Biogen Inc. (BIIB) Earnings Call Transcript & Summary

Awesome. Thank you, everybody, who has been joining all morning and pinging me with questions and making this a fun event again this year. Really happy to have Al Sandrock on, the head of R&D from Biogen here for this panel, where we're going to talk about a lot of stuff across Biogen's late-stage pipeline and just kind of thematically across neuroscience. I'm going to kick it off for a quick second by disappointing people, and hopefully, this will save my inbox from random aducanumab regulatory questions. And unfortunately, that just can't be the parameter of this discussion and Biogen can't really comment at this point in depth. So. But I do think we're still going to have a great chat because there's a lot of late-stage readouts this year. There's a lot of different things we can talk about in neuron. So with that, Al, thank you again for doing this panel.

And I think I wanted to kind of start with 1 thematic question about CNS, and that's really that -- the hot topic in this space now has been geneticitation and becoming targeted oncology. And you've obviously been through a lot in this area and some successes and setbacks as well. Where kind of are we on the kind of CNS genetic medicine, new wave approaches in this space? How close are we to really kind of rational development and subsets versus are we still kind of in the phase where we're an early target validation for even a lot of things that might look compelling on paper?

Yes, interesting question. I think in some -- in many ways, we're already here. It depends on the disease. And let me talk first about single-gene diseases or at least segments within diseases that are driven by single genes, so-called familial forms versus sort of the more sporadic, if you will, less obviously genetically based. But in single-gene diseases, we're already here. In the case of SMA, for example, I've recently found out that 70% of infants are now identified at birth, based on newborn screening, as to whether or not they have SMA and we also know the copy number at the same time. And the standard of care has moved to treating right after birth based on compelling data from studies such as NURTURE which was -- which not too many people know about because it's not in our label but at least not in detail. But NURTURE is a 25-patient study where 15 of the patients had 2 copies of SMN2, 10 had 3 copies of SMN2. So most of the patients are predicted to have Type 1 SMA, in other words, which is usually fatal by age 2. And in that study, which is now going on to about 5 years after the initiation, all the patients are alive. 90% are walking with or without assistance. So it's a remarkable shift in the treatment of a neurological disease that prior to 2016, we had no treatments whatsoever. So I think it's already happening with single-gene diseases. I think that -- and in the case of diseases like ALS, Alzheimer's disease and Parkinson's, there are people with familial forms of these diseases that are driven largely by single genes. And I think -- so the same thing will apply there. Of course, the timing of when you start treatment is trickier. Can't start right after birth as you do an SMA, but there are markers that are coming that help us determine when to start, when neurodegeneration begins. For example, in ALS, people with SOD1-mediated ALS, it's known now that neurofilaments starts to rise months before first symptoms. And so you can see how that goes. Now in the case of these more complex polygenic diseases, it's not as clear but I think we see evidence that we can start to identify these patients. First, I think you're going to get a polygenic risk score, if you will, based on many of the alleles that we know will determine risk of disease. And since it's many genes, you'll get a risk score out of all the different alleles, sort of added up, if you will. The second is the emergence of blood tests. So at the ADPD meeting 2 weeks ago, I was amazed at how far we've gotten with respect to blood tests. So for example, we now have 2 blood tests for phosphorylated-tau, p217 and p181, both of which are the ROC curves relative to, say, pet imaging, are like -- the area under the curve is like 0.9 or greater, meaning very high sensitivity and specificity. So you don't have to do the pet imaging. You're going to be able to do the blood test to know what's going on in the brain. You add to that some of the a-beta tests that are coming up, the a-beta 42/40 ratio, as well as, as I was mentioning, things like neurofilament that indicate neurodegeneration, I think we're going to be able to identify these patients based on molecular diagnosis, if you will. And I think that's going to be important because the treatments we have are going to be related to the toxic proteins that are being built up in the brain, a-beta, tau, hopefully, 1 day soon, alpha synuclein, [ TDP40 ] and the list goes on and on, right? So I see that coming. I think it's already here in many ways but it's very exciting.

Yes. Yes, that's awesome. And plasma neurofilament, you're preaching to the choir but that's going to be a great -- maybe even a great surrogate endpoint someday that can be big for companies and just small samples in Phase Ib. I guess if we talk a little bit about your ALS, 2 subsets you're pursuing, SOD1 and C9orf72, I guess what's appealing about SMA, right, is the underlying biology is unequivocal, and you've got really high or 100% genetic penetrants and a gene dose effects. For SOD1, C9orf72, is the biology as clear cut here that knocking down these proteins really truly should confer a clinical benefit if you herd in early enough?

Yes. I think for SOD1, it's pretty clear, after decades actually of work, that it's a toxic gain of function. So I think everybody would agree that knocking down the SOD1 expression is the way to go. The antisense that we published last year in the New England Journal that the antisense that we have now in Phase III does actually do that we measured CSF-SOD1. Moreover, we saw effects on other measurements, including neurofilament, by the way, but also some clinical measurements such as the ALS FRS or the Functional Rating Scale, the slow vital capacity as well as muscle strength as measured with a handheld dynamometer. So all -- it looked very promising, very small study, big error bars. And we've been here before with ALS, as you know, with dexpramipexole a few years ago. We had a very promising Phase II trial that we could not reproduce in Phase III. So we have to do the Phase III trial. It's a -- it will read out this year. But -- so I think SOD1, I think it's pretty clear we have to decrease the expression of that mutant gene. In the case of C9orf, it's a little less clear, right? I mean there's 3 potential ways C9orf mutations can cause disease. One is haplo insufficiency. The other is RNA-mediated toxicity and third is dipeptide or protein-mediated toxicity. And our antisense, by the way, was designed to actually accommodate all 3 hypotheses. So as you know, C9orf is due to a hexanucleotide expansion in the noncoding region between axons 1A and 1B. So some of the transcripts contain 1A and some contain 1B. Those that are containing 1B are not going to be affected by our antisense because the antisense is just that stream of the hexanucleotide at peak. And so it will only affect the 1 -- the repeat containing transcripts and preserve the transcripts that contain axon 1B as well as, of course, from the wild-type allele. And so -- and then it will affect -- and you don't have to rely on either RNA-mediated toxicity or protein-mediated toxicity as the key toxicity because the antisense will reduce the RNA and it will also reduce the dipeptide. In fact, what we've measured in the clinic is the -- as a reflection of the reduction in the messenger RNA is the dipeptide levels in spinal fluid. We're not measuring that because we believe that is necessarily the toxic -- the way you get toxicity, but it's certainly a good readout of the expanded repeat containing messenger RNA. So I think that in C9orf, I think we have to be a little cautious about the potential for haplo insufficiency being an issue. But I think -- but I do think the predominant evidence is that there is some gain of function toxicity. And frequently, in these diseases, we have mixtures of gain of function and loss of function type things. So -- but I think antisense is designed to accommodate all 3.

Yes. Okay, that's great context. On tofersen, I guess from your prior experience in ALS and based on this early data, what do you see as the 1 biggest risk to replicating this early setting where you had a pretty big effects on us?

Well, I mean as I said, the Phase II trial, the trial we published in the New England Journal, it was a small study and a lot of it was based on the rapid progressors. And they were talking about very small Ns of like 4 patients. And so as you know, anytime you have small Ns, the confidence intervals are large. And we're cautiously optimistic. But one never knows when you go to Phase III.

Yes, yes. Okay. All right, fair enough. Let's talk about neurocycle a little bit, which is something that is awesome that I can finally talk about now with Biogen. I guess, first, a random one. It's a pretty bold move, stepping in front of a major depressive to start at Phase III. I guess, what gave you guys the confidence to do that? And that in this WATERFALL study, Sage is poised to get it right?

Yes. So listen, I think that we've been tracking this field for quite some time. When you go to neurology meetings, can't help but hear about psychiatry at the same time that you hear about neurology. And of course, in many of these diseases, you have psychiatric type symptoms, agitation psychosis in Alzheimer's disease as well as Parkinson's disease, right? And then in some diseases like Huntington's disease, psychiatric, the behavioral aspects are just as common or just as -- the reason why people go to the doctor as much as the neurological symptom. So -- and look, we know a lot of the people at Sage. They're just down the street. And we've been following what's going on. I think -- so we -- and we greatly admire them. I mean we know them and we admire them. They know the science. When we looked at zuranolone, the key thing was do we understand the MOUNTAIN study and why it was negative on the primary endpoint? And when you look at that study, first of all, at earlier time -- I missed on the day 15, which was the primary point, but if you look at earlier time points, there was actually separation statistically significant from placebo. And the key thing is -- and by the way, it's not uncommon in depression trials. I mean frequently, you have to do 6, 7, 8 trials to get 2 that are positive, so there's a little bit of sort of a roulette phenomenon, if you will, in depression. But I think that the key thing was do we understand why MOUNTAIN was not positive. And I think we do. And most of it relates to exposure, whether it's due to noncompliance or other reasons. We think -- and so we had a very thorough clinical pharmacology review, if you will, of the data. We did our diligence. Moreover, if you think about it, I don't think PPV and NPV are all that different in terms of the final time of pathway. There's been a positive NPV trial and a positive PPV trial with zuranolone, and then it's very similar molecularly to brexanolone, which is an IV drive, as you know, that was already approved for PPV, and that's positive there. So we have this class of drugs, very similar, both GABAA modulators, synaptic and extrasynaptic, and multiple positive trials. And we have 1 negative trial, but we understand -- I believe we understand quite well why that trial was negative on the primary endpoints. So I'm pretty excited. Obviously, we wouldn't have done the deal if we didn't think it was exciting. But I -- so I think the WATERFALL trial with a higher dose, 50 milligrams, takes into account and then also enrolling patients whose baseline score is above 24 I think takes care of some of -- it accommodates the learnings from the MOUNTAIN study. And so I'm very hopeful that it will be positive.

Yes, yes, yes. Okay. Well, I certainly am, too. But 1 thing I do wonder with this drug is about, the PRN dosing approach and treating patients acutely and then looking at placebo-controlled follow-up at these later time points. And maybe, Al, be curious what your kind of views are on this paradigm-shifting dosing approach. And when you look at these trials reading out later this year, how you think about the importance of demonstrating clinical durability? And how exactly that needs to look, if it needs to be shown statistically or qualitatively? It feels like it's a little bit of a complicated question. Do you feel like I'm barking up the right tree? Or is this something that is really more noise and maybe doesn't ultimately matter in the end game?

No. I think it is a paradigm shift. And anytime you have a paradigm shift, you have to worry about whether or not it's going to be adopted by the prescribers, right? I mean -- and we've talked to many of them. We talked to them actually before we did the deal and we continue to talk to them. So I think it's going to -- there's going to need to be a strong, data-driven medical education that we have to do. And by the way, Biogen, we're comfortable with that, right? I mean we had to do that multiple times in the past for multiple sclerosis as well as in arguably SMA. We just talked about newborn screening and paradigm shift in the way we treat, even regard SMA as a treatable disease. So we're comfortable with that. But I think that the data are pretty strong. If you look at the SHORELINE data, where even with 30 milligrams, most patients either needed just 1 single treatment for the year or 1 additional. So 70% of patients required 2 or fewer treatments. I think that's the kind of data that's pretty compelling. And look, the whole concept of as-needed treatment would not be possible with the current treatments, right? You wouldn't do that. When it takes 4 to 6 weeks for a drug to start working again, start to working, you're not going to take somebody off a drug with the risk that, "Oh my God, they have a relapse. You have to treat again." 4 to 6 weeks is a long time for -- to ask the patient to be depressed and it can be dangerous, right? I mean some patients commit suicide. So the rapid onset of action of zuranolone, you get efficacy within 72 hours. I mean you see the p-values. And so that allows for us to think about this concept of as-needed treatment. If you didn't have that, you couldn't think about it, plus the durable effect that you see based on the SHORELINE data. Both of those things allows for one to think about an as-needed. But it is a very different way of thinking and the current treatment for depression is more chronic. But that underscores the unmet need. I mean a lot of people who are on current SSRIs or SNRIs tolerate the side effects. I think up to 40% have sexual problems. There's weight gain. There's all sorts of things that people tolerate because, again, because the drugs are slow to act and they don't have a durable effect necessarily. So I think that there's going to be a strong -- I think the data is -- we need to produce the data and we need to educate the prescriber. And I think -- and then for me, I think having a mechanistic understanding of why you get that durable effect in terms of biology would be very, very helpful as well. And of course, there's some data that suggests that it could affect the trafficking of receptors. But I think -- I suspect that there are more effects that have to do with synaptic strength, either due to post-translational modification of ion channels or receptor expression on the surface of the synaptic membranes, or maybe even -- we know that spines can grow and in response to -- which underlies long-term memory, for example. So the nervous system is designed to be this elastic and have durable effects, if you will. That's what memory is, right? So -- and so I think that it's fascinating biology. I'd love to learn more about how that happens and I think -- because we see it clinically.

Yes. No, it's -- I mean look, it's kind of incredible and ahead of their Phase II study, I thought that what we saw at brexanolone was more driven by the acute nature of postpartum depression rather than something that was biological about the mechanism. But I guess, maybe taking a step back, there's so much to point to in neurodegeneration about genetics and new tools like antisense and gene therapy. And like it still kind of feels like it is what it is, right? Like it's really hard and it's a lot of trial and error. Is that why Biogen has been less active in psych? Or -- and I guess do you look at the Sage deal as kind of a pivot point for you where psych deals that get breaking open scientifically? Or is this more of a one-off and opportunistic type thing?

Yes. It's a very interesting question, Paul. And I'm -- and my colleagues and I at Biogen are really thinking that through right now. First of all, we're learning a lot from our colleagues at Sage. They're experts in psychiatry and -- but is this a turning point in psychiatry? That's a key question you're asking, and I have a feeling it is. And I'm very interested, not only -- I mean -- and I talk to people like Steve, Paul and others who have done this for a long time. It's interesting that rather than genetic -- for me, human genetics is a way of starting with the human disease and working backwards to find validated targets. That's how we've been thinking about neurology, right? In psychiatry, it's human experience. So a lot of these things come from clinical observations made by drug. And sometimes, in psychiatry, you make a drug against a channel or a receptor, and you do the Phase I trial and you don't get the effect you're looking for. You may be testing it as an antiepileptic. But then you notice that there are other psychiatric-related effects. And so you start with that clinical -- the human observation rather than human genetics, it's more of like clinical observations and then you work backwards. And so that's been the history of psychiatry. Now that requires a little bit of serendipity, which is not a great -- I mean relying on serendipity is problematic as a strategy. But on the other hand, I think that there is -- first of all, there is growing genetic understanding of diseases like schizophrenia. There's a lot going on in terms of neuro-developmental changes. And by the way, they can incur pretty late even during the adolescence. And so could we be entering a period where we have human genetics driven more disease-modifying type treatments for psychiatric disease? I think that's possible. And I do think that many of the measurements we can make, for example, quantitative EEG and even imaging, I mean many of these diseases have effects you can see on cortical thinning, for example, by MRI. So I think that the tools are there, if you will. And I think that the targets are there. And so I think it is potentially a very promising area. But of course, at Biogen, we have to prioritize. There are so many promising things going on in the field of neuro but you can't do it all. So the key is like, well, okay, how do we -- what do we choose to do? And what do we choose not to do? And to me, it's about where are the validated targets? How much can we believe based on human validation, as I was talking about? Is there a way to affect the biology the way we want? And luckily, as you pointed out, we have multiple modalities now. We have small molecules. We have antisense, all the new nucleotides and we have proteins. And then we're going to have gene therapy. And by the way, in terms of proteins, our Denali deal, I mean what they just showed with essentially enzyme replacement therapy for the brain in MPS II is pretty exciting. Antibodies, we know, can get in and do things already but improving that would be tremendous. So I think that -- and so I feel like sometimes a kid in a candy store in terms of all the wonderful things one can do. But the key question is what do we choose to do and what do we choose not to do. And that's trickier.

Yes, yes. Okay. Fair enough. Well, let's maybe move on to a couple of the antibody programs you have and talk a little bit about tau. And I guess, one of the main questions with the tau antibody approach, in my mind, has been accessing the target, right? And pathogenic tau is largely intracellular, and can you really inhibit it? And fair or unfair, I kind of felt like the failure in PSP maybe read negatively on to 80, just from a pure mechanistic perspective. Is that unfair? How did you view the read-through to 80? And how are you thinking about the prospects for an antibody as a good tool for the job or not in the TAS phase?

Yes. So we continued our Alzheimer's program because as you pointed out, tau is intracellular as opposed to a-beta, which is extracellular, right, or it's cleaved off of membrane protein that -- and so the epitopes that we're after are extracellular. The tau is intracellular but we were relying on this spread hypothesis. So it's very -- I think it's pretty clear that you can get a prion light spreading mechanism that altered forms of tau can induce the normal tau that misfold and cause toxicity. I think that's pretty clear. I think the spreading hypothesis, if you will, is better in Alzheimer's disease. The evidence for that is better in Alzheimer's where you have -- this is known since the early days of broad staging, right, when you did -- looking at human pathology that there's clearly spreading going on. And now with the advent of pet ligands for tau, we can see spreading in the longitudinal studies in individual patients. So the idea was that there must be an extracellular phase of tau as the misfolded hyperphosphorylated tau spreads from cell to cell and can we catch it as it's spreading. Now catch it, I say it as if it's 1 protein. It's complicated, right? There's 6 isoforms based on alternative splicing of 16 axons of -- the gene. Then there's multiple -- there's multiple protein modifications that are made. There are different confirmations of taus, so-called strains of tau. And then there's proteolytic cleavage products. So it's a gamish, essentially, in the extracellular space. And so -- and each of these antibodies bind to certain subsets of those, that gamish, if you will. And so that's the other complexity. But we felt that it was worth continuing the B92 program in Alzheimer's even in the face of the PSP failure because of spreading hypothesis has better evidence in Alzheimer's. And also, we can look at it better with tau -- we have tau tracers that are working in Alzheimer's and they don't work in PSP yet. So I think for both of -- but I do -- I tell you what makes me worried is the fact that other companies with anti-tau antibodies fail and our own alpha synuclein which has the same concept, alpha synuclein is also largely made of cellular protein. It has -- there's good evidence that it spreads by a prion-like mechanism, and we didn't have a positive result. So I'm worried, but that's why we have our antisense program. We have an antisense with tau in collaboration with Ionis. Because there, you don't have to worry about which isoform, which protein species is going to reduce the expression of tau. And you might think, well, that seems dangerous. But if you look at knockout animals, they're pretty well off. I mean I wouldn't say they're totally normal. But there's redundancy in the system. There are other microtubule associated proteins that seem to increase their expression and perhaps substitute for what tau does normally. So the key question, is it safe to reduce tau? How much can you reduce it? And can you reduce the toxicity related to tau while maintaining -- while preserving some of the key cellular functions associated with tau? We'll find out. We're pretty excited about our antisense. We'll be announcing results of that in the coming months and to a year or so.

Yes, okay. That's great. I want to ask -- can I ask 1 follow-up about the antisense program? And that's the...

Which program?

The antisense program. The antibody data, I think, are coming relatively soon. For the antisense approach, I think 1 fallout question from this setback in Huntington's, right? And look, it's speculation but it's fair, right? That was a great genetic target and it didn't work out. Is really -- is an intrathecally administered ASO going to get good enough biodistribution to treat some of these broader brain diseases? You have a lot of experience with SPINRAZA. How do you think about pursuing Alzheimer's with an antisense drug?

Yes. So we not only have human data in terms of distribution of antisense, because some of those children died in the early days of the SPINRAZA drug development. But we also have a lot of nonhuman primate data and other large species data with respect to distribution. And look, I think we know that antisense can get to high enough levels for spinal cord disease, right? Otherwise, we wouldn't get the effects we see in SMA. Even in older children and in adults, we see effects of SPINRAZA, so. And then we know that if you -- we know that it also affects -- that you get good drug exposure in the cerebral cortex. So the gray matter, the cortical mantle gets very good exposure with the antisense. The area that's most challenged in terms of exposure are the deep gray structures, which is exactly where much of the disease is early on anyway in Huntington's. So that's always been a concern of ours is when you get to diseases that are more related to deep gray structures, caudate, putamen, that you might not get enough drug in there. Or you might have to expose other parts of the brain to very high levels of drug in order to get in. I think the other thing is you have to think about the potency. I mean these -- all the nucleotides vary quite widely in terms of potency, and of course, always therapeutic window as with any drug. So I think that these things are important to consider. But -- so yes, so I think that for cerebral cortex diseases that are in -- so tau is mostly cortical, right? I mean whether we're talking about Alzheimer's disease, or tau-opathies like FTD. And obviously, ALS is a spinal cord but also in people who have FTD, frontal dementia-related symptoms, it's cortical. So I think we're pretty comfortable. It's the subcortical regions that we need to do more work on.

Awesome, okay. Makes sense. So maybe in that vein, on the whole biodistribution and delivery discussion, I want to talk a little bit about this BBB platform deal you did with Denali. And I mean I agree that early data they have in Hunter syndrome is awesome. I guess, as a company that's pursuing targets in neurodegeneration that aren't yet biologically validated, what does this BBB platform get you, right? Because like at the end of the day, you can get an anti-amyloid antibody in the brain and have incredible data on pet, right? Or for tau, right, the question is less about getting a drug into the brain, it's like, does it get into neurons? So when you think about this delivery platform, how much risk does it really mitigate when ultimately, the underlying target biology risk is still a lot of the problem?

Well, I mean listen, I think that we only get 0.1% to 0.5% of antibodies into the brain, and we're lucky that, first of all, that our antibodies are potent and very specific. But also, in the case of aducanumab and other antibodies that we have, it doesn't bind to anything in the periphery, you see? So if you have an antibody, even though it's potent, it binds to antigen in the periphery. It may never see the target. It may never get into the brain because it's already found its target, if you will, in the periphery. So there are -- so the question is, can we improve on that, increase from 0.1% to 0.5% and not rely so much on potency and specificity and maybe we can use other types of proteins? The other thing is, I mean -- I mean we talk about proteins but it's macro molecules in general. Could we use the platform to improve the penetration? We just talked about oligonucleotides and some of the challenges of distribution. Could we use that platform to get better distribution of oligonucleotides into various spaces? And that's something we're very interested on working on as well. So for that -- so I think that it opens the door. But I think that the MPS II data they had was really cool because they took patients who were already on enzyme replacement therapy, and they moved them over to the ETV, right, enzyme transport vehicle. You see no change in the urinary glycosaminoglycan levels heparan sulfate. But you see a 76% reduction in CSF, I think, is pretty -- I mean that's pretty good proof-of-concept as far as I'm concerned. So I congratulate my colleagues at Denali for doing that, and I'm hopeful that we could work together on something really cool for Alzheimer's disease and potentially other diseases.

Yes, yes. That's great. So in the last 10 minutes here, Al, I want to ask kind of 1 question about a few different programs for the sake of completeness. And maybe 1 that's imminent is this Phase III readout from the NightstaRx program. So maybe can you kind of speak to that, some of the early data that gave you the confidence to move into Phase III? And how you sort of think about the clinical setup and the key question heading into this readout?

Yes. So it's a Phase III trial in choroideremia that will read out next. And we did the deal because we were -- small Ns again, but relative to the natural history, that they had a wonderful natural history data set in the U.K. to compare to. And there were patients who had improvement. I mean that's just -- you just don't see that in a natural history improvement by multiple lines. And we know -- not to malign on the visual acuity scale. So I think that -- and we know the regulatory hurdles associated with that, right? So we know the 3 lines of improvement is enough to get regulatory approval. That's a precedent that's been set. So I think that's the key thing is do we hit on the primary endpoint on that. And so it's a single-gene disease. We're replacing the missing protein, if you will. It's subretinal so we get -- as opposed to -- I think gene therapy for CNS is still in early days. One of the key things is, can you get the vector to the right place for the right cells? And then what's the transduction efficiency and things like that? Here, we deliver right to the area that we need it. We know that the protein is important for both motor receptor and pigment epithelial cell function. So I think that we got -- so I'm pretty hopeful, again, that based on the earlier trial, small Ns but relative to natural history, there were patients that improved their vision and still more than stabilize their vision and natural history in this disease is steady worsening. And so -- and we have -- we know we have a registration quality endpoint so we'll see.

Yes, yes. Okay. Great. You also have data from a program in stroke, I think, coming in the first half of this year. It's a huge unmet need, such a tough indication, right? Well, how does this approach, the 007 program fit into kind of some of your prior efforts there? And how would you kind of lay out the case for cautious optimism?

Yes. So stroke is not an area that we've dealt with too much. We did do a trial with TYSABRI, as you know, 2 trials. But we haven't done too much in stroke in the past. What's interesting about TMS-007 is a natural compound derived from a fungus. It actually is a small molecule that binds to plasminogen and changes its confirmation so that endogenous plasma engine activators can be effective, can't have more activity. And so it's different from just giving plasminogen activator, which goes everywhere, right? This will induce a confirmational change so that just at the site of the clot, the plasma-endogenous plasminogen activators can work on it and activate plasminogen. And so that's different. The other thing we're doing in the study is expanding the window of treatment up to 12 hours. That is a big -- right now, one of the reasons why the current fibrinolytic agents are not -- or thrombolytic agents are not being used is the patients don't come in within the 3- to 4-hour window, right? For some reason, I never understood this. People come right away to the hospital when they have chest pain. But they can be aphasic or have weakness of 1 side of their body and not come to the hospital for hours. I never understood that, but that's the way it is. But what's cool is that in the mechanical thrombectomy world, we're learning that you can remove clot up to 24 hours after the stroke and get efficacy on the modified Rankin scale. That is quite a revolution going on here in the sort of device world, if you will. So for me, if it works with mechanical thrombectomy, why wouldn't it work with medical thrombectomy, if you will, or thrombolytic approaches? The other thing about TMS-007 is that also is the -- epoxide hydrolase. And as soon as you open up -- if you relieve the ischemia, you're going to get inflammatory cells. And that was a whole inflammation. That was a whole concept behind TYSABRI, right? Can we reduce the inflammation that occurs in the setting of acute ischemic stroke? TMS-007, theoretically, anyway, has the ability to not only be thrombolytic in a different way from the current treatments but also reduce the inflammation at the site of the ischemia. So I'm looking forward to seeing the data. The other thing is we -- I think we have good MRI measurements. We can look at re-cannibalization now in patients with large vessel ischemic stroke. We can look at modified Rankin scale. And so I think we have better scales. We have better imaging endpoints. I think we have a pretty interesting molecule. And I think the mechanical thrombectomy is teaching us that we can expand the window. So for all those reasons, as you said, cautiously optimistic.

Yes, okay. Sounds pretty cool. All right. 2 more. Your lupus program, kind of totally under the radar. You got 2 approaches there, have some early data for 1 that you advanced into a late-stage study. How does this kind of fit into the whole Biogen portfolio? And how do you kind of think about that -- those 2 assets going forward?

Well, I think Biogen historically is a strong immunology company, right? I mean that's sort of how we got our start in [ the massive ] well. I mean our founders figured out a way to clone interferon, right, I mean interferon alpha and beta. And so that's our history. And so we still have these programs based on our strong immunology research and development. And we have 2 programs, CD40 ligand, anti-CD40 ligand, which is partnered with UCB. That showed, I think, promising data in early phase trial. It didn't hit on the primary endpoint, but boy, if you look at -- but it was a small study. And if you look at the results of that trial, there's very promising data there. As you know, CD40 ligand is an important co-stimulatory -- that CD40 allele and CD40 is an important co-stimulatory pathway. And so it can affect not only cell-mediated immunity but also antibody production, ultimately. So pretty cool pathway and we're in Phase III with UCB. The other program, the anti-BDCA2, is a homegrown program. If you've seen some of the published data and some of the stuff you've seen in meetings, I think you're right, it's under the radar. And I don't know why we're talking about it at CNS day but that's people. But anyway, I think it's a pretty cool program. The data -- so this BDCA2 is on plasmacytoid dendritic cells. These cells get into the tissues and they are like a little type 1 interferon production factory. I mean -- and what we -- when we use our antibody, we can decrease the production of type 1 interferons with this antibody. We showed that very clearly in actually early phase human studies. And we have efficacy now in cutaneous lupus as well as systemic lupus in cutaneous when we looked at the skin scores. And by the way, in an early phase study, there was a very nice correlation between reduction in type 1 interferon expression in skin based on MxA histology with the clearing of the skin. And so very nice sort of PK/PD efficacy correlation, which I always love to see. But also the -- in SLE, we saw an effect on joint count. And so I think it's pretty cool. I think the unmet need in lupus is very high. We have BENLYSTA, but we need more therapies. And so I'm pretty -- I'm excited about our entire pipeline, Paul. So I'm just going to keep saying I'm pretty excited about it and I hope you don't mind.

No, it's all good. Your IR team wanted me to ask about lupus when I forgot to and asked about all the CNS. I'm definitely interested in talking about it. All right, last 1. Tremor. How big of -- like how key was that asset in the Sage deal? It's earlier, it's riskier and I think the obvious question is really therapeutic index in the elderly. What do you think about Sage-324? And this was the Sage deal really largely a zuranolone deal and then we also like this asset? Or am I being kind of too cautious on the prospects with tremor?

Well, we like GABAA modulators, and we were interested in the fact that 324 had a very different pharmacokinetic profile that allowed it -- that the idea was to keep a more even, if you will, drug exposure in plasma. And the early phase trial that they did showed very nice correlation between serum levels of drug and tremor reduction, right? I mean it was like a mirror of each other. And so this drug maintains a very nice -- I think the half-life is like 72 hours. And by the way, it's not just in elderly. ET is the most common movement disorder. I used to treat it myself with drugs like tremolol and a drug called mysoline and old -- these are old drugs that work, but they have side effects. And sometimes, we would also use benzodiazepines like clonazepam. And they all work and also alcohol works, by the way, ethanol. So people will tell you that their tremor goes away when they have a drink. So I think it's pretty clear that there's a role for GABAA in the treatment of a central tremor. It's the most common movement disorder. And we'll see. I think we'll have the readout this year in our ongoing, I guess, it's sort of a Phase IIa trial, if you will, where we can answer that very question, which is, is the side effect, I mean well tolerated? So we're testing more than 1 dose. And is it tolerable and do we get a nice reduction in tremor as measured by clinical measures but also accelerometer measurement? So we'll find out. But I think it's still -- yes, it is earlier than zuranolone and so it has all the caveats associated with an early phase program.

Yes, yes. Okay. Last question, I promise, 1 from an audience member that hit me. And that is about a topic I'm super interested in, the medicine [ neurotholamin ]. And I think the question is really about [ neurotholamin ] is talked about a lot but it seems nonspecific. Is it something that you view as exciting and broadly applicable? Or does it need to be individually validated for every neurology?

Yes. So it is nonspecific. We see it go up in MS. We see it go up in ALS and SMA. It goes up in various neurodegenerative diseases as well. And so it's nonspecific. But in clinical practice, we use a lot of nonspecific measurements. For example, the ESR, the erythrocyte sedimentation rate is nonspecific. It goes up in a variety of inflammatory conditions [indiscernible] if you will. But we still use it clinically as not only in the setting of diagnosis but also in treatment response. So for example, in temporal arteritis, giant cell arteritis, we gauge our treatment based on how much the ESR goes down. And so I think it can still be useful. Look, I think it's very clear, it's helpful in group settings, like for example, if you look at Phase II trials or whatever. But the key question is, is it useful in individual patients? Can we still use it to make decisions in the clinic in individual patients? I don't think we know the answer to that yet. It's still early days. But I think as a drug developer, I'm pretty certain I can use it in group settings and clinical trials that we look at group behaviors. And so -- but look, I think [ neurotholin ] is just the tip of the iceberg. I think there are going to be other CNS proteins that leak out, if you will, into the blood that we're going to be able to measure with super sensitive precise assays. So we're just at the beginning. We're going to have a whole array of proteins we can measure in the blood that will help us understand what's going on in neurological disease, help with diagnosis but also help with treatment decisions ultimately.

Awesome. Well, thank you, Al, so much for taking the time. I really appreciate it. And it was sweet to cover so much ground. So thank you, again, for a very candid discussion. And yes, thanks a lot for everyone joining, and our next panel is at 1 p.m.

Take care, Paul. Thank you. Bye.