Alector, Inc. (ALEC) Earnings Call Transcript & Summary

Hello, everyone. I'm Pete Stavropoulos, biotech analyst at Cantor with another webinar on the blood-brain barrier crossing technologies. It's an exciting time for the field. The blood-brain barrier has been a bottleneck in CNS drug development, preventing many promising therapeutics from reaching their targets in the brain. However, a wave of delivery platforms designed to actively shuttle large therapeutics across the barrier is gaining momentum. These platforms are already demonstrating potential to dramatically expand the reach of CNS drug development, and they will lock a lot of value from neuro indications. We're excited to have with us Alector, who has developed the Alector Brain Carrier, a differentiated blood-brain barrier crossing platform, developing a pipeline of brain penetrant therapeutics. With us, we have Arnon, the CEO; Giacomo the CMO; Neil, CFO/CBO; and Eric Brown, Director of Antibody Discovery and Protein Engineering. So with that, please give a brief introduction of yourselves as well as a snapshot of where Alector is at the current moment.

So thank you for hosting us, Pete, and welcome, everyone. Alector was created with the mission to eradicate neurodegenerative disorders that constitute about 10% of the general population and over 25% of individuals over 65. So we are focusing on large diseases like Alzheimer's disease, Parkinson's disease, Lewy body dementia as well as small diseases. And we are integrating in-depth biology to understand genetically validated target, and we are developing technology, as Pete mentioned, blood-brain barrier technology specifically to develop drugs. At this point, we have 5 drugs that we are advancing either in the clinic or advancing to the clinic. We have a Phase II drug, which is an antibody that elevates progranulin. Progranulin is a risk gene for Alzheimer's disease. So there is significant genetic validation. This is a Phase II drug with interim analysis will be read in the next few weeks. It's a placebo-controlled double-blind trial with a completely novel mechanism of action. We have a brain-enabled anti-A-beta antibody that we are going to discuss in more detail during this seminar. We have brain-enabled Tau siRNA program. We have brain-enabled alpha-synuclein siRNA program, and we have brain-enabled GCase enzyme replacement therapy for Parkinson's disease and Lewy body dementia. So most of these therapeutics were not enabled until we developed the blood-brain barrier technology. And our blood and barrier technology is tailored for different drug modalities. And as you see here, we are developing antibodies, siRNA and enzyme which are enabled by our [ blood and brain ] technology. And the integration of the blood and brain technology and our sort of insightful drug, we think will lead to really novel breakthrough in neuro-degeneration.

I have a brief introduction by everybody in terms of backgrounds and so forth. Arnon, I want you start, then we can go to Giacomo.

Yes. So again, I'm Arnon Rosenthal. I'm the CEO and Co-Founder, was at Genentech for 16 years, then founded Rinat Neuroscience that developed AJOVY, the migraine drug that is now marketed by Teva, then co-founded together with Ben Barres, Annexon Therapeutics that's now is in Phase III with complement therapeutics for Guillain-Barre syndrome and geographic atrophy. And for the last 10, 12 years, Alector is my life.

Hello. I'm Giacomo Salvadore. I'm the Chief Medical Officer at Alector. I've been with the company for over 3 years. I've been in pharma for more than 15 years between Johnson & Johnson and Acadia and then Alector more recently. Prior to that, I was at the NIMH for 5 years, and I'm a psychiatrist by training.

I'm Eric Brown. I've been at Alector for 9.5 years now. And really the bulk of my time here has been spent the last 7 or 8 years building up the ABC platform.

And Neil Berkley, as mentioned, CFO and Chief Business Officer, been doing corporate development, business development, strategic financing, strategic marketing and things for about 25 years. Been started my career in very small companies, a couple of companies I helped found and ultimately exited and worked in companies such as GSK and other large companies as well. So kind of a broad breadth of experience across small and large pharma.

All right. Thanks for that. Arnon, you've been around for a while in neuro. Eric, you just mentioned 9.5 years at Alector. So you've also been involved in drug development for neurological indications for years. How do you think about the blood-brain barrier and its role in CNS drug development? How has this barrier historically limited the development of therapies for neurological diseases?

So the blood-brain barrier largely prevents the entry of large molecules to the brain. So if you look at antibodies, only 0.1% of the peripherally delivered antibodies can enter the brain. If you look at nucleic acid therapeutics or enzymes which have shorter half-life than antibodies, they do not enter the brain at all if you inject the delivery. So the blood-brain barrier was a significant impediment to drug development in neurodegeneration and neurology in general. And I think once we are able, as you mentioned earlier on, once we are able to overcome the blood-brain barrier, we will be able to significantly expand the type of drug modalities that we can use to include nucleic acid therapeutics, enzyme replacement therapy for multiple lysosomal disorders that affect the brain and even antibody therapeutics that without the blood-brain barrier technology enter very poorly to the brain and do not distribute well to deep brain regions. So I think that the blood-brain barrier technology could and would revolutionize sort of therapeutics for brain disorders.

Yes. Can you just briefly walk us through the mechanisms, the brain actually uses to transport molecules across the blood-brain barrier?

Sure. I mean this slide highlights it quite well. There is a variety of passive diffusion mechanisms, but those are only available for gases and very small molecules. Then there are solute carriers that move things like sugars and amino acids. And these are proteins that are on both sides of the barrier that facilitate passive diffusion. And then you have the type that we utilize for blood-brain barrier transport, which is receptor-mediated transcytosis, which transports larger cargoes such as the iron carrying protein transferrin. And in this case, a receptor at the surface of the blood-brain barrier actually picks up a cargo on the blood side, translocates it physically through the barrier and then releases it on the other side.

All right. Eric, you spent many years working on antibody discovery and protein engineering to basically enable biologics across the blood-brain barrier. And Arnon, you also had a major role in thinking about the ABC platform, Alector's platform for crossing the barrier. Can you just give us a quick overview, and we're going to get into detail shortly of how you and others are sort of exploiting receptor-mediated transcytosis to enable efficient delivery of these large drugs into the CNS?

Yes. Really, the key innovation here is to take the blood-brain barrier itself, the endothelial cells and turn them from a barrier into a highway into the brain. And the way we do this safely is not by just breaking the barrier open and allowing nonspecific transfer, but by utilizing these natural and very specific transfer mechanisms such as through the transferrin receptor mentioned above. And this allows us because the brain is so highly vascularized, if you can actually translocate through the capillary network, then you're now able to deliver to every part of the brain, even the deeper brain regions that are minimally accessible to other types of brain delivery such as IT.

So what specific elements and sort of behavior of this crossing technology that we've seen to date sort of enable improved brain distribution versus conventional antibodies?

Yes. I mean I think the biggest part is really getting that broad bio-distribution. And again, so this relies on identifying a receptor that's not just highly expressed on brain endothelial cells, but on the capillaries itself. The capillaries are the smallest type of blood vessels in the brain that really vascularize the entire neuronal network. So it's basically said that like every neuron in the brain is fed by its own capillary. So if you can enable transport through these specific cells and you're really delivering drug directly to the target cells throughout the brain. And what you can do with this is you can take a drug and the canonical example of this would be gantenerumab, right, which was ineffective without a blood-brain barrier technology attached to it, put the blood-brain barrier technology onto it. And suddenly, you have trontinemab, which is a very effective seeming drug in the clinic.

All right. So historically, neuroscience drug failures have been attributed to both target and biology limitations and challenges with CNS delivery. So how do you view sort of that balance between those 2 factors? And do you see new delivery technologies meaningfully shifting sort of the odds of success sort of going forward?

So traditionally, delivering antibodies to the brain has always been a challenge with less than 0.1% entering in the CNS. And the use of blood-brain barrier shuttles can help us testing validated targets and enhancing delivery in a different way and increase the probability of success. Eric just mentioned the example of gantenerumab and trontineumab. Gantenerumab, it's a drug that showed 55- to 65-centiloid reduction measured with amyloid PET. And with the use of blood -- brain shuttle, trontineumab was able to induce a reduction of 84- to 99-centiloids as early as day 78 to day 196, while the gantenerumab studies were fairly wrong. They were 2 years duration. What's even more striking is the percentage of subjects who become amyloid negative after treatment because gantenerumab was -- with 25%, 30% of the patients becoming amyloid negative. Trontinemab, which is gantenerumab with a brain shuttle, it may increase this number to 91% in 6 months. So this is a perfect example about how the use of brain shuttle can turn the fate of a drug that has failed earlier because of poor brain penetration.

So antibodies still work somewhat without the blood and brain technology, although as Giacomo mentioned, they do it poorly, but other types of drug modalities like lysosomal enzymes do not enter the brain at all if you inject them peripherally. Nucleic acid does not enter -- do not enter the brain at all if you inject it peripherally. So in these cases, [ blood and brain ] technology clearly transform the therapeutic potential. And if you inject either enzymes or nucleic acid to the brain directly, it's a medical procedure. The drug do not distribute to the brain evenly. So in the case of drug modalities beside antibodies, the blood-brain barrier technology is clearly the main solution to the drug and the potential of therapeutics.

All right. So we see an approval here in the U.S., one in Japan for the -- for a drug that has this technology. But what do you see overall as sort of the next main target, which we can all guess what it is, but what is also sort of the low-hanging fruit?

Yes. So the low-hanging fruit, I think, are, again, anti-A-beta therapeutics that show some efficacy without blood-brain barrier technology, but significantly more sort of more rapid and more profound benefit with the blood-brain barrier technology and importantly, elimination of the main adverse effects of anti-A-beta therapeutics, the ARIA that's associated with blood-brain barrier technology. So I think the lowest hanging fruit is anti-A-beta antibodies, and there is multiple companies are developing anti-A-beta therapeutics with blood-brain barrier technologies. I think the second low-hanging fruit are siRNA against the hallmark misfolded proteins of neurodegenerations like tau and alpha-synuclein. Both of these proteins, the pathology is intracellular. So antibodies may not be as effective and the better way to counteract these misfolded proteins is with siRNA -- and again, if you can peripherally deliver nucleic acid therapeutics and the nucleic acid therapy will reach the brain. It's a lot more convenient that the drug distribution is much better with peripheral delivery. It's scalable. It's much safer. So I think that going after misfolded -- intracellular misfolded proteins would be the second sort of lowest hanging fruit. And then lysosomal enzymes, there are over 50 lysosomal enzyme diseases. Many of them affected the brain. For example, GCase is one of the major risk genes for Parkinson's disease and Lewy body dementia, it's a lysosomal enzyme that -- whose loss of function leads to, again, lysosomal pathology and the disease. And until now, it was not possible to develop enzyme replacement therapy for lysosomal enzymes that affect the brain. And now with this brain shuttle, we are able to do this.

Right. So there are a number of active transport technologies that are emerging to shuttle these biologics across the blood-brain barrier. What are the key design principles that you believe will sort of determine whether these platforms succeed?

So I think there are several key principles that we look at. I mean, as we discussed, having the right receptor that has a strong expression on brain capillaries is important to get that widespread biodistribution. And then optimizing the affinity, which is the binding strength of your binding domain to -- the receptor is very important and not just the affinity, but actually the kinetics we found optimizing how fast it binds on and off because really, this is an event where you need to bind on at one site on the blood-brain barrier and then be able to release on the other end to get it into the brain parenchyma. And importantly, what we've noticed with our studies across multiple different cargoes is the optimal affinity you need differs quite significantly between different types of cargoes. But to your question about how translatable the platform can be within certain classes of molecules and especially classes that are chemically similar to each other, such as siRNAs, we have found that it is starting to be more and more plug and play where if you have a carrier that's really optimized for one type of cargo like an siRNA, it translates very well to another siRNA.

Okay. But overall, when you think about from antibody to siRNA or to enzyme, is it still sort of a plug and play or...

It's less plug and play. The enzymes in the siRNA have an important second mechanism where the carrier has to get it across the blood-brain barrier, but then it is also responsible for driving intracellular localization into a cell type in the brain such as neurons. And so this often means you need like a somewhat higher affinity. You still need a balance, right? It still needs to be able to go across the blood-brain barrier and release, but then it also needs to drive entry into the neuronal cells and into specifically the endolysosomal system. So there's definitely a different optimal affinity and optimal format and characteristics needed for these different types of cargoes.

All right. I mean though you just sort of did touch on affinity, but how should we be thinking about the relationship between the receptor binding affinity and valency at the blood-brain barrier and sort of the efficacy of transcytosis into the brain?

Yes. I mean there's been a lot of studies on this, identifying that there's sort of an optimal maybe moderate affinity range where if you have too low affinity, you're not getting enough engagement at the barrier, you're not getting enough transport. If you have too high of an affinity, it's not coming off. It's doing things like driving receptor degradation. And since you mentioned valency, often bivalent binders also tend to drive this degradative phenotype, which is a major safety risk in and of itself. I think the key thing that we've identified as we've undergone moving multiple studies from rodent models into the non-human primate system is that this optimal affinity window exists in most systems, but it's different. And that's where there's been a lot of, I think, translatability gaps in our internal programs and within the field in sort of having to refine-tune the affinity as you move into non-human primates, which is why we really think it's very important to focus on non-human primate data to sort of validate that your platform works.

So the technology is largely plug and play, but you need to have a toolbox like a range of transferrin affinity that you can optimize for the drug modality for the requirement of whether you want an active or an inactive effector function for the half-life of the drug. So it's plug and play, but you still want to have a toolbox to really optimize the shuttle technology to the drug modality and the specific requirement of the target.

Do you think different receptors have different sort of optimal affinity windows based on their sort of intracellular trafficking?

For sure. I mean the affinity window is definitely going to be based on how fast the carrier itself translates through the blood-brain barrier and sort of the requirement for the antibody to be able to stay bound for that amount of time. Definitely, yes.

So Arnon, you just said that non-human primate data are important. How well does sort of the affinity exposure relationship observed in rodents model sort of translate to non-human primates and then ultimately to humans?

I mean I would definitely have to say not well between rodents and non-human primates. Like without getting too much into detail, we've taken multiple molecules forward that worked perfectly well in rodents, moved them into non-human primates and the performance was substantially worse in terms of brain uptake. And this is where really we've had to do like 7, 8 different NHP studies across different cargoes to now be at the point where we have molecules that are more or less plug and play for these different modalities.

So in other words, it's not so straightforward to develop this technology.

So rodents are not predictive at all for non-human primate and for human, like rodents have leaky blood-brain barrier, a very broad range of transferring affinity works really well in rodents. There is no difference in the -- brain entry between different affinities in rodents. And it's very different. The picture is very different than non-human primates, and we do think that non-human primate is predictive for humans. So you can't really say anything about your technology before you really validated in non-human primate in the relevant dose. I think the relevant dose is also important. A lot of publications and reports are, for example, these antibodies are at very high doses of 10 mg, 30 mg, 50 mg per kg, which significantly mask and nullify the blood-brain barrier technology effect. You really need to go to low doses like 1 mg to 3 mg per kg to start seeing the benefit of the blood-brain barrier technology and the differentiation between technologies.

Okay. Is there sort of a need for different PK profiles for different targets, meaning drug targets? And how do you think about that?

Yes. I mean I think the need for PK -- it's all dependent on the mechanism of action of the molecule, right? And the characteristics of the cargo. So if you have a cargo like an enzyme or an siRNA that will itself rapidly clear within the periphery, then you don't need a long-lasting blood-brain barrier technology. You need something that drives really fast uptake. If you have, say, a blocking antibody, something that relies on the trial concentration, you need that to get into the brain, then you would want to pair it potentially with a blood-brain barrier technology that lasts longer in the periphery, maybe you pair it with half-life extension technology, a lower affinity binder. But again, it's all based on sort of the mechanism of action of your cargo.

And also the safety of your cargo like that means we didn't discuss it yet. But as you know, one of the on-target risks of the transferring based blood-brain barrier technology is anemia because there are 10x more transfer receptors on reticulocytes than on the blood-brain barrier. So a lot of your drug goes to reticulocytes. And if your drug is, for example, an effector function that can recruit immune cells, you basically bring activated immune cells to reticulocytes and that's what damaged the reticulocytes and leads to anemia. So you can deal with it with either low affinity or as we dealt with it with identifying an epitope that reduce the ability to co-bind the immune cells and reticulocytes. But the safety is also a factor for the PK.

Okay. We will get into safety shortly with your molecule specifically, but so next slide, please. How do you think about sort of the selection of isotype backbone and Fc function? Is it something that can also be 100% modular and sort of plug and play?

So again, as Arnon just introduced, the safety really plays a role in how you think about what Fc component to use. So if you have a cargo like an siRNA where there's no need for effector function to drive the mechanism of the cargo, then you can go with a plug-and-play silenced Fc, and it really is, again, more on the plug-and-play side. But if you have a cargo like anti-A-beta where to drive phagocytosis of the target in the brain, you really need the ability of the Fc to engage those Fc receptors then it's a little bit less plug and play, right? So then in that case, you want to utilize that full effector function molecule, and we have to modulate the affinity and especially we get into a lot of the epitope of TfR that we're binding in order to still drive strong brain uptake but to get safety with using that fully active Fc.

All right. So what are actually some of the nuanced details and sort of mechanisms that may lead to an optimal or more optimal from a safety perspective and efficacy profile?

Yes. I mean, definitely, selecting on the epitope and the binding kinetics are 2 of the most important ones. If you're talking about a transferrin receptor binding blood-brain barrier transfer technology specifically. So in our case, we sort of engineered a unique epitope on the transferrin receptor that still doesn't interfere with the native function, still drives best-in-class brain uptake, but also sort of decouples a little bit the ability to co-bind the transferrin receptor and then still have the Fc engage innate immune cells such as like natural killer cells in the periphery, which drive some of those hematological safety risks that Arnon already introduced.

Different companies are dealing with the safety efficacy profile differently and especially it comes across with the anti-A-beta antibody. So some companies completely use an inert effector function. Basically, they engineer the constant region to not to be able to bind immune cells to the Fc gamma receptors at all, and they rely on other like what's called nonclassical phagocytosis mechanisms to remove A-beta. Other companies cripple the effector function and try to thread the needle between safety and efficacy and basically partially mutagenizing the constant region in different ways. And again, we found a different solution. We think that a full effector function is essential for full efficacy and that you can't really modulate the effector function without impacting efficacy and you need to find another solution. And again, our solution was to find a different epitope that can disconnect anemia from removal of A-beta plaques.

Is there a possibility -- and I would assume the sort of modality specific as well as target of marrying this tech with an extended half-life to further decrease dosing?

It's certainly possible, but it really is based on the properties of the cargo. If your cargo itself is driving strong target-mediated clearance, then half-life extension technology isn't really going to give you much of an effect. So this is obviously the case with these enzyme and siRNA cargoes. It's possible for like an antibody cargo that has low target-mediated clearance in the periphery. But again, if you're using like a high affinity TfR to drive really strong brain uptake in a short period of time, you're still, again, not really going to be able to rescue that with half-life extension technology, because half-life extension technology only really rescues you from like the native background level of clearance that happens, not from specific target-mediated clearance.

So again, the transferring is a very potent target-mediated clearance. So half-life extension could have limited impact. And also specifically for anti-A-beta antibodies, the half-life extension could have both cons and pros like the cons are that if the antibody is present longer in the periphery, it has a higher possibility of interacting with reticulocytes and leading to anemia. It has higher possibilities of interacting with vasculature A beta plaques and leading to ARIA. So you really, again, have to weigh the cons and pros of longer half of half-life extension, which is dependent on the target.

Okay. So what are the key biological properties that sort of make a receptor well suited for ferrying cargo across the blood-brain barrier? And do you see the field sort of converging around a few optimal receptors? Or will different receptors be sort of required for depending -- depending on the cargo type or therapeutic type and indication?

Yes. I think one of the key properties, obviously, is that high expression on the brain endothelial cells. Ideally, you want to have very specific expression, so high expression in the brain and low expression in peripheral tissues. Obviously, it needs to be able to drive strong transcytosis. And I think one of the most important ones is that it actually has minimal or manageable target-based adverse effects. And this is where transferrin receptor does have a known class effect with these hematological side effects. But again, it's known, it's monitorable, it's manageable. Any new target introduces a whole new suite of potential safety risk that needs to be mapped out and derisked both in non-human primates and in the clinic. And that's why we think for the next -- both short and probably medium term, transferrin receptor is the most validated and the field is converging around transferrin receptor, and it will dominate the field again, in the next probably 3 to 5 years because of that clinical validation. Proof of concept for these new targets is, I would say, years away, especially for the ability to really utilize them more broadly. Again, when we're talking about these internal cellular targets, things like siRNA and enzymes, the transferrin receptor has a very favorable dual mechanism where it both is able to drive entry into the brain, but also drive entry into cells in the brain. And we've shown this now for neurons, astrocytes, microglia. So all the key cell types in the brain are actually able to deliver intracellularly after you cross the blood-brain barrier. And this is something that I don't think has been mapped out nearly as well for any of these other receptors.

So the transferrin receptor is responsible for that, isn't?

Yes.

I guess what sort of drives into my next question is what biological properties make it particularly well suited for receptor mediated transcytosis?

Yes, exactly. That ability to drive strong uptake across the blood-brain barrier, strong uptake in the cells. The transferrin receptor drives entry through the endolysosomal system. So again, if you're delivering a lysosomal enzyme, that's perfect because you're getting into that system. The same thing for siRNAs, therapeutic siRNAs need to be delivered into the endolysosomal compartment at the end of the day for the antibody part to be degraded off and the siRNA to undergo endosomal escape and get into the cytoplasm where it ultimately has its effect. So that is a very favorable property of transferrin receptor.

So how does the sort of receptor-mediated uptake between transferrin receptor differ from the other sort of transferrin receptors, which is currently being evaluated IGF-1R or CD98 heavy chain? And in terms of intracellular trafficking and sort of distribution into the brain.

IGF-1R has a different non-endolysosomal mechanism for uptake. There is a nice publication recently that highlights the mechanism. So that can be favorable for delivery under the parenchyma, although anti-IGF-1R has never shown as high of absolute brain uptake as a TfR-based shuttle. And then CD98 has a much different kinetics of uptake, which can make it favorable in some applications. But again, that one has not really undergone the validation in non-human primates and in the clinic to show that it can both be translatable and safe for delivery.

So there are multiple groups engineering these blood-brain barrier penetrant therapeutics. Sort of what metrics should investors focus on when sort of comparing the different platforms?

Yes. So basically, the parameter should be like the fold penetrant into the brain and like this should be measured by absolute concentrations of the drug in the brain, not by relative concentration. Again, a lot of the publications are looking at relative like fold elevation in the brain compared to the periphery and compared to naked drug. If your naked drug is degrading quickly or is a lousy drug that doesn't enter the brain, the fold elevation could be misleading. So you want to show with absolute concentration of drug that you have like meaningful elevation of drug in the brain. It has to happen at a low concentration like that enable like subcutaneous delivery that basically enable the advantage of blood and barrier technology to reduce the drug dose for both safety and convenience of use. So you want to see high brain penetration at low drug peripheral dosing. You want to see manageable safety at least, again, starting in non-human primates, not just in rodents. You want to see that the shuttle does not impact the normal function of the receptor like in the transferring receptor case, for example, you don't want to disrupt iron transfer to the brain or iron transfer to reticulocytes. In CD98, you don't want amino acid transport to be impaired. So you want -- again, basically, you want to see good safety or manageable safety, good brain delivery at low dose and durability that basically you don't down-regulate the receptor, you don't degrade the receptor, you don't saturate the receptor very quickly. So you want to see -- and you want to see also the drug configuration that you have is commercially or clinically viable that you can manufacture it at high level, that you can store it, that you can basically give it to human.

Okay. So next slide, just can you sort of give us a high-level overview of the ABC platform in terms of target selection, modalities, even though you did touch on it, that can be leveraged. And just what gives you confidence at a high level that you have candidates derived from the ABC platform that are sort of ready to enter the clinic?

Yes. So Eric, I don't know if you want to answer that. But yes, as Eric mentioned, we tested our blood and ABC technology in multiple like in over 8 non-human primate studies with 3 different drug modalities, enzymes, nucleic acid and antibodies. The shuttle configuration, the blood-brain barrier that transferring valency, that transferring affinity, that transferring binding kinetics are really tailored for each drug modality. The constant region of the antibody is tailored for each drug modality, the location of the blood-brain barrier, the transferring binding epitope is tailored to each drug modality. So we really engineered our blood-brain barrier shuttle for siRNA specifically, for antibody specifically and for enzymes specifically to really optimize all the aspects of the combined drug. And as a result, in non-human primate, we do see very sort of durable and broad brain distribution with low dose delivery and we see drug efficacy. We see with siRNA, for example, we see profound reduction of the target siRNA, for example, tau. With enzyme replacement therapy, we see significant elevation of the enzyme in multiple brain tissues with enzyme -- with antibodies, for example, anti-A-beta, we see 4- to 12-fold higher brain concentrations compared to competitors. So basically, we have multiple validating data in -- primarily in non-human primates with multiple experiments with different dosing regimens, different drug modalities, different durability that really validate our platform.

The key thing I would add there, just to this not being a simple endeavor, we didn't get clinical candidates out of our first NHP study or our second NHP study or our third NHP study. We got clinical candidates out of our sixth NHP study, our seventh NHP study and our eighth NHP study. And that's sort of the depth of knowledge it took to get to the point where we didn't just have drugs that are good enough, but like really have the potential to be best-in-class. And again, it really relied on that validation in the NHP system, not in anything we were seeing in murine models.

All right. We're going to get into details shortly, but just from a high level, again, what is differentiated about your platform versus other transferrin-based approaches?

Again, just the breadth of tailored shuttles that pair with each of these different modalities. the different kinetics we can access, the different Fcs, the different formats that we use. On A-beta side, sort of that novel epitope that we'll talk about in more detail later, that allows us to utilize the full effector function and still have a manageable safety profile. And then at the end of the day, the validation in all of those different NHP studies covering all those different modalities.

Okay. And when you say a novel epitope, you mean epitope on the carrier...

Carrier -- on the transferrin receptor.

Okay. From an engineering perspective, how is transferrin binding sort of incorporated into the ABC platform? Is it binding via a Fab fragment engineered into the Fc domain? Or is it ScFv, VHH or peptide? Will it be a combination of all or just sort of multiple interchangeable modules?

So we utilize multiple types of modules. The key design philosophy is to minimize interference with the function of the cargo. So in the terms of an antibody, we're appending something like an ScFv on the end of the Fc as far away from the active fab end of the antibody as possible. And then with enzyme and nucleic acid cargoes, we can utilize an anti-Fab or a Fab anti-TfR binder paired in different ways. And again, it all -- you have to test like multiple combinations to figure out how the molecule goes together in a way that both works therapeutically, but is also like highly manufacturable and stable. We did try utilizing engineering binding into the stock domain itself, and we found that it was really significantly different to engineer the breadth of affinities and epitopes that we wanted. And so we kind of actually gave up on that approach and stuck more to the antibody fragment-based approach.

Okay. We did touch on it a little bit before, but again, what are sort of the key trade-offs between monovalent and bivalent transferrin receptor engagement? And what factors, whether it be the cargo, modality, safety or something else drive the decision on which sort of format to choose?

Yes. I would say for brain delivery monovalent formats are strongly preferred. Bivalent formats really have a very strong potential to drive receptor degradation by cross-linking receptors and driving them directly into the lysosome. So unless you're a company like Avidity where you're actually trying to deliver into target cells in the periphery, we would encourage the use of monovalent because you really need that antibody to be able to bind to the target and then come off at some point. And bivalency really drives towards binding and then staying on or even cross-linking the receptor.

Awesome. Yes, a bivalent invariably increase the binding stress, the receptor to a point where, yes, it's hard to dislodge the shuttle from the receptor and this irreversible or almost irreversible binding leads to receptor internalization, transferring receptor degradation and in some cases, would lead to either anemia, is another way to lead to anemia if you degrade the transferring receptor. And if you degrade the transferrin receptor on the blood-brain barrier, you cannot really transport drugs through the blood-brain barrier. So bivalency is really primarily effective when you want to transport drugs to the muscles where receptor degradation doesn't really matter.

All right. A little bit more details about your candidates. So you show a very wide range of transferrin receptor binding affinities. How do you select these binders? And do they provide a sort of an affinity response curve? And although we touched on it before, sort of a broad question for the ABC platform, how do you think about optimal affinity window for brain delivery?

Yes. I think the interesting thing here is you see binders across a 1,000-fold range of affinities. All of these binders worked in the murine system to drive very strong brain uptake and actually not that much but surprisingly consistent brain uptake between even single-digit nanomolar and single-digit micromolar binders. It was really when we moved in the non-human primates on the right side of the slide, you can see 3 binders that work equally well in the mouse have up to a fourfold difference in performance in the non-human primates. So we really had to sort of fine-tune that window in the non-human primate system to really give us the best sense of what the optimal affinity window is. And then again, it is, again, different when you move from an antibody cargo to something like an siRNA, where, as I said, you have that dual mechanism where it needs to translocate into the brain, but also drive uptake in the cells. So that itself requires a different affinity window.

All right. Can you just actually walk us through the middle figure, sort of what does the experiment sort of tell you about affinity and receptor binding and internalization and how this sort of will translate to in vivo brain exposure?

Yes. In a broad sense, the middle figure here is showing the first step of the transcytosis process, which is the ability to engage the receptor on the surface of brain endothelial cells. And that process tracks very well with affinity. So a higher affinity binder will drive stronger uptake into the cell. So basically will drive the first step of the transcytosis mechanism. But you really have to go in vivo to elucidate the entire mechanism where it actually has to translocate and then fall off and translocate into the parenchyma.

Okay. How many modalities sort of have you tested with the -- in either mice or non-human primates? And do these -- do you see modality-specific optimal affinity ranges, which I assume you sort of touched on before, so I assume yes. For example, siRNA conjugates sort of require different affinity windows versus antibodies and enzymes?

Yes. So we've tested actually 13 different cargoes in the murine and non-human primate system, including the 8 non-human primate studies I mentioned before. And I think we really are seeing that for certain cargoes like the siRNA, there is really an optimal affinity or even just an optimal binder and format that you can make much more plug and play and you can just take a new siRNA and put it on the same binder and get a very similar performance.

So it's really -- you have to test it experimentally. Affinities that work well for antibodies don't work at all for siRNA or for enzymes. You really have to test each drug modality and identify what's the optimal bell curve for a given drug modality and drug configuration. And again, the reason is that antibodies have significantly longer half-life in the cell. You can get away with lower affinity. There is more time for the drug to enter the brain. siRNA and enzymes have short half-life in the cell. You have to really push them through the blood and barrier much more rapidly and then you need much higher affinity. So the affinity ranges of antibodies, enzymes and siRNA are very different like the optimal affinity is.

All right. Next slide. Can you just sort of walk us through what is structurally distinct about your ABC transferrin receptor, I guess, binder versus other platforms? And does the epitope matter?

Yes. So we identified an epitope that is on, I'd say, the side of the transferrin receptor molecule. So it's still far enough away from the binding side of the native ligands like transferrin or ferritin, so it doesn't interfere with their binding, but it's not located on what we call the tip region of the apical domain, which is where some of our early antibodies and some of the competitor antibodies such as Roche and Denali bind. And we find that binding this epitope allows us to partially decouple antibody effector functions such as ADCC, which is the innate immune system-mediated mechanism that drives the hematological toxicity. We can sort of uncouple that from the affinity of the TfR and enable us to utilize higher affinity TfRs to drive stronger and faster brain uptake while still minimizing the possibility of use of downstream effector functions such as ADCC.

So I'm just going to make a comment here because people look at this and if you're not a structural biologist, you probably have no idea what you're really looking at. And so I guess the green domains are the ligand for the receptor, correct?

The light green and the yellow domains are the ligand itself bound to the receptor, so they're blocking sort of the region of the receptor that they bind to. And you want to make sure you're binding to a region that's not where that light green or the yellow domain is to avoid interfering with native function.

So the receptor is the dark and light blue, the transferring receptor and the transferring ligands are the green and the yellow. And you want to bind your shuttle, basically, you want again Trojan Horse that use the receptor, but don't disrupt its function. So you want to bind in a region that is away from the ligand binding domain. So there is really no physical interaction and no disruption of the physiological function of the transferring receptor.

How tightly coupled are our TfR affinity or epitope and ADCC risk, assuming that the Fc effector function is not silenced?

Yes. So that's a great question. Within an epitope, so within antibodies that are binding the same epitope, the ADCC window really correlates directly with affinity, right? So stronger affinity will give you higher ADCC at lower antibody concentrations. But what the figure on the right is showing is the ability to uncouple it by using a different epitope. So the blue and the green binders there are much higher affinity antibodies that bind to the ABC epitope. And then the red is a much lower affinity antibody that binds to that more exposed epitope on the tip of the apical domain. And you can see even though the antibodies binding the ABC domain have much higher affinity, they drive a significantly lower ADCC response. So again, within an epitope, it is affinity driven, but you can break that relationship somewhat by utilizing a more optimized epitope.

All right. What are the safety -- I mean, we sort of touched on it again, so briefly in 1 minute or less, what are the safety considerations or theoretical safety concerns with chronic TFR engagement with repeat dosing. Next slide, there we go.

Yes. So one of the main concern is related to the anemia. We have to remember that anemia is very frequent in elderly subjects and even more so in patients with Alzheimer's disease who display a rate of baseline anemia that is between 20% and 30% and half of it is a moderate anemia degree. And any drug that worsen baseline anemia that can make it symptomatic or more severe, I think it's going to be problematic, especially when the drug is administered outside the rigorous setting of a controlled clinical trial when it's administered to patients in the general population, I think it's going to be particularly challenging. And now the drugs that are in development, they require careful monitoring of hemoglobin, red blood cells periodically and to look at particularly drop threshold compared to baseline or levels of hemoglobin. I think a drug that doesn't have these liabilities that can be administered safely in this vulnerable population with high rates of undiagnosed anemia will be really transformational.

Okay. Sort of skipping a little bit forward, we did touch -- Giacomo did touch a little bit on trontinemab. But from a safety perspective, there were a lot lower rates of ARIA for trontinemab versus other plaque reducers like lecanemab and donanemab. Why do you think this is the case? And will this actually translate broadly across all A-beta therapies that have this blood-brain barrier crossing that?

Yes. We believe that one of the reasons why underlying the lower ARIA rate observed with trontinemab is the fact that there is not much time when the drug is actually present in the large blood vessels, therefore, allowing the drug to bind to the amyloid that is present in the large blood vessels and then therefore, damaging and causing edema, ARIA-E or hemorrhages. So I think this is going to be a broad characteristic of drugs which have the blood-brain barrier technology. And it's important to note that there is a background ARIA rate in Alzheimer's disease that is typically less than 5%. So the absence of ARIA, I don't think it's an attainable goal, but what will be important to show no increase over background ARIA rate in patients with Alzheimer's disease, therefore, waiving the need for repeated MRI assessments as part of the safety monitoring. Approved anti-amyloid treatments require 5 serial MRIs with a different time, especially at the beginning of treatment and then more MRIs if needed in case subjects become symptomatic and there is suspicion of them having developed ARIA. So -- and this comes with significant challenges for patients as well as the healthcare system. So I think that's another big advantage of drugs, anti-amyloid treatments that are paired with brain shuttles in terms of their ability actually to be prescribed more broadly to the patients with Alzheimer's disease.

All right. Moving specifically to your candidates, AL137 or 037, specifically against PyroGlu3 A-beta, similar to donanemab, how much derisked the target to a degree. What's your sort of rationale for choosing that target versus other A-beta species?

So I think that donanemab data with all the shortcomings that we discussed in terms of safety, et cetera, they show very strong amyloid clearance, much stronger than lecanemab. The number of amyloid negative subjects at the end of the Phase III trials is double for donanemab versus lecanemab. And donanemab is a PyroGlu3 antibody, which is species that is present in the plaques. And so therefore, even the greater clinical efficacy and greater amyloid reduction of drugs with -- that target this PyroGlu3 epitope, I think it makes perfect sense once we decide to develop an amyloid antibody with the brain channel to choose this kind of target.

Do you believe full effector function is actually required for meaningful plaque clearance once you achieve higher brain exposure? And the reason why we ask is because there are a couple of candidates out there that don't have effective function at all.

Yes, we do think that it's critical. Basically, all the antibodies that have shown efficacy in the clinic, donanemab, lecanemab, aducanumab has full effect of function. Antibodies with not a full effect of function like ranibizumab did not show efficacy, was an IgG4. So we do think that very aggressive removal of A-beta plaque is essential. Also, you have to know that, yes, we talked about gantenerumab. Gantenerumab was able to reduce A-beta plaque by 60%, 70%, and it didn't lead to clinical benefit. You need to extensively remove A-beta plaques beyond 70%, closer to 80%, 90%, 100% to get clinical benefit. And if you cripple your effector function, your likelihood in our view of being able to do this is significantly reduced. So we think that every drug that modulates, that cripples the effector function to achieve safety against anemia is going to pay with efficacy. And again, time will tell, but that's really what the data tells us so far.

Okay. Moving on to the next slide. What are sort of the exact differences between 037 and 137? And how should we think about functional differences between the 2 molecules?

Yes. Really, the only functional difference is the strength of the binding to the transferrin receptor. So they utilize the same anti-beta domain, the same Fc, but they utilize 2 different anti-transferrin binders that have about a 15-fold difference in affinity. And as you can see on the serum PK curve, 037 -- sorry, 137 utilizes a higher affinity anti-TfR and that drives stronger brain uptake at the shorter time period, but also drives faster clearance. And then AL037 has a weaker affinity TfR binder. So you see that the peripheral clearance is much more similar to a naked antibody, but you have a little bit lower brain uptake at the initial time point, which will probably translate to higher brain uptake at later time points because you have more retained in the serum. And then there's also different potential for safety with the 2 different TfR affinity. So that's something we're extensively mapping out. We're doing GLP studies on both of these molecules to determine which one has the best balance between efficacy uptake and safety.

So you have faster clearance with 137, but I'm assuming this is measurement of serum. And so that's likely going into the brain because you can see to the right-hand side, you have higher exposure levels...

Exactly. So with the higher affinity, you're driving higher uptake into the brain and also into TfR expressing tissues in the periphery. So that leads to a faster decline in the serum. But again, if the trade-off is getting the drug to where you want it to go, then that can be a trade-off worth taking.

Okay. And are you able to make any type of sort of comparisons to exposure levels for trontinemab -- to trontinemab?

Yes. I mean we mapped this out for both of these molecules. So actually, in their published NHP data at 10 mg per kg dosing, we still see higher -- significantly higher brain uptake for both these molecules at a significantly lower dose of 3 mg per kg. And we calculate that if you normalize it out, we're seeing a 5 to 10-fold increase in brain exposure in the brain itself, in the parenchyma over trontinemab if you sort of normalize for the equivalent dose.

Okay. So next slide, how can sort of a short systemic half-life for 037 or 137 than a naked antibody sort of help improve safety?

So there's 2 points. As Giacomo mentioned, the ARIA risk is really driven by exposure of the antibody to those central amyloid deposits in the larger blood vessels. So if you can decrease both the dose you need and then the clearance, that will reduce the amount of time that large levels of antibody are present in those large vessels to interact with the central amyloid deposits and drive ARIA. And then hematological tox is similar. It's reliant on the presence of a large concentration of antibody within the blood vessels to interact with the reticulocytes and actually drive that damage. So as soon as the antibody concentration, the serum drops below the level where you're inducing that damage, then the system can begin recovery. So if the antibody clears faster, you reach that recovery phase more quickly.

Awesome. Is there anything else you want to sort of disclose about this program specifically before we just briefly ask you about your 2 other programs?

Yes, again, that's sort of -- we think that we have a really good combination of efficacy and safety with this antibody. We are not sacrificing one for the other. So we think that overall, we could really have a very potent and safe drug and that would be very competitive. And our goal is to deliver it subcutaneously. Based on our modeling, we think that 1 mg to 3 mg per kg once a month, even with subcutaneous delivery will be more than sufficient to lead to complete removal of A-beta plaques. So we have a drug which we think, again, it's very potent because it has a full effector function, has manageable reticulocyte effect because of the unique epitope and enters the brain to very high concentrations and could be very competitive in the field.

Very awesome. So just briefly touch on your other disclosed programs or at least 2 of them, siRNA, Tau program. I know you've generated a lot of preclinical data, but just can you provide a brief overview of this program and sort of key data you would like to highlight?

Eric, do you want to address that?

Yes. I mean I think one of the key things and even just showing on this slide is that with our delivery platform, we're seeing very consistent delivery across different brain regions. And this is stand and start contrast to things like IT injection where you can see up to a 30-fold difference in the actual siRNA concentration between different brain regions, we're seeing at most like a threefold difference. And at the same time, we're seeing both very high absolute level of siRNA accumulation. So in this case, well over 100 nanomolar accumulation of the siRNA cargo, and that leads to significant knockdown across all those multiple brain regions.

Did I hear you correctly, 30% variation when you have intrathecal versus 3% variation?

30-fold variation and 3-fold variation. So it is quite extensively -- because you're injecting into the spinal cord and you're relying on sort of a passive diffusion mechanism. So based on how close the brain region is to the injection site, you can get much more or much less of the drug.

Yes. And the reduction in sort of the knockdown of mRNA really translates to reduction in phospho-tau protein, both in brain tissue and in the CSF and it's durable like we measured in tissue up to 98 days after like 3 weekly injections, we see around 60% knockdown of the phospho-tau protein. So we have a very sort of a durable effect. I mean we have an siRNA drug that can be injected peripherally, and we are going again for subcutaneous delivery that reach the brain, that distribute in the brain very broadly, that leads to significant knockdown of the target mRNA and resulting in significant and durable knockdown of the phospho-tau protein in brain tissues. So we are, again, very excited about this program. We think it could go either as a stand-alone drug to enhance the efficacy in Alzheimer's disease or potentially could go sort of in combination with anti-A-beta therapeutics to really push again the clinical benefit that is limited in anti-A-beta antibodies on their own.

Okay. That is the future combinations. So next slide, I know you have generated again preclinical data, but can you just sort of provide an overview of the GCs program and sort of key data that you just would like to highlight?

So GCase is a lysosomal enzyme that's basically loss of function of this enzyme is associated with a significant percentage of Parkinson's disease and significant percentage of Lewy body dementia. And although the concept of enzyme replacement therapy is very validated in the periphery, like in Gaucher disease, for example. It's -- until now, there was very little ability to elicit enzyme replacement therapy in the brain, like there is sort of a drug for Hunter disease that sort of, again, it's a lysosomal enzyme that affects brain that leads to brain pathology. But there is really -- until now, there's no ability -- there was no ability to deliver GCase enzyme to the brain for Parkinson's disease and Lewy body dementia. And we were able to overcome 2 major issues with this approach. We were able to engineer GCase. The natural GCase is very -- has very short half-life and mediocre activity. We were able to engineer the enzyme itself to have up to 50-fold higher activity and stability, and we were able to enable delivery to the brain following peripheral injection with our [ blood and barrier ] technology. So we see, again, with peripheral injection, we were able to show in non-human primate broad distribution in multiple brain tissues, including relevant brain tissues like the substantia nigra and the putamen that are affected in Parkinson's disease. We were able to show more than doubling of enzymatic activity. And in rodent disease models, we were able to show very durable effect, a single injection is effective in reducing the toxic substrates that is accumulated in the absence of the enzyme. So we are able to reduce the toxic substrate for over a month with a single injection. So we think that we have a really exciting enzyme replacement therapy for Parkinson's disease and Lewy body dementia that are caused by GCase deficiency. And we think that this therapy could ultimately also be beneficial for sporadic versions of the disorders because Parkinson's patients, for example, that have lower than normal level of GCase and higher than normal level of the toxic substrates progress significantly faster than other sporadic Parkinson's patients. So we think that, that could be the first and best-in-class GCase enzyme replacement therapy for -- again, for these 2 disorders.

All right. Thank you very much for that. And we are a little bit over, but not a big deal. Last 2 questions. Neil, if I've not heard from you, so let's hear something. How are you thinking about BD? Are you looking to sort of partner out specific candidates? Or if large pharma wants to license the platform for specific targets, are you open to that? Or will it be a combo of both?

Yes. Thank you very much. We're very -- we're focused on asset-focused deals. So we're primarily focused on discussing the potential of partnering around specific assets where we can capture our full value of those programs. I mean as part of one of the relationships around an asset, we could consider broader platform access, but it would be in the context of an asset anchored deal. We're not and have historically not pursued stand-alone platform type just enablement of someone else to use our platform to develop their drugs. That is not something that we have pursued.

Okay. I'm going to ask you how active have you actually been on the BD side? And like the reason -- one of the key reasons why I'm asking is that when I was in San Fran in January, any company that had an asset that needed to go into the CNS, they were looking for some type of platform to link to. And then companies like you that I met said you were quite busy with BD. And so just how busy have you been? And can we expect to hear some type of deal within the next 12 months or so?

We've been very engaged with pharma partners. Very strong engagement, very strong appreciation for what we're doing, very strong interest. And they know that we've been working on this for a long time. We've been working on this for 7 years. So there's a lot of knowledge and credibility behind our platform and what we're doing. So partners are very engaged. We will -- we're being very selective about what we want to do. We're excited about our pipeline, and we're being selective about how we're going to partner those, and we're focusing on transactions that reflect the value and asset of those platforms -- of our platform and our assets. Can I do -- as a BD person, I formally don't make predictions on deals, but I do think the engagement could lead to something over the next -- definitely 6 to 12 months as you suggest.

All right. That's great to hear and excited to see what comes out. Last thing is next sort of key milestones for the platform over the next 12 to 24 months?

Arnon, do you want to take that?

Yes, sure. Yes. So we are advancing our anti-A-beta antibody. It's going to be in the clinic in the next sort of definitely less than 12 months. And we know now how to really look at efficacy in Alzheimer's patients with a small short clinical trial. We are going to sort of look at profound reduction in A-beta plaques with PET imaging. We are going to look at multiple biomarkers. We are going to look at minimal or no ARIA, sort of very manageable anemia risk. And we are going to also look at minimal infusion reactions, which is one of the liabilities of some of the competing drugs. And we are going to, again, do all of this move very quickly to subcutaneous delivery for maximal convenience of use. So the anti-A-beta drug is going to sort of have initial proof of concepts in the next 2 years. We are advancing our Tau siRNA to IND and into the clinic in the next 2 years and also possibly the alpha-synuclein and the GCase programs. So in the next 2 years, we'll have multiple sort of drugs going to IND and beyond.

Awesome. Thank you very much for that. Thank you very much for your time. Really appreciate it. I think a lot of insightful sort of nuance and information. So thank you, and also thank you to the audience for listening in.

Thank you very much for having us. We really appreciate it. Thank you.