Good morning, everyone. My name is Chris Yu. I am on Terence Flynn's team covering biopharma and biotech companies. I am delighted to host Allan Reine, CEO of Prime Medicine. Before we start, for important disclosures, please see the Morgan Stanley Research Disclosures website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative. Allan, thank you for joining us. Thank you for having me here. Maybe we can start with some high-level questions about a company. In the past two decades, we have all witnessed the amazing evolution of gene editing technologies, and you guys are at the forefront of that. Maybe talk about how Prime Editing is different versus the earlier generations of gene editing technologies. Yeah, and I think you said it right. It really is a revolution. I think we are all very lucky to kind of be alive to see these technologies and others as they advance. I know they are just incredible as you think of where we were 20 years ago and where we are today. I always think about the evolution of CRISPR kind of nuclease-based editing kind of to Prime Editing. In nuclease-based editing, we are looking at Nobel-winning technology, really incredible. For the first time, you could, with very high efficiency, go to a very specific point in the DNA, and with a double-stranded break, really make that cut, and the result being very effective at knocking out targets, right? Very effective at knocking out proteins or changing regulatory elements like they do in sickle cell, and an incredible technology. We have an approved drug now. We have others that are likely looking like they are close to approval. This is becoming sort of that dream of 10+ years ago is now becoming a reality for patients today, which is obviously what this is all about. That nuclease-based editing with double-strand breaks, there are some liabilities there. Obviously, some of the both on-target but also off-target consequences that can happen, chromosomal rearrangements, translocations, some off-target editing that can occur. But in general, really, really promising technology. Then out of David Liu's lab came another type of gene editing called base editing, and now, for the first time, you can not just knock something out, but with high efficiency, actually correct something, right? Make a change to the genome, a very specific change. With base editing, they use what is called a deaminase enzyme, and you can change kind of one base pair to another. You can do an A to a G to an A, C to a T to a C. Those are the four base changes that you can make with base editing, and very effective for certain diseases where you have that very specific type of mutation. But limited to really only correcting those types of mutations, or like CRISPR, you can use that to knock things out as well. It also uses, instead of a double-stranded break, you are using what we call kind of a Cas9 nickase, which is creating a single-stranded break which is a lot gentler on these cells and ultimately does not lead to the same rate of chromosomal rearrangements, translocations, and other off-target issues. A nice step forward in doing things that nuclease-based editing could not do. With Prime Editing, we can do everything the kind of first-gen genetic technologies can do. We can do everything that base editing can do, but not just those four base changes, we can do all 12 changes, but we can also do so much more. We have the ability to, instead of using a deaminase, also out of David Liu's lab, he came up with the idea of, well, let us use a h im and Andrew Anzalone, who are our scientific co-founders, came up with, we can now use a reverse transcriptase and actually create a template that goes right into the guide, that gets transcribed right into your DNA. Instead of changing one base pair, we think about this as a word processor. Instead of it going after one letter, we can put in whole new sentences. For the first time, we can make very much larger kind of full insertions of anything from 10- 15 base pairs, but with our PASSIGE technology, we can actually do almost full gene insertion. It is a really powerful technology. I like to say it is the most versatile gene editing technology that exists, given the number of different changes we can make. But it is also the safest as you think about the limited kind of off-target, on-target, and other consequences that you see with some of these other technologies. That's a great way to start this conversation. Certainly, there are a lot of potentials for Prime Editing based on what you just described. The way I see it, like with gene editing, a lot of efforts have been focused on delivering to the liver. Can you kind of talk about what are the remaining hurdles to kind of expand the technology kind of outside the liver to other organs and other diseases? Yeah. So you're asking the billion-dollar or hundred- billion- dollar question. I think the initial promise as we looked at gene editing was, there were just so many different diseases, both genetic and non-genetic, that would benefit from this type of technology. It was just a matter of, okay, let's get the delivery technology towards all of these different tissue types. Well, that's proved, I think, a lot more challenging than probably initially thought, and I think today we've shown that these. It's not just gene editing, it's other technologies too that are RNA-based that require delivery. But I think we can very effectively today get delivery into the liver for gene editing. We've seen that proof of concept for both nuclease-based editing, base editing, and soon-to-be Prime Editing. We've obviously seen this done with these technologies, both prime, base and nuclease-based editing, ex vivo, so it can obviously be done in cells. We've shown that in our chronic granulomatous disease program. Obviously, it's been shown in sickle cell disease with those two other technologies. But as you go beyond that, it gets more challenging. I think we've seen good delivery to the eye. I know Editas had an earlier program where they could at least effectively deliver, and we've got some early preclinical data showing that we got really good eye delivery as well. We think we can get delivery to the ear, so it becomes like are there real opportunities there to go after from a commercial standpoint. But where I see the biggest opportunity for Prime Editing would be if we can get successful delivery to the brain. I think there are a couple of maybe AAV-based approaches today that may work. But I don't think there's anything today that we've seen to say we think this is a solved issue. I would say we continue to scour the world for advancements in these types of delivery because I do strongly believe this is a case of when, not if. But I don't know if that when is in one to two years or in 5-10 years. But as soon as that gets solved, we're going to be there to explore those areas because, again, there's a tremendous amount of unmet need in the brain where Prime Editing can make significant impacts in the lives of patients. If in 5-10 years, I hope to be treating many of those diseases with Prime Editing. Beyond that, obviously, there's heart, there's muscle, there's kidney, there's other organs that we'd like to get to as well, and there's varying levels of advancement. Again, we're staying very close attention to those so that when we do see something that looks promising, we can jump right on it. I think the lung would be a good example, where in collaboration, obviously, with the Cystic Fibrosis Foundation here, we've got programs in cystic fibrosis that we're developing where we think lung delivery looks very promising. We've seen AAV, we've seen HSV with other companies, and we've seen LNP be successful, at least in delivering that cargo. We think Prime Editing has the potential to be a best-in-class therapy as we think about cystic fibrosis. So that's an area we think is in the process of being solved, where others we think are probably more in the future. Again, tons of opportunities that delivery question gets solved. It's a great question, and one we spend a lot of time talking about and thinking about. Yes, that sounds good. Recognizing that in the past year, there has been a lot of changes at the FDA. Now I think we have seen some stability there. Just how supportive in general is the FDA on gene editing? That's part one. Then second is how has your interactions been with them recently? Yeah, I think in general, the FDA, and I know that we've all been through a lot over the last year in terms of what's gone on there. I think even over that entire period of time, even with all the changes, it's been in the right direction, I think when it's at least come to cell and gene therapy for the most part, especially when it's come to, I think, gene editing. I think they seem very supportive. There's a lot of different measures and endeavors going on at the FDA, where they're trying to help these companies create more products for patients in need, especially when you go down to the rare and ultra-rare indications. I think that's going to continue. These efforts are both on the CMC side, the ease of getting to the clinic. For us, we went and the FDA essentially gave us alignment that we can file for approval on two patients worth of data. That's incredible flexibility. Obviously, it's an unmet need. We showed a really strong data set in curing two patients. That's showing a lot of flexibility and that real desire to get these drugs to patients in need. In terms of recent conversations, we don't comment on specifics on our dialogue with the FDA. What I would say is we continue to have constructive conversations across the board with the FDA. For the most part, it seems like business as usual. I think it's nice that we have leaders that are not interim, that are now in place at CBER. Well, for us, CBER, but also CDER. I think Karim's going to do a fantastic job there. We're sort of excited that whole agency can now go in the right direction going forward. That's great. Let's kind of talk about the specific programs that you are running right now. Wilson's disease is your most advanced in vivo program, and the IND recently got cleared. Can you kind of walk us through the phase I/II design specifically, and then what are the key biomarkers that you would be measuring? Yeah. The IND was cleared back in July. We also had a CTA cleared back in New Zealand in June. This is going to be a global study, so we'll be enrolling in other geographies ultimately as well. The design of the study is going to be a dose escalation. We'll be looking at a few, call it initially planning for three cohorts. You can always add cohorts if needed. We'll be looking at three different dose levels. The lowest dose is potentially a biologically active dose as we start to get into the second and third dose levels. There could be a lot of biologic activity if the preclinical data translates well to the clinical data, and I think from everything that we've seen from other programs across different gene editing approaches, we've seen fairly good translation as you look even just mouse to human. Within those studies, we will obviously be assessing safety. I think safety is paramount. With our LNP, this will be the first time this LNP has gone into the clinic, we appear to have a very wide therapeutic index, much wider than some of the LNPs that have gone into the clinic that we have actually benchmarked doing experiments internally. So excited about the prospect there. Then obviously looking at efficacy. I think the good thing with gene editing, if you are in the right indication and if you have a good biomarker, you can get a really early read. If you are making that precise correction and you are doing it with high enough efficiency, and you have a good biomarker, you know pretty early if your drug is working. For Alpha-1, which I am sure we will talk about, which investors are very familiar with because there is a number of programs out there, obviously you can very simply look at blood levels of AAT and look at the differential between M-AAT and Z in the systemic circulation. Wilson's disease does not have a biomarker that is that easy because most of these patients are on standard of care, so we cannot just look at copper. What we can do, another way to do that is we are going to be doing radiolabeled copper PET scans. We will be doing those scans both at baseline and then about six to eight weeks after treatment. We have really nice images that are in our deck that I encourage people to go look at, which are these mouse models showing what the copper PET looks like. In a diseased mouse, which you would expect to see similar in a diseased human, and there is some publications that actually show what this looks like in a human as well, you could see these livers completely lit up with copper because they lack the enzyme and the ability to effectively shuttle it into the bile for normal fecal excretion. As you look at what a wild-type mouse or human would look like, you have very normal copper metabolism. You do not see that copper lighting up in the liver because it has been normally metabolized. If you look at even just a lower dose of 0.4 mg/ kg and a somewhat higher dose of 0.8 mg/ kg, if you look at the images that we have, we look very much like that wild-type mouse, like we are almost completely normalizing copper metabolism. You really only need to get to a heterozygote level. You only have to get to that level, a heterozygote level within a copper PET study, which is not quite normalized copper metabolism, but those patients do not have disease. So even if you can get to that level, you could feel very comfortable removing standard of care and knowing that these patients will not have a copper burden that is going to continue to add to their disease. Just to follow up on that point, obviously everything is going to depend on the data, but is it reasonable to assume that if the data looks good, the radiolabeled copper biomarker could potentially serve as an endpoint for a therapy approval, or it is still early to say? Yeah, I think it is still early to say. I always like to say everything is always data dependent, and the better your data, the more flexibility that might get awarded to you. I use a base case as, "Hey, let's get patients off of standard care for some period of time as a potential registrational endpoint." If your data looks incredible, that you are getting all these patients to vastly improve copper metabolism, and that is very predictive of getting patients off standard care, there is always an upside scenario where the FDA can be a little bit more flexible there. But we are using our base case as some period of time off standard of care ultimately for registration. Depending on strength of the data, there could be upside to that. Okay. Just maybe two other things to add. The other things we will look at, it is not just the radiolabeled copper PET, but we will look about an enzyme called ceruloplasmin, which is very low in patients with Wilson's disease. It is something that helps to bind the copper and shuttles it through the circulation. In the absence of this ATP7B enzyme that is mutated, that does not occur, and that ceruloplasmin degrades very rapidly. As you treat these patients and you have normalized enzyme function, you should see those levels of ceruloplasmin go up. We will also be looking at urinary copper, which in diseased patients is very high because the main mode in the diseased patient of copper excretion is through the urine, not through the feces. In these patients, we will also test if you see that urinary copper come down. There is a mix of different endpoints that we can look at, not just the radiolabeled copper PET, and they should all start to point in the right direction if we are having the right impact. Okay, sounds good. Then you guided proof of concept clinical data in 2027. So what kind of initial clinical data do you expect to disclose? Do you know how many patients that you would disclose those data for? Yeah, I think at a minimum, we'll be looking at those three endpoints that I talked about, radiolabeled copper PET, ceruloplasmin, urinary copper, safety obviously as well. This is standard where you do a few patients per cohort. We've talked about at least three cohorts. So you can do the math on what a potential first data set could look like. You always have the ability to backfill in certain cohorts if you feel like you're at a good dose level. So we'll wait and see where we are. If things translate well, then you'd expect somewhere between six, nine, 12 patients essentially as we get to that first data set. Great. Thanks. With that POC data that you're going to get in 2027, what kind of data will be sufficient to give you the confidence to move the program forward? Yeah, I think what we're going to want to see to gain. So I think at a minimum, are you improving copper metabolism? That's the bare minimum, right? If we're not improving, if we're making that edit, you should be improving copper metabolism. So are we able to achieve the level of editing efficiency that we've seen pre-clinically, or even some level that's similar to that, which is not that high to see a really strong effect, right? Even at high 20s whole liver editing, we're getting to almost normalization of liver copper metabolism in a mouse model. So, as long as we can see vastly improved copper metabolism and these other markers are going the right way, we know we have a drug. Then I think the secondary question to that is this something that is effective in every person, right? Is there any reason why we're not getting widespread efficacy, which would be the expectation based off animal models? If you have one of six people responding, then you know you can have a response, but there's something wrong there. You'll want to see pretty consistent responses across the majority of patients. I think if you have that tells you you've got a drug. Okay. That makes a lot of sense. For Wilson's disease, your current candidate only covers one mutation. I think it's the H1069Q mutation. There are other mutations for Wilson's disease, so talk about how you're going to broaden that program to cover other mutations and the timeline to do so. Yeah. The H1069Q mutation covers about, we say 30%- 50% in the Caucasian population. We think about predominantly, the U.S. and European populations. Other numbers we've looked at, maybe that's closer to 40%, 45%, but somewhere within that range. We think ultimately with a handful of additional editors, we can get to about 60% of the population in the U.S. and Europe. The ability to get those studied is we know, or at least we have alignment that we can do additional mutations under the same IND, and we can leverage all the information, or the majority of the information that we have from the first mutation for subsequent mutations. It's a lot faster and it's a lot cheaper to do each additional mutation. Hopefully, given genotype to phenotype is pretty similar across Wilson's disease. There's not a lot of not significant difference. Hopefully, it's going to be a small number of patients that need to be studied as well, to ultimately get those to licensure. As we go into Asia, it's a little bit different. The mutational backdrop's a little bit different. The prevalence actually seems to be a little bit higher. The most prominent mutation there, R778L, is even potentially 40%, 45%, or even more percent of those patients. With a handful of editors, we may even be able to get to 70% of that population. We think we can pretty quickly get to the next mutations. R778L is pretty close to ready to go. There's some additional off-target and other work we'd have to do to get ultimately to the IND, but we've got a pretty good editor right now. We're trying to determine the right time to layer that in to this current study, even though it's a prominent mutation in Asia. There are patients we've identified in the U.S. that do have that mutation, so you could start studying that outside of Asia first before you go into Japan. That would be the plan for that program. Got it. A competitor is developing a gene therapy for Wilson's disease. What are you watching in this space when it comes to gene therapy? Yeah, look, I think first off, I mentioned that we're probably going after 60% in these geographies and 70% in Asian geographies. So there's a market beyond what we're doing, so there is a place for other therapies to play. We think a gene therapy and a gene editing approach are very different. A gene therapy, you can't fit the full gene in an AAV, so it's a truncated version of that gene, which can or can't have the exact same type of functionality. It's something that's not under endogenous control, so you're under the regulatory controls of the AAV that you're putting in there, where when we're making a permanent change at that mutation location, at that gene in that genome, so you're under the normal endogenous or the normal regulation of that gene. We always say, you don't want zero copper, right? It's always going to be some balance that you're trying to get with many of these things, and being under endogenous control is the right answer. The other thing is, every daughter cell that has the correction is going to have that correction. So this is something that's durable and gets maintained over time. Some data suggests that you even have a positive selection bias over time, but even without that, this is something that's a durable fix. For gene therapy, hepatocytes, they're not dividing like epithelial cells, but they do divide over time. So it is something that will get diluted out over years because every daughter cell, only one of the daughter cells will have that vector. So there's a long-term durability question there, too. I really feel for this specific disease, or most diseases where you are going after the liver, a gene editing approach where available is a better approach than a gene therapy approach. But yeah, and we are hopeful that we will see data to validate that. That makes sense. I would like to talk about AATD now. It is another very important program of yours. Recently, the arbitration panel ruled in favor of you guys, so now you can move ahead with your program. I think Beam said that that decision was somewhat narrow. They still have some rights in developing AATD with Prime Editing. Just can you help us square away what each party's rights are in this space? Sorry, what? Yeah. Each party's right- Yeah. your right and their right. Yeah, we've never contested whether Beam has rights to Prime Editing in Alpha-1. As part of the collaboration agreement that we did with Beam back in 2019, they have rights to Prime Editing for doing certain types of edits, which are transition-only edits, essentially what you can do with base editing. What I think this decision, and we always felt, okay, we're operating under the Prime field, where we're not just doing transition-only edits, we're also doing non-transition edits. What this decision says is it clearly defines what the Beam field is and what the Prime field is, which is, again, back to what we believe the Beam field and the Prime field were before. As long as we're operating within the Prime field, which means, we are doing non-transition edits as well, then again, that has a material impact. That's within the Prime field. Mm-hmm. Okay. That makes sense. It's a clear definition across the fields. Okay. AATD is a somewhat crowded field for gene editing, and for good reasons, just because there's a really high unmet need, and there's a large patient population. There are two companies that already have some clinical data for AATD. Where do you think your candidate can potentially differentiate from those two? Yeah. There's two base editing companies that have had data. Beam's got probably the largest collection of data to date, and I think they're showing somewhere in the mid-teens level of systemic levels of Alpha-1, which is kind of right square into what you'd expect for a potential heterozygote patient, which typically don't get disease. I think it seems, looking at kind of MZ levels, that they're probably near levels of saturation. With that other company out of China, they've released a few patients worth of data. I think the first patient, which had a higher baseline, that looked much higher, and I think the next two patients that came out kind of look maybe in a similar range to where Beam is, maybe a touch higher. It looks like the data that we've seen to date from those two companies looks very similar. From what I'm told, the company out of China, Yol Tech, has a lot less bystander protein than Beam has, which we can debate if there's some theoretical potential risks there. But it seems like everyone's kind of getting to saturating levels and then potentially getting to kind of similar levels of Alpha-1. I think what becomes really important as you go out is not just what levels of Alpha-1 you're getting to, but also it's going to be safety. How safe is your LNP? It's one thing to dose five patients, 10 patients, even 30 patients. What are things going to look like when you've dosed 100 patients, 200 patients, 300 patients? Again, we think we've got a LNP that is a pretty wide therapeutic index, and from a safety standpoint, we think we're going to be in a good place. From an efficacy standpoint, I think we're going to have to wait and see. I still believe a Prime Editing approach where you're taking a patient back to wild-type protein, where there's no bystander edit or anything else going on, that you're fully at wild-type protein, to me, it should be the preferred therapy here. Even though we're not first, I think there's still the potential to be best for that reason. We'll see how all the data plays out over the next couple of years. As you said, it's a crowded space, but it's a large market, and there's a lot of patients to treat out there. Mm-hmm. That makes sense. Is the LNP used in the AATD program the same as the one used in Wilson's disease program? It is. Okay. It is, and for that reason, we've been able to leverage a lot of what we did in Wilson's to Alpha-1. For future liver indications, we think we can take that to another level, leverage even more, scale even more, et cetera. Okay, that's great. I know you guys are working on an IND right now for AATD. What are the remaining gating factors in there? We are still reiterating our guidance of a regulatory filing this quarter. Okay. All right. Soon. Similar to what I asked about Wilson's program, you expect to provide initial proof-of-concept data in 2027. That's correct. Again, what kind of data will you provide? For how many patients? Yeah, I don't know patient numbers yet for that, but what I would say is we've kind of seen from others, you can get pretty quick readouts. Looking at serum Alpha-1 levels and percentage M versus percentage Z, you could see that as early as a week to two weeks. You can get pretty quick readouts there. Probably not dissimilar to Wilson's disease as we think about patient numbers potentially, depending on how many cohorts we need to get to get to the right dose level. Expect to have that data sometime next year as well. Okay. And then similarly, what kind of proof-of-concept data would give you the confidence to move the program forward? Yeah, I think we're not recreating things here. I think we've seen what other companies have been able to do. I think similar types of development paths are likely for all of us, where you're going to do some small number of patients, as long as you're seeing pretty consistent levels, response rate in these patients, and levels that you can very quickly transition to a registrational trial for accelerated approval, where you're using serum levels as your main endpoint. Right. Okay. All right. Perfect. Now I want to switch to your CGD program, your ex vivo program that you had two patient data. We have all read in the news how those two patients are doing so much better thanks to your drug, so kudos to you guys. In June, you talked about that FDA is willing to accept those two patient data as a filing package. Where is that conversation now? Any update on the timing? Yeah, so we're planning for a filing sometime in the first half of next year. There is just some small CMC requirements that are needed, and that's what's gating to that filing. Yeah, we think we're in a good position there. Okay. If it's approved, what is your commercial strategy there? Yeah, there's not a lot of patients out there with the disease. It's an important therapy, so we want to make sure it's available. We're not looking at this as some massive commercial opportunity for us. But we think with a very minimal amount of expense, we can get this drug to the patients in need. Okay, great. In the last two minutes, I want to touch upon the other programs that you have in your portfolio. Cystic fibrosis, you talked about that earlier. Maybe talk about just an overview of the status of that program right now. Yeah. We are continuing to make good progress there. That is an area where we are looking at predominantly LNP as the delivery technology to the lung. We have gotten to a number of editors that look like we are getting high efficiency editing, definitely seeing really high efficiency editing as we look at ALI cultures and other things that kind of people look at to try and de-risk this. Next step is to really getting to really good in vivo data and getting to a drug candidate to ultimately, hopefully, get this into the clinic in the next couple of years. You touched upon this earlier, delivery to the lung, it has been a challenge for the industry. Can you talk about your progress there and what have you seen so far in that area? Yeah. We are evaluating both external and internal LNPs. As you look at LNP to the lung, I think that, and there is some mRNA and other therapies and gene therapies, that is more been HSV or AAV, but there has been some therapies where I think you can get the drugs delivered, right? Then it is a question of do you have the right cargo? Are you getting the right amount of expression of the CFTR gene in the right cells? There is a couple of different approaches here as I think about cystic fibrosis. There is one, can we get enough editing in the right cell type and sort of those basal lung cells that are going to give you that sort of that long-term benefit, and that would be sort of a great result. But this is also an indication where even if you can hit those bronchial epithelial cells that are going to turn over at some rate, this could be a possibility of, hey, maybe I have a little bit more frequent dosing, and I am dosing this every twice a year or a few times a year instead of a once and done. I think there are different approaches here that can ultimately lead to success, and I think we'll probably test a number of them and figure out what's best for these patients. Because we can go after all the patients that really are not amenable to TRIKAFTA and other treatments because of the mutation set that they have. But there's no reason we also can't go after that F508del population in the future as well. Right. Yeah, that makes sense. When should we expect to see some preclinical data in that area? Yeah, I think we hope to share some additional data later this year or sometime next year as we get to proof of concept. Okay. Our time is up. Allan, thank you so much for joining us. Yeah, thank you. Great questions, and thank you for having us.