All right. Good afternoon, everyone. Thank you once again for coming to the 46th Annual TD Cowen Healthcare Conference. I'm Steven Holtzman. I'm the biotech team, and I'm joined by the management of Alector, Inc. We'll be hearing a presentation first from Eric Brown, the Head of Antibody Discovery and Protein Engineering. We may take some questions. Eric, thank you for coming. Take it away.
In addition to my role leading antibody discovery and protein engineering, I've also been heading up our ABC platform for the last seven or so years, and that will be the primary focus of our talk today. If you look at our portfolio here at Alector, we do have one non-ABC-enabled program, AL 101 , which has an interim futility analysis in the first half of this year conducted by an independent monitoring committee. I will not be talking about that particular program today because the focus is entirely on our ABC platform. I will be talking about the rest of these programs, including our lead antibody candidate and Aβ antibody enabled by ABC, our lead enzyme candidate, which is a GCase enzyme. Both of these programs are headed towards IND.
I'll introduce our siRNA platform, which is also being enabled by the blood-brain barrier technology as exemplified by our tau ABC molecule. There are a couple other lead molecules or follow-on molecules that I will identify in our siRNA platform. To get into our ABC technology, we started on this, as I said, about seven years ago, casting a very wide net, looking at a number of blood-brain barrier receptors, including transferrin receptor, CD98, IGF1R. Through several years of work, we've sort of honed in on TFR as the one most ready for platform applications.
Even within TFR, there's a very wide range of potential ways to go, so we honed in on both the specific epitope on TFR and a very wide range of affinities to build a broad toolkit to apply across a wide range of cargos. Really we spent the last several years on what actually turns out to be the trickiest part of this, which is applying the specific ABC of interest to the specific cargo, putting them together to make a molecule that both crosses the blood-brain barrier, is efficacious on the other side of the barrier, and then lastly, obviously is a highly manufacturable and scalable therapeutic. In the course of this, we've actually taken, it says 12, but it's actually 13 different cargos that we've applied antibody, Alector Brain Carrier technology to. This is a mix of antibody's enzymes and siRNAs.
In that case, we've actually taken all of these in vivo, many of these into NHPs, and we've finally gotten to the point, I think, where we really have excellent ABC technology to deliver all sorts of cargos, antibodies, enzymes, and sRNAs. As you'll see, there are slight subtle differences for each of them that we'll get into in each section. One of the unique properties of our ABC platform versus other TFR-based platforms is the epitope that we're using. If you look on the left or on the left side of the screen, you see a schematic of the transferrin receptor in blue, and then appended to that on, in green and yellow is a schematic of the native ligand transferrin.
Initial efforts to utilize a transferrin receptor as a blood-brain barrier receptor focused on what I'm calling the Alector epitope B, but you see is also highlighted as where the Denali ATV epitope and the Trontinemab epitope from Roche are. This is at what we call the tip of the apical domain. This is a epitope on TFR that's very available for binding, does not interfere with binding to the native ligand. But the problem in a sense of this epitope is if you bind an antibody there, then the Fc of that antibody is very highly exposed in order to bind to innate immune cells, which is what causes the downstream safety functions such as reticulocyte depletion and anemia that are commonly seen with anti-transferrin receptor antibodies, particularly those utilizing an active Fc.
This is very important for our case because we are utilizing an anti-Aβ antibody that we feel requires fully active Fc in order to have the highest possible Aβ phagocytosis. In addition to looking at antibodies at that well-characterized epitope B, we also searched for an alternative epitope that would drive the strongest possible brain uptake but also ameliorate the safety liabilities. We found what we call the ABC epitope here on the side of the transferrin receptor. Antibodies are able to bind here. They don't interfere with transferrin binding. They don't interfere with the native function, but they do show very strong brain uptake. Now we've somewhat disconnected the ability to drive brain uptake with the ability to drive these downstream effector functions. We can highlight that in a number of different ways.
If you look at the middle of this slide, you'll see an in vitro ADCC assay where we show an antibody in red against the exposed epitope B, where we can drive a strong productive ADCC response through the transferrin receptor versus antibodies in green and blue that bind at our ABC epitope. Even though these are significantly stronger higher affinity antibodies, they drive significantly less of an ADCC response. This translates in vivo on the right-hand side of the slide to decreased reticulocyte depletion in the murine system.
In this case, we took two antibodies with the same affinity dosed at a relatively low therapeutic concentration of 3 mg per kg, and we see the antibody binding at epitope B in red shows very significant reticulocyte depletion, whereas the antibody binding our lead epitope in blue, the ABC epitope, shows very minimal reticulocyte depletion. We actually carried this forward into non-human primates to show that we see significant amelioration of the safety. I'm gonna show on the next two slides, two match studies where we did tox dose evaluation of an antibody that binds at the exposed epitope, which would be this slide. Even with a partially active Fc, this is the cis LALA-PG type of mutation.
In this case, we see both reticulocyte damage, more importantly, over time, you see in the middle and right side of the slide, sustained depletion in both red blood cells and hemoglobin, which is what leads to an active anemic phenotype. On the next slide, we'll show a very similar study design also in the NHPs, in this case with a full effector function where we utilize our proprietary ABC epitope with a similar affinity binding to TFR. In this case, while we still see some transient decrease of reticulocytes, you'll see over the course of the study between each dosing interval, the reticulocytes have time to recover, thus because they're recovering between each interval, you don't see any decrease in either red blood cell counts or hemoglobin over the course of the study. Again, this is a tox dosing study, it's very high dose repeat.
This is sort of the worst case scenario for this particular antibody. Getting back to the ABC platform as a whole, we've actually validated on the left side binders anti-transferrin to transferrin receptor across a 1,000-fold range of affinities. All of the binders on the left side of the slide actually work to drive significant brain uptake in the murine system. What we've really noticed, I think one of our biggest learnings from, you know, seven years of working on the system is how disconnected the murine and the NHP results can be at times. When we took antibodies on the left side that work equally well in the murine system, and we put them in the NHP system on the right side, you can see that there's actually substantial difference in the amount of brain uptake we're able to drive.
All of them work in the sense that they drive significant increase, but there's, you know, working 10-fold and then there's working 30+ fold, and we're looking really for the antibodies that work 30+ fold. That's where we're honing in. This is where, you know, the affinity uptake relationship just seems to be different between the murine system and NHPs, and this will matter a lot when we get into the specific applications. The first specific application I'll talk about is an anti-amyloid antibody that's enabled with ABC technology. In this case, we combine both the best in class anti-Aβ epitope with the pyroglu Aβ with our ABC technology. In this case, we found it's very important to have fully active effector function in the constant region. We're basically using wild type IgG1.
We're not trying to ameliorate or cripple the effector function because we feel like basically every clinical molecule that's worked to clear Aβ in the human patient population has had a fully active IgG1. Antibodies that do not have the fully active IgG1 have not been effective in patient populations. Whereas some of our competitors are trying to differentiate by fine-tuning or tweaking the Fc, we're going full rip on potency, and we're again fine-tuning the safety using our ABC epitope. Just again, when we're highlighting the molecular features here, all of these antibodies are designed as highly manufacturable therapeutics, so concentratable to a very high level, up to 150 mg per ml in this case. Subcutaneous administration is definitely the ultimate goal with this antibody.
Given the amount of brain uptake we're able to drive and the very low dose that supports, we think this is the highly likely way to go. In our clinical trial, we'll go immediately into both IV and subcutaneous administration. Just to talk just a little bit about the Abeta side of our antibody. It is able, obviously, to drive significant level of phagocytosis by microglia against the pyroglu Abeta, which you see on the left side. On the top middle, you see iMicroglia phagocytosis of, again, the pyroglu Abeta. If you can squint and notice, you'll see when we apply the ABC technology to the Abeta antibody, we don't see any decrease in phagocytosis. We actually see a slight increase, a slight amount of additive phagocytosis, which is nice for us.
The bottom middle, you can see light sheet imaging of a 5xFAD mouse that has been dosed with a murine surrogate of our AL137 lead antibody. It has the same anti-Aβ Fab, but it's using an anti-murine TFR surrogate ABC. You can see there's no vascular staining. The antibody's not stuck. It's really distributing and modeling exactly where the plaques are located across the brain. On the right-hand side, you can see on a PD sensibility of our antibody to drive reduction in Aβ42 species in an Alzheimer's mouse model, in this case, the 5xFAD. This is again with the murine surrogate TFR. At the time we ran this study, we didn't have the mice crossed with the human TFR with a 5xFAD mouse .
Skipping ahead a little bit to the non-human primate system, we did a two-dose study in the non-human primates. This case, we're looking at CSF uptake. CSF is not the perfect proxy. CSF is not actually where we want our drug to go. We want our drug to go into the brain parenchyma, but it is the measurement that we're able to look at, rapidly in the clinic, so we wanna make sure we're able to see this rapid uptake. You can see even 24 hours after dosing, you're seeing a greater than 12-fold increase in antibody levels in the CSF with AL137 compared to the Aβ antibody without the ABC technology. The difference is even more stark when you look at where we actually want the antibody to go.
This is brain uptake in what we call a vessel depleted brain fraction. We've actually isolated the brain's different brain regions from these animals, did differential centrifugation to remove the vessel fraction, and we're only looking at antibody that's fully crossed the barrier and is now in the brain parenchyma. In that case, we see, you know, approximately 30-32 fold increase in AL137 over the naked antibody. The important thing too here is in addition to the vessel depleted fraction, we also looked in the whole brain fraction in order to compare to some of our competitors. We did absolute brain uptake. On the top, left here, you can see we benchmark our antibody after a 3 mg/kg dose as getting in at 8.4 nM in the frontal cortex.
To put those numbers in context, we did literature search, we looked at some of our competitors were doing, we pulled out the Roche NHP data, and they're showing 2.2 nM brain uptake after a 10 mg/kg dose. We're seeing four fold more antibody getting into the brain at a 3.3-fold lower dose. We're really driving a significantly stronger increase in brain uptake, not just compared to a naked antibody, but compared to other BBB-enabled antibodies. We did similar comparisons for Denali's Transport Vehicle module for the J&J antibody and for BioArctic. In all cases, our antibody is driving significantly higher absolute level of brain accumulation, specifically at the lower dose. We think this has a lot to do with the affinity of the anti-TFR module that we're using and its specific sort of on-off kinetics.
We also, not shown on this slide, have a backup to AL137, AL037, which also benchmarks stronger than any of the other competitors that we looked at, but has a little bit different affinity to TFR, a little bit different performance and safety in serum PK metrics. Importantly, because we're looking at a fully active effector function antibody, relatively high affinity to TFR, the safety readouts are pretty critical. In this study, we dosed AL137 not just at the 3 mg/kg therapeutic dose, we also dosed a 30 mg/kg tox dose. AL137 was well-tolerated at both the therapeutic and the tox dose up to 30 mg/kg. As expected, we saw a transient reduction in the reticulocytes. At therapeutic dose, the reticulocytes had fully recovered by...
for the next dosing interval, which is a week. In the course of the study, we did not see any decline in red blood cells or hemoglobin, and we overall saw no test article-related adverse findings in any of the groups in this study. This is a molecule we are moving forward to IND by the end of this year, maybe slipping into early next year, and we hope to dose first in human relatively soon after that. To switch over to a second example, this is an ABC-enabled GCase enzyme for treatment of GBA and PD. In this case, we're delivering an enzyme GCase, which is implicated in pathogenesis of Parkinson's disease and Lewy body dementia. This is actually in terms of a lysosomal enzyme or lysosomal storage disorder, a relatively large patient population.
It's a very attractive genetic target where, you know, a double knockout of the enzyme leads to Gaucher disease, and a single knockout leads to the GBA-deficient PD. It's a genetically defined target characterized by loss of enzymatic function, relatively clear path forward. The reason it hasn't really been addressed to date is there's two main challenges. One is just delivering an enzyme across the blood-brain barrier and into the lysosomes of cells where it's needed. The other one is for anyone who's worked on it, GCase is a very finicky, difficult to produce enzyme. We had to sort of overcome both of these at the same time.
In my group, we engineered a version of the GCase enzyme that's approximately 50 x more active than the wild-type enzyme, and that has a stability in serum-like conditions that's increased from less than 6 hours for the parental enzyme to over a week for the engineered enzyme. We took that and paired it with a specific flavor of TFR ABC that's designed to both drive uptake into the brain and into neuronal cells in the brain to put together our final AL050 molecule. In this case, we can get an additional layer of safety by pairing our ABC TFR epitope with inactive Fc. To look in the murine system in wild-type mice, these are mice on the left that don't have any deficiency in GCase with again, a murine surrogate.
We're already able to see a 100% increase in GCase activity. This is already more than the 50% that we would need to see to normalize a heterozygous GCase patient. We're able to see significant reduction in the accumulation of toxic substrates, in this case, glucosylsphingosine, in the GBA homozygous knockout mouse system. Obviously, you can't see knockdown of the toxic substrates in the wild-type mouse because they don't have accumulation of toxic substrates. The most important data to me is on the right side here, where after a single dose of the AL050 surrogate, we see sustained reduction in the toxic substrates far beyond what the PK of the molecule would look like. This is an enzyme.
It can clear relatively quickly, but once it gets into the cells, into the lysosome, it's able to do its activity for far longer. The reduction in substrates in this case lasted out well beyond four weeks, right? You can see at four weeks, we're still maintaining a 40% reduction. We took this molecule forward as well into the NHPs. In all of these case studies, I'm kind of simplifying everything, and a lot of cases, it took multiple rounds of murine and NHP studies to get to the final point. In this case, we're looking at enzymatic activity in the plasma just to confirm the increased stability and activity of our enzyme. Other enzyme replacement therapies for GCase that utilize a wild-type enzyme have an activity half-life in the serum of minutes. We're looking at over six hours.
This is where you see that 50-fold increase in activity and stability. When we look in the brain, we are seeing substantially increased brain uptake of the GCase enzyme when we apply the ABC technology to it. This is compared to what we call naked AL050, but it's just it's the enzyme, it's on the same backbone. It has the Fc, it just has an isotype control Fab so that we're not seeing binding to TFR. In this case, we see increased uptake from five to 18-fold, but I think the really important data here is the amount of enzymatic activity we're able to drive. These are healthy wild-type monkeys. They have no deficiency in GCase enzymatic function, so they're already starting at 100%.
In this case, our bar would be to see a 50% increase over the 100% wild-type function in order to give us what we would need to, you know, increase a heterozygote patient up to wild-type levels. In every brain region we test, we see significantly higher than this. We're actually seeing between 63% and 132% increase in enzymatic activity. The increase in enzymatic activity is actually higher than what we see in the increase in just enzyme level when you compare both total endogenous Sino GCase and our introduced GCase, which is, I think, reflective of the higher enzymatic activity of our engineered enzyme. This molecule is also well-tolerated in the NHPs. No AL050 related clinical signs. There was no impact on hematology.
In this case, there was not even any transient reduction in reticulocytes. In this molecule, we have double protection with our ABC epitope and the silenced Fc. Basically, there were no testicle-related adverse events at all. Just to quickly introduce our ABC platform. We realized, you know, quite quickly that our platform, as well as enabling protein-based therapeutics, is also a great platform for delivery of siRNA therapeutics. I'll show a little bit of data in a murine study we did with a SOD1 proof of concept, but then we very rapidly moved to applying this to therapeutic targets, including tau, alpha-synuclein, and NLRP3. This utilizes a format relatively similar to the GCase.
In a lot of ways, the mechanism is similar between the enzyme and the siRNA because we're using ABC to drive it across the blood-brain barrier, but also into a cell type of interest. This requires a certain flavor of TFR, relatively high affinity. It again allows us to use the silenced Fc to get a little bit better safety. In this case, on the siRNA cargo side, we partnered with an industry leader, Axolabs, to allow us to very rapidly develop siRNAs that are very highly potent, very highly specific. Again, obviously, the main point of using the antibody delivery of the siRNA is to get around the current brain delivery of siRNAs, which relies on inconvenient routes of administration, such as intrathecal.
All the data we'll be showing are with peripheral amounts of administration, such as IV and subcutaneous. In our proof of concept murine study, we did a multi-dose IV study. This was modeled after the OTV study in Barker et al. 2024. Multiple IV doses compared to an ICV dose. It's not labeled on the side, but the ICV dose siRNA also included the C16 modification to make sure that we were doing, you know, an apples-to-apples comparison and showing the strongest brain knockdown. In every brain region we tested, the IV-dosed, ABC-enabled SOD1 siRNA actually showed stronger brain knockdown than the ICV-dosed siRNA with the C16 modification. We thought this was pretty good proof of concept.
Again, we at the same time were developing anti-tau, anti-synuclein, anti-NLRP3 siRNAs that are, as you see on the left side, highly potent. This has a IC50 of about 21 picomolar in a transfection model. It knocks down all the different isoforms of tau in a neuronal setting. It is also active in Sino cells. On the right side, you can also see activity in neurons in a passive uptake. The first three on the left are with the unconjugated siRNA. The one on the right side is actually with the final conjugated molecule with, in this case, using the ABC to drive the cell uptake instead of using an artificial transfection type system. The molecule is active in both settings. When we look at, again, performance in the NHPs, we're able to see very strong accumulation.
This is a repeat dose, 3 x 3 mg per kg siRNA equivalent study, modeled after an Arrowhead study. In this case, we're seeing uptake of 40 to 130 nM of siRNA into different brain regions. This is one of the advantages over, you know, other delivery routes such as intrathecal. When you dose siRNAs intrathecally, you can see very high local siRNA levels in some brain regions, but you can see a difference of 10- to 30-fold in other brain regions. In our case, because we're delivering through the capillaries, you see a much more widespread biodistribution. This leads then to a knockdown of the tau mRNA of up to 70% in different brain regions. Looking at how mRNA level knockdown corresponds to what we're actually looking for, which is protein level knockdown.
On the left-hand side, you can see knockdown of phospho-tau 217, which is one of the toxic tau species now at the protein level. We're seeing knockdown of 43%-64% at day 28. Day 28 is probably not where we're seeing the full knockdown effect because if you know the tau system, the turnover of the protein itself is around 20 days-22 days. You're starting to see the protein being knocked down here. On the right-hand side, we can see some CSF tau data. Obviously, the CSF tau is a lagging indicator, but as we took the study out longer, out to day 49, out to day 70, we see up to about a 50% decrease in the tau at the protein level in the CSF. Then as well in this study, we did both therapeutic dose.
We also did a very high 30 mg per kg siRNA equivalent tox dose, again, to move this molecule forward as an actual therapeutic. Even at the high tox dose, we saw no adverse events. We saw no effects on hematological tox parameters such as reticulocytes or red blood cells. Again, this molecule is very well tolerated at doses up to 30 mg per kg siRNA, which when you convert to a total dose, is very, very high. This is a really safe platform, we think, for delivery of siRNAs. We also saw no increase in liver enzymes or inflammatory proteins, so very safe across the board, and I believe I've left exactly five minutes for questions.
All right. Well, thank you so much, Eric. Lots of really interesting science here. Maybe we'll address that with a couple of questions, and then I do wanna ask a couple on the clinical side as well. You know, you've got two or three cargos here. One, an antibody, which, you know, can bind via, you know, one of two catalytic arms. You've got an enzyme which kind of has small molecule, you know, catalytic activity, and then you've got an siRNA which targets within, you know, like intracellular nucleic acids. All of that is limited by brain penetrance. How are you thinking about dosing across three very different modalities and how does that, you know, how might that look from a potential, you know, you mentioned a sub-Q goal with the tau antibody?
Sorry, with the amyloid beta and then, yeah, and then, in terms of the other platforms as well.
I should've said, subcutaneous administration is a goal for all of these programs. For the A-beta drug, I think we have a very clear path to that. We've already mapped out, Giacomo's team has done some modeling studies to show we should be able to get sufficient coverage of our drug even at doses well below the 3 mg per kg dose that we showed here, and that would give a total dose level that's very amenable to subcutaneous administration. For, say, the GCase cargo, it is still a goal. We've worked, you know, even though it's an engineered enzyme, we are able to concentrate it quite high, but the ability to dose it subcutaneously will really depend on the final efficacious dose that we need to move forward into the clinic.
As well for the siRNAs, we're working hard to enable a subcutaneous dose.
Gotcha. Then maybe I might ask Giacomo a question, if you don't mind. Please. You know, in terms of maybe the path towards the IND and the plan to submit that IND in potentially Q4, maybe Q1 of 2027, what does that timing look like? How many patients are you thinking in a dose escalation and expansion arm and, you know, maybe what you might present as an initial readout?
Sure. Thanks. As you said, the timeline to IND is Q4 this year or Q1 2027, depending on the availability of the clinical supply. We plan to start in healthy volunteers with a single ascending dose study, which will also have a subcutaneous arm. In this subject, we plan to enroll our standard number of subjects in any other SAD. We want to quickly pivot to the MAD that will be done straight in patients with the early AD, so MCI and mild AD. An important readout will be the degree of amyloid clearance as measured with amyloid PET. We know this to be a very sensitive marker and with an effective drug, we believe we are gonna be able to show an effect only with 10 patients - 15 patients.
The MAD part of the study will heavily use the subcutaneous formulation because the goal of the phase 1 program is to be able to show decreased amyloid in the brain in patients with early AD, with subcutaneous delivery without significant ARIA above the background level of ARIA and without clinically significant anemia. That's the goal of the overall phase 1 program, and we think we can execute it fairly quickly. We know where to go in terms of the sites. We have done already on many other studies in Alzheimer's disease, so we are looking forward to starting the studies in healthy volunteers and patients in 2027.
Makes sense. How long is the follow-up?
The follow-up will be pretty standard. I think, you know, meaning that, if you look at donanemab data, they're able to show almost maximal amyloid clearance by 6 months and significant degree of amyloid clearance as early as three months, and, you know, we the studies will be long enough to show maximal effect in the double-blind portion of the study, and this will be followed up by an available extension, where we keep collecting biomarkers as well.
Thank you. Maybe one more in the last minute. You have these, interim analysis, futility analysis for the AL101 candidate coming up. you know, it's either happened already or it'll happen soon. Is that going to come with a data readout? What are we gonna see from that? If the futility analysis is in H1, is that going to be early H2 that we're gonna see that or is that still, you know, around the May-June timeframe?
I start talking about the timing. The interim analysis will happen in first half of this year, and we will update when it happens. This will be done, the interim analysis, by an independent data monitoring committee who will look at clinical efficacy measures as well as biomarkers according to pre-specified futility criteria. The outcome of the interim analysis will be binary, meaning the recommendation to stop the study for futility as the pre-specified criteria have been met, or the recommendation to continue the study as planned, as specified in the protocol, until to completion, until the last patient out. That is planned for the end of 2026. We will have no visibility on the magnitude of the effects or on what drives actually the decision to continue the study as planned, if that's the case.
If the study is deemed to be futile, independent data monitoring committee will communicate to the sponsor the recommendation for futility, and then we'll stop the study and look at the data in detail.
Okay. Thank you very much. Thanks for joining us.