Metagenomi Therapeutics, Inc. (MGX) Earnings Call Transcript & Summary

Good afternoon. Thanks for coming to another session at the 43rd JPMorgan Healthcare Conference. I'm Brian Cheng, one of the senior biotech analysts here at the firm. Next up, we have Metagenomi who will be presenting their slide deck. And then after their presentation, we'll move on to Q&A session. Brian, the floor is yours.

Thank you. I'll make sure everybody can hear me. I appreciate it. I'm very pleased to be here at the 43rd annual JPM. I'm excited to tell you about some updates from Metagenomi. Please note that this presentation will contain some forward-looking statements. So at Metagenomi, we are using microbial genomes found in the natural environment around us across the planet to really develop a comprehensive set of next generation gene editing capabilities. We believe that these gene editing technologies will be the cornerstone of curative genetic treatments. The success of gene editing systems in curing disease is going to be transformational. This will be a paradigm shift in how we think about healthcare. As the name of our company suggests, the science that underpins what we do at Metagenomi is called metagenomics. Metagenomics is an amazing science that I've spent many decades working in, but it allows us to tap into natural environments around the world and then to study genomes from organisms that are not cultivatable and have never been looked at before. We have built a proprietary database of the genomes of these unknown organisms, and this is what we use to mine for novel gene editing systems. Many of these systems are extremely rare, and that complicates our ability to use conventional bioengineering and other methods to develop these for human application. The way that we get around this is to leverage artificial intelligence and high throughput robotics to allow us to very rapidly screen through that database and find enzymes with novel characteristics that give us a couple of key points. They are systems that are highly efficient, they're systems that are highly specific and they're targetable where we want them to go in the genome. This becomes very important when you start to think about the landscape of genetic disease and how you apply these systems towards correcting genetic disease. The result of our efforts to discover and develop these next generation gene editing systems has resulted in a powerful toolbox of gene editing systems, and you can see those on this slide here. So the core capability is at the left here, and that's our nuclease platform. There are thousands of nucleases that we have uncovered from the natural environment that all have characteristics that make them unique and uniquely suited for the application in a human therapeutic development. These are programmable CRISPR enzymes, so we can use them to target throughout the human genome. They share characteristics that really allow us to go quickly towards translating these into therapeutic applications. And again, those include the ability to target anywhere in the genome at very high efficiencies and showing little, if any, off target activity. So this really becomes a screening exercise for us, where we have been able to leverage 4 billion years of microbial evolution to help us identify which of these enzymes is best suited for a particular application. We've shown time and again that we can use this process in order to rapidly go from a discovered novel nuclease in an environmental sample to having something showing in vivo proof of concept for human therapeutic application. The other thing to keep in mind about our nuclease platform is it's very extensive, but you can also now start to think about it as a chassis, and in that -- on that chassis, you can then start to add other functionality. This, again, though, is where we can leverage our metagenomic database because we can go into that database and find other enzymatic activities such as deaminases. We can tether that to the chassis, and that allows us to form a base editor as an example. We can go into that metagenomic database and we can find novel reverse transcriptases, and that allows us to tether that to a chassis. And now you have an RNA based integration system also known as prime editing. A perfect example of that is our base editor developments, where we leveraged a chassis in our genome that's ultra small. We have a variety of these, everywhere from 30% to 50% of the size of a typical Cas9 type enzyme, and that allows us to build highly targetable and efficient base editors, but also with the smallest known size. So this is important, because when you're talking about gene editing, one of the key things that you focus on is how are you going to deliver that. Having a very small base editor using this small chassis allows us to now think about other means for delivery such as AAV. Similarly, we've been able to find novel reverse transcriptases that allow us to do RNA mediated integration, and we can do this in small and large scale. And then finally, on the right, you see our CAST system. This is also a naturally occurring system. It's very promising in that it is a combination of a programmable nuclease, so you can target it to go where you want it to go in the genome. But it's naturally evolved to be defective so it doesn't cut the genome when it binds. Instead, it relies on its counterparts of the transposase components to do an active integration. So the reason that we developed the toolbox to have this variety of capabilities is really so that we can accomplish what you see at the bottom here, which is, any genomic correction that you need to make in order to solve disease, we have it in the toolbox. We have the ability to do knockouts and knockdowns. We also have the ability to do single nucleotide changes as well as small corrections, deletions and insertions. And with Cas, really the ultimate goal of being able to integrate very large pieces of DNA. So now I'd like to tell you what we're doing with this toolbox and how it's driving our pipeline and our translationary efforts. So we've built a pipeline of both wholly owned as well as partnered programs, and the strategy that we used in order to develop that pipeline is really around a core topic that I've mentioned, which is delivery. One area -- one organ that is readily available for delivery is the liver. We know how to deliver there and that technology is proven. That's driving our lead programs in hemophilia A, and as well, we'll talk today about our expanded use of that platform to go after other diseases of secreted disorders in the liver. So those are -- that's our knock-in wholly owned program. We also have a partnership where we're focusing on cardiometabolic disease, and that is driving our efforts in the knockdown space. You can see here, we've moved several of our targets in that partnership forward into lead optimization. And then as you go further down the table, you see that the gene editing becomes more and more complex. And we're initially targeting the liver, but at all times, really looking for opportunities where delivery supports our ability to go into other organs such as the brain, the kidney and the lung. Before I dive a little deeper into the progress that we've made on some of our lead programs, I'd like to tell you about what 2024 resulted in for some of the milestones and key developments that we've had. Our wholly owned Lead Program in hem A, for this program, we declared a DC, MGX-001. This was the result of several years of experimental data that we collected and it piggybacked on the success of our long term durability study in non-human primates. That was another key development that was published at -- presented at ASH here just a few weeks ago. This study was a very important one, and we're going to talk more extensively about what that result looks like, and it's really supported our push with hem A towards the clinic. In addition, we had a successful regulatory engagement with the FDA on this project. And in addition, in support of this approach to our IND, we built a GxP facility many years ago, and that has been ramping up in order to support the efforts around our Hem A Lead Program. And I should also mention here we had another successful milestone by filing our first drug master file with the FDA this last year. As I mentioned, the strength of the progress we've made on our Lead Program in hem A has allowed us to realize we have an opportunity to use that as a platform, and we're going after several secreted protein deficiency disorders there and we were able to achieve proof of concept -- in vivo proof of concept with several of the initial targets there. In our partnership programs with Ionis, these are all, as I mentioned, in cardiometabolic programs. All four of them made it into lead optimization, and we also have attained in vivo rodent proof of concept for all of these targets. And then finally, for other therapeutic programs and technology development, we had several opportunities this last year to demonstrate how we're using artificial intelligence to really support our development of our core gene editing technology. We had a presentation at Cold Spring Harbor Labs, where we were able to show that using AI and ancestral state reconstruction, that we were able to take an enzyme that was effectively inactive but had characteristics we wanted to pursue and then alter that using this information to bring that system to be a robust gene editing enzyme. Similarly, with our base editor, we were able to demonstrate that one of the key activities in cell therapy applications is doing multiple knockouts at once, so multiplex editing. For this, we were able to show that our base editor is able to knockout simultaneously 3 different targets in the same cell completely, and this was presented at the European Society for Gene and Cell Therapy. And then finally, we had a publication in Nature Communications on our ultra small nucleases, where we use structural biology and, again, ancestral state reconstruction to really understand the mechanism of this novel gene editing system. Okay. Now I would like to turn to our Lead Program in hem A. So hemophilia A is a life altering bleeding disorder. It's caused by a variety of mutations in the Factor VIII gene, which lead to a loss of functional protein. This disease leads to joint damage, which is a major cause of morbidity, and importantly, it can also lead to intracranial bleeding, which is a major cause of morbidity and mortality. Diagnosis of hemophilia A usually occurs, at least in the Western world, early in life. And then determination of the disease severity is seen in the graph in the upper left. So anything -- any Factor VIII levels that are below 1% is a severe condition, between 1% and 5% is considered moderate disease and then between 6% and 40% is considered a mild disease. Our target is really to get as high as possible, but above 10% effectively gives us -- takes the patient from a serious condition to one of just mild disease. There are about 26,000 patients in the U.S. and more than 500,000 worldwide. So the mechanism that we've chosen to pursue with hemophilia A is seen here. It's a 2 component system. It consists of an AAV that carries a donor Factor VIII gene, and that Factor VIII gene, importantly, does not have a promoter. So as the AAV infects and goes to the target cell, there's no expression that occurs. It just exists extra chromosomally without any expression that occurs. This is followed then at a point later in time, usually 2 to 4 weeks, by a dose of lipid nanoparticle that carries our enzyme as well as the guide RNA. So this is where we are targeting that nuclease to cut in the first intron of the albumin locus. You can see that on the bottom of the schematic on the right. This is an important intron because the exon immediately upstream of it, exon 1, is the secretion signal. So as the nuclease comes in and gets turned into an active gene editing system, it cuts within the intron. And many possible outcomes can occur. It can cut and then the host cell can repair that damage, in which case nothing happens. You just get potentially a deletion or insertion at that location. But it's in an intron, so it doesn't cause any effect on the albumin gene. You can get insertion of the target in the reverse orientation. And in that case, again, you've got nonsense DNA inserted into an intron and it doesn't impact. When the gene inserts in the correct orientation, you then get a splicing that occurs, that it puts that secretion signal at the beginning of our Factor VIII gene, and that's how the Factor VIII is able to exit the cell and make it into the bloodstream. So this system allows us to leverage both of these components in order to get the Factor VIII donor gene integrated. And this is a permanent integration. And so one of the concerns with Factor VIII therapies is durability. As we've seen with gene therapies, durability is a key issue. And especially when using an AAV where you have a limited number of times that you can use that AAV on a patient, you really only have a single shot. And so for us, durability was key, and that's why we initiated one of our first studies about a little over 16 months now ago to really assess this durability and ensure that we were getting the permanent edit that we thought. What you can see here on the left is the current readout from our durability study in non-human primates. This is now at 16.5 months and we're about to conclude this study and we'll be concluding that this year. What you can see is we've dosed 3 animals. And each of the different lines there, the blue, the purple and the green, represented a different non-human primate. You can see that we're able to achieve rates of Factor VIII activity at the lowest levels, approximately 8%. So again, going back to the degrees of severity of the disease, we're already getting to a stage where you get an alleviation of the symptoms. The middle animal is at 30%, and then the upper animal is at 75%. The variability here is key to understand. It's directly related to the degree of integration. The lower graph line has a degree of integration of about 0.7%, whereas the middle one is 1.3% and the upper one is 3%. So again, all those different variations on how you can integrate into the locus of albumin, not all of those are successful. And that actually works in our favor here. We have very low levels of integration, and that directly ties to the level of expression we see. We've also made the important assessment that optimizing this system is critical, and we have -- in the meantime, while the durability study was running, we've also been optimizing the components of the system in order to really increase that Factor VIII activity. What you can see on the right is a dose dependent study of Factor VIII activity with AAV. Our goal is really to reduce the level of AAV that we need to deliver for a variety of reasons, but we want that dose to be as small as possible but also to be able to deliver us high enough activity to cause a alleviation of the disease. And so what you can see on the right is that we're able -- with an engineered variant of this Factor VIII system, we're able to achieve very high levels with a much smaller dose of AAV. For these types of treatments where AAV is used, we have to be very sensitive to the toxicity associated with AAV administration, and that's why we've really sought to have the highest degree of activity of our engineered variant. Another aspect of our hemophilia A program and really any gene editing program is around specificity. As you start to modify the human genome in a permanent way, what you quickly realize is that any off target activity becomes a significant concern. We use an industry standard method for assessing off target, which includes multiple methods for predicting where those off targets can occur. In the case for the hem A albumin intron 1 platform that we have, we identified 481 potential sites using a combination of in silico biochemical and in cell assays. We then interrogated all 481 of these sites extensively with amplicon sequencing in primary human hepatocytes at both saturating editing as well as twice saturating editing levels, and were able to see only on target editing and no off target editing. So this becomes very important for us and it also is a sign that the type V enzyme that we're using for this system has a very high degree of specificity. Okay. So as shown on our pipeline slide, and as I've mentioned already, MGX-001 gene editing as a platform is also very appealing to us. We've targeted secreted protein deficiency disorders as our next place to focus that platform. And you can see the schematic on the right is very similar to the hem A diagram. The only difference is that we've swapped out the cargo from hem A to be one of these other secreted protein deficiency disorder targets. Everything else about the system is identical. The guide is the same, the LNP formulation is the same, the nuclease is the same, the AAV is the same. The only thing that's different is that target. We really think that we can benefit from this because the hard work that we put into hem A and including everything down through our ability to manufacture this system can now pay off with multiple other targets that we can go after. The graph on the left here shows you that we've been able -- in 3 different targets in this secreted protein space -- secreted protein disorder space, we've been able to achieve greater than 200% of normal activity using this system. So we plan to continue the process of moving this MGX-001 gene editing platform forward. Partnering is a very important aspect to Metagenomi. It's been part of our strategy from the beginning. It's been something that I really feel strongly is important. Whenever you have novel technologies such as gene editing, having strong partners helping you move it forward is really critical. Our partnership with Ionis has been going on for a few years now and we've made really tremendous progress in Wave 1 of that project. One of the diseases that I want to show you as an example today really exemplifies what Metagenomi brings to a partnership in the gene editing space. This is around a target called refractory hypertension. This is the angiotensinogen gene. This is characterized by uncontrolled hypertension despite multiple drugs to control that. It's a significant risk for cardiovascular events, and it's got a large patient population. We begin first with showing that knocking down a particular target has an impact on the -- both the mRNA and the protein expression of that target. And you can see that in the first graph on the left. The blue line indicates the degree of editing as you increase the dose of the editing system across the graph. As you can see with the green line and the gray bars, you can see that we get a corresponding reduction in both the mRNA and the protein for this target. At the middle column, we then move on to our optimization, where we're optimizing the mRNA, the guide RNA chemistry as well as the lipid nanoparticle formulation. And we move that into an AGT transgenic mice. And you can see that in -- with the various constructs that we have, we've been able to even further increase the effectiveness and the potency of this system. And then finally, we can move that into a spontaneously hypertensive rat. And that model really helps us to say, "Is this going to have an impact on blood pressure reduction?" And that's planned for a future study for us. But again, you can see that we're able to achieve 95% protein knockdown in these rats. So looking ahead, I want to go over some key milestones across the company that we're focusing on. First, in hemophilia A, we want to complete the ongoing NHP durability study that I mentioned, that you've seen some data for. We also want to have our pre-IND meeting as well as ex-U.S. regulatory meetings. That will happen this year, and that will be in support of our IND, CTA filings the following year. In the secreted protein disorder space, we will disclose our lead indications for this platform as well as we plan to achieve NHP proof of concept for this platform, and that will lead to a DC nomination in the following year. With our partnership with Ionis in the cardiometabolic space, we will nominate 1 to 2 DCs this year and as well we will also disclose the remaining indications of the Wave 1 targets. There are 2 waves in our partnership with Ionis, and the second wave will be a future endeavor. In 2026, we hope to initiate IND-enabling activities for these programs as well as nominate DCs for the additional Wave 1 targets. And then finally, in other activities that the company is involved in, we want to continue to lead the innovation space in the gene editing technology development and really do everything we can to support future INDs using a novel set of gene editing tools. In summary, 2024, I think, has been a very significant year for Metagenomi. 2025 and 2026 promises more to come. Across our hem A secreted protein deficiencies, Ionis collaboration and our metagenomics platform, we continue to execute -- we continued to execute in 2024 with targeted plans to drive towards the clinic in '25. And then finally, today, we also had the pleasure of announcing the addition of a new Board member to the company. We're pleased to welcome Eric Bjerkholt. He's currently the CFO at Mirum Pharmaceuticals, was previously at Chinook. Eric is a veteran in biotech and brings considerable experience in business strategy and capital formation. So thank you for your attention.

Thanks, Brian.

Yes, let's begin the Q&A session. Yes. So I guess just more of a broad sense on your portfolio approach. Can you talk about just how you're thinking about prioritizing the different tools in your toolbox? It's a pretty wide range of gene editing approaches. So how do you think about prioritizing and which one to allocate resources? Maybe we'll start there.

Sure. And I'd like to introduce my Chief Medical Officer, Sarah Noonberg. And I'm going to let her answer that question.

Okay. As we think about our toolbox, which is really advancing toward our goal of addressing any genetic mutation anywhere in the human genome, we're always striving to match the optimal tool to what an indication may need. So for instance, alpha-1 antitrypsin disease, there are currently therapies in development that are looking to insert -- just a copy of A1AT for expression would address pulmonary, but not liver. There are currently base editors that could form potential bystander effects. We have looked across our pipeline and said we're going to use the optimal tool with small rigs or prime editing approach to do so. So in each case, we're going to say, "What is the best approach for addressing a given disease?" So for hemophilia A, it's replacing a fully functional copy of hemophilia Factor VIII gene. For other disease areas, we're going to do what's best for the patient population. So it gives us enormous amount of flexibility.

And I'd just add to that, Brian, that a lot of the efforts that you put into -- even though it's a variety of different gene editing modalities, you still have to do off-target analysis, you still have to kind of do the same core things. And where Metagenomi, I think, really excels is we've built a company that can really take advantage of the work that we put into, as an example, our Lead Program with hem A and then leverage that in other areas, both with our knockdown programs with Ionis as well as with our expanding to the MGX-001 gene editing platform.

So as you think about hem A as sort of the stepping stone to do more things and you also have the partnership in the background, too, right? So how do you think about just balancing the risk of development? I mean, if we -- let's look forward into the late 2020s, and how should we kind of think about what the -- how the portfolio is going to look like? And as you kind of build out all these different pieces, is there going to be one that you think is definitely going to work and then there's one that is less going to -- less likely going to work but has much more unmet needs. Like just how do you think about balancing both risk and also reward?

Sarah?

Sure. Our portfolio is based on advancing the most advanced technologies that we have, with delivery systems that are readily available now, the biomarkers that are readily available now and a regulatory path that's becoming increasingly clear. So our first wave of programs, we feel very confident. We are looking at disease areas where there's clinical validation already achieved. So we've derisked some of that biology. And we have increasing confidence that we understand what regulators are looking for and can get into the clinic. So I would say that first wave is the strongest level of data available that we have now to put us in a position with success. As we start to get into more novel technologies where the regulatory path is still being developed, where there are disease areas where there's less clinical validation, of course, we're going to increase that risk. And I'd say that our overall approach is to start where we have clinical validation, we have a readily available biomarker that will rapidly tell us whether we have proof of concept without having to dose many patients or wait very long and where we know that gene editing specifically can make an important advance in the standard of care and support payer and reimbursement discussions. That's obviously going to lay the framework. And in a number of our indications, particularly with our partnered program with Ionis, we can start in smaller indications, more rare niche indications, but also they have larger cardiometabolic areas that they can then grow into. And I think that's another strategy to start in well-defined smaller indications that have the opportunity to grow into much larger cardiometabolic indications.

And I would just really add -- I think Sarah said that perfectly, but I would just add one thing, that there's also the delivery component. Some of these other technologies really are waiting on perfect delivery because the target that we would want them to go after is a more complex gene edit that's not available with a base editor or not available, for example, with a nuclease.