Beam Therapeutics Inc. (BEAM) Earnings Call Transcript & Summary

Good afternoon, everyone. My name is Gena Wang. I'm a SMID Cap biotech analyst at the Barclays. Welcome to Barclays 27th Global Healthcare Conference. Together with me, we have John Evans, Chief Executive Officer from Beam Therapeutics. So thank you very much for giving us the opportunity.

Thanks. Great to be here.

I know yesterday, you just announced -- today is Tuesday, right? Yes. Yesterday morning, you just announced a very big data update. So maybe give us a high-level overview, and then off the data, we can dive into the questions.

Perfect. Yes. So it's great to be here. So Beam, as you know, is working on a next-generation gene editing called base editing. It uses the same power of CRISPR to do precision gene targeting. But once we get to the target spot, we're not doing the traditional cutting that you get with nuclease editing, the first-generation tools, but we're doing a much more precise single base change. This gives us precision and control. It's also gentler on the cells. So when we thought about how to apply this tool, we can actually use it to do things like knocking down genes, turning them on, activating them, our own BEAM-101 in sickle cell disease activates the fetal hemoglobin gene more robustly than others have done before. So it's a very versatile tool. But at the same time, one of the most exciting applications of the technology, because we can make a precise change without disrupting the sequence in the chromosome, is to actually go in and rewrite the genome and correct mutations. And I can tell you the top of our list from the beginning was to go after alpha-1 antitrypsin deficiency for that application. So AATD is a very severe disease. You have a single point mutation, one-letter misspelling, in all patients who have this -- almost all patients have a severe disease. There's 100,000 people in the U.S. who have 2 copies of that 1 point mutation, and then they have the disease. And they have low levels of alpha-1 in their bloodstream, which causes lung progressive emphysema because they're missing that alpha-1 protein to protect their lungs when they get an infection and the immune system mounts a response. They also have liver failure because the mutant form of the protein is building up in their liver, and it's causing damage over time. And so it's really a 2-sided disease, and there's really very little for these patients. So with BEAM-302, we designed a base editor that can go directly to that one spot in the genome. And for the first time in history, actually rewrite a broken mutated gene back to normal, and that's what we're doing. And so yesterday, as you noted, we shared the very first glimpse of data into that drug. It's, of course, an open-label trial. We're still in the dose escalation phase, so we're just in the middle of the beginning of this. But nonetheless, we've been through 3 dose cohorts so far. We've had 15-, 30- and 60-milligrams given to 3 patients each, and we were able to show really the entire range of outcomes we were hoping to see with this drug. We saw, first, safe and well tolerated across all dose cohorts. This is LNP in vivo delivery. It's a simple infusion. We saw dose responsive increases in total AAT. And at the top cohort, we were indeed above the therapeutic threshold of 11 micromolar, showing that we've really moved these patients out of the disease category into something that looks much more like a normal carrier of the disease, who only has maybe 1 mutant copy rather than 2. We are raising AAT because we're creating the normal form of the protein, because we've rewritten the gene. And at the same time, we're dropping the amount of toxic protein in the body quite significantly, up to 78%. And the ratio of the good-to-bad protein is reaching very high numbers. And finally, we were able to show that all the protein we're creating is functional. There's a variety of assays we can do that show that the new protein we're creating does what it needs to do in the body. So that was all the things we were hoping to have happen with this drug. And from here, we have lots more to do. We're still developing it. And we're going to continue to expand the trial and push farther. But already, I think we have a potential therapeutic in view.

Right. And I totally agree with you on that part. Yes. And so, for me is given such a good safety, why can't you dose higher? Why so cautious, only dose up 20% to 30% for the fourth dose cohort?

Yes. So we can dose higher. The safety will allow us to. We jumped 30 milligrams between the second and third dose cohort from 30 to 60. I think that from there, we do want to shade back down the increments that we go in, right? And that's just because as you get the therapeutic in higher doses, you just want to be seeing the safety evolve in a slightly more careful way, rather than jump too far and get surprised, right? So safety is really important here. And again, the enrollment has been so fast, we don't need to be in a rush on that. We're enrolling very swiftly. We have a great clinical site network. So it should be fairly easy to get that next cohort enrolled. And again, depending on what that looks like, we'll make a decision of where we go from there.

So now it goes back to the AAT protein expression, right? I think the stock reaction and the feedback, of course, the market yesterday was horrible. But some of the investor feedback, what I got is also they think they are looking for even higher, even though the first 3 you already certainly hit the minimum threshold and actually pretty promising direction, right? So what they are looking for is what is the potential to even dose higher, right? And the goal is, would that be possible, reaching close to 20 micromolar. And so, maybe first question is based on your feedback, maybe physician or scientifically, how that can translate to clinical benefit and [indiscernible] also the lung functioning later in the year. And so how do you think about the numbers? Should we be -- like, are we too fixated on, say, 11, 12 micromolar, 14 micromolar versus 20 micromolar? How much actually that in terms of the protein level differences can translate to incremental functional benefit?

Great questions. So I'll answer it in 2 ways. One is I think we probably are going to end up being too fixated on specific numbers because I think there's a very strong argument that we are already out of disease, right? We're into the normal range. So there is no diagnosed person with AATD who is the ZZ genotype who has 11, right, or 12. So they just don't exist. So I think there's a lot of good evidence that we've already achieved -- we think we're in the range of the carriers in this case. So someone who has just one copy of the mutant protein, the Z protein and the other copy is normal, we call that M. And so, MZ is the category there. And they don't have disease. They don't have progressive disease, right? They may have very low risk of emphysema if they smoke. And so, that's maybe one consideration. But other than that, they're never going to progress. And so, if we've really gotten patients there, I think that we've had already a transformative effect. And it would be very difficult actually to show a difference between where we are now and a higher number. Now, all that said, we will go higher. I think there's no reason not to. And we have that capability. So in fact, there's a few different things you can look at in our dataset. So one is, even at the 60-milligram dose, we're probably marginally underreporting our efficacy because we had a fairly low baseline level. Patients were about 4.4 micromolar, versus you often see 5s and 6s. And we were delivering almost 3x fold change in the alpha-1 levels, 2.8. And so, it's easy to imagine gathering more patients, having more patients at that slightly higher level, 3x times 5 to 6 would be up in the mid- to high-teens. So I think that even the 60-milligram dose may get us in those directions. And then if you look at that fold change chart, we're also really at the steep part of the dose response curve, right? So with LNPs, you have a sigmoidal shape of the curve. And we think we're right in the middle of that curve right now. So incrementally higher doses, we expect, will continue to unlock more efficacy here. And so, the dynamic range, I think, is going to be important. So the bottom line is we're going to keep going. I mean, this is the very first data set. We couldn't be happier with what it shows. And we're going to keep on going because the safety is so strong that it really enables us to push for those higher numbers, and I think that they'll be achievable.

So I think regarding the 2.8, close to 3-fold increase from the baseline, right? I have a slightly different view, like is that because the baseline is so low, so it's easier to show higher multiple, the increase. If baseline was higher, I think 30 milligrams with a 5.4, and we saw like 1.9, increase. So what is the right way to think about? My calculation using linear increase correlation, if using the same baseline, 4.4, the 15 milligram and a 60 milligram. And if you do the linear correlation and then roughly translate to, say, 75 milligrams will translate to like 14.3 micromolar.

Yes. So I think that I would not do that is what I would say. So I think if you look at the fold change chart that we provided, the error bars are very tight, right? And so it shows you that there's a real effect there. So I think that what we should be thinking about is a certain amount of LNP dose is given. The first step for the LNP is it has to saturate the liver. So the reason you have a flat dose response in the beginning, at low doses you may not have fully saturated the liver, okay? But once you've done that, now you're just driving up the concentration in all the cells, all the hepatocytes, and that's going to drive your editing rate up, okay? And so, we think we're in the steep part of that curve. So the reality is patients -- the reason the baseline matters, is patients just do have different set points for alpha-1. They may be low or medium or high. And then you apply the fold change to those patients, and they get that robust change. So we're controlling for the baseline levels with that fold change chart. So the bottom line is, I think that we don't think the baseline matters for the fold change. We think fold change is really about how many alleles have we edited in the liver at the end of the day. We have a reasonable proxy for that, right? So if we look at the proportion in circulation of M versus Z, we know from clinical genetics, again, that carriers, and I keep pointing to that group of people, they have one copy of M, one copy of Z. And so they are 50% of the gene is M, 50% of the gene is Z. But in circulation, because their Z doesn't get out of the liver very well, they're actually more like 80:20, right? And so we are actually at that level or above in our proportions in circulation. We reported 88% M in the latest follow-up point. So that would suggest that we are getting maybe about 50% of the alleles being edited. We're in that MZ genotype range. But that very clearly also suggests that there's more room to run here. So I don't think we're close to the end of our curve in terms of dynamic range of adding more dose.

Okay. So if you don't have a safety issue, what would be the ideal your goal to reach, say, M AAT protein level?

I think, I wouldn't want to set a number. I think that we -- because I think we already have an approvable drug, but I think that as you go higher, you start to think about numbers in the 15% to 20% range. And I think that those are certainly in view. And now you're into the high MZ or basically into normal MM levels.

Okay. Very helpful. And I think even before data update, and we discussed in the past, we still got tons of investor questions on the lipid-nanoparticle and yours compared to the Verve, especially given the previous safety issue. So maybe can you take this opportunity to talk about your lipid-nanoparticle and how similar or different your current one, and you also have internal other.

Great point. So the LNP is, of course, critical here. This is in vivo editing, where it's a simple infusion. It's hard to overstate what a powerful thing that is. Cost of goods is low, they're redosable, titratable. It's really an amazing system for accessing hepatocytes. And so we did really our own work on this. The ionizable lipid is in-licensed from Acuitas. We also have some of our own ionizable lipids that we can use for the future. But then it was all Beam work to put the other components together, process development, formulation, and manufacturing. Manufacturing these carefully, it really does matter in terms of how they perform in the body. So I think the team did a really superb job. And so, this is a beam proprietary system here. And it is now, we think, fully derisked. So we have a great safety margin here, very little in terms of LFT changes and very dose responsive and clearly reaching full efficacy. So that is an asset both in alpha-1 because it allows us to do what we were just talking about, which is push the dose and explore higher levels of efficacy in the near term, which we're excited about. But also now it's a platform that we've really validated Beam's approach to in vivo delivery. And so, we can now do many other things targeting the liver, and that's, of course, a big strategic part of our portfolio. So certainly, immediately with BEAM-301, our glycogen storage disease program uses exactly the same LNP, has a low bar for editing. I think that you have to say that has a high probability of technical success at this point. And then we are ramping up additional liver programs to follow. I'd like to think of it as almost a conveyor belt at this point of preclinical programs that now can move forward. And this is when, in my eyes, it gets really fun because the platform leverage that we now have can apply. We've done all the work. And so a new filing for a new program, you can just have a tweak to the editor, a new guide RNA, same LNP, same manufacturing, same off-target set of assays, and you have a new product. And so, the ability to now get leverage from our investment to date and create a string of new programs, all of which should have high probability of success, at least of getting to Phase I safely and getting to a certain amount of editing, is now in place. And so, that's a big part of our strategy going forward.

So for this lipid nanoparticle, is that the unique part is ionizable lipids there that make a difference, like better safety, the delivery efficiency?

I think, no. I think it's one part of the puzzle for sure. And we like this one, and we're using it for 301 and 302. But as I said, we have our own as well that are at least as safe and potent, so we can swap those in and out over time. But I actually think the special sauce, if you will, is much beyond that. I think that LNPs are validated, but they're not trivial in terms of how to make them work. You really have to have a lot of know-how. And it's all about how well do they come together, how well is the payload packaged? Is it uniform? Are you getting small particles, so the biodistribution is predictable. These things make a big difference, and they are completely functions actually of the process development and the manufacturing process. And that's all being proprietary. In fact, these LNPs that you're seeing with BEAM-302 were manufactured in North Carolina at our facility by our team. So that's all in-house know-how.

Okay. Good. Going back to the AATD program. So there are 2 assays you have conducted to test both the total protein and the functional protein, right? Do you think that, that will be sufficient for the FDA? That by-stander editing generated functional protein. Can you foresee any additional data or testing that FDA would like to see, given you did introduce additional mutation there?

Yes, I can't. I think that, that is the definitive assay that you want to check. And so, what you're noting is so base editing has this signature. So we turn the Z mutation into the normal allele at that position. It's called M. And then with the base editor, this version, we get some amounts of what I call the canonical sequence. And then we create some amounts of that change plus another amino acid change. And what we've been able to show is that, that is normal. It's a silent change. And there's actually people who have that SNP in the world, it's a site that has several SNPs in the population. But most importantly, it's functional, and we've shown that several times preclinically. And so, now we've shown that clinically as well. And so, the functional assay is really the critical point, where the whole point of having alpha-1 levels in your blood is that they are functional, that they inhibit neutrophil elastase and stop it from degrading your lung tissue. And so, I do think that is the important assay. But really, it's the constellation of all the assays. So what the FDA will want to see is, okay, total AAT is going up past these thresholds. What's the composition of the AAT, right? So as we've shown, we're reaching 88% M. So out of the 12.4 micromolar, that means we've got 11 micromolar of M and just 1-plus micromolar of Z left over. That's a huge change in the quality of the AAT that is circulating. So really important to think about that as opposed to, for instance, augmentation therapy, where you're adding some extra on top, you still have all the Z there. And Z, we do think, is a bad actor, not just in the liver, but systemically. It creates polymers. It interferes with the action of the M protein. So there's a lot of problems there. So the FDA wants to see total AAT. They want to see how much M are you producing. We're in the double digits now. They want to see how much Z is left. We're down to maybe 1 micromolar of Z left, and we can drive it further down. And then they want to see functional. And so, in our case, at the top dose level, the functional assay readout was overlapping with the total, which shows you that we have, again, converted nearly all of the total AAT to M and all of the M that we're creating with and without any variant is functional by those assays. And so, I think that all of that together is exactly what the FDA would want to be seeing.

Okay. And then regarding the regulatory path, and this is a relatively new space, right? So do you think it's possible to use biomarker? Or do you think they're like, say, lung function, FEV1, which should be pretty quick to measure? And like what could be the likely path there?

Yes. So we -- so the beauty of precision medicines like this is they show themselves early, which we've just done. This is, we think, clearly a potential medicine. And then because you're fixing the root cause of the disease, any endpoint is available to you because you know that you will work on those endpoints. So obviously, we favor those that get this drug to patients quicker. And I think that this is a perfect setup for something more accelerated. And there's different flavors of that. But at the end of the day, in my experience as a drug developer, you do find that when you're on the core driver of the disease, and you're reversing it and you have a lot of biomarkers that are telling the same story in a coherent way that you have reversed the disease, and especially when you have clinical genetics like we do, which show you that the phenotype we're now creating is a nondisease phenotype and is therefore predictive of clinical benefit, that is usually enough to drive that accelerated approval conversation. So we will absolutely be exploring that. And I think that we're cautiously optimistic that, that sort of thing, will be available to us. Nonetheless, we also would think about running other kinds of studies to generate those other longer-term endpoints. We want to create a strong value story here for the product. And again, we are confident that the drug would succeed in those trials because we've reversed the disease at its root cause. So for the lung, that would be things like, can you measure lung function over time and show that there's no more progression of disease, things like CT densitometry are available, literally measuring destruction of lung tissue over time. Those are achievable endpoints. In the liver, there's been a lot of work with the RNAi knockdown field to show resolution of aggregates and change over time in liver function. The liver, in particular, can regenerate. It's a fairly dynamic organ, and we would expect even, again, already the amount of knockdown we have here to start alleviating the pressure on the liver and allow it to start healing itself. So that's another angle that we can create. So I think overall, the answer is, yes, we do think there's more accelerated pathways here, and I expect us to be able to generate important functional data over time.

Okay. Regarding the Part B study, I think you did mention that dose could be lower. Maybe any thoughts on the initial dose?

Yes. So Part B, we have no reason to expect a different tolerability or efficacy signal in Part B. When we've dosed animals that have cirrhotic livers or have late-stage disease, we didn't see any differences there. But with an LNP, which is delivered to the liver, we just felt that it was prudent to start by excluding those patients. It's the minority of patients, maybe it's 15%, 20% of them who have this very heavy liver involvement. But you wouldn't want to confound your initial safety signal with the questions about whether the liver was playing a role. So I think it's just cleaner this way. So we now know that Part A has delivered a very clean safety profile, very well tolerated. So that's great. And now we get to go and check, is there any difference in patients who have heavy AATD-driven liver involvement. So formally, it is another dose escalation. So you will step back down and titrate up just the way we've done in the Part A. That said, we may not need to completely repeat all of the doses, I think, especially with this tolerability profile. So that's -- we'll work with our safety committee and decide exactly which doses to test.

Okay. So that will be one of the 15, 30, 60, or it could be even lower?

You wouldn't need to go lower. I'm sort of arguing we may not need to go all the way back down to 15 even.

I see. 30...

Right. So that'll be something we have to debate.

Okay, good. I know we only have 1 minute left to maybe quickly give update on the sickle cell program.

Sure.

You did show very impressive data.

Thank you. Yes. So BEAM-101 is our other program that I feel confident is the potential medicine. It showed, we think, best-in-class profile at ASH with superior hematology readouts with 60:40 ratio, reaching that, again, carrier trait profile and also fully resolving anemia in those patients and having fewer cycles of mobilization and faster time to engraftment. So that trial is fully enrolled on the adult side. We are adding adolescents now, and they are on study. But that puts us in a great position to move quickly. We'll have new data at EHA, most likely at ASH as well. So I think if we can see continuation of those strong benefits, do we see with larger n. Are we still getting good mobilization cycles and fast time to engraftment and that same hematology profile? Does that best-in-class profile continue to firm up? We'd like to see that. And then the other metric we've been tracking there is the 30th patient dosed, somewhat of an arbitrary midpoint checkpoint in terms of doses. But the reason we chose it is that the Vertex approval for CASGEVY was ultimately on 30 patients, with the 15 months follow-up to check for how many vaso occlusive crises were given. And so, if we were to get the same package from the FDA, that 30th patient will start the clock for follow-up, after which point we could write it up and bring it to the FDA for a filing. So we're not so far off on that program, and that's moving quite quickly as well.

Okay. Will you announce when you dose the 30th patient?

Yes, we'll probably give -- maybe I don't know if it's the day of, but we'll certainly update the guidance when that's happened, yes.

Okay. Great. Well, thank you very much, and we look forward to the multiple updates later this year.

Thank you very much.

Okay. Thank you.