Hi, everyone. My name is Maury Raycroft, and I'm one of the biotech analysts at Jefferies. Thanks so much for joining us today. I'd like to welcome our guests, Brian Thomas, the CEO of Metagenomi, and Sarah Noonberg, the Chief Medical Officer. We're gonna do hybrid format. So, Brian's going to start off with some slides, and then we'll do some questions at the end. So thanks again for joining us.
Thank you. So I'm really pleased to be here, very excited to tell you about Metagenomi. We're a genetic medicines company, and we are leveraging a platform of discovery that allows us to go into the natural environment and find novel gene editing systems. We do this using a science called metagenomics. Metagenomics is the ability to go into the natural world, and instead of isolating an organism and sequencing the genome of that organism and then looking for novel cellular machinery, we actually sequence everything that's in the environment. It's a very exciting science because it tells you a couple of key things. It tells you that there's a lot of life on this planet that we had no idea was out there, and two, it tells you that we know what these organisms are capable of doing.
We know what their genes are now that we have their genomes. We use these reconstructed genomes as a database to go out and find novel gene editing systems, and we have done this on a really grand scale. We have thousands of novel programmable systems, as well as supporting enzymatic machinery that allow us to create other gene editing modalities. This can be seen here in our toolbox of gene editing capabilities. The gene editing capabilities are really grouped into two groups: the small edits as well as then large integrations. But it all starts with that top row of programmable nucleases. These are CRISPR programmable nucleases. They allow you to target a specific location in a genome and then cause a double-strand break there.
The uniqueness of our, of the enzymes that we're using are that they have a very small size, they're highly efficient, they show little, if any, off-target activity, which is really important when you think about, the safety of these enzymes and their application. And then finally, given the fact that we have so many of these systems that we're translating, it actually gives us a much broader targetability, and I'll be showing you an example of how that targetability really, plays out. As I mentioned, we can go back into these metagenomic databases and find novel activities such as deaminases. These deaminases, now can be used with the chassis of nucleases to create base editors, and so I'll give, I'll give you some updates on our base editing progress here today. In addition, we can go out and find novel reverse transcriptases.
That allows us to hook that activity onto the chassis to create a what we call an RNA-mediated integration system. It's also for small edits, known as prime editing. Where we're really starting to differentiate, however, is in the ability to do large targeted gene integrations, and I'm gonna show you an example of our lead program and how and the progress that we're making with that. We also have multiple platforms here, and I'll be telling you a little bit about how these are differentiated and what our plans are for continuing to develop those. This is an example of our prime editing of our base editing system. Just to remind everybody, there are two different types of base editing. There's adenine base editors and cytosine base editors.
One thing that you can think about when you look at the fact that these are man-made systems that use a programmable nuclease and then a deaminase activity is where can you target the system? And so if you look at a base editor built on an SpCas9, you can see that you can theoretically only hit 18% of the A's in the genome. Being able to leverage the chassis that Metagenomi has built, we now have essentially a dialable targetability. We can target much more of the human genome of the A's present there, up to 95%. And this is really important, and it demonstrates why the chassis that we've spent so much time and effort characterizing continues to pay off for us as we develop these other technologies.
Obviously, we wanna make sure these systems are also working at a high efficiency as well, and the graph on the right really accentuates this. What you're seeing there is three different locations where we've targeted these base editors to knock a gene out, and we've compared it to just a programmable nuclease knockout. And why this is unique, first of all, you can see that we're able to achieve levels of knockout efficiency that are starting to meet or exceed that of a double-strand break nuclease. But importantly, in an application where you need to make multiple edits within the same cell simultaneously, this is where base editors really shine. You can do multiplex knockouts, and that's what you're seeing here, is that we're able to do three knockouts simultaneously in these cells.
If a double-strand break were to do that same exercise, you would end up with a significant risk for translocation. In addition to building the toolbox, we've also spent quite a bit of effort building a company around translating these technologies towards the clinic. That really can be seen in this slide, where we've broken it down into how we can support our discovery and characterization efforts using automation. So we've invested early and significantly in high-throughput automation, not only to screen for novel enzymes, but also once we move an enzyme forward, we wanna be able to screen a variety of guides in a variety of cell types, and we leverage our platform of automation for that.
In addition, as you translate these systems, you need to demonstrate that they have a high level of editing in a human environment, in a human cell environment. So we've spent quite a bit of money, and time developing our own mRNA and guide RNA optimization capabilities. In addition, whenever you talk about gene editing, the first thing that comes to mind is: how are you gonna deliver these? We developed our toolbox in a way that we could really try to be agnostic to the ways that we're delivering it. We wanna be able to take advantage of any delivery capability that is available to us.
Having an assortment of tools gives us that flexibility, and so we've got quite a bit of internal exploratory efforts in LNP and viral, but we also have licensing to support our lead programs in the LNP space as well. Then finally, a few years ago, we invested in a GMP facility so that we could actually produce these materials at scale and at the quality that we need in order to support our upcoming preclinical and clinical activities. The combination of our toolbox and the company that we've built has allowed us to articulate a pipeline that we think is very compelling in its ability to really demonstrate applications of all of these different tools. You can see that there are a couple different categories that we're focusing on.
One of the core strategies we've had from the beginning is to move technology that's ready now with a delivery capability that's ready now, forward. And so that means, our most significant progress has been made in, liver as the target organ because we, have the LNP, and we have the mRNA capabilities to, create those, modalities. Our lead program is in hemophilia, and I'm gonna be providing you an update there. We also have a partnered, a partnership with Ionis. This partnership, is, a very broad, multi-target, partnership, making fantastic progress here, and, the partnership there continues to accelerate.
As part of that partnership and as part of our interest, we also are interested in getting outside of the liver, and so we do have efforts to use our nucleases in, for example, viral delivery to get into the neuromuscular space. And ultimately, we really would like to get into other organs, for example, the lung or the kidney. So to give you an update on our heme program, this is a disease of blood clotting. We are replacing a Factor VIII gene. The system is schematically diagrammed here. You can see there's a virus particle on the right that delivers the Factor VIII payload. This is a promoter-less Factor VIII, so there's no expression from this DNA upon delivery.
And then we follow that with a lipid nanoparticle that includes the nuclease and the guide. The guide is targeting to a location in the human genome, at the first intron of the albumin gene, seen at the bottom of this diagram. This is a really interesting design because we're inserting into the intron. So if an edit happens, and there's no DNA donor to integrate there, then the intron is just closed up. There's no impact to the albumin gene. And then the donor DNA can integrate either in the forward or the reverse direction. If it integrates in the reverse direction, again, you get no activity.
If it integrates in the forward direction, however, it takes advantage of a splice acceptor site that we have added to the system and a poly-A tail that signals to the termination of that transcript. It also co-ops the exon one of the albumin locus, which is the signaling peptide that causes excretion of the factor VIII molecule. So it's really elegant that way, and it also takes advantage of the constitutively on and highly active albumin promoter. So even with low levels of integration of this donor DNA, you get therapeutic levels of factor VIII being produced. We're in the middle of a durability study in non-human primates.
We've done numerous non-human primate studies with the lead enzyme that we're using for this system, showing repeatedly that we can get indels that are useful for allowing to edit the genome so that we can get the integration levels. And you can see here in the third column, these are three different animals in the table and the corresponding graph. You can see here in the third column that the integration levels are anywhere from a little less than 1% to up to 3%, and that's causing us to achieve clinically relevant levels of factor VIII expression. So this, you know, anything from 10%-150% is the target that we're going for. As I mentioned, this is an ongoing study.
We're very excited about the progress and looking forward to a second half of this year, where we're gonna be able to present more data on this program. Excitingly, though, this proof of concept in the heme program with non-human primates really accentuates the fact that that system in itself can be seen as a platform for large targeted integration. Because you can easily imagine that the hard work of characterizing the nuclease, characterizing where it cuts and the off-target activity there, that is all can remain constant, and we just then replace the payload of the AAV. And so we could go after other secreted diseases in the liver, really setting us up for a series of programs that could move very quickly once we get a little further along with our heme program.
That's an important platform using our nucleases, but we also have our RNA-mediated integration systems, and one area we're really distinct from other companies in the space is we have reverse transcriptases that have a very high fidelity and a very high processivity. What that means is they can go further on a transcript without falling off, and further on that transcript without making an error. For the first time, we've been able to show that these systems are capable of integrating 900 base pair piece of DNA from a cDNA template. We're excited about the potential applications here, again, because this can be packaged completely as mRNA inside of an LNP, and we're thinking about our first applications here in the liver, taking advantage of that.
Finally, a year ago, we were excited to present for the first time that we were able to get a Cas system active in human cells. Cas are CRISPR-associated transposases. They're a two-component system. The CRISPR gives you targeting to a specific location in the genome, but then the transposase is responsible for the integration, not the nuclease. So a year ago, we were able to demonstrate for the first time activity of our system in a human cell, and I wanna dive in a little bit more about how our system is unique from some of the other systems that you may have heard about. Our system is depicted here in the upper left, and it's in comparison to the lower left, which is the Cascade system.
This is a very common system used in academic fields. One thing that you'll see is the mgCas system is much simpler. It has less number of proteins, and the proteins themselves are smaller than the Cascade system. We purposefully chose this system just due to those characteristics and have made all of our progress with that system. The progress includes very excitingly demonstration that we can integrate a therapeutically relevant cargo gene, integrated at a safe harbor site. We've also been able to do this in cell lines, as well as now liver-derived cell lines, and it really sets us up to achieve our goals of getting animal proof of concept in these systems later in the year.
Finally, my last slide, almost last slide, is about something that is a very exciting topic. It's been in the news quite a bit lately, and it's AI. AI is clearly going to make a significant contribution to biotechnology. I think we're just at the beginning of getting a glimpse of what that could look like. I just would like to tell you about how we're using AI at Metagenomi to really drive our discovery and translation process forward. We had a publication in 2022 on an enzyme, a programmable enzyme system called SMART. This is a really unique system because it's half the size of SpCas9, so it's only about 750 amino acids, and it's incredibly rare.
The first panel really is, it demonstrates a phylogenetic tree, but the point is, there's very few representations of this system in nature. We were able to find one and identify it and predict that it would be a programmable CRISPR enzyme, and then we showed that we could achieve some levels of activity with it, but it was very low. And so we had to spend quite a bit of time trying to understand the structure of this system in hopes to learn something about it. We had a collaboration with the Taylor Lab at UT Austin. We now are actually seeing the first crystal structures of this enzyme, which is very exciting.
But I wanna tell you that the problem that we hit here was really around the activity levels being very low, and given the fact that it's such a rare enzyme, we couldn't go back into the evolutionary history of this enzyme to find relatives that would help us to inform how we would do the protein engineering to boost that activity. So here, what we did is we employed two different AI methods. One of them is called ASRs, which stands for Ancestral Sequence Reconstruction. So this is, it's the blue of the graph that you can see in the third panel, this sort of small blue in the middle. These are sequences that don't exist in the database.
They're unknown, but yet we were able to impute what these sequences are and help us to bridge this gap of this phylogenetic tree, and that was then useful in helping us to think about ways that we could modify the 3-D structure of the enzyme to improve activity. But in addition, we've also been able to take advantage of the metagenomic database that we've built to leverage. We have over 7 billion proteins in that database, and these are of novel life forms. So you can imagine, in a generative AI setting, this is a very valuable dataset because these generative AI systems, they need to feed off of that dataset in order to come to solutions on their own. And so we've leveraged this generative AI to generate the orange sequences in that third graph.
These are de novo created from the generative AI algorithm, and the important thing is, where did this get us? It allowed us to take an enzyme that was essentially less than 10% active and now increase the activity to almost a 100% editing at the location that we're guiding it to. So again, we've been using this for a few years now, and it's a really amazing example of how we're gonna continue to leverage AI to accelerate our discovery process at the company. Then finally, I'll just end with this has been an exciting year for us. We have a very strong cash position. We're excited about our second half milestones around our lead program in HemA, including we're planning our DC nomination.
We have an important 12-month durability readout coming in the second half of the year, and we also have some planned presentations on some of our technology coming up at ESGCT that we're excited to talk about. So, happy to answer questions from Maury here between myself and Sarah Noonberg, my Chief Operating Officer, and appreciate the time.
Great. Thank you very much, Brian. A great overview of the company and what you, and what you guys are doing. Maybe just starting off with some of the things you mentioned at the end there with the smart enzyme that you discovered. Where's that coming from? Is that from bacteria? I guess, it-
It's actually, it is a bacterial enzyme, yes, so.
Got it. Okay. And you also mentioned the 7 billion proteins that you have in the database. Is that something that you continue to add to?
It is. So we actively run field trips where we're going out to novel environments, and we are taking samples, and I think this, again, is paying off in its value because we're seeing that we're able to come to solutions that the public databases just don't have present.
Got it. And just a question about strategy. When thinking about the diverse platform and pipeline potential outside of your selected lead program for hemophilia A, what are gating factors for how you prioritize investment in the rest of your R&D capabilities? And is there a certain editing technology that you're bullish on relative to the other technologies? So I guess, what's your favorite technology?
Yeah. That's a really good question, and, you know, I mentioned it a little bit in my presentation. One of our early strategies was to really move forward with technology that's ready now. But what we've been hesitant to do is sort of look at those future technologies or technologies that maybe aren't as developed and sort of say, "Oh, let's put all of our bets in that one particular technology." And so as a result, as I mentioned, we built an organization that can really move a variety of technologies forward to generate the important data that they're all gonna need. So we haven't seen really the need to limit our technology capabilities. And I think very importantly, the core strategy that we've been pursuing is bringing the right tool for the job.
And so being able to look at a particular genetic disease, look at the right mechanism for fixing that, and then combining that with the state of the art in terms of delivery, and then being able to really match the right tool and the delivery to go after the right indication. I don't know, Sarah, if you want to add anything.
Sure. Maybe I'll address the first part of your question around gating factors that we look at when we think about pipeline prioritization. So, you know, we are really interested, first and foremost, where gene editing can make an important advance for the standard of care for patients. So there has to be a clinical unmet need, there have to be value-supporting endpoints that are gonna be attractive to payers, clinicians, patients, but also where we understand the biology really well, where we have a readily available biomarker for rapid proof of concept. So those are some of the factors that we're always looking for, in addition to an addressable market that we can then develop.
So those are sort of some of the checkboxes, and, and when you say, "Well, what are we looking for in the future?" I, I think some of the key themes that you'll see are, are really this large gene integration platform. Our hemA program is moving along extremely well. That gives us a readily made platform for addressing deficiencies of other secreted proteins, because all we need to do is just drop in another donor DNA template, and we can address diseases that aren't available right now through siRNA or antisense technology, where, where knockdown does have another clinical solution. So I think large gene integration is gonna be a theme, both from the platform we've established, but also this all RNA-mediated system, our Big RIGS technology, that's an area that we're really enthusiastic to, to push forward.
Got it. Yeah, I think that's a great segue to Hemophilia A, which is sort of the tip of the spear for you guys, where this is gonna establish a lot of technology and proof of concept with what you're doing. And you're gonna have data second half of this year from the non-human primate studies that you're doing. How can you set expectations for what we can see in this update, and how that would compare to some of the competitors out there, including BioMarin and then Pfizer and Sangamo?
Sure. I think, you know, what we've learned with BioMarin, and I think you're seeing the same trends with Pfizer's Sangamo, is that durability with gene therapy is limited, and therefore, the initial goal of achieving a curative approach, at least for adults with hemophilia A, is not going to be met. So for us, with a permanently gene-edited solution driven off a native promoter as opposed to an exogenous promoter, we believe that durability and no loss of, of factor VIII expression over a 1-year period is, is critical. If we see loss of durability, then that's something that we would not see ourselves as, as really being able to differentiate.
And I do think it's really limiting the uptake of ROCTAVIAN in the commercial setting, as you can have a transient reprieve from giving yourself Factor VIII, but you're essentially taking yourself out of the potential for future curative approaches. And what we've learned from our outreach to international KOLs, patient advocacy groups, is they are primed, they are waiting for a cure. They thought they had that at hand with the development of ROCTAVIAN, and they're not letting up on that goal and that push for a curative approach.
Got it. Makes sense. And, anything more that you're saying about the data for second half of this year? How much follow-up you could have?
We will have at least 12 months of data.
Got it. Okay, and you've shown the kinetics of Factor VIII expression across the three non-human primates, and definitely seeing early durability with that, in that data set. But there's some variability across the three non-human primates with the 15, 30%, and then 75% of normal. How can you explain that, some of the—what are some of the reasons for that?
I would just say, important to keep in mind, this is our very first study, too, and so this is not a dose-finding study, and so I think Sarah will get into the specifics here.
Sure. You know, I think that there are some levers that we can pull to bring in some of that variability, but what's more important is hemophilia A or B, but hemophilia A is a disease that can really accommodate wide range of expression of Factor VIII without really any clinical impact. So if you're above 10%, if you're below the upper limit of normal, 150%, the clinical difference between these patients is really minimal. And so even if we see that degree of variability, we don't believe that that's gonna have a large clinical impact. Where we don't wanna go is having some patients with 2% and others with 250%. That's a degree of variability that would be problematic.
The degree of variability that we've seen in our current non-human primate studies, and as Brian said, we've yet to even kind of done all the things that can bring down, that variability and improve precision, would still be very acceptable within the clinical norm, and, and certainly exceed what, what has been seen in the gene therapy experience.
Got it. Can you talk about when you anticipate completing the IND-enabling studies for Hemophilia A and bookend timing for CTA or IND filings? And then you're not the first company, first gene-editing company that's getting an IND or CTA, and so you can learn from some of the other companies out there, too. What kind of learnings do you have and feedback from FDA based on this?
Sure. We've had early interactions with the FDA that were extremely informative. They've given us a clear blueprint of what they're looking for, and I think our expectations are very much aligned, so that was a big de-risking first regulatory interaction for us. We are targeting an IND in 2026. There's always the potential to accelerate, but at this point in time, I think that's our best estimate of a first IND CTA.
Got it. Okay, thanks so much for joining us today.
Thank you, Mark.
Thank you.
Appreciate it.