Great. Thanks, everyone, for being here. I'm Yanan Zhu, one of the biotech analysts here at Wells Fargo. I'm very delighted to be joined by management team from Editas Medicine. With me here are Gilmore O'Neill, CEO of the company, and Amy Parison, CFO of the company. Gilmore and Amy, thank you for being with us. Thanks for having us. Thank you. It's great to be here. Great. I was wondering if you can kick us off by providing a company overview, and then we can jump into questions. Yeah. I'd be happy to. Editas Medicine is a company which is entirely focused on the development of in vivo delivered CRISPR edited medicines. Our philosophy centers around the creation of medicines that are highly differentiated for efficacy. That is that they will make a substantial and meaningful change to the standard of care. The second is that they exploit the unique properties of CRISPR editing. CRISPR editing can do many things, but there are certain things which are absolutely unique. Its ability to agonize or augment or increase the expression of potentially disease mitigating proteins as well as potentially repair them. I think the third thing is a focus on ideally disease areas where there are markers or biomarkers that became rapidly and easily accessed to allow rapid and meaningful, robust proof of concept in humans. Finally, particularly with this in vivo focus, we want to be sure this very powerful technology is accessible and available, so that it can really have a significant impact on public health. That means a very deep dive into ensuring that we control the cost of goods and get it down. It's one of the reasons we don't actually do cell-based therapies at Editas anymore. This basically has resulted over the last four years when we first articulated that strategic plan and vision, to the manifestation of that strategy in our lead asset EDIT-401, which is a CRISPR edited medicine delivered by single IV infusion, that basically reduces LDL cholesterol, Lp(a), and ApoB by the mean of 90% in non-human primates. We are tracking to be in the clinic this year, with top line data next year. Great. Let's dive into this program, EDIT-401. Could you lay out what problem are you trying to solve with EDIT-401? Yeah. It is easy sometimes for us to forget, I don't know why, but because this is so common, that cardiovascular disease still remains the main cause of mortality and morbidity for humans on this planet. While we have made substantial progress over the years in managing cardiovascular death and cardiovascular complications, particularly around the control of blood pressure as well as cholesterol and LDL cholesterol, there is still a substantial proportion of patients who sit at very high risk for cardiovascular complications who cannot get their LDL cholesterol, which is the key bad cholesterol, down to a target level. That is notwithstanding statins and statins combined with PCSK9 inhibitors. Our approach is to essentially directly increase the synthesis of the LDL receptor in liver cells, and that receptor basically pulls LDL cholesterol out of the blood. We have found a way by directly upregulating it that we can reduce LDL cholesterol in non-human primates by a mean of 90%. What does that mean in the context of non-human primate? Well, the non-human primate has been very successful in predicting biological effect sizes, both for cholesterol-lowering agents, as well as more recently for predicting the effect sizes of in vivo gene edited. That combination, we feel substantially de-risks that approach. There are a few other things that are relevant to how we have de-risked it, and if you do not mind, I might just go ahead and talk a bit about that. Oh, yeah. Sure. The LDL receptor was a target that we identified again, through the application of our disease or therapeutic selection philosophy. If you want to truly differentiate and change the standard of care, you cannot really go after pharmacologically validated targets because others have actually done that. We are actually using an approach where we leverage human databases and literature where gain of function or naturally occurring gain of function variants are identified in humans. In the case of LDLR, a gain of function variant in the LDLR was described a few years ago in an Icelandic kindred. In that Icelandic kindred, people who actually were heterozygous for this gain of function mutation had LDL cholesterols, which were ranged from about 13 mg per dL to about 60 mg per dL. And then more recently, a French kindred has described a similar association, again with a similar gain of function variant. In both cases, this is a deletion in the three prime untranslated region of the LDLR gene, which results in the production of a messenger RNA transcript, which is stabilized and has a longer half-life because that deletion in the three prime UTR removes a number of microRNA binding sites. All that detail doesn't really matter. What really matters is that the message for the LDLR persists for longer, which means that more copies of the protein can be produced from that. And certainly in our hands and in the lab, we have demonstrated that we're achieving at least a six-fold increase in LDLR expression in the liver of our non-human primates, and that is associated with the significant, and dare I say, unprecedented reduction in LDL cholesterol that we're seeing. Got it. Yeah. Thanks for going over that background. I was wondering then let's talk about your editing strategy and how does it replicate or aim to replicate- Yeah that natural mutation. That natural mutation or natural occurring gain-of-function variant that was described in the Icelandics really pointed us to a region at three prime UTR and basically a series of microRNA binding sites whose deletion or removal would benefit and lead to beneficial effects. We essentially walked across that region, our lab, I didn't, but the people in our lab walked across that and basically tested and validated a number of guide RNAs and found an optimal, not an identical deletion, but an optimal pair of guide RNAs that create a very comparable deletion, which actually in fact manifests even a greater potency than the gain-of-function variant that we saw in the Icelandic kindred. I see. Hence the 6x or the six-fold, because in the published data, those families were seeing maybe a 3x upregulation of LDLRs. We would actually get even more potent. Okay. Your increase, is that just due to your biallelic or even if you're heterozygous you still- No. Actually it doesn't have to be biallelic by any means. In fact, I'll return to that in a second. Oh. We're seeing that effect based on essentially the selection of the guides. Basically a combination of the guides and essentially we believe the deletion they created. Okay. Yeah. Got it. Got it. But in your human application- Yes Is the goal to just produce a heterozygous genome type or? Well, that's a great question. A couple of things. From a goal point of view, our goal is to reduce LDL cholesterol by a very high level, and certainly at a level that has not been seen before. With regard to how we are doing that, we have found in the non-human primate that we can achieve those high levels of LDLR expression and those very low levels of LDL cholesterol with editing as few alleles as about 15% of alleles. Okay. That's the first thing. That's from an objective point of view. That actually matters because we want to be able to help patients who have multiple reasons for elevated cholesterol, and one important group are those patients with heterozygous familial hypercholesterolemia, and they will have one copy of a mutant LDLR receptor. So where essentially most of the benefit they are going to get is from upregulation of their wild-type copy. Does that make sense? Yeah. Right. But would the guide also edit their mutant allele? No. It may upregulate the allele because in the end it will basically create the edit we are making, the genetic change we are making is in the three prime untranslated region. That's non-coding DNA. Yeah. Right. Not relevant to the loss of function that these patients have. Okay. Got it. Yeah. Great. One question that I'm sure you're getting is are these people normal or how low can LDL be without- That's a great question, and it's certainly one that we actually have looked at very closely as we were identifying this target, determining to go after it. First thing is that the people, I shouldn't call them patients because these are very healthy humans within these kindreds actually are very well, both in the Icelandic kindred and the French kindred. They actually do very well. As I said, they have ranges of LDL cholesterol as low as 13 mg per dL up to about 60 mg per dL. That's one piece of evidence that says you can go very low. The next piece of evidence is that in certain families or kindreds who are homozygous for loss of function for PCSK9, they actually can have very low LDL cholesterols. That's the second piece. Those are two pieces of genetic evidence. The third piece of evidence really comes from interventional trials. Over the last 10 years particularly, a large number of cardiovascular outcome studies have been done combining PCSK9 and cholesterol. When they actually, for example in the FOURIER study from Amgen with Repatha, they are actually able to look at quartiles. So lowest quartile against the highest quartile from the point of view of levels achieved of cholesterol. They had patients or patients in their trials who had achieved levels as low as and lower than 10 mg. So basically about to single digit, which really is below the lower level of quantification. They found two things. One, these patients did very well from a general health point of view. Secondly, they actually had a greater benefit from point of view of outcomes relative to patients who achieved some reduction but were in the highest quartile. So they basically had a benefit from a cardiovascular point of view, and then from a general health benefit point of view, the EBBINGHAUS sub-study, for example, looked at cognitive function, and these patients did very well. There was no difference between them over a long follow-up versus patients with higher cholesterols. Indeed, looking at homeostasis for fat-soluble and cholesterol-based or sterol-related hormones was actually also very good. So overall, the data altogether suggests that driving LDL as low as we can be tolerated. Indeed, as I've said before, and again across these multiple studies, the lower you drive it, the greater the cardiovascular outcome. Indeed, there's almost a linear relationship between LDL cholesterol levels reduced or achieved and the cardiovascular outcome, with possibly a 20% reduction in five-year cardiovascular outcome risk or risk of cardiovascular events at five years for every 40 mg or so per dL reduction. I think all the data point towards driving LDL low and long. Okay. Yeah. Maybe going back to the editing of one or both alleles. Sorry, I- No worries. I wanted to clarify a little bit. So you mentioned in animals, if you do 15% editing- You can get to 90% knockdown of LDL. Yes already, right? In the non-human primates. In the non-human primate. When you get that low level of editing, right, in those cells, should I understand this as you are editing probably 30% of the cell? Yes. Each with only one allele. Yeah edited. Yes. Overall that's like 15%. Yes, correct. That's how it- That's about right. Basically you don't have to edit the entire cell population of hepatocytes to get this effect size. It's not like you're editing 15% of the liver, which each cell biallelically edited, right? Can you distinguish the two? Yeah. So we have looked at that. We have a set of data and obviously a substantial amount of modeling data using that data to really suggest the majority of these are monoallelic. Okay. Yeah. Got it. Yeah. That's great to learn. No, it's very exciting. I think one of the reasons that you can do this, obviously, again, it's an augmentation strategy. So by increasing the levels of LDLR, if you're trying to do an antagonist and try and block something, you have to edit almost all the cells and all the alleles. Right. Exactly. That's an important potential upside. There are many potential upsides by using augmentation, not least is the differentiation, but the other is that you may not need to edit as much because you're not trying to block something almost entirely. Right. Does that make sense? Yeah. Yeah. That makes total sense. Let's touch on Lp(a), because I think- Yes Earlier you did mention Lp(a). Yes. Obviously, we got the news about the HORIZON study. Yes. Does that make any difference to you or? We don't think so. I think there are two key points or takeaways from the HORIZON study that we saw very high level announcement on Friday evening. This is something we've talked about before in anticipation of these readouts. I think the first thing is that the Lp(a) is, you could regard it as just a sort of a no-cost additional upside for our approach. But the key thing, it doesn't change anything about the risk reduction one can anticipate with the achievement, should we replicate in humans what we've seen in non-human primates of a 90% or really significant reductions in LDL cholesterol. I think that's the first thing. I think the second thing is that I think the story for Lp(a) is far from done. I think the robustness of the data that connect the genetic risk, or rather the genetic description of Lp(a) with the risk of cardiovascular disease is extraordinarily robust across multiple genetic association study and even Mendelian randomization studies. I think when you make or bridge the gap between genetics and clinical risk, there are a few other things that have to be taken into account that we don't fully understand. There's the biochemistry of the Lp(a) particle. Not all Lp(a)s are the same. They are defined by the conjugation of apo(a) to ApoB. But the content of the lipid within that lipoprotein particle may differ. And one big thing that's really under a lot of discussion and focus is for oxidized phospholipids. The second obviously is structure. What is the impact of a lifelong risk? Because the Lp(a) risk probably starts at birth or even beforehand. The LDL risk kind of builds over time. Your risk exposure sort of develops maybe later in childhood, but the Lp(a) appears to be fully exposed. So how long do you have to follow a patient and intervene? There are many other factors about the trial design that obviously we can pick apart. I think kudos to Novartis, kudos to the investigators and the patients for participating in trial because even if they haven't definitively answered the Lp(a) question, that data's going to help inform others on how to design studies for Lp(a) and looking at that interdiction. But again, just returning to one key thing, the key thing for us is that our approach is reducing LDL cholesterol substantially. Associated with that is a reduction with Lp(a) and ApoB. There is a very good mechanistic explanation for that, not least of which is that LDL cholesterol, which we're significantly reducing, is the key substrate for synthesis of Lp(a). Yeah. Got it. Yeah, no, that makes sense. So maybe, you're about to start a phase I trial, right? Correct. Talk about where you are in the Yeah I guess the CTA clearance. Yes. We are progressing well there. We have submitted our documents to the HREC. We have said that we are running our studies in Australia, the Antipodes, and we have submitted to the HRECs, and we actually are in dialogue with the HR ECs. We are progressing well, and that is going according to the timeline that we have outlined, which keeps us on track to dose this year. Oh, okay. Great. Yeah, definitely looking forward to that. Yeah. So are we. Okay. Got it. Can we talk about the dose translation Yes from your non-human primate to human? Yeah. Dose translation is something that has been really increasingly better understood over the last couple of years. And what we've seen consistently is that the experience has lined up with the FDA guidance. And essentially we're seeing a crude 3 to 1 dose translation from non-human primate to human. So a 1.5 mg per kg dose for non-human primates will translate into approximately 0.5 mg per dL or 0.6 mg per dL in humans. And we've seen that consistently play out across a number of now in vivo editing programs in the hands of multiple sponsors. We feel very confident again about that translation that we're making. Right. Does that mean, I'm not sure from monkey to humans who have maybe higher LDL levels than healthy monkey, right? Yeah. But if we ignore that, it sounds like it's like 1.5 in monkey could translate potentially 0.5- 0.5 or 0.6. Yes as your efficacious dose. Yes. Right. Yes. Yeah. Would you think the increasing LDL levels at baseline may require a little higher dose or not so much? Listen, you can never be sure, but the data that we have seen to date would suggest that we would expect comparable efficacy no matter what the starting baseline is. That is because we looked very carefully at that. Obviously, the non-human primates have relatively low baseline LDL cholesterols, but we actually also looked at this construct in both wild-type mice on a high-fat diet, as well as mice that were heterozygous for a loss of function for an LDLR allele. We saw very comparable efficacy or similar efficacy all with them achieving mean reductions of 90% in LDL cholesterol no matter where their baseline was. That gave us a confidence that we should see a robust reduction no matter what the starting baseline is. Got it. Yeah. For the phase I study, could we touch on trial design? Sure. Yeah. The trial design is going to look very similar to what people have seen before for first-in-human studies in hypercholesterolemia. We are targeting a patient population with heterozygous familial hypercholesterolemia. I should say the study has a part one and a part two, again, very comparable to what you've seen over the last 10 years. The first part will comprise a bunch of serial dose escalating cohorts. Each cohort will be at a dose, and every cohort that follows will actually have a higher dose. We currently have a design which is a 3 + 3, so we anticipate at least three patients per cohort with an option to add an additional three in the event that we need to actually look at something more closely, and we anticipate at least four cohorts in the clinical trial. At the end of part one, we will select a dose to bring into part two, and that part two will then expand the patient population, and the patient numbers exposed. Let me spend just a little more time talking about the patient population. We're basically targeting, in part one, heterozygous familial hypercholesterolemia patients. These are very high-risk patients who are not getting to target. They're not getting to target despite intensive use of standard of care. Because that's basically the patient population that we actually want to target to start with for this medicine, and the patients with the highest unmet need. Right. I'm sure you looked at this in a preclinical manner or in a clinical matter that for the patient who has. Obviously, these patients have a defective allele. Yes. Editing that allele wouldn't do anything or- That's a great question. We looked very carefully at that. We both looked at it from a modeling point of view, and we actually really have, in modeling, considered all the possibilities of increasing expression of a mutant allele, and we see actually very low risk of that. We see a substantial benefit, obviously, accruing from upregulating the wild-type allele. Right. Got it. Yep. So, in terms of starting those, I think Verve used the 0.1 mg per kg. Yeah. Is that something in a ballpark? Yeah. We haven't shared what the starting dose is, because ultimately that will be a matter of agreement with the HRECs. We were very excited about the data we saw. We actually liked our tox data. Obviously, there's been a substantial experience now. Multiple sponsors have taken multiple LNP formulation for in vivo CRISPR editing into the clinic. Obviously, we want to get the balance right of potential risk, obviously protecting patients, but at the same time, making sure that even if we start at a dose that is not relevant to the effect size that we have seen in the non-human primate at the target dose, that we start at a reasonable dose, because these patients are receiving an edit, and we'd like to see if they could get some benefit. But I think the key thing is for our starting dose, we just want to get that balance right. But it obviously will be a dose that we anticipate would achieve a therapeutic level much less than what we expect to see at the higher doses, and certainly what we saw in our non-human primates at 1.5 mg per kg. Right. Got it. In your work, I was wondering, when you dose escalate- Yeah Where do you determine how low you go in target LDL level? Oh, yes. Well, what we're actually targeting ultimately is we want to see a significant reduction. It is worth calling out that we've thought a lot about, is there a level that's too low? I think we believe that multiple strands of evidence converge on the conclusion that one can drive LDL cholesterol very low without adverse effects. What I would say is that it's clear that the evolution of the space, we've seen evolving data over the last 20 years. Recently with the guidance beginning to catch up with it, would suggest that certainly very high-risk patients who are at very high risk of cardiovascular disease with uncontrolled or poorly controlled cholesterol, they need to get their LDL cholesterol down to about 40 mg per dL or lower. Indeed, the lower you can get it, the greater the benefit. Again, the data continues to show that, both in individual cardiovascular outcome studies, and if you look at the totality of those brought together in a number of really elegant meta-analyses, you see this very strong linear relationship with no ceiling and no floor seen on the dynamic range yet for risk reduction with LDL reduction. Okay. Essentially, you might be guided with safety so that- Mm-hmm. Yeah you reach the highest tolerable dose. Yeah to derive the best benefit. Exactly, yeah. Got it. Yeah. Data is expected as early as the first quarter 2027? We are targeting sharing safety data in Q1. The reason that the safety focus is obviously, I've said we are starting with the dose we'd anticipate being in that sort of very therapeutically relevant range. Obviously, that data is still, though, a significant de-risking event. Because we will have a much better sense of the behavior of our lipid nanoparticle. Yeah. Which I think, again, is an important de-risking event. Then we look to having top-line data for the part one later in 2027. Okay. Got it. I see. Essentially, although LDL lowering can be observed very quickly. but you will be only at lower doses Correct as of 1Q 2027. Yeah. Got it. Yeah. Makes sense. I was wondering at what point you would consider bringing your trial to U.S. We have actually had very good interactions with the agency. We had a very good pre-IND meeting, at the beginning of this year. What was wonderful beyond just the feedback, which was very good, was that a number of areas that we had really asked specific questions were very clearly answered, and actually even removed some obligations that we thought because the data really supported the positions we took. That was pleasant, not just because of the feedback we got, but that it was comprehensive and detailed. It was great to see that depth of expertise, experience, and bandwidth at CBER for that going forward. That basically means we are tracking towards entering the U.S. probably in 2027. The rate limiting step is really building the documentation around the manufacturing process and so on, which is a significant lift from bureaucratic point for U.S. as opposed to other jurisdictions. Got it. Yeah. That's very helpful. Can we talk about, because safety is very important. Yes Let's talk about your LNP. Yes. I think 0.5 mg per kg is a very safe dose according to other companies' experiences. Yeah. It sounds like that's where you probably target. Can you give us any additional color on- Yes all of the lipids, components- Yeah whether some of them have- There are a couple of things. Obviously, we think based on the scaling for right, it's going to be about 0.5, 0.6. It could be higher. Obviously, the empiric data in the clinic will tell us. With regard to the LNP, we are working with Genevant, which is a very experienced LNP company. The LNP that we're using is unique to our program. However, most of the components in it have been in humans before. Then I think very importantly, the non-clinical tox data have actually been very compelling. Certainly at the therapeutically relevant doses around 1.5 mg per kg and around there. We actually saw minimal increases in liver tests, particularly the transaminases, and those were really no different from what we saw in the control animals. You may say, "Why?" Well, the control animals were administered saline. It wasn't LNP, it was saline in the control animals. They saw bumps, too, and I think the key need for that bump is that they're all pre-treated with high-dose dexamethasone, which is the standard protocol for prophylaxis against infusion-related reactions. The transaminases essentially were comparable to vehicle at that level. At higher doses, we have seen some bumps that would actually exceed that of vehicle. Again, those resolved within days and would be in normal range in a week or less. Got it. I was wondering, could you also talk about your cash runway? Yes. How you think about capitalizing the company for Sure further studies. Yes. We have cash into the second half of 2028. We ended Q3 with $212 million of cash, and as Gilmore just stated, that will provide us ample runway and progression to progress EDIT-401 through our phase I to part one and two. When we think about capital allocation, that is really to progress this program and to really prove the proof of concept in humans. Great. Is there additional gain-of-function editing programs that you are looking into? Yeah. There are, and we have disclosed one in the past, which was our HSC program. I know you are very familiar with our old cell-based therapy. That payload, which with human validation, we have now actually placed into a targeted LNP, and we have shared data at past scientific meetings showing really very efficient editing of the HBG1/2 promoter in hematopoietic stem cells, which is actually also durable. We had to make a choice. We are a small company, and we want to be very clear that we applied our capital in as responsible manner as possible without spreading ourselves too thin. We made a very tough choice last summer to prioritize the EDIT-401 program. The in vivo HSC program still remains in discovery. The team has the opportunity to continue to optimize it, which is something that no researcher ever refuses to do. We have that. We have a number of other gain-of-function programs that we have in early discovery. We can continue to work on those, and then at the right time, with a value inflection, with the appropriate inflow of capital, then we could actually move those forward. Certainly in the context of the liver, we can leverage much of the work we will have done from process and analytics development, to actually even potentially move certainly liver targeting programs a little more rapidly because they can leverage the LNP and messenger RNA formulation work we do for 401. Right. Because when you think about it, the only thing that is different for your target gene, if it is in the liver, is going to be the guide. Guide. The messenger RNA encoding the CRISPR- Cas9 is the same, the lipid nanoparticle that you are delivering is the same. That guide RNA change is a 20 nucleotide change. So you can leverage a lot of the process that you have done and a lot of that investment we will have done to really move that along. Got it. Great. Definitely looking forward to the initiation of the first human study. Sounds like pretty soon. Yes. Also, of course, to the data next year. Yeah. And we're very excited and looking forward to sharing it with you. Thank you. Next year. Yes. Thanks very much for your time. Take care. Thank you. Thank you, everyone.