GSK plc (GSK) Earnings Call Transcript & Summary

Welcome back. Delighted to introduce our next session today. We have Hal Barron, the Chief Scientific Officer at GSK. We've got Jeff McLaughlin and other members of the GSK Investor Relation team. Pleasure to have you with us, Hal.

So look, I'm looking forward to a broad-ranging conversation. Given GSK is one of the preeminent vaccine manufacturers in the world and given we had news overnight of the suspension of one of your competitors' Phase III trial programs for COVID, as a way of segmenting into your COVID efforts, perhaps you can talk to how you would see the hurdle for safety to COVID. And I guess, general concept, because, obviously, we don't know the details and there's been some information that trial may restart. How many episodes of demyelination associated with a vaccine in a patient without preexisting conditions is enough to kill a program? That would be one question. And then another question would be, what is the standard operating process, let's say, happened within your trial, which obviously is a very different platform that's very well-established and likely lower risk. But what is the practice for addressing these signals when they arise? I'm assuming history examination, MRI, looking through subacute cases. But anything you can add, and then we'll kind of segue to perhaps more relevant of your programs, both on the vaccine side and through the therapeutic monoclonal side.

Yes. Thanks, Andrew. I don't know a ton about the case, so I'm not going to comment on the specifics. But I think the challenge for drug development and vaccine development for that matter is that drugs and vaccines offer a potential benefit with some degree of risk, I mean that's the nature. No drug is perfectly safe. They offer a degree of benefit for given associated risk. The purpose of drug development is to both quantify the benefit, but also quantify the risk and make sure that there's a net benefit to patients. And that net benefit somewhat depends on the available therapies and the disease that they treat. So a disease that is life-threatening, potentially can kill you, has a different profile for a drug or a vaccine in terms of the benefit risk that is needed to make that a viable alternative to essentially nothing. So I think it's a very complicated equation whenever you're doing this. I think for a specific adverse event, because you asked to comment on how we approach that, whenever you see an adverse event, the first thing you ask is, how likely is this to be related to the drug? Natural history, when you're doing large studies, things happen in both the active and placebo arm. They're just things that happen to people every day anyway. So you have to ask yourself, is it likely to be caused by the drug or the vaccine? Of course, it depends on which arm were randomized to. Obviously, an adverse event in the control arm is extremely unlikely to be due to the vaccine of the drug is in the other arm, and that can actually tell you something. So you sometimes need to unblind to figure out which arm it was in. You also look for other evidence that maybe cases similar that might have been missed, might be similar enough that you're starting to see a signal. You ask from a biologic perspective, is it likely be drug related? And I think you also ask things like, is it dose related? Is it there -- are there covariants of patients that might be at risk for them? Is it something seen in the elderlies? Is it seen in -- whatever coverage you look for. It's very hard to figure out, of course, if you have only one case because there's -- it's not a big data set. So those are all the things that go through someone's mind, I think, when you're doing drug development. But now it's important to remember also that, sometimes -- and I think TESARO was a good example of this, sometimes the event that occurs might occur after a long duration of therapy. That's obviously not the case with the vaccine. But in which case pausing a study to see if it's after treatment, say, for 4 years of a drug, and you're starting to see one case, that's very different that if the one case was acute. In other words, you might need time to watch the other patients who are being treated for 2 and 3 years. As they come up to their fourth year, maybe a signal starts to emerge like what we're seeing to some. So lots of different things go into how to interpret an adverse event. And I think all of us in the pharma industry are very, very focused on patient safety and doing the right things for patients. And again, not to comment on the specifics, but I'm quite confident that AZ will handle this thoughtfully and understand what's available to be understood and do the right thing.

And then just in the -- within the context of COVID, where large numbers of well patients, uninfected patients are going to be vaccinated if the drug vaccine reaches the market, what is an accessible -- an acceptable number of events? I mean, obviously, it depends on the magnitude of the benefit. But given the sheer numbers that are being vaccinated, 1 in 8,000, is that a pass to an 8,000? Is that a pass? How do you -- what's the tipping point here?

Yes. It's a great question. And frankly, I don't know the answer. Obviously, has a lot to do with how bad the pandemic you're treating is. It has a lot to do with what available therapies are out there. And just to back up, I think the reason that we took what we'll call the tried and true approach, is that we have a lot of experience with antigens and adjuvants. And so when you ask what safety database do you need, you have a lot more comfort when you're doing something that has been in people for long periods of time with many different approaches in terms of vaccines. But the antigen-adjuvant approach that we're taking, it has a long track record of being safe and effective. And that is something that we use -- we used the term prior or prior for seeing new safety events wasn't 0 because anything is possible. But that gives you a degree of confidence that the tried and true method will be able to have a well appreciated efficacy profile like all of them, but they have a safety profile that makes them more confident. So that's -- that was the underlying driver. But the specific incident, the frequency, it's a very tough thing to put together, and I'm not sure I can comment.

And then maybe just specifically on your vaccine, could you talk to the importance of CD8 levels expressly, which are not typically found with protein subunit vaccines that can be elicited with certain adjuvants? Could you talk to them? Shame on me, because I haven't done the cross-trial comparison with some of the oligonucleotides or the viral vector based vaccines. But can you say what you've seen and what you think are the clinical indications of that?

Well, with COVID, I don't think we have enough data to answer that. It's an important question. And what we're really talking about, I think, is the contribution that tumor immunity plays versus cellular immunity. And I think in some vaccines and some diseases, tumor immunity is all you need. And others, it's absolutely not accurate. And as you point out, the vaccine -- when an adjuvant is used in a vaccine, you oftentimes induce some more robust tumor, but oftentimes, depending on the antigen actually, you can induce some more T cell immune response as well. And that can affect the duration of response, that can affect the magnitude of the response. And we have an enormous amount of experience with adjuvants and these biomarkers, if you will, these cellular assays that give us confidence in preclinical models as to what is needed and why. And I think that's an underappreciated area of science and one that we think we're leaders in, both from the assay development perspective, but also an understanding what these assays mean in terms of the ideal dose, the ideal adjuvant, and how to think about tumor versus cellular response. We've used this to our advantage to create better vaccines in the past. And I'm hopeful that we'll have data from our trials that will give us comfort in our program. But it's a great question, and I think the jury is out on the need for cellular. My guess is it will be important, but we'll have to wait and see. And I think if it is, we're well positioned. We also have an antibody program. So I don't want to speak out of 2 sides of my mouth, but I think that there's, I think, a very strong possibility that neutralizing antibodies could actually have a protective effect if given proactively. They might actually have a protective effect if given in the early stages of infection in the pre-hospitalized setting to prevent supply such as hospitalization or even worse, respiratory failure or death. And we also know that there's -- and so we're very excited about the collaboration with Gilead. This antibody, I think it's really important to note it's bound from looking at convalescent serum from COVID-1, SARS patients and hundreds and hundreds of antibodies were screened, and we were looking for one that had a very [indiscernible], so we could visualize where you probe the antibody, but what also that was binding in Hepatol that was found on SARS COVID-1 as well as COVID-2, COVID-19. And we thought that would be important because over that period of time, there had been no genetic drift away from an epitope, but that probably suggests that the epitope does not have -- has a disadvantage probably, and therefore, would be less likely to mutate in that spot. Therefore, as the virus starts drifting or mutating looking for fitness advantages, that we would be less likely to be susceptible to being inactivated by those mutations. And so far, that preclinically is borne out with many of the mutations that we've observed. So we're very excited about the antibody. It's also very potent in the dosing use. It's extremely model relative to some of the competitors who are not only given high brand like doses, but also some who are actually feeling the need to put 2 antibodies together to avoid this resistance problem and to optimize potency. And also, we've engineered in a mutation in the antibody to live a long half life. So this lower dose, longer half life, highly potent antibody that is potentially recognizing an antigen less likely to drift, we think it's very unique. In addition, it also has -- this antibody has pretty significant effector function. And this is where the cellular immunity pieces may come in. They turn out to be that the Fc portion of these antibodies can start to have impacts with the cellular response. So how they induce their effect, it's not really fully understood, but this antibody clearly is unique in its factor function, which kind of gets to your earlier point. So a very exciting program with Gilead. And there's even some interesting backup stuff I won't talk about yet that leverages in understanding the CBA response.

And all that engineering hasn't translated into neutralizing antidrug antibodies?

No. These are pretty tried and true mutations that extend half life that have been studied in the clinic, and we think that's neutralizing antibodies as rural as possible where we think that's real likely.

Okay. And then the other question is, if it's as successful as you hope it is, obviously, it begs a question of manufacturing given the demand. Roche is obviously committed to Regeneron until Regeneron shows itself not to be effective, and they have access to a big slug of manufacturing. There are other sites. But where are you in your scale-up for manufacturing?

We have a very robust approach and a lot of manufacturing capacity, which was committed to at risk to enable, exactly what you said, to supply, given what might end up being high demand. I think it's also really important. And again, this is where -- it's not just the number of patients treated and the dose really matters, right? So if you're studying 500 milligrams and somebody's studying 5 grams, so that's 10x more manufacturing capacity needed. So I think dose is really important. And I think soon people are going to start looking at that more carefully and realizing the implications of a very potent molecule and what that means. It's rarely the case that you ever have more need than the manufacturing capacity. But when you start thinking about the problem, and again, I'm very hopeful on lots of levels that nobody needs this thing. I hope the vaccines are perfect. But if that's not the case. It's actually a pretty staggering amount of drug that's going to be needed. I mean some calculations when we did, you're looking at metric tons of antibodies, which is going to be quite significant. That's why we're very hopeful that we're not being on the antibody as well because I think the world is going to need a lot of drug, should antibodies work and be needed, I don't think the entire demand is going to be met by one company. I think it's going to have to be considered as sub-collab. There's multiple trials ongoing on the antibody group. I'm glad that we have others at low dose, but -- lower dose, but we'll have to wait to see all this of these different data points as the emergency, what is the real unmet need and what antibodies will fill. I'm also excited about the antibody program, the GM-CSF, where the so-called cytokine release syndrome, this macrophage activation syndrome. We were pretty confident that cytokines were playing a role in this and that modulating would work with steroid, I think, it supports that, but the anti-IL-6 data probably suggests that inhibiting just one cytokine is not enough, particularly IL-6. And so what we're excited about is that since GM-CSF really sort of suppresses macrophage monocyte activation and lowers the cytokine levels for IL-6, but also [indiscernible] IL-1 beta, et cetera, that this could actually end up being potentially effective therapy and more specific than steroids, but less focused than IL-6 condition.

So I will come back to infectious disease, but maybe just a segue for a bit on to oncology where you're probably the most active in late-stage in terms of business development, both with TESARO, but then with the Merck KGaA as well. Intra-fast gated decision, we're expecting sometime this year. And what I'm trying to understand is, assuming that you -- that the hurdles that you have set are met and surpassed and you expand the trial, I'm assuming that given the sites are open, you'll be able to complete that trial incredibly fast because accrual of the additional patients will be required is going to happen very, very quickly, particularly with -- if there's evidence of treatment effect from the nondisclosed GSK commentary or Merck commentary. But then your readiness to expand in combinations in potential registration trials beyond that, and obviously, in non-small cell, the money is at the combination of PD-1 and chemo. And so are you ready, if the data is supportive, to start dosing in a Phase III trial Bintrafusp on top of chemo? Or do you need to do a Phase II trial to look at safety tolerability dosing? Others have gone straight ahead into Phase III without that, but I'm interested. And then the final question is, obviously, there are -- Merck is exploring [ PARP's ] maintenance therapy post PD-1 chemo tablets. And you have PARP. You've spoken about maintenance settings. So there's lots of potential avenues to go. And I'm just interested, your stage of readiness, how quickly you can move forward if the data is supportive?

Yes. I think if the data was very robust, we'll be very ready for a really quite aggressive program. I mean if you think about it, we have the Z1 program. We -- the idea of using PARP as maintenance in lung cancer with the idea of using potentially Bintrafusp in lung cancer, the idea that we have dostarlimab as a PD-1 that could potentially be introduced. There's a number of ICOS could go into lung cancer. We have lots of different IO and synthetical legal approaches. And if you look at what's common across all of them is I think there's the potential for activity in lung cancer. So we have lots of goals in lung cancer, very different purchase, all could be complementary. Some might even be synergistic. And we've got this all mapped out in terms of how data points will read out and what to do with each. And frankly, even micrograms we might not pursue elsewhere in the pipeline to afford this. So I would say we're very ready. Obviously, data reads out for all these programs at different times in different ways, and we have to make sure that combinations are safe. But one of the things we really liked about Bintrafusp data set was unlike many programs that initiate a sort of Phase II/III program. 400 patients worth of safety data gives you a lot of confidence that you're probably not going to run into something that was unexpected. You never know when you do new combinations with chemo. But I think we're ready to move quickly should we see activity of the magnitude that you were suggesting.

So you can initiate a Phase III trial?

Yes, absolutely. Yes.

Yes. And the mucosal bleeding would not be a concern in the TGF mediated minor bleeding in combination with chemo is not.

Well, we'd have to look at the entire package to make sure that we thought that was the case. But is there a scenario? For sure, there's a scenario we could be very aggressive. There's obviously other scenarios where we couldn't. But I think it is -- what you said was in the optimal state, would they be ready for an aggressive program? The answer is absolutely, yes.

Got it. Belantamab, which -- congratulations on the approval. The market is circumspect about the commercial possibility, understandably, I think given the current restrictions and current labeling. But obviously, that can change with time as the drug moves upstream in head to head trials versus active comparators as you seek to address the adverse events through dose fractionation or GSI. What's your commitment to belantamab in a very, very dynamic and competitive segment? And when do we have readouts? And you've got a number of programs ongoing. You've got the GSA combination going. You've got the PD-1 combination ongoing. I'm sure you're going to be looking at dose fractionation and low dosages. When are we going to get some of this incremental data that can help inform us, at least from your side, if not from the competitor side, as to where this drug fits in?

Yes. Well, let me just start by saying we're extremely committed to the drug. We think this is a very active drug, a very important class. BCMA is probably the most important target in myeloma today. We believe that for lots of reasons, this is the best approach for hitting BCMA in terms of its activity, it's -- the antibody drug conjugate. The immune activation could end up being very synergistic with PD-1. So we think that, that could be an opportunity. But as you point out, the challenge is the ocular toxin. We think that this is -- more data is needed, but we think that moving more approximately is going to allow us to do things, as you say, like looking at lower doses, looking at different schedules. I'm optimistic, but cautiously so that the gamma secretase inhibitor combinations by preventing and clipping of BCMA plasma cell. If you think about how that will add value, it's particularly useful for an antibody or a conjugate where the antibody binding soluble BCMA, that's probably not the ideal thing to have happened. We would want the BCMA to be internalized to plasma cell. So the more you can keep the BCMA on the plasma cell itself that you probably will get not only more efficacy, and therefore, be able to lower the dose. But you might actually even have less toxicity because you avoid the systemic mission. I do think there's opportunities to lower dose and to change schedules. We took a very -- I think it was a smart risk to take in terms of going very fast. I mean going from 30 patients in Phase I to a pivotal study and approval in almost record time didn't allow us to do all the dose reading that one might want to do in terms of schedule. And to some extent, we're a little surprised that 3.4 and 2.5 were identical almost in terms of efficacy and 2.5 is probably a little safer. We don't completely know what -- how much lower you could have gone and certainly a different schedule. We don't know how much synergy there will be with other concomitant medications that we won approximately. So when you think about all the possibilities that this program affords in terms of ways to optimize the [indiscernible] toxin, I'm optimistic. I don't know how long it will take to figure this all out. That's the question. It may be simple. In which case, we'll find out sooner. It may be more complicated. It may require both the gamma secretase and a lower dose and various combinations. I think the other thing that I would just highlight is what we learned in DREAMM-2, and this is very important. I don't think it came out as clearly as it could up to the community, but was highlighted by some of the germ people at the advisory committee. The [indiscernible] that develops affect the corneal epithelial cells. And the reason it gets better is because the corneal epithelial cells turn over, over time. They rejuvenate, even in the elderly. And the half-life of the rejuvenation is well-characterized in ophthalmology. And in fact, people predict pretty confidently for other prototypes when the symptoms will result, and they usually result pretty classically around the time that one would expect from the rejuvenation of the corneal epithelial cells. What's interesting about the [indiscernible] program in the later lines of therapy is, because of how sick the patients are, you don't have that kind of follow-up time, they would want because people are very sick and they die. And what we have afforded is the opportunity when we move into earlier lines is to see what is the natural history after you stop the drug when we have 6, 12 months or whatever, a longer period of time to actually wash out the drug and see the [indiscernible] result of it. And so the natural history, I guess, is what I'm trying to say, of this toxicity may be better than people think. We don't know that for sure. But I can't say that there was a trend dependent improvement. And I don't think we gave the patients -- we didn't have enough data on the patient's recovery time to know forwarding how well this resolves. So I think dose reductions, schedule reductions, gamma secretase inhibition, ability to be potentially additive, if not synergistic on other concurrent, all made me think we can find a window of that, plus longer follow-up. I'm optimistic we'll find a way to move forward in the earlier lines, given how profoundly active this drug is and how excited the field is about the target.

And then on another one of your platforms, but doesn't get much airtime, but obviously, GSK has invested significant CapEx over the years, is in cell therapy. And you're on the Board of Juno from memory. So you have lived through the early days. And I know that you're instrumental in the Lyle deal. The initial [indiscernible] so I think the response rate was lower than one might have hoped, partly due to the admission of the population for patients who were likely to respond in antigen expression. But there are moves that regulators making in order to accelerate development of these drugs, and one of the initiatives is the parent-child IND, which enables sponsors to explore multiple iterations of a cell therapy in N1IND. And I think GSK is the earliest adopter of that. So perhaps you could talk to that and just sketch out some time lines, because if you -- once you identify a patient population, if you see a response, you know have a cell therapy because of the durability, and you've seen that with [indiscernible]. But so it's not on investors' radars right now. Solid tumors have not so far been shown to be responsive to cell therapy. But cyclones in melanoma, but there's -- I'm optimistic things will change. So perhaps you could just share with us your thoughts.

Yes. No. Thank you. It is a very exciting technology. The -- I think there's no doubt that cells will be therapies, they already are. And the chance that we've peaked out with our understanding and our opportunity for this is very well. There's going to be advances that make it more effective. I think that the likelihood it works in solid tumors is higher than most. Very, very tiny numbers of patients, but the adaptimmune data recently suggesting activity in solid tumors is directionally consistent with what I just said. I think the question really boils down to, who's going to be able to make the commitment? Because it is pretty profound from a capital expense to make the manufacturing available, to be putting technologies to work to advance this, and at the same time, understand the biology. And we thought that with our deep understanding of what it takes from a manufacturing perspective, a lot of understanding biology, actually pretty profound data with NYE cell and the emerging data on the tills, but also with the adaptimmune solid tumor data, that this was worth being the disruptive technology we focus on. Now there are 2 things you mentioned a while ago. The reason we're really excited about the Lyle deal is that we believe, and I think it's consistent with their approach, that's the biggest liability of cell therapy in terms of working in solid tumors is understanding why they exhaust. Why did these cells seem to be very liquid different? But when they get in the tumor microenvironment, some process occurs that basically exhausts them and makes them inactive. And we can understand that, we have the opportunity because there's a lot of opportunity for ex vivo manipulation of these cells. And we do start to prevent exhaustion might that be transformed. And that is really the focus of the a lot of collaboration, and we're making a lot of headway. I won't go into the specifics. But if that was to be the case, this would be a very transformative because it's possible that, by doing that, the dose is lower. There's lots of second order implications of making this more active that could reduce some of the other liabilities, which is being freezes and costs, et cetera, to make this as a much, much more viable approach. The importance of the parent-child IND approach is that if each time you made one tweak to the manufacturing process or which genes you're knocking out or adding in was a complete new IND and you had to restart. The speed with which you could do these small experiments in humans is limited, but the FDA working with us and Lyle and others has created this path forward, we call them parent-child IND, and these things that are evolving, if there's just more, I won't call them subtle changes, but minor changes to the process that allow you to explore various different components of what the impact of knocking off the T cell receptors or changing the various things. You can do that in a very streamlined way. You can end up learning in the clinic in a way that I think has been unparalleled in cell therapy. And I think it's one of the biggest reasons we have even gate even more advances? So I'm excited about the technology and excited about what's been observed in the world. I'm really excited about what Lyle and GSK can do together because I do believe that exhaustion may be one of the more important components here to solving what may be a big problem that could result in a very transformative and effective cell therapy on cancer patients.

And so obviously, enhancing efficacy is one side of it as well as minimizing safety issues, which perhaps are more worrying with CAR T than they are with TCRS. The other thing is the financial constraints, and you kind of alluded to that. And a few years ago, I think Axle put together the [indiscernible] deal with [indiscernible] IMAX, the closed manufacturing system, which seems to play in that with transfection inside the bone marrow transplant center itself with rapidly accelerated time lines and presumably cost advantages. So how far are we from that becoming a reality? I know that's only one system. You may well have others you've engaged. Is this 5, 10 years down the track? Or is this something that you're working to somewhat nearer?

We -- look, obviously it's going to be data-driven. But we not only have that process, but we have a very deep scientific effort on developing stable cell lines that we think could transform costs to some extent. I think -- I really do believe, and I just want to reiterate, that if we were able to figure out which cell subset is actually responsible for the effectiveness of cell therapy, it's quite possible that a lot of the problems, both lower efficacy could be improved or lack of efficacy in solid tumors could be improved, I think the safety profile gets better because some of the safety profile may be driven by, I'll call them bystander T cells that are just infused that aren't related to efficacy, but we don't know that yet, and they may be old T cells or cells that are just more likely to pump out cytokines. I think being able to understand what's driving the effectiveness might actually reduce its toxicity. And again, I think if you can find the cells and make sure that they don't become exhausted, you may find a plasmapheresis process to be a lot more efficient. The vein-to-vein time is more efficient. The cost goes down because the number of cells being injected goes down. I think all of this technology I'm referring to is related in the sense that making a dent in exhaustion could actually help in so many different ways. That's why our focus is really trying to figure out how to infuse these T cells in a very efficient way.

So you've got, I think, a parent-child IND trial being initiated shortly with multiple tweaks against [ MI ]? So is that post control yet or...

I don't know.