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

Okay. Great. Thank you. My name is Gena Wang. I'm a senior biotech analyst at Barclays. Welcome to our first in-person Global Healthcare Conference post pandemic. It is really nice to see everyone. I would like to thank all the participants, investors, companies, also especially our event team and corporate access team who made this event possible. With that, I would like to introduce our next presenting company, Beam Therapeutics. With us today, we have Pino Ciaramella, Chief Scientific Officer and the President. Pino, so maybe before I turn things, could you give a brief overview of about Beam?

So first of all, thanks for the invitation. Finally, to see people live rather than through Zoom is a wonderful experience again. So my name is, as you can see, Giuseppe Ciaramella, but everybody calls me Pino. It's an Italian thing, but feel free to call me Pino. I'm the President and CSO at Beam. Beam Therapeutics is a next-generation gene editing technology that primarily is based on a technology that was originally developed by Professor David Liu at Harvard and now at Broad Institute, which is called base editing. And base editing is still a CRISPR-based gene editing technology, but with 2 fundamental changes. The first one is that the CRISPR protein is modified such that it can no longer make a double-stranded break, which we believe is a significant stress to the cell. In fact, the protein has then been further modified by attaching to it a human enzyme called the deaminase, which is capable of catalyzing a direct chemical reaction, which is called deamination, direct on the gene and can convert 1 nucleotide to another simply by eliminating essentially an amine group from the binding surface of the nucleotide. And depending on the deaminase that we use, we can actually catalyze a C to G change or an A to G change. As you know, a C or an A are present in all of the base pairs in the human genome. And therefore, if you target the leading of the ligand strand, you actually can achieve 4 edits with the 2 editors. And of course, there are many things we can do with this editor. Of course, we can correct direct point mutation that causes disease back to a wild-type sequence. But actually, because base pairs carry essentially many of the different biological functions, we can change other things as well. So we can change regulatory elements like we are doing in our BEAM-101 program, whereby in this case, we're actually inserting point mutations in the promoter region that prevent the binding of repressor proteins. We can change amino acids in the active sites of enzymes, making them more or less powerful. We can change phosphorylation sites, therefore, modulating regulated sort of signaling cascade. And then finally, we can also introduce stop codons as well as disrupting splice acceptor donor site so that we can knock out genes very effectively. So it's a very versatile tool that we've now deployed it across more than 12 programs in the 4 years or so that the company has been active.

Thank you. So your lead candidate BEAM-101 will enter clinic pretty soon. So what is your thoughts on the initial Phase I trial that is on?

So in this case, in BEAM-101 dimension, we are modifying the promoter of the gamma 1 and gamma 2 genes. These are the 2 genes that are responsible for expressing the fetal form of hemoglobin, which in adulthood, soon after birth, are typically switched off by the binding of reverse proteins. By making this one mutation, those reverse proteins can no longer bind to it. And at the end of last year, we filed a successful IND, which is now open, and we plan to basically select the first patient in the second half of this year. The trial design would be very similar to probably what you've seen with CRISPR Therapeutics. It will require -- it is U.S. based at least initially, and it will require what is called a sentinel cohort, which is a cohort of 3 patients who are sequentially dosed. So the first patient is selected and dosed, and then we need to wait for engraftment to take hold, making sure that there are no safety concerns. And then at that point, we'll dose the second patient, and the same process will be repeated including the third one. And at the end of the third patient, we will then have an expansion cohort. And then subsequent 2 expansion cohorts that have the potential to become fileable, pivotal, particularly in the last cohort, and we will see obviously how CRISPR is able to actually file. They have mentioned that they will probably be filing at the end of this year. And so we will see whether we can follow that similar pattern.

Okay. So then can you give us a sense at what point we will be able to see the initial data and after how many patients?

Yes, we haven't, frankly, neither guided, nor decided, to some extent, whether we want to disclose just a single patient or whether we want to have the opportunity to have evaluated a number of patients, and in particular, have follow them for a certain period of time, I think remains to be seen. The important thing to think about as we are up regulating the fetal form of hemoglobin, just the procedure in itself can activate and elevate the fetal form. It's called hematopoietic stress. And it's important to wait at least 6 months in order to ensure that the stress has come down and the results that you see are truly attributable to the therapy rather than the procedure itself. So that's why we are not yet committed to a rapid disclosure of data until we have had the opportunity to evaluate the circumstances. So we will let you know as the program progresses.

Is it fair to say like sometime, 2023, you will be able to see the data?

Well, it's fair to say that it could be possible sometime later in 2023, but whether we will be willing to do that or not, as I said, depends on the totality of the data. We really want to provide a robust assessment of what we are doing rather than rush into a single patient data if we can.

Okay. And you mentioned that you will start selecting patients in second half this year. Several questions on why it will take so long? And the second question is, given the blue experience of non-transfusion dependent anemia into patients, and how would you be thinking about patient selection?

Yes, I'll come back to the second part. So the first part, the timing is, frankly, a technical timing. They're essentially transplants, that's the therapy. So it requires, first of all, to select centers that have experience of that, and we have already set with several in the U.S. You need to go through ethical committee as well as IRB and contracting out that typically takes about 6 months also just in that process. We then are allowed to start screening the patients. The patients that have been selected will need to undergo at least 1 month period of flushing out hydroxyurea from their system. They're typically chronically treated with hydroxyurea and that's not something that we want to have on the study. So they will have to receive transfusion during that period of time. They will then need to come back to mobilize their cells from the bone marrow in order for that collection of the cells to take place. Typically, they receive at least 2 mobilizations in order to generate sufficient cells. And then at that point, those cells will be manufactured. And eventually, the patient will be conditioned in order to create essentially the niche for the edit itself to be engrafted. And so you can see how this is a pretty long process in order to do that. The good news, though, is that once you start the clinical trial, actually, the potential path to pivotal and licensure is actually very fast. If you look at CRISPR, it will be roughly about 5 years from starting of the clinical path, and we tend to follow similar time lines as well. In terms of your questions around the Bluebird experience, we are not planning currently to genotype for alpha globin. However, we do collect samples at baseline prior to treatment for all the patients. And we will have the opportunity to evaluate whether something, obviously, occurs like the anemia that has been seen in the beta globin Bluebird trial. The other thing to say is that we would exclude any patient that would have had thalassemia prior to the study.

Okay. Another very technical question. When you say the -- you did mention you added both promoters, alpha 1 and alpha 2 -- in the preclinical or the human cell lines you show like edit efficiency like over 80%. Can you give a little more color, will those be monoallelic or biallelic? And you also have 4 sites, and when you say over 80%, under what context?

Yes. Actually, for BEAM-101, typically, we see more than 90%. It's with BEAM-102 that we see about 80% to 85%. We have done single cell studies, although these have some technical challenges, because we need to differentiate the cells. But from those studies, it's clear that more than 90% of the cells actually are quadruply edited. So all 4 loci are changed to the formulation that we want. And then we have about 3% to 5% of cells which have 2 to 3 edits. So they're really very robust and you would expect that once the editor actually goes into the cell, it's capable actually of making the edit on all of the 4 loci as part of that. And then, again, we measure the hemoglobin F activation. Again, in vitro, you need to differentiate and even that procedure itself can cause some changes in the fetal level itself. So it's not the most reliable of a study, but we consistently see very high levels of hemoglobin f activation. In engraftment studies in the mouse, where I would argue it's probably the more relevant and reliable data points in terms of upregulation of F, we see the highest level of upregulation that anybody has reported clinically, typically 60% to 65% upregulation in cells isolated from the bone marrow of mice after 16 weeks of engraftment, which is the only time at which the stem cells eventually now have been repopulated the bone marrow. And the concomitant decline of the hemoglobin S as you reach those very high levels of F is probably equally clinically relevant, and we typically see less than 40%, which is the levels that are seen in sickle trait individuals. These are individuals who are heterozygous and do not have symptoms of sickle cell disease. So we believe that with that approach, we're actually getting almost like 2 for 1 highest level of F, which outcompetes the S protein as well as the concomitant decline of S. And nobody has reported that. Not even the clinical data that have recently been disclosed through the CRISPR study. There is still about 50% to 55% hemoglobin S, which as you know is about the sickle trait individuals.

Okay. So would that highlight the differentiation regarding the promoter area approach, alpha 1 and alpha 2 versus BCLA and BCL11A enhancer approach?

Yes. So I would say, obviously, the highest level of hemoglobin F, as well as the declines in hemoglobin S, we do believe that the higher that level, the better is the clinical outcome. Of course, you can start to see some benefit below that level. But actually, there are many parameters that can hopefully more deeply and more quickly resolve, that ultimately contribute to what is sickle cell disease. As you know, there is a very noticeable and obviously impactful, as far as the quality of life of the individuals is in crisis that the sickle cell patients actually suffer from, they're called the vaso-occlusive crisis. And it looks from the CRISPR data and the Bluebird data that those are relatively easy to fix. In fact, they get about more than 90% improvement over the prior 12 months or so. However, really, these patients are prematurely killed, not by vaso-occlusive crisis, but by the progressive damage to many organs, including the brain and typically the lung and the kidney. And it's this progressive damage that we are looking to at least stop, if not regress, in order to achieve a curative therapy for sickle cell defects. And so as part of our clinical trials, we will monitor many of those parameters that contribute to that progressive organ damage. Typically, this is an inflammatory milieu, which is generated by the fact that the blood cells in sickle cell individuals are lysed very frequently. When we look at the half-life recovery, we will look also at the hyperproliferative consequence of the fact that these cells are essentially lysed very frequently. This will contribute also to aspects of the quality of the blood, like the viscosity, for instance, the prevention and removal of hemolysate. So all of those factors will be able to give us a clinical picture, which is much more complex than just VOC and would hopefully tell us that we are on the way to prevent this progressive organ damage.

So for you to share the initial data, what would be the threshold for the numbers you quoted?

Yes. As I say, we haven't decided. It's probably not even more of the numbers, but it's more the follow-up time. As I mentioned before, we want to make sure that what we have seen, particularly the upregulation of hemoglobin F, is not the consequence of the transplant itself. It's not a consequence of the stress that the cells would undergo as part of the process, but actually it's truly attributable to the therapy. And so it's really more about how many patients we have with a reasonable progress post-transplant that gives us confidence that basically we actually have a therapy that's what is supposed to be done.

And you also have another program, Makassar program, in sickle cell. Maybe can you lay out how do you see the initial program differentiate from, say, CRISPR's approach and have the second program will have a further improvement?

Yes, in the Makassar approach, which we call BEAM-102, we are taking the ability of the base edit and its precision actually to convert from the single combination that's causing sickling, into a naturally occurring normal band called the hemoglobin G-Makassar. And this approach actually is moving slightly behind BEAM-101, but making very good progress. And we love the fact that we are uniquely placed to actually do a clinical study, the upregulation of hemoglobin F, where we do not completely eliminate the sickling form of hemoglobin, but you are reducing it significantly and you are outcompeting it with hemoglobin F versus the actual elimination and the curing of the cells from hemoglobin F itself. In that case, we directly convert an F into a normal form. The plan is to essentially take them both to clinical studies and then, frankly, compare the clinical data between the 2 mechanisms and see which 1 has the better impact on the organ damage progression that, as I said, is our aspiration to generate a treatment for sickle cell disease. And it will probably likely take 2 pivotal trials on 1 of the 2 approaches in order to move forward. The other point to make there is that BEAM-101 with the hemoglobin F upregulation also has the ability to treat beta-thalassemia. And so there may be a differentiation also at the level of the indication that we actually can present.

So I think there is quite some pushback from investors regarding sickle cell disease indication. Is it such a crowded space. You have several candidates ahead of you and they're setting a pretty high bar. And how much do you really think the differentiation -- the additional benefit you can provide in order to have some meaningful market opportunity there?

Yes. As I said, the good news for sickle cell disease patients somewhat is that even though, unfortunately, they do die prematurely, the disease is not an acute killing and therefore, it allows the opportunity for these patients basically to be warehoused and wait for the best possible treatment that they can actually receive. And as I mentioned earlier, we do believe that the highest level of upregulation of F for us will be complete elimination pretty much of the sickling of form of hemoglobin are going to provide the best clinical outcomes for these patients. And so really, the treatment here is about cure and is not just about fixing VOCs without having a positive impact on stopping the progression of the organ damage. So that's why we believe that actually even under the initial paradigm, which will utilize busulfan as a conditioning agent, we have 2 potential very best-in-class sickle cell disease. And we do believe that through that package of clinical information that I mentioned earlier that we will be able to demonstrate the ability of this deeper resolution of sickle cell to have a positive clinical impact on patients. Having said that though, we do think that the patient population of sickle cell disease has the ability to expand very significantly as there is further progress in this treatment paradigm that moves forward. We've recently articulated that we see this treatment paradigm evolving in 3 waves. Wave 1 is the one that we've just mentioned here, which is essentially a transplant, which is aided by busulfan, which unfortunately, even though it is the standard of care, it is a somewhat toxic regimen and leads to sterility. So we and others are actually actively developing improved conditioning regimens that do not cause sterility and can, frankly, increase the patient population that can actually and is willing to undertake this treatment paradigm. And eventually, we are developing and we are innovating significantly in delivery technology including lipid nano-particles that actually are capable of going directly into the bone marrow and the stem cells in the bone marrow, therefore, completely and potentially eliminating the need for transplant, and that would be wave 3. And once we achieve that paradigm, we do think that we have not only now it's against a population that is much, much wider than the current population exists. Typically, right now it's about 10% of the sickle cell patients are the ones which are eligible for the transplant. By going through with 2 and 3, we see a very significant expansion of that patient population. So we can be competitive, both in terms of the best-in-class as well as having potentially a very significant portion of an increasingly large market.

Thank you. We have a few more minutes left. I wanted to touch on in vivo approach, and you do have a corner where you already have a lead indication in targeting PCSK9, and you also identify your own first one, that's a GSD1a going after R83C mutation. So maybe give us a little bit more color that what is the main reason choosing GSD1a specific target? And what is the limiting factor there? Is that the editing efficiency in vivo, that level of the protein that you require to be generated?

The choice of GSD1a is typically driven by the fact that it is a very significant medical need and base editing can essentially correct that point mutation causing the disease back to wild-type. And we have shown in transgenic mouse models a very significant correction of not only at a high level of editing efficiency, in heterozygous mouse we are typically seeing 60% correction, which actually in the liver is getting close to saturating levels of editing, because the bad sites, which is where this editor is targeted is about 60% to 70% of the entire liver volume as part of that. And in terms of -- but obviously, that's not the only 1 that we have. We have 2 other programs. One is actually the correction of the single point mutations in the alpha-1-antitrypsin, E342K mutation, and that's also progressing well in preclinical study, as well as a second mutation for GSD1a, which causes the premature stop codon at position 3479. At Beam, the way we prioritize the portfolio is we are actually prioritizing strategic buckets and then each 1 of those strategic buckets as a resource to move at the speed at the pace that they can, studies that will just continue at a pace. There is a difference between, let's say, alpha-1 and GSD1a. GSD1a has a lower bar in terms of editing efficiency, but needed different transgenic animal models to be able to understand the disease. The biomarkers are more complex. Alpha-1 actually as a biomarker is essentially an easy protein that one can look at readily, so it has a little bit more of an edited efficiency. In both cases, actually, we are achieving that in preclinical animal model. What is the more challenging aspect of correction as opposed to knockout, which is what you simply say is [indiscernible] is really how you predict the clinical starting dose. Of course, the translation between a mouse model versus human is not necessarily straightforward as for a knockout. So really what the preclinical package focused on is creating a PK/PD relationship between mouse, nonhuman primates, and hopefully human that allows us a good prediction on the starting dose, and that's basically what we are in the process of doing.

Okay. So in 2022, you will identify or select a second candidate, liver targeted. Any color you can share? Is ATV a good guess or it would not be A...

It could be either.

Okay.

Its alpha-1 and GSD1a, both of them are moving, frankly, at the same pace. So we're not prepared to say which 1 is going to happen quick. We'll just push them at the speed that they can go. And of course, I would also remind you that we have 2 partnerships, 1 with Pfizer, the other 1 with Apellis also with some liver targets, and those are making good progress as well.

So since you mentioned Pfizer partnership, so how is the progress with muscle and CNS targeted or directed lipid nano-particle profile?

Yes. So we've chosen 3 targets. So we know exactly what the targets are, although they have not been disclosed. I would argue this is up to Pfizer to disclose what the targets are. One is for the liver, one is for CNS, and the other one is for muscle. For the liver, obviously, the LNP technology is already there. For the muscle and the CNS, part of the work is actually to discover the LNPs that can go to those tissues. Thanks to the Guide technology that we acquired last year, we can actually barcode each individual formulation; literally, dose hundreds of this formulation in the same animal at the same time. And then at the end of the study, we can look at the different organs and the barcode will tell us which formulation has actually gone to this tissue. And so we have a rapid way of developing LNP technology that can actually have the bar distribution that we really are after. And that's in part why Pfizer was interested in creating basically this relationship such that you would stimulate further discovery work in CNS and muscle LNP, and that's what we actually do.

Thank you. Thank you, Pino. We are running out of time. Thank you.

Thank you very much.

Okay. Thank you, everyone.