to detecting double-stranded DNA and reacting to it. And sure enough, for the first four or five years of our work here, we were confronted over and over again by the reality that when you deliver double-stranded DNA, such as a double-stranded DNA construct in an LNP, it is delivered squarely into the cytoplasm of immune cells and somatic cells. And those cells in different degrees, but largely in the same way, they detect that double-stranded DNA with cGAS-STING, one of two main pathways that are oriented towards detecting double-stranded DNA. And so as a result, what we would see over and over again, this is your question about the panel, across a range of cytokines, you would see large, and I mean large, not like 50% elevation, but thousands of fold elevation in these biomarkers, IL-6, interferon beta. There's a suite of about seven of these cytokines that are typical of an innate immune reaction to double-stranded DNA. The consequence of that is that you would get a flu-like kind of reaction in the animal, and you would have a relative suppression of the ability to produce protein. So bad on tolerability and bad for expression. So that's the thing that has held non-viral DNA back for so long. So our thinking was closed-ended DNA had some features that we knew to be very important, such as these inverted terminal repeats. We knew that, obviously, for the molecule to be functional in the nucleus, it needed eventually to become double-stranded to support transcription and then translation. But what our team started to grapple with is, well, maybe we can make the molecule substantially single-stranded with those ITRs while it's in the cytoplasm, get it into the nucleus, and then have it become transcriptionally active. If we did that, then substantially all of the molecule would be single, not double-stranded. As it turns out, cGAS-STING has a very, very low affinity for single-stranded DNA. We just didn't think it would be possible to make a structured molecule like that until we had invented something that we call Rapid Enzymatic Synthesis, which is a cell-free way of making structured DNA. Because we had that tool and that idea, we could kind of put them together and try it. And what resulted, the structure that we disclosed at ASGCT, is what has survived this interplay between having a very flexible production platform and five years of experience of what it takes to get DNA into the nucleus. And iqDNA is what resulted. And to your point, we're able to show that in a variety of formulations in monkeys, it's immune-quiet. They don't get these cytokine elevations. And because we have that kind of flexibility around the scaffolding, we can start to work our way towards a path of engineering on the molecule from a molecular perspective so that we can engineer in aspects of the molecule we want to support a greater or steeper potency curve. That's the kind of stage we're at now is we're quite sure we've eliminated or made the innate immune reaction low in a way that's comparable to purified or modified mRNA. Now it's about moving through the engineering cycle of manufacturing, purification, and structure activity changes to get it in the right place from a potency perspective. Got it. Got it. Okay. All right. So I think a lot of the data that we get from mice and through NHP is a little bit kind of earlier construct. I guess some of the luciferase iqDNA data that you presented as well. I guess, help us contextualize where the actual Factor VIII iqDNA construct is, maybe next steps there, and if you have any sense of rough timeline. Yeah, absolutely. So it's on our list this year to be able to hit a potency target for Factor VIII to allow us to advance the program. And what that looks like for us is at about 0.5 mg/kg, around 50% expression in mice. That's about the threshold where we feel confident about the translation to NHP. We've shown that we're in the kind of mid-single digits, 5, 6, 7% in mice with a dose at that level with the current or the last initial generation of iqDNA. We now have what we're referring to as a second-generation scaffolding that we're testing for Factor VIII. So as that data matures, we'll communicate it to the market and in medical meetings as well. Got it. Okay. All right. I know just in the last minute here, I do want to ask you because you touched on the Rapid Enzymatic Synthesis here. So I know it's been a big focus for you guys the last couple of years now since you kind of made the shift here. Just maybe help us understand, and I know we don't have time to dive into all the logistics of this, but why this was the right investment right now for manufacturing, not just to get you into the clinic, but kind of for the years to come for Generation Bio? Yeah. So this has just been one of those very lucky and very powerful inventions because we made the invention initially because we didn't want to deal with the impurities and the inconsistency of biologic manufacturing. So we're trying to solve quite a narrow problem there. But in inventing RES and indeed having solved that problem, we get almost 100% single-species purity and very quick turns. It led to the invention of this molecule we've just been discussing. And also, it's kind of an R&D engine by itself because it allows us to further tweak SAR and so forth. And to your point, exactly over the long term, what it unlocks is the ability to produce non-viral DNA cargo at scale at really incredibly low costs of goods. I think if I reflect on my last 30 years in the industry, every cutting-edge technology I've worked with would have gotten to 10- or 100-fold more patients had it been affordable, had there been an X factor on cost and scale. I think what RES unlocks for us as the platform really crosses the horizon into a series of portfolio programs, it means we can aim big from the beginning and to do that in a way that can support access. That's something that's always felt really important to us here. All right. Fantastic. I think that means we are right at time. I want to thank you, as always, Geoff, for catching up with you. A lot of exciting updates here. I want to just remind everybody tuned in, we have one last fireside chat to close out the Genetic Medicines and RNA Summit. So hang in there for one last session. Thanks so much, Geoff. Thank you, Brendan. Enjoyed it.