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

Good morning, ladies and gentlemen, and welcome to Alector's conference call and webcast highlighting its progranulin franchise and Alector Brain Carrier programs. [Operator Instructions] Now I would like to turn the call over to Katie Hogan, Senior Director of Corporate Communications and Investor Relations. Please go ahead.

Hello, everyone, and welcome to our event. Before we begin, I will go over a few housekeeping reminders. There will be a moderated question-and-answer session following prepared remarks. [Operator Instructions] The webcast replay of this event will be available tomorrow after 12:30 p.m. Eastern in the Investors section under Events and Presentations on our website, www.alector.com. I'd like to note that during this event, we'll be making a number of forward-looking statements, and you can find our disclosure here. Turning now to the agenda. We'll begin with an overview from Dr. Sara Kenkare-Mitra, our President and Head of Research and Development. She'll share Alector's perspective on our multistage pipeline and our approach to driving value in treating neurodegeneration. Sara will then provide a review of our progranulin elevating franchise in FTD-GRN and Alzheimer's disease, highlighting latozinemab and nivisnebart, formerly AL101, which we are developing in collaboration with GSK. From there, our Chief Executive Officer, Dr. Arnon Rosenthal, will discuss advancements in Alector Brain Carrier and introduce our lead candidates for Alzheimer's and Parkinson's. With that, I'll now turn it over to Dr. Sara Kenkare-Mitra. Sara?

Thank you, Katie. Alector is dedicated to developing first and best-in-class disease-modifying therapies for neurodegenerative diseases with urgent unmet needs. We're building an integrated biotech that brings together expertise in genetics, immunology and neuroscience with deep capabilities in discovery, development and manufacturing. Our therapies are designed to address the root causes of disease. Our 3R strategy aims to remove misfolded proteins, replace deficient proteins and restore dysfunctional immune cells and neurons. This strategy is powered by advanced technologies, including our proprietary Alector Brain Carrier, or ABC, which improves therapeutic delivery across the blood-brain barrier. With our clinical late-stage progranulin elevating antibodies in partnership with GSK, a growing pipeline of ABC-enabled programs, strong cash resources and an experienced leadership team, we are positioned to drive both near- and long-term value. Our progranulin elevating programs form the foundation of our late-stage portfolio. Latozinemab in development for FTD-GRN has received breakthrough therapy, fast track and orphan drug designations, and top line Phase III data are expected by mid Q4 of this year. Alongside it, nivisnebart in Alzheimer's disease is fully enrolled in a Phase II trial. In parallel, we have selected lead candidates for our ABC-enabled anti-amyloid beta antibody program for Alzheimer's disease, and our ABC-enabled GK's enzyme replacement therapy for Parkinson's disease. We are also progressing ABC-enabled siRNA programs for TAU, alpha-synuclein and NLRP3 for multiple neurodegenerative diseases. This portfolio provides an important near-term milestone within our late-stage programs. It also includes the potential for the initiation of first-in-human trials from our ABC-enabled pipeline in 2026 and 2027. Let me now turn to our GSK partner progranulin elevating franchise, which targets frontotemporal dimentia caused by granulin mutations and Alzheimer's disease. Frontotemporal dementia, while Alzheimer's disease is the most common form of dementia, you may be less familiar with frontotemporal dementia due to a granulin mutation, or FTD-GRN. FTD-GRN is an aggressive early onset dementia. Patients often present with compulsive behavior, lack of restraint, apathy, anxiety or aphasia. Tragically, life expectancy is typically less than 10 years, and there are no approved treatments to slow or cure the disease. Heterozygous loss of function mutations in the granulin gene reduce the levels of a protein called progranulin by 50%, directly causing the disease. In partnership with GSK, we are developing latozinemab, which is designed to elevate progranulin levels back to physiologic levels. FTD is rarer than Alzheimer's disease, but it is the most common cause of dementia in individuals under the age of 60, and most cases occur between the ages of 45 and 64. In the U.S., the prevalence is estimated to be approximately 50,000 to 60,000 people. And in Europe, the number is closer to 110,000. FTD-GRN accounts for approximately 5% to 10% of all FTD cases and represents about 8,000 to 17,000 cases in the U.S. and EU alone. Importantly, the overall economic burden per patient for FTD is nearly twice that of Alzheimer's disease, underscoring the urgency of developing effective therapies. FTD-GRN is frequently misdiagnosed as Alzheimer's disease, Parkinson's disease, Lewy body dementia, vascular dementia or unspecified dementia, as you can see in the bottom half of the table. This underscores the importance of genetic testing to ensure patients are properly identified and can access future therapies. As we have seen in other therapeutic areas, once disease-specific treatments become available, the uptake of genetic testing typically increases. That means the landscape for FTD-GRN diagnosis could shift meaningfully as potential therapies move closer to approval. Both latozinemab being developed for the treatment of FTD-GRN and nivisnebart, or AL101, being developed for the treatment of Alzheimer's disease are designed to increase progranulin levels by blocking sortillin, a receptor that binds progranulin and directs it to the lysosome for degradation. Progranulin encoded by the GRN gene is a secreted glycoprotein which is primarily expressed in neurons and microglia within the central nervous system and has several activities, including being an immune and neurotrophic factor. Human and mouse data shown on the left side of the slide demonstrate that higher sortillin levels mean lower progranulin. By inhibiting sortillin, our antibodies increase extracellular progranulin levels. The rationale for progranulin elevating drugs in FTD is clear. On the left, we show the 50% reduction in plasma and CSF progranulin in granulin mutation carriers and patients compared to healthy controls. This deficiency triggers a neurodegenerative cascade, neuronal cell death and microglial dysfunction leading to TDP-43 accumulation, lysosomal impairment, complement activation and inflammation. These processes result in destruction of brain structures, ultimately driving the cognitive and behavioral deficits seen in FTD. Latozinemab is designed to elevate progranulin levels in the brain, addressing the underlying deficiency that contributes to neuronal loss, inflammation and cognitive and behavioral deficits in FTD-GRN. The rationale for progranulin elevating drugs in Alzheimer's disease is also compelling. Human genetics show that loss of function mutations in progranulin increase Alzheimer's disease risk, while preclinical studies demonstrate that elevating progranulin can be protective. On the left, human genetic data highlights granulin as a risk stream for Alzheimer's disease. In the center, immunohistochemistry shows progranulin embedded within amyloid beta blocks in Alzheimer's disease brain tissue, emphasizing its association with disease pathology. And on the right, data from AD mouse models shows that increasing progranulin improves disease-relevant outcomes. These findings together provide strong genetic, pathologic and preclinical support for evaluating a progranulin elevating drug, nivisnebart, in Alzheimer's disease. The genetic and biologic rationale for blocking sortillin to elevate progranulin in FTD-GRN is grounded in multiple lines of evidence. Loss of function mutations in SORT1 lead to chronically elevated progranulin in humans and mice with minimal or no discernible adverse effects, providing a strong genetic foundation for this approach. Biologically, progranulin that does not bind sortillin can still enter lysosomes through alternative receptors, where it remains partially active and supports neuronal survival. You can see this in the images on the upper left. Even in SORT1 deficient cells shown in the bottom panel, progranulin continues to traffic to lysosomes, similar to wild-type cells in the top panel. The schematic next to it illustrates how multiple receptors facilitate lysosomal entry in the absence of sortillin. Importantly, progranulin that is mutated to bypass sortillin appears more potent than wild-type progranulin in rescuing microglial pathology, reducing NfL and correcting lipid abnormalities in mice. Experimental data further support this strategy. The graphs on the bottom left show results with nivisnebart, our antibody that blocks sortillin to elevate progranulin. In this first panel, nivisnebart nearly eliminates functional sortillin, confirming target engagement. This is accompanied by significant increases in progranulin levels in plasma and CSF, as shown in the middle and right panels. And these findings demonstrate that blocking sortillin effectively elevates progranulin and restores key functions of progranulin. Together, these genetic, biological and experimental data establish a compelling case for sortillin -- targeting sortillin to increase progranulin and address neurodegenerative disease mechanisms. It's important to note that latozinemab and nivisnebart target distinct regions or binding epitopes on the SORT1 protein. The PK/PD profile distinguishes nivisnebart from latozinemab. While latozinemab is being developed to be a treatment for FTD-GRN, nivisnebart properties could make it suitable to address a broader spectrum of neurodegenerative diseases, including Alzheimer's and Parkinson's disease. Both latozinemab and nivisnebart have demonstrated a two to threefold increase in progranulin levels and have been generally well tolerated in clinical trials to date. Turning now to clinical data in our progranulin elevating antibodies. I will begin with our INFRONT-2 Phase II study of latozinemab, which was designed to gather data on safety, PK/PD clinical outcomes and biomarkers. In the trial, 12 symptomatic FTD-GRN patients were treated with latozinemab at 60 mgs per kg every 4 weeks for 49 weeks. To determine whether there was treatment-related slowing of disease progression, we used historical data from the GENFI2 and ALLFTD noninterventional registry databases to generate a matched control cohort that would allow us to make comparisons to FTD-GRN participants in our open-label interventional cohort. In INFRONT-2, we evaluated 3 areas: target engagement through progranulin levels in plasma and CSF; biomarkers of disease activity, including lysosomal function, inflammation, brain health and atrophy, as well as clinical progression measured with CDR plus NACC FTLD, a tool designed for FTD. We'll now review some of the key findings. In INFRONT-2 study, latozinemab increased progranulin in plasma and CSF two to threefold, restoring levels to those seen in healthy controls and sustaining them over 49 weeks. Also, plasma and CSF concentrations were strongly correlated, supporting the use of plasma progranulin as an important biomarker for use in clinical trials. Here, we show the INFRONT-2 data for GFAP, or glial fibrillary acidic protein, a marker of astrogliosis that is elevated in symptomatic FTD-GRN and correlates with disease severity. These figures show that GFAP levels are elevated in symptomatic FTD-GRN patients at baseline and that at treatment, GFAP levels decline over the course of the study towards the range seen in asymptomatic GRN mutation carriers. We also looked at disease progression using the CDR plus NACC FTLD, a measure of clinical progression in FTD agreed upon by the FDA and EMA. In matched controls from the GENFI2 registry, patients declined by 6.4 points over 12 months. By contrast, participants treated with latozinemab declined by 3.1 points over the same period, which represents an estimated 48% slowing of clinical progression compared to matched historical controls. In INFRONT-3, our pivotal Phase III trial, we are measuring the same clinical measures and core biomarkers that we assessed in our INFRONT-2 Phase II trial. INFRONT-3 is a 96-week randomized double-blind placebo-controlled global trial evaluating latozinemab in 103 symptomatic and 16 at-risk individuals with confirmed GRN mutations. Participants received 60 mgs per kg of placebo via intravenous infusion every 4 weeks. The primary analysis will be conducted in symptomatic participants, and we plan to include at-risk participants in a sensitivity analysis. The clinical co-primary endpoint is the CDR plus NACC FTLD sum of boxes. And in the U.S., the primary -- the biomarker co-primary endpoint is plasma progranulin. Additionally, we are measuring key secondary outcomes, assessments and collecting fluid and imaging biomarkers, including plasma NfL, GFAP and volumetric MRI. We believe this positions us to deliver a clear and well-aligned data package later this year. Our Phase II study of nivisnebart in early Alzheimer's disease is ongoing. In partnership with GSK, we completed enrollment of that study in April. An independent interim analysis is expected in the first half of 2026. PROGRESS-AD is a global, randomized, double-blind, placebo-controlled Phase II clinical trial enrolling patients with early Alzheimer's disease. The study is designed to assess the safety and efficacy of 2-dose levels of nivisnebart compared to placebo. Participants are randomized to receive nivisnebart or placebo intravenously every 4 weeks for the duration of the 76-week trial. The primary endpoint of the study is disease progression, as measured by the clinical dementia rating sum of boxes, or CDR sum of boxes. We are also measuring key secondary endpoints and biomarkers, including amyloid PET, Tau-PET and biomarkers in CSF and plasma. Our progranulin programs, latozinemab and nivisnebart, are being developed in partnership with GSK. This partnership included $700 million in upfront payments and includes a $1.5 billion in potential development and commercial milestones, a 50-50 U.S. profit share and tiered double-digit royalties ex-U.S. Potential milestone payments include $160 million for the first commercial sale in the U.S. and $90 million for first commercial sale in at least 2 of the EU countries. With that, I'll turn the call over to Arnon to discuss our preclinical programs enabled by the Alector Brain Carrier.

Thank you, Sara, and welcome, everyone. I will now guide you through to a tour to our imminent future, which is propelled by Alector Brain Carrier. Alector Brain Carrier is using the transferrin receptor on the blood-brain barrier endothelial cells to deposit and transport cargo to the brain. There are 3 features of Alector Brain Carrier that are worth mentioning. First is that our brain carrier module is completely independent. It can be placed in multiple places on the target. On the left here, you see replacing the brain carrier on the constant region of the antibody with a linker. The linker can be at any size or no linker at all. In the middle figure, you see us placing the brain carrier model as part of the FAB configuration. And on the right, on the bottom, you see that we can place the brain carrier as part of a single-arm antibody or as part of an FAB. Our brain carrier can be bivalent, as you see on the left bottom, or monovalent. So this flexibility really enable us to tailor the drug to the different drug modalities, be it antibody, enzymes or nucleic acid. The configuration of the drug really dictate there, the drug half-life, immunogenicity and brain penetration. So this is an important tool to optimize blood-brain barrier propelled drugs. The next feature of Alector Brain Carrier is the large variability or the large range of affinities that we can deploy. As you see on the left table here, we have almost 1,000 full range of affinities from 5-nanomolar to almost 5,000 nanomolars. And again, the affinity dictate hematologic adverse effects, drug half-life and brain penetrants. And this large range of affinity enable us to tailor the affinity to each drug modality and specifically within each drug modality to each drug and gives us a really unique tool to optimize ABC-propelled drugs. The third unique feature of our platform is the epitope that they are using to engage the transferrin receptor. Whereas many many blood-brain barrier technologies are using an epitope which we consider to be exposed on the transferrin receptor, we have identified a different epitope that is depicted on the left side of the structure of the transferrin receptor. And this epitope is between 2 lobes. And the uniqueness of these epitopes is that it is reducing the ability of ABC propelled drugs to induce transferrin antibody-dependent cytotoxicity. This is apparent on the right graph. And you've seen in the red line using the exposed epitope, what we call an epitope B, which many other companies are using. And this enables the antibody to facilitate ADCC that is transferrin receptor dependent. In contrast, when we use our own epitope, that's depicted on the graph in green and yellow here on the right, and you see a significantly lower ADCC. And this really provides a unique safety feature for our drugs. And we think that reduced ADCC would translate to reduce hematologic adverse effects and increased safety of our ABC module. With these ABC features, we are developing, as Sara mentioned to you, 3 types of drug modalities. We are developing antibody drugs that are propelled by ABC, enzymes that are propelled by our ABC, as well as multiple siRNA that are propelled by ABC. I will start by describing our antibody against the beta amyloid that's ABC enabled. Every aspect of our anti-A-beta antibody, which we designate as AL037, was engineered to optimize efficacy and tolerable pharmacokinetics and safety. So the anti-A-beta binding epitope was engineered to recognize the pyroglutamate version of the A-beta peptide. As was shown by Lilly, the targeting the pyroglutamate beta was the most effective in reducing A-beta plaque rapidly. And there is now A-beta pyroglu in the circulation in the serum. So this is minimizing sync effects and retention of the antibody in the periphery. We undertook or chosen to retain a fully function -- fully functional effector function. We think that recruiting myeloid cells through the effect of functions is critical to fully remove beta-amyloid plaques and that effector function cannot be substituted by the transferrin receptor or by a crippled effector function. So we are targeting to maximize removal of A-beta plaques with a full effector function. We then added to this drug, our optimized ABC technology, which is designated to maximal brain penetration and minimal hematologic adverse effects as well as maximize pharmacokinetics in the serum. And I will now show you some data with this drug. The first thing that we did is to confirm that AL037 can really transcytose through brain endothelial cells. And on the left side, you see a cell culture assay with brain and endothelial cells where you can put the drug on the top on the epical side, the blood side, of the culture. And then you can measure how much of the antibody can move to the basolateral sites, the brain side. You can see in the middle panel here that naked AL037, this is an antibody that does not have the ABC module, that is not able to cross through the endothelial cells. In contrast, AL037, which does have the ABC module very readily being transduced through the endothelial cells. And this is quantified on the graph on the right. A very significant part of the antibody reaches the basolateral of this culture -- of this transwell culture, suggesting that it will be able to translate those to the brain in vivo. The next thing that we confirmed was the ability of AL037 to phagocytose A-beta peptides. And on the left side here, you see images of human microglia phagocytose in fluorescent A-beta and internalizing it. And again, you see that AL037 is very effective in phagocytose in A-beta. On the right side, there is a quantification of this phagocytose is efferent. You see that naked AL037 in orange here can phagocytose A-beta plaques but the blue, which represents AL037 with the ABC module can phagocytose's A-beta plaques even better. This suggests that ABC module can contribute to phagocytosis, but it's still incremental and you still require a full effector function recruiting the A-beta gamma receptor for full phagocytosis. We next move in vivo. We first tested AL037 in mouse models of Alzheimer's disease. This is the 5xFAD mice. You see on the left picture, this is a light microscopy. You see that the naked antibody primarily get stacked in the ventricle, in the large blood vessels and is not distributed well into the brain. In contrast, the 3 right graphs shows antibodies that are enabled by our ABC. And in all cases, you see that there is no longer stickiness into ventricles and blood vessels, there is homogeneous distribution in the brain. On the 2 right panels you see that our antibodies can detect A-beta plaques very well and really bound to them. The most right pictures show a 3-dimensional constitution. You see antibodies recognize practically all the plaques in these brains or many of the plaques and can sort of recognize and bound to them. AL037 not only bound to plaques, it also stimulate microglia dependent removal of plaques, and this is seen on the graph on the right, you see that even after 3 to 4 injection of the antibody, in a very short time frame, you see significant reduction in the level of A-beta 42 in the brain. So antibody at fairly low doses is able to penetrate the brain, distribute homogeneously and remove A-beta 42 in a very short duration. With the mouse data at hand, we advance to nonhuman primate studies. The first thing that we looked at was the half-life of AL037, and you see in 2 doses, 30 mg per kg and 3 mg per kg. We see that the half-life of AL037 which is linked to our ABC is about 106 hours. This is significantly better than reported. Brain shuttle linked antibodies, and it's actually not far off from a naked antibody like a normal antibody. So we do show a very decent half-life of the antibodies in nonhuman primates. The next thing that we looked was -- at was hematologic side effect. As you know, there is 10x more transferrin receptor expressed on the reticulocytes compared to endothelial cells. So all brain shuttle technologies that are using transferrin receptors have an inert risk of hematologic side effects. So we measure that. As you see on the left graph, we do see transient reduction in reticulocytes, but the reticulocytes recover very quickly within a day. And even after 2 injections at day 1 and day 8, we don't see a meaningful reduction in red blood cells in the middle graph or in hemoglobin in the right partner. So we see a very tolerable safety profile for AL037. We next looked at brain penetration in the nonhuman primate. And to our delight, we saw very potent brain penetration. As you see on the left panel in the frontal cortex, even at 3 mg per kg, which would be around the range of probably clinical doses, we see 18-fold, an elevation in brain penetration, and we see it in every brain region that we tested. And here, we show just the front of cortex on the right. On the left, and hippocampus on the left, but there is a homogeneous distribution of AL037 in the nonhuman primate brain. If you calculate the molarity of brain level, you see that our AL037 even at 3 mg per kg can reach a 3.8 nanomolar in the brain. And this is, according to our calculations, at least 5x higher than what was reported for other brain shuttle antibodies that are currently in the clinic. So we see a very potent high concentration in the brain that suggest that this antibody would be very effective at removing A-beta plaques in human. Given the significance of Alzheimer's disease as an unmet medical need, we are developing a second anti-A-beta antibody, which we designate as AL137. This antibody is deploying a different ABC modality with high affinity. And we see that in this case, we can achieve 32-fold increase in brain level. This 32-fold increase can translate to 8.4 nanomolar even at 3 mg per kg in the brain. And this is over 12x higher concentration that what according to our calculations was reported for clinical brain shuttle anti-beta antibodies. So we have, again, an even more potent anti A-beta antibody, which have still very, very good pharmacokinetics and tolerable safety features. To summarize what I've shown you on our anti A-beta programs, we have 2 anti-A-beta antibodies, AL037 and AL137, that target that pyroglu anti-A-beta peptide, which in our view and as was demonstrated by Lilly, is the most potent epitope for removal of A-beta plaque. We retain a fully active Fc region to enable full function of recruitment of myeloid cells to remove A-beta, and we have an optimized ABC module that can lead to high brain penetration, but still good hematologic safety and good pharmacokinetics. So we think that we have 2 uniquely important and safe antibodies that could really be best-in-class in anti-A-beta therapeutics. And we are targeting first in human in 2026. The next problem that I will describe is our brain carrier-enabled GK's enzyme replacement therapy for Parkinson's disease and eventually Lewy body dementia. As you know, GKs or GBAs, the lysosomal enzyme that remove toxic lipids like glucosylceramide and glucosylsphingosine form cells. And if you don't have functional GKs, these toxic lipids accumulate and cause diseases. There are up to 10 million to 15 million people that have Parkinson's disease that carry the GKs mutations. This translates to up to 1.5 GBA mutation carriers with Parkinson's that are up to 2.4 million GBA mutation carriers that suffer from Lewy body dementia. And there is -- there are over 100,000 gaucher disease patients with the GBA mutation. There is -- as you know, currently, there is enzyme replacement therapy for gauche disease which display peripheral symptoms, but there is no enzyme replacement therapy for Parkinson's disease or Lewy body dementia because current enzyme replacement therapy cannot enter the brain. And even if it enters the brain, the way it enters cells that does not allow us to enter, does not allow it to enter as nerve cells or myeloid cells in the brain. Even though we are starting with a genetic mutation carrier, there is really good evidence that GK's enzyme replacement therapy could be also beneficial for the sporadic form of Parkinson's disease and likely Lewy body dementia. The reason is that if you look at all-comer Parkinson's patients, if you measure the level of GK's activity versus the rate of progress disease progression, you see in the yellow line on the left graph that people that have high level of GKs activity show low progression rate, whereas people that show low level of GKs activity in the dark line display very high progression rate. And consistent with the level of enzymes, you see on the right graph here, in blue, people that have low level of the toxic lipids show slow progression rate, whereas people with Parkinson's, this is regardless of whether they carry the genetic mutation or not. People that have high level of the toxic lipids display very happy progression. Right? So this really suggests to us that even in sporadic form of Parkinson's disease, GK's brain penetrant and GK's enzyme replacement therapy could be beneficial to slow disease progression. So as we did with AL037, we did for AL050, we engineered every component of the drug. We first engineered the enzyme itself, the wild-type GKs enzyme has very short half-life. It is very unstable and it's hard to manufacture. So we engineered GKs that is almost 30-fold more stable than wild-type item. You can see it on the table on the left bottom, you see the wild type enzyme has a half-life at 37 degrees of 6 hours versus 7 days of our enzyme. Likewise, we engineered the enzyme to have almost 50-fold higher activity. You see this is a logarithmic scale. You see the activity of the wild type enzyme versus AL050. So we designed a very potent and stable engineered GKs as the first component of the drug. We then linked it to an optimal ABC that has a higher affinity that enables faster removal, a fast transport to the brain. And as I'll show you also, transport to cells and to the lysosomes in brain cells. And finally, because we don't need to recruit immune cells, we silenced the effector function of this drug to minimize or eliminate hematologic related adverse effect. So we took this drug initially to testing in cell culture. And the first thing that we did was to confirm that AL050 can enter lysosomes in cells. As you may know, the peripheral enzyme replacement therapy is using macrophages, mannose receptor to enter cells. So current enzyme replacement therapy for GKs, the protein can only enter macrophages in the periphery. It cannot enter nerves cells, and it cannot enter microglia cells or other cell types in the brain. So first, we wanted to see whether we can substitute the mannose receptors with the transferrin receptor. And we saw that this is indeed the case. You see here that our GKs that is propelled by ABC can enter the lysosomes and you see overlay of staining -- fluorescent staining with lysosomal markers like LAMP1. So on the left side, we show that our GKs is propelled by our ABC can enter cellular lysosomes, and these are neuronal cells that the naked -- the regular enzyme replacement therapy for GKs will not be able to do. On the right side, we also looked at activity of the GKs in this lysosome. And we see that there is very potent GK's activity in this GKs-deficient neuronal cells. And the level of activity is transferrin affinity dependent. So -- and the higher affinity the ABC model, the higher the activity. And you see that with affinity that we are using in the -- for the clinical program, we are easily exceeding the wild type level of GKs activity, which is marked in the gray horizontal bad on the right graph. So based on the cell culture activity, we found out that we have a very potent and active enzyme. And we have a drug that can enter lysosomes in neuronal cells and can retain activity in neuronal cells. With this information, we took our drug to nonhuman primate. We first test the half-life of our drug in the nonhuman primate plasma. And as you see on the left graph, AL050 display half-life of 5 hours. And this is compared to current enzyme replacement therapy that's sort of illustrated on the right side of this slide, current enzyme replacement therapy displays a half-life of less than 30 minutes. So in this experimental part in at least AL050 display tenfold higher, longer half-life in the plasma compared to current enzyme replacement therapy. We then looked at the enzymatic activity in the nonhuman primate plasma. And we see on the left side here that AL050 delivered at 10 mg per kg peripherally, display half-life of activity of 6.6 hours. And again, this is almost 40-fold longer than enzymatic activity that was reported for current GK enzyme replacement therapy, where the serum half-life of enzymatic activity is less -- is about 10 minutes at the maximum even in human. So at least with these experiments, we show that both in, in vitro biochemically in cell culture and nonhuman primate, we have a stable and an active drug that retain enzymatic activity extracellularly and will -- and sort of this will allow time for the drug to enter the brain and possibility retain activity in the rain. And these are the things that we looked at next. And before we did that, we sort of looked at the safety of AL050. Again, all brain shutters that you're use transferrin as the Trojan horse have a risk of hematologic side effect. So we wanted to make sure that this is not the case with 050. You see on the left graph here, AL050 does not lead to any meaningful reduction in reticulocyte count after 2 injections. Likewise, there is no effect on red blood cell count and no effect on hemoglobin in this experiment, suggesting that AL050 will be safe with regard to hematologic adverse effects. We next looked at what happens in the brain. So the first thing that we looked at was whether our ABC propelled AL050 can actually enter the brain, basically cross the blood-brain barrier and enter tissues in the nonhuman primate brain. And we see that this is indeed the case. You see that in every brain regions that we looked at, the frontal cortex, hippocampus, as well as brain regions that are relevant for Parkinson's disease and Lewy body dementia, the substantial nigra where the dopaminergic cell, cell body design and the putamen where the dopaminergic and nerve endings are, we show very good level of AL050 that's sort of somewhere between 6- and 20-fold elevation of enzyme. We then looked at enzymatic activity. And again, for enzymatic activity, our AL050 has to enter the brain, has to enter neurons and support us in the brain, has to enter lysosomes in these cells and then has to retain activity. So this is a very significant demand from drug, and we were not sure that this would happen. But to our sort of delight, it did happen. And you see that in all brain issue that we looked at, again, including the substantia nigra and the putamen, which are relevant, brain regions for Parkinson's disease and Lewy body dementia, we see at least twofold elevation in GK's activity. Parkinson's patients and Lewy body dimensions have modest reduction in the GK activity, somewhere between 15% and 50% reduction. So doubling the level of GKs activity would be more than sufficient to fully repair enzyme deficiencies in these diseases. And moreover, we think that there is some negative regulation feedback on lysosomal enzymes. And so in healthy brains like in the nonhuman primate brain, high level of GK enzymatic activity could lead to reduction in the level enzymatic activity of the endogenous GKs. So we think that in patients, this level could be even higher. Although, again, as I mentioned, twofold elevation is significantly more than what is needed to fully restore GKs deficiency in Parkinson's patients and Lewy body dementia patients. We then wanted to see whether our drug can actually correct disease pathology. And for this, we went back to mice. These are mice that carry the -- one of the Parkinson's GBA mutations, the D409V changes. And this mutation reduces GK's activity by 85%. So on the left bar graph, you see the level of GKs in wild type mice. This is in the gray. You see on the right side of the left panel you see that the mutation care -- the mouse mutation scares that were treated with PBS do not show hardly any GKs enzymatic activity. However, once we inject AL050, surrogate, we see almost complete restoration of GK's activity. And this is just 24 hours after 1 injection, you see almost complete restoration of enzymatic activity. Now we went further and looked whether the increase in enzymatic activity is also associated with reduction in toxic substrates, and this was done initially in the periphery in liver. I'll show you data in the brain in the next slide. So again, you see on the right panel, the right graph, wild type animals do not have toxic substrates because the endogenous GKs get rid of them. However, GBA mutation carriers have a very high level of toxic total, as you see from the blue bar graph here, but even sort of 2 injections of AL-050 that reduced these toxic substrates by almost 90%, suggesting that at least in the periphery here, our drug is very active, can enter cells and can restore enzymatic activity and can remove the toxic lipids that cause the disease. We then wanted to see if the same thing can happen in the brain of this GBA mouse mutant. And what we found out is that this is indeed the case. You see on the left panel of bar graphs, this is still a wild type mice that carry the human transferrin receptor, but they have a wild type level or normal level of GKs. You see with PBS, this is the in gray. This is the normal level of GKs. But if you inject AL050 surrogate, you almost doubled the level of GK's activity in the mouse brain. We then went to GBA mouse mutants. Again, in the middle graph, you see that wild type mice injected with PBS does not show any toxic substrates because the endogenous GKs take care of these. In contrast, mutant mice show high level of toxic substrates, as you see from the blue bar and single -- or 2 injections of AL050 surrogate reduce the toxic substrates by 80%. And again, the enzymes will continue to work. So 2 injections is a very short and acute time line. We think that over time, the reduction will be even more profound. Although 80% reduction in the toxic substract we think could be significantly therapeutic. We wanted to look at the durability of the effect on the right panel here, and we see that even after a single injection, the reduction in the toxic substrate can last for over 14 days, suggesting that once our enzyme getting to the lysosome, it retains activity for multiple weeks, way beyond the residents in the plasma. So just to summarize what I've told you, we have engineered AL050 to have multiple beneficial drug features. It has an engineered GKs that is 30-fold more stable and 50-fold more active than the wild type GKs that shows long half-life and long enzymatic activity in the nonhuman primate serum. We showed that AL050 can enter the nonhuman primate brain, enter nerve cells and support cells in the brain, enter the lysosomes and retain enzymatic activity. So we think that we have a pretty potent drug that could be beneficial for, again, Parkinson's patients that carry the GBA mutations, Lewy body dementia patients that carry the GK mutations and eventually sporadic forms of these diseases. And we are targeting first in human in 2027. I will just now describe our progress with Alector Brain Carrier-enabled siRNA programs. As you know, siRNA is becoming a very successful drug modality. I think that sort of there are over 20 approved sort of siRNA or ASO drugs now. Most of them are for peripheral indications. There are 2 nucleic acid drugs that are approved for central indications for ALS and for spinal muscular atrophy. The issue for nucleic acid or siRNA for brain disorders is that you have to deliver it by IT or ICV injection, this is a surgical process that is not safe. It's not highly sort of scalable for large diseases like Alzheimer's disease. And also with IT or ICV delivery, the nucleic acid is not distributed equally throughout the brain. So some brain regions which are close to the injection sites receive more nucleic acid drugs, whereas regions which are further away from the injection site receive more deep in the brain and receive maybe insufficient drug. So we are undertaking to change that by enable peripheral delivery of siRNA using our ABC technology. So we screened very extensively, as you see on the second from the left column, we screened very large range of transferrin affinities with different types of linkers, with different drugs and types of drug modalities to identify a good ABC carrier for siRNA. And this was really enabled by the versatility of our technology for -- where we can put the ABC module everywhere practically on the drug. We can use linkers or not use linkers, we can use cleavable or noncleavable linkers. We can use different affinity and we can use different valency. And this has really enabled us to optimize the siRNA delivery. As you see on the third column from the left, we are able to achieve over 40-fold increase in siRNA delivery in the brain following peripheral injection. And the first thing that we did was to look at superoxide dismutase as a proof-of-concept strategy. And again, with siRNA to superoxide dismutase, we see with different ABC modules, we see 30- to 40-fold elevation of siRNA in the brain. And we also see, as depicted in the right column, very good long plasma half-life. We then wanted to see whether our ABC propelled siRNA enters the brain, enters cells in the brain and enter the cytoplasma of the cells and is really able to suppress, to downregulate siRNA. Again, these are very demanding requests from a drug. We inject the drug peripherally. It has to go through the endothelial sales, like the 2 membranes of the endothelial cells to go through membranes on nerve cells and support sales and to enter to the right subcellar localization to access the messenger RNA for, in this case, SOD and to down-regulate it. And to our delight, we saw that our drug is able to do that. If you see every brain region that we looked at, the thalamus, cortex, hippocampus, brainstem, cerebellum, spinal cord and striatum. In all cases, we see reduction in the SOD mRNA by up to 80%, and this is depicted by the purple and blue graph that represents -- as our peripheral injected siRNA with 2 different ABC modules. The purple graph represents ICV injected naked siRNA. And you see that in all cases, peripherally injected siRNA with ABC is at least as good as ICV injected siRNA. And in most cases, it's significantly better if you're seeing that cortex, in the spinal cord, in the brain stem, you see that our peripherally delivered siRNA is better than siRNA injected ICV. So this really tells us that we could peripherally deliver siRNA, achieve homogeneous distribution in the brain, convert surgical procedure of ICV or IT into either -- easily deliver the infusion centers or ultimately even at home delivery with IV or even subcutaneous delivery. So we think that there is a significant potential for this technology to increase ease of use, safety and also efficacy. With this technology, we are now developing 3 programs. We are developing TAU-siRNA with ABC for Alzheimer's disease and frontotemporal dementia that is caused by tau pathology. We are developing alpha synuclein siRNA for Parkinson's disease and Lewy body dementia. And we are developing NLRP siRNA for multiple neurodegenerative diseases. As you know, NLRP is an inflammatory mediator that was -- that is start to be involved in practically every known degenerative disease, from Alzheimer's disease, Parkinson's disease, ALS, Huntington disease. And so far, small molecules for NLRP were not as effective in the clinic partially because of off-target activities and partially because of incomplete blockade. So we think that all of these 3 programs have a very profound potential in very large diseases. Just to summarize what I've told you today and what you heard from Sara, we have a good mix of late stage and late preclinical programs at Alector. Our late-stage programs have a program in Phase III that is a pivotal Phase III where we received breakthrough therapy, orphan designation and fast track designation, and we will have data in the middle of Q4 2025. And because of the designations that the drug get, we think that we can proceed if the drug justifies that we can proceed to BLA rapidly and 2 commercializations. And we have commercial rights in the U.S., so we will be leading commercialization of this drug in U.S. as our partner will lead commercialization ex-U.S. Our second drug, as Sara told you, is in early Alzheimer's disease. It's sort of completed recruitment. It's a 76 in April, 76 week long trials. So if you calculate, you see that trial completion is in 2026, and we expect to have data shortly after. In addition to our promising late-stage programs we have, as I described to you, multiple late preclinical programs that involve antibodies that are propelled by our ABC. I described the anti-A-beta antibody, but we also are advancing anti-tau antibody. We have enzyme that are propelled by ABC, for Parkinson's disease, Lewy body dementia and eventually gaucher disease with neurological pathology. And we have an emerging platform of siRNA that are propelled by our ABC. These include currently siRNA for tau, for alpha-synuclein, for NLRP. But if successful, there are many more targets that would benefit from our ABC modules. Again, as Sara described to you, we are completely focused neurodegeneration, and we are going after the core of the disease by developing tools to replace damaged or mutated proteins with enzyme replacement therapy, removing these other proteins with antibodies and siRNA and restoring damaged neurons and supported by antibodies that stimulate signaling in these cell types. We have still over $300 million in the bank that will enable us to complete the 2 late stage programs and to take at least 2 of our preclinical programs to first in human. And we are targeting, again, to have multiple catalysts in post late and early-stage programs in the next 1 to 2 years. So thank you, everyone, for listening, and we will now open the webinar for questions.

[Operator Instructions] First question comes from the line of Pete Stavropoulos with Cantor Fitzgerald.

Thank you for hosting this event. Very informative, and great to see you all, this activity across the pipeline. First question has to do with in INFRONT-3 as you have the Phase III reading out and it's top of mind. Can you just discuss some of the conversations that -- and data that you provided to the FDA that helped enable progranulin as a co-primary endpoint? And the sense that you received that they could possibly lean on it for an approval if clinical outcomes were trending in the right direction, but not to that take?

Thanks, Pete. I think I'll have Giacomo address this question. Giacomo?

Yes. Thanks for the question. The ask by the FDA to move progranulin as co-primary endpoint came as part of the review of the statistic analysis plan that we submitted several weeks ago. We didn't submit any new data to justify this recommended change by the FDA. Previously, in 2024, we discussed the use of progranulin as confirmatory evidence in addition to a single pivotal Phase III study, and the FDA agreed upon, and this is change that stemmed from this original discussion. But again, no new data were submitted by the company.

All right. Have you previously shown them, the correlation between plasma? Because specifically, the co-primary endpoint is plasma progranulin versus CSF.

We hadn't shown the correlation, but we had presented as part of the breakthrough designation package, the results on relation of progranulin in plasma and CSF that were comparable in magnitude.

All right. And 1 question, please, on the amyloid-beta ABC program, first in human trial in 2026. And I know it's early, but not in clinic yet. But curious to hear how you're thinking about clinical development sort of leveraging the data generated by approved amyloid data antibodies and those in development that can cross the blood-brain barrier easily with the technology. What would a proof-of-concept study sort of look like? And what would you like to see to move forward into late-stage programs? Would it be just imaging data or clinical data? And when it comes to safety, there were very low levels of ARIA for trontinemab. Does that molecule -- do you know if that molecule has an active Fc region? And what are your expectations for your amyloid beta ABC, will that translate to low ARIA rates in safety translate?

Maybe Giacomo can address the first part of the question, particularly in terms of what we believe a proof of concept would be for our anti-amyloid beta ABC.

Sure. The most recent data presented by other companies have shown that using biomarkers such as amyloid PET, it's possible to derisk anti-amyloid treatments relatively faster relatively early in the development program. So the current thinking is to use a similar approach and use extensively biomarkers, amyloid PET, as well as the [indiscernible] species in plasma to have information about the target dose and have an early read on pharmacodynamic effects that are important for therapeutic benefit. Regarding the other question was about what we are expecting to see and will be. We will design studies to provide evidence of very meaningful and fast amyloid clearance but using both PET biomarkers as well as food biomarkers.

Yes. So just to add to this, there's a very clear path for anti-beta drug now, as Giacomo said, like we want to see profound reduction of A-beta plaques within 3 to 6 months, we want to see minimal ARIA. And Roche is also using a fully fund, so the full effect of function. So we don't think that, that will impact the ARIA. I think the ARIA is impacted by the way the ABC or the blood brain shuttle enters the brain. So we expect minimal ARIA, and we also are hoping to see a low level of infusion reactions, something that was not reported with sort of competing antibodies. So we think that there is very clear sort of proof of concept for a drug that can happen with a fairly small clinical trial fairly quickly.

Our next question comes from the line of Myles Minter with William Blair.

The first one on INFRONT-3 again. I know you've only enrolled I think it's 16 asymptomatic patients in INFRONT-3. It's not part of the primary analysis, but is there a path for those asymptomatic patients to get them on label if you do see positive data from INFRONT-3 in the symptomatic cohort? That's the first one. The second, I'm curious as to the comment that the progranulin efficacy on lysosomal function is actually enhanced if it gets up taken through LRP1 versus sortillin one. If you could just expand on that because that is a source of investor concern, that you wouldn't actually get active granulin concentrations in the lysosome itself if you block SOD1? So that's the second one. And the third one is just coming back to Pete's question, if you're going for fast amyloid removal with your amyloid antibody with the ABC carrier technology, is it safe to say that you'd want to see greater than 91% amyloid clearance at 6 months that we did see from the trontinemab data out of AAIC with the high dose?

Thanks, Myles. Maybe I'll address the first question and then pass it to Arnon to answer your progranulin biology question and to Giacomo on the anti-amyloid A-beta. In terms of INFRONT-3, as you asked, you're correct, Myles, that we have 16 asymptomatic carriers that we'll have data on. And as of now, we have not had any direct discussion with the agency in terms of label and what that data -- the impact of that data will be. As I said, we will be doing a sensitivity analysis based on data from these patients. And then depending on what we see, we would have further conversations on the label with the agency. I will pass it to Giacomo to speak about -- to Arnon to speak about the progranulin biology question first.

Yes, Myles. So this is a recent publication by an academic lab, where sort of they compared actually with sort of gene therapy progranulin. And that's mutated that does not bind sortillin and progranulin that can bind sortillin, and they show that the program -- like in mice that are deficient in endogenous progranulin, and they show that the progranulin that does not bind sortillin is actually more potent in multiple aspects. It can -- reduces neurofilament, whereas the wild type progranulin was not able to do that. It can reverse multiple lysosomal pathologies more potently that the wild type progranulin. It is surprising, but that's what they reported, and I'm happy to send you the publication. And sort of -- that's a sort of second publication from this group showing the progranulin that cannot bind sortillin can restore both intracellular pathology primarily, again, lysosomal pathology in microglia and overall disease pathology.

Giacomo, do you want to address the A-beta question?

Absolutely. So Myles, we are very pleased to see the results as Arnon presented about effects of AL037 in the nonhuman primates, which shows very strong brain penetration. We aim with this program to show evidence of fast and very meaningful reduction of amyloid in the brain, along with the safety profile, which is favorable, meaning very low frequency of ARIA, no significant concerns around anemia and potentially a lower number of infusion-related reaction. And as well as developing a compound lens itself to subcutaneous administration as part of the development program. Regarding [Technical Difficulty], we don't think that we necessarily need to be at a greater efficacy in removing amyloid and 91% that you cited with trontinemab. Alzheimer's disease is a very large population with a very significant unmet needs. And I think we -- I mean these are the goals of the program, not necessarily showing superiority, which I don't think is needed at this stage.

Our next question comes from the line of Tom Shrader with BTIG.

Thanks for the very informative session. I mean just to the last point, doesn't sortillin -- shouldn't a sortillin -- progranulin that can bind sortillin be more active, it's essentially stabilized. So doesn't it make complete sense?

Yes. In our mind, it does make sense. But as sort of Myles just alluded to, I mean, there is still raging debate in the field whether how essential sortillin is for the function of progranulin despite the fact that there are -- there is a wealth of evidence that progranulin can enter the cells without sortillin, enter lysosomes, retain functionality, both in cell culture and in mouse models in vivo. Despite our like Phase II data, the open-label data showing that all the parameters sort of going the right direction, cognition, biomarkers, imaging. So for us, it makes sense, but yes, it will -- hopefully, the actual Phase III data will convince the doubtful.

I get it. I get it. I get it. Okay. And then with all your work in shuttles, you must -- or it seems likely you've made either a shuttle direct progranulin or a shuttled sortillin receptor. Did either of them behave better in the gazillion assays you must be able to run? Or do you think 001 and 101 are, for you, going to be good enough forever?

Yes, we are sort of developing second-generation program. So everything that we do so far, we think that 001 and 101 are looking really good. Means the issue with enzyme replacement therapy for progranulin is controlling the dosing. Like progranulin is a growth factor. We have a built-in safety feature that we cannot elevate progranulin more than two to threefold, which we think is a sort of safe domain, like if you do enzyme replacement therapy, you can elevate progranulin in the periphery way more than that, and that could really lead to adverse effects. Means -- so we think that there is a safety issue like -- one of the companies that develop enzyme replacement therapy for progranulin had to sort of show significant adverse like immune-related adverse effects and have to use like immunosuppressive to deliver the drug. Our drug seems to be very well tolerated, to the point that we think it will be applicable for prevention therapy if the data justified that. So, so far, we think that we have a really good drug, but we are always developing second-generation trial. We think that the target is [indiscernible].

And then the last 1 on the siRNA, you gave a relative amount of delivery relative to systemic to 40-fold. But do you know you have enough? You kind of know how much siRNA you need outside of cells in primates or something very close to humans. Do you know you're getting enough siRNA in? And in some sense, are you out of the woods with PK for siRNAs?

Yes, in the rodents, all the experiments we've done with equality with the ICV and peripheral injection. So we think that we get -- and we see over like up to 80% suppression of the SOB mRNA. So we think we are getting enough siRNA to the brain. We are sort of in the means of sort of not even primate studies, we'll see how it's translatable. But so far, what we see is that we can get enough siRNA both to peripheral tissues and to the brain.

Our next question comes from the line of Yaron Werber with TD Securities.

Great event. This is Steven on for Yaron. A few questions here. So for the anti-amyloid beta ABC program, you mentioned first in human by 2026 and you discussed 2 molecules, AL037 and 137, which seems to be new. And you highlighted a few differences in the half-life. So what accounts for those differences? Is this -- are these 2 different molecules with different epitopes or is this same antibody with a different ADC carrier that might affect the PK/PD? And then secondly, just to clarify, are you still deciding between 037 and 137, which look pretty similar? And how would you make the decision of what to take into the clinic?

Arnon, go ahead.

Yes. Yes, the difference between the 2 antibodies is the ABC module, I mean, they have different affinity and the affinity dictate both the sort of level of brain penetration and the pharmacokinetics. So the difference in pharmacokinetics is sort of target dependent clearance in the periphery. So they have different pharmacokinetics. It's not clear what's the optimal peripheral pharmacokinetics for anti-beta antibody is. But we think that we have 2 very, very potent and promising anti A-beta antibodies that are driven by 2 variants of our ABC. At this point, we are advancing both of them. And if there will be -- we are continuing to explore if there will be data that will clearly point to 1 of them, we will advance one. But at this point, we think that both of these antibodies are very promising. They have different features that are worth examining, and we are committed to advance both of them at this point.

Understood. And maybe finally, you mentioned an interim data update on the business and PROGRESS-AD in the first half of next year. What can we expect to see from that in terms of patients followed or a time of follow-up?

So I can respond to that. The interim analysis planned in the first half of 2026, we haven't actually diverged the exact details of the interim analysis, and we shall be doing that in the future.

Our next question comes from Paul Matteis with Stifel.

This is Julian on for Paul. I appreciate you hosting this very informative event. Could you just describe -- have you ever shared beyond the 12-month data that you've disclosed for INFRONT-2, we've had 12 patients compared to the 10 patients in the longitudinal FTD registry data set. And if you have continued to follow patients, is there anything even qualitatively you can share about how these patients are doing or how the disease has progressed within them relative to natural history? And then I guess, briefly, a second question. You talked about 5% to 10% of FTD patients having GRN mutations. If you were to guess, on which end of that range do you think most of the evidence actually supports? And how confident would you be in driving patient identification efforts? You talked about misdiagnosis in this patient population. So anything you can talk about that would be great to hear.

Great. Perhaps, Giacomo, you want to address the first question?

Sure. So as you mentioned, the Phase II study at 12 patients treated with latozinemab open label, and we have 10 mass control from [indiscernible] at 1 year. We -- the sample size is very small, and there are not the data, and -- there are not many data 2 years in the registries to support any kind of analysis as we did in 1 year. We have looked at biomarkers after 2 years. And the sample size is small, as there are -- not all the subjects who provide data at 1 year, were available for 2-year follow-up. This a very severely progressive disease. But what I can tell you is that the drug appears to be safe, well tolerated. The biomarker effects that we saw 2 years in a smaller number of subjects are consistent with what we have shown at 48 weeks. And by mid-Q4 this year, we will present the results in a [indiscernible] number of patients, 103 symptomatic patients treated for 96 weeks. So we -- within this data will be the one that will be the most informative.

Arnon, do you want to add anything?

Yes, I can address the -- the second -- what's the second question? You can...

I think the second question was around the FTD range, the 5% to 10%.

Yes, the prevalence.

The prevalence. Chris, would you mind re-asking that question specifically?

Yes, relates on the range. So the fact is we don't exactly know, but as Sara showed you, up to 40% of progranulin mutation carriers are misdiagnosed. There's other types of neurodegeneration. So you think there is a fairly high level of misdiagnosis of progranulin mutation carriers that sort of don't fall into FTD. There was -- means a recent Los Angeles Times article on FTD that sort of put FTD as a very wide range of somewhere between 50,000 and 250,000 patients. So that's something we think that we are looking at medical, Medicare sort of claims. And so we will have a much better idea on the prevalence, means that we will disclose at a later stage. But in general, again, as Sara said, that once there is genetic testing like available and once there is a drug, I think that the real prevalence will really come out.

Our next question comes from the line of Alec Stranahan with Bank of America.

Thanks for hosting the R&D call today. Maybe first on AL137 versus 037. You mentioned the reduction in reticulocytes between these 2 assets. Following up on a previous question that was asked, is this also a driver for which asset you'll take forward? And I guess, did reticulocyte levels recover in your preclinical model?

Yes, in general, safety is a parameter like what sort of was have shown with trontinemab is that sort of anemia is less of an issue than was initially thought. It's still, I think, up to 10% of the patients suffer from some level of anemia. In our nonhuman primates, yes, the reticulocytes are very sensitive to both multiple blood draws and to transferring dependent drugs. So there was a reduction in reticulocytes, but there was a roughly recovery. And the real measure will be long like chronic number of red blood cells and hemoglobin. And so far, we don't see changes in these. But absolutely, the safety will be a component of drug selection.

Okay. That makes sense. And then I guess, how do you think about 001 and 101 as monoclonal antibodies versus other MOAs such as gene therapies and development for FTD-GRN and maybe Alzheimer's as well? Obviously, noting that the others are a bit earlier in development.

Yes, sorry.

Arnon, do you want to?

Yes, I can take this. So as you said, both in terms of gene therapy, the small molecules, there is enzyme replacement therapy. They are both very early. So we are comparing some hypothetical potential to an actual like we are going to get pivotal Phase III data in a few weeks. So it's really hard to compare. But conceptually sort of the 2 gene therapy companies are struggling, means 1 of them could not show sustainable elevation of progranulin. Sort of basically progranulin was elevated for a few months and then sort of went down. The second company had sort of significant AAV-related adverse effects and again, has to use immunosuppressive drugs and to significantly reduce the dose. And so the technical issues with gene therapy -- also, gene therapy is still, in fact, a small percentage of the nerve cells, and you assume that this small percentage of nerve cells will be a depot like will produce a large amount of progranulin that will then diffuse throughout the brain for the other brain regions. It's not clear how effective that is. And again, there will be a large disparity between some regions of the brain that have very high level of progranulin, other regions of the brain that have may be insufficient level of progranulin, it's not going to be a homogeneous distribution. And we offer the correct level of elevation. We offer homogeneous brain distribution. So we think conceptually at least that our drugs could be superior. So the 2 gene therapies, again, have safety issues, I think, have durability issues. And I think we also could have efficacy issues that didn't really show any benefit outside of, again, transient progranulin elevation. Means the enzyme replacement therapy is still early. It was on hold for several months because of adverse effects. It's requiring again, immunosuppressive drugs to delivery. It's not clear how viable it will be. And again, it will be a big discrepancy between very high level of progranulin in the periphery versus -- we'll see if sufficient level of progranulin in the brain. And there are also small molecule drugs that try to activate transcription from the progranulin gene. In the past, sort of these small molecules were tested and failed. And there is a small molecule drug that sort of blocks sortillin that's conceptually what we do with an antibody. And again, it's very early, so you don't really know the off-target activities. Progranulin and sortillin is a part of a very large family of sortillin receptors with very similar structure. So it's not clear what the off-target risks or adverse effects of a small molecule will be, what the durability will be and what the ultimate efficacy will be. So we think that at this point, we are in a very different place from competitors, both with how advanced we are and the scientific rationale for the program.

Our next question is from Graig Suvannavejh with Mizuho.

I just wanted to revisit INFRONT-3 real fast, just on the co-primary endpoints. Could you just remind us if you need to hit a p-value that's 0.025? Or do you -- is there room to hit 0.05 on both? I assume the former, not the latter. And as a follow-up, if you do end up seeing positive data, what are the gating factors in terms of next steps to be able to file a drug? If you could just remind us whether there are any CMC-related issues? Obviously, you'd have to meet with the FDA, but just trying to get a sense of how "quickly" you could actually indeed file for approval.

Maybe Giacomo can address the first question, and I can take the second part.

Sure. So we don't split the alpha for the trial to be positive. Both co-primary endpoints need to be met, and they are tested at the same time, not sequentially. So they both the CDR plus NAC as well as plasma progranulin, each, they need to be -- show a p-value that is less than 0.05.

Okay. I think in terms of just the BLA, et cetera, if the data are positive, we are already poised to be able to start working on getting the BLA filing going, and we are on track for any of the commercial manufacturing activities to do that in a rapid manner.

And just 1 last follow-up. With regards to any potential next opportunities with latozinemab, would you think about potential label expansion besides just potentially, the asymptomatic patients? I know a long time ago, there are interests and other indications as well for AL001.

Giacomo?

We are currently having those discussions, both internally and with our partner, GSK. So this is a very timely topic, and we'll disclose further details in the future date.

And we have a question coming from the line of Pete Stavropoulos with Cantor Fitzgerald.

A question on the scales to be used in INFRONT-3. For FRS, is this a more or less sensitive measure compared to CDR, sum of the boxes? Especially for certain, let's say, where you are within the disease stage, early versus later? We will be able to see different rates of improvement on FRS versus CDR over the 96 weeks of treatment? And is one better for asymptomatic versus symptomatic? How should we be thinking about that when the data comes?

Yes, great question. So most of the data that we have around the natural history, the natural progression of the disease are the one generated using the CDR plus NAC. The FRS, obviously, it's a scale of interest is an exploratory end point. The sensitivity and retail progression on the FRS are not very well known or described. So the -- hence, the decision to use the CDR as primary outcome measure, which is being validated in FTD as a whole and is the one that is recommended by the FDA. Will be interesting to see the results on the FRS, which is an exploratory outcome measure. And we also, obviously, as you said, investigate the effect, I mean the change over time in presymptomatic as well as symptomatic subjects and the sensitivity to change over time. So it will be an important analysis.

Okay. And now I will transition back to Katie Hogan for any webcast question from the other side. Back to you, Katie.

Thank you. I think we have time for one more question that we received online. And this question comes from Sean at Morgan Stanley. And Giacomo, this question is for you. What are the key differentiators of latozinemab compared to other therapies and development for frontotemporal dementia due to a granulin gene mutation? And how do you see its potential impact on the standard of care? And we're specifically talking about Prevail Therapeutics and Passage Bio.

Thank you, Katie. I think Arnon already addressed these questions. And -- but I would like to highlight the fact that there are a lot of unknowns around gene therapy, what we know is the safety profile that doesn't look ideal. One company has reported only transit increase in progranulin letters in blood and in most patients, not on patients, which is very different from what we observed. And then the other point is the rationale that a 50% decrease of progranulin levels are enough to elicit the disease frontotemporal dementia, mutation carriers. And with our approach, we are able to restore progranulin levels to normal levels, while it's a little bit more unclear, the rationale around the need to elevate, several fold higher than the normal levels. Then going back to the safety, the 1 company showed increase in NfL that reflects the dorsal root ganglia toxicity, then the other company had immune adverse reaction. So there is quite some safety liability for the use of gene therapy currently, and that pertains around progranulin treatment, but also other forms of gene therapy.

Great. Thank you, Giacomo. That's our final question for today. And operator, I'll turn it back over to you.

Thank you so much. And as I'm not showing any questions at this time on this end, I will conclude the conference, and thank everyone for participating, and you may now disconnect.