All right. Good morning, everyone, and welcome to the Goldman Sachs Annual Healthcare Conference. Delighted to be joined by the team from Alector here today. Maybe I'll kick it off. I'll let you guys introduce yourselves and maybe provide a brief overview of the company.
Sure. I'm Arnon Rosenthal. Welcome, everyone. I was the CEO and co-founder of Alector. Before that, I was at Genentech for 16 years as part of the research team. Then I started Rinat Neuroscience that was acquired by Pfizer. One of the drugs that came from this company is Ajovy, the migraine drug that started the CGRP migraine therapeutics. I then started Annexon Bioscience. It's a publicly traded company. For the last 12 years, Alector is my life. Alector was created really with the vision and mission to really eradicate degenerative brain disorders, to really make degenerative brain disorders similar to smallpox now. In order to do this, we really created, we built an integrated research organization that involved human genetics, protein engineering, deep understanding of cellular and animal models for neurodegenerative disorders. We built an extensive clinical team, extensive manufacturing organization.
We are building a large integrated company that can go from ideas all the way to approval. We currently have two programs in late-stage clinical development. We have a phase three program in frontotemporal dementia, which is the second largest dementia for people under the age of 60. The readout is expected to be by the end of this year. It is a pivotal study. If it is positive, we will sort of apply for approval, and hopefully we will proceed to commercialization. In addition to the phase three program, we have a phase two program in Alzheimer's disease. We completed recruitment for a placebo-controlled double-blinded 18-month-long trial, and we expect data sometime in 2026. In addition to the clinical programs, we have a significant portfolio of preclinical programs that target Alzheimer's disease, Parkinson's disease, Lewy body dementia.
We are integrating our programs with our blood-brain barrier shuttle to facilitate the entry of both antibodies, enzymes, and nucleic acid into the brain. These preclinical programs are expected to enter the clinic next year also, at least some set of them.
Maybe we'll start with latozinemab in frontotemporal dementia. You could maybe just set the stage for us on what is frontotemporal dementia. You mentioned this earlier. What do we know about the kind of the pathogenesis of the disease, and what's the current treatment for those patients?
Sure. Frontotemporal dementia encompasses subtypes that are genetically determined. Among those, one of the most common genetic forms of frontotemporal dementia is associated with mutation of the progranulin gene. These are loss of function, heterozygous loss of function mutations. Roughly 5%-10% of the cases with frontotemporal dementia are characterized by these kinds of mutations. Our drug latozinemab binds to sortilin, which is the major enzyme responsible for trafficking of progranulin and degradation of progranulin. Through the binding and neutralization of the sortilin receptor, our drug latozinemab is able to increase extracellular levels of progranulin by two- to threefold, both in plasma and CSF.
Great. I guess what do we understand about the clinical benefit of increasing progranulin once people have disease? Is that sufficient to kind of drive clinical meaningful results?
Abnormalities and mutation of the progranulin gene are associated with a 90% probability of developing frontotemporal dementia by the age of 75. There is almost complete penetrance of the gene of those mutations. Frontotemporal dementia in patients with mutation of the progranulin gene is associated with a roughly 50% decrease of progranulin levels. Our drug is able to increase progranulin level two- to threefold in CSF and plasma, respectively, restoring progranulin levels to normal levels. We have seen this kind of data both in healthy volunteers as well as in patients with frontotemporal dementia in our phase two study.
You referenced the phase two. Maybe you could spend a bit more time on what you guys saw in that phase two trial and how did the results there kind of inform your phase three design?
Yeah. The phase two study was an open-label study where we enrolled different types of genetic FTD patients. Among those, we enrolled 12 patients with frontotemporal dementia and mutation of the progranulin gene. These are the same population that we are enrolling in the phase three study. We saw at 12 months of treatment, after 12 months of treatment, besides increases of progranulin that are already described, we saw changes in lysosomal markers as well as change in volumetric MRI. We also measured CDR NACC FTLD sum of boxes, which is the gold standard measurement to track disease progression. We compared the change in the CDR sum of boxes, the FTLD version, to the natural progression of disease in matched controls from the registry data from the GEMFI study.
We observed an improvement of disease progression by 48% at one year using these matched controls. We also saw a positive effect compared to the natural history of the disease on other biomarkers. We saw a decrease in brain atrophy, in frontotemporal cortices atrophy, as well as a decrease in ventricular enlargement compared to these natural controls.
Okay. You talked about this in that response, but there was a natural history kind of matched control data set. I guess how do you think about the fidelity of that natural history kind of matched population in terms of the variability across this patient population? I guess how much confidence do you have that that will be kind of recapitulated in the placebo-controlled study you're doing here?
These controls were taken from one of the two largest registry data sets available. The name of this registry is called GEMFI. We matched the historical controls to our phase two patients according to certain baseline characteristics, such as severity of disease measured with the CDR NACC FTLD sum of boxes, as well as matching by age, gender, as well as neurofilament light chain baseline levels. They were very comparable. We use propensity score matching, which is a statistical technique used very often for these kinds of comparisons. Our study is in the phase, the population in our phase three study is very similar to the phase two study. We do not, it is a placebo-controlled trial, pivotal phase three study.
We do not expect the placebo response to be particularly impactful because according to other studies in neurodegenerative diseases, the placebo response is not very large, and it tends to plateau after the first few weeks of treatment. Moreover, we model the magnitude of placebo response in our statistical power assumptions.
Okay.
Just maybe to add two things. First, the FDA reviewed the phase two open-label studies and based on the phase two studies awarded us breakthrough therapy. They accepted the statistical analysis and the open-label studies. All the data in the open-label studies were consistent. Again, we saw normalization of multiple lysosomal and neurodegenerative biomarkers. We showed slowdown in brain tissue loss. We showed slowdown in cognition by 40%. Our phase three study is 10 times larger and twice as long as the phase two study. We think that we have a very large margin of error.
Sure.
Means even if the phase three is not exactly as the phase two, we still have a lot of room to.
The cushion.
Yeah.
There have been some other progranulin-directed agents that were discontinued. What distinguishes latozinemab in the way it works versus some of these other programs targeting the same mechanism?
Basically, what discontinued are transcriptional activators, like what's called HDAC activators that were supposed to increase expression of the progranulin from the progranulin gene. They ended up not really working mechanistically. They never really worked. Nobody was really able to show in human or even in non-human primates that there is a meaningful elevation of progranulin. I think that just the mechanism of action of the mainly one previous drug was just not working. Now there are other therapeutic approaches, as you know, there's gene therapy, there's enzyme replacement therapy. We think that what's unique in our approach is that we are elevating the endogenous physiological progranulin in the right places and exactly in the right amount. Like progranulin is a mitogenic growth factor. If you elevate it more than you need, there's risk of cancer, basically.
If you look at other growth factors like insulin, insulin like growth factor, human growth hormone, all these factors, there is a sort of Goldilocks optimal level that you want to elevate. If you elevate too much, you start getting diseases.
Sure. Okay. So you've got the phase three results coming up here soon. Maybe speak to us about the key endpoint to watch and what you're powered to detect in terms of difference between the treatment and the placebo arm?
Yeah. Our phase three study is powered to detect the 40% slowing of disease progression in the active treatment arm as compared to placebo. However, if we observe a slowing of disease progression as low as 25%, according to our statistical simulations, this study will be positive.
Okay. What comes after kind of a positive result in this trial?
We are well positioned to deliver meaningful clinical data by the end of 2025 in four quarters. Our study is adequately powered to detect a meaningful effect by the drug and a clinically meaningful effect by the drug. If the study is positive, we plan to proceed with this regulatory submission based on the single pivotal phase three study.
Okay. Understood. In addition to latozinemab, you also have AL-101. It's also targeting progranulin. How is it designed to differ from the latozinemab agent?
Both AL-001 and AL-101 target the enzyme sortilin.
Sure. Yeah.
AL-101 differs from AL-001 for PK as well as PKPD characteristics, which makes it better suited to treat more common diseases such as Alzheimer's disease as well as Parkinson's disease. As Arnon said, we just completed the enrollment in a relatively large phase two study in April 2025, and we are expecting the results by the end of 2026. This is a 76-week-long study, and the primary outcome measure is the CDR sum of boxes. We have enrolled patients with early Alzheimer's disease.
Okay. Maybe sorry to.
Yeah. Just it's interesting that sort of progranulin is one of the few universal risk genes for neurodegeneration. Like most of the risk genes are specific for given diseases. There is a set of risk genes for Alzheimer's disease, Parkinson's disease, ALS.
Okay.
Progranulin is one of the very few genes, maybe one or two additional genes that appear in all. Its loss of function is a risk for frontotemporal dementia, for Alzheimer's disease, for Parkinson's disease, for late, which is another type of late-onset dementia typified with CDP43. It is a risk gene for ALS. It is a universal risk gene. Every time you see even a modest loss in the level of progranulin, even 15%-10%, you increase risk of disease. We are developing, again, a franchise of multiple drugs that elevate progranulin for multiple indications. That is really what was the excitement around.
You mentioned that the difference is on PK and PKPD. I guess what were some of the key things you needed to optimize for as you think about setting this up for larger indications?
We tried to optimize for convenience of use. The 101 is a two to three times longer half-life compared to 001. You can either deliver it less frequently. 001 is being delivered once a month, so we can deliver 101 every two or three months. Or you can reduce the dose and eventually convert it to subcutaneous delivery.
Okay. Understood. In terms of the ongoing Alzheimer's disease study, I guess what was the decision for Alzheimer's disease first? Could you talk about some of the preclinical and clinical data you've reported on this agent that supported that decision?
As Arnon just said, mutation of the progranulin genes are ubiquitous in many neurodegenerative diseases. Through our collaboration with GSK, we thought that besides frontotemporal dementia, Alzheimer's disease would be a very good indication to pursue, especially in light of the progress made in the past decade and the identification of very valuable biomarkers that can help with decision-making and speeding up the development process. In the ongoing phase two study, we used several of those biomarkers, including fluid biomarkers that look at markers of Alzheimer's disease such as the different pTau species, but we also have amyloid PET and tau PET. We decided with GSK to pursue these larger indications.
In frontotemporal dementia, you're stratifying or you're selecting for patients who have specific progranulin mutations. In the case of Alzheimer's disease, are there biomarkers or any other features that help you select for the patients most likely to respond?
The population enrolled in the phase two study is very consistent with the population enrolled in other Alzheimer's disease trials. These are patients with early AD who have evidence of amyloid pathology. We do not specifically select for patients with any specific genetic mutation because the penetrance in Alzheimer's disease is much lower versus frontotemporal dementia, where, as I said, a patient with progranulin mutation has more than 90% chances to develop the disease. In Alzheimer's, the link is different. The population is, as I said, patients with early AD with presence of amyloid pathology.
I mean, the rationale for going with Alzheimer's is that, for example, in animal models, overexpressing progranulin beyond normal level is protective in multiple Alzheimer's disease models.
Okay. In terms of the phase two study, what would you look to see? Maybe also this is a question sort of for how the partnership works, but what will you look to see to proceed forward with phase three studies in this indication?
The primary outcome measure is the CDR sum of boxes, which is the gold standard, one of the accepted clinical and functional measures to determine whether a drug is effective in Alzheimer's disease. We also collect a number of meaningful biomarkers. Overall, it's a study which is an adequate length to detect an effect, it is one and a half years duration, which is, according to prior data, enough to see a meaningful effect of drugs in Alzheimer's disease. At the end of the study, we review all the clinical data as well as all the biomarker data and make decisions about the next steps in the program.
Yes. Again, we are measuring PET imaging for Aβ, PET imaging for tau, CSF and serum biomarkers, Aβ40/42, multiple phospho-tau, neurofilament, GFAP.
The gamut.
Yes. Like five different clinical readouts. If there's a movement in any of them that's meaningful, we proceed.
Okay. Understood. How do you envision the Alzheimer's disease landscape will kind of evolve with things like new mechanisms coming to bear? Do you expect this would be used in like a polypharmacy sense and sequential therapy? How does that inform the way you think about phase threes?
Yeah. So the data from anti-amyloid treatments that have been approved are very encouraging, and they represent a significant milestone in the treatment of Alzheimer's disease for the medical community. However, there is still lots to be done. This treatment improves disease progression by 25%-35% and does not, there is room for even larger improvement. We also know that there are some patients, especially patients with more advanced disease, which do not benefit as much from anti-amyloid treatments. Moreover, as we all know, there are significant hurdles for the implementation of those treatments in the field, which are driven by the frequency of ARIA, especially in patients who are ApoE4 homozygous. There is definitely a lot of room to improve the current treatment in Alzheimer's disease. Our drug AL-101 has a different mechanism. It modulates progranulin, which aboves physiological levels.
We think that progranulin is a very important modulator of the immune function as well as neuronal survival. We think that it may be as well as a complementary mechanism to existing anti-amyloid treatments, which target only the amyloid plaques. Regarding the mode of use, as you asked, sequential treatment combinations, I think the best approach will be data-driven. As we can see from the Alzheimer's field, there are different ways of administering drugs with different mechanisms of action that are being tested. If we think about the way that anti-amyloid treatments are combined with anti-tau treatments, we still do not have the best answers, but what is the optimal way to combine treatment with different mechanisms of action?
However, we think that given the diversity of AL-101, there is plenty of room for the drug to coexist with anti-amyloid treatments, either in a sequential way or even as a combination treatment, as our drug does not, so far as I am shown, or other side effect profile or similar side effect profile as anti-amyloid.
Okay. Maybe last question on this program, just remind us the parameters around the GSK collaboration and what gets triggered with positive phase two results?
The whole franchise is in a way 50/50 profit-share collaboration. When the collaboration was established, we received $700 million upfront payment. It is a total of $1.5 billion milestones, which includes for 001, like first commercial sale in the U.S. is $160 million. Commercial sales in the EU is $90 million. We have commercial leadership in the U.S., so we are the commercial lead in the U.S. for the 001. There is a similar sort of situation for 101. It is a 50/50 profit-share in the U.S. and royalty, which mimic 50/50 ex-U.S.
Okay. Great. Maybe we could spend some time on your pipeline. Last summer, you introduced a suite of Alector Brain Carrier ABC candidates. Maybe you can speak to the brain carrier technology that you've developed first?
Yes. We are developing brain carrier that should enable us to deliver drug more efficiently to the brain. We are going after antibody drugs, enzyme drugs, and nucleic acid drugs. It is a pretty versatile technology. We are using Trojan horse technology that sort of other people are using, basically hitchhiking on either the transferrin receptor or the CD98, which is an amino acid transporter. What is unique in our technology is the sort of the degree of versatility. I mean, we have a very large range of affinities to the receptors. The technology in general is not plug and play. You have to really tailor the technology for the given drugs. You have to play between efficacy and safety. Because specifically, for example, for the transferrin-based technology, the target-mediated adverse effects are hematologic because there is a lot of transferrin receptor on red blood cells.
There's actually 100 times more transferrin receptors on red blood cells compared to the blood and very endothelial cells. If you are somehow binding red blood cells and recruit the immune system to attack red blood cells, you lead to anemia. That is sort of an intrinsic problem in this technology. You really have to play with the affinity, with the binding epitope for each drug to optimize the safety and efficacy. We have, I think, a unique ability to do that. We have a 1,000-fold range of affinities that we can play with. We have unique epitopes that actually may exclude simultaneous interactions with red blood cells and immune cells to increase safety. We can use multiple configurations to optimize for the different drugs. We have a different technology for antibodies.
Even this depends on the antibody, like whether if you need a full effector function, if you want the antibody to recruit the immune system, you use a certain type of affinity and epitope of transferrin. If you do enzymes and you do not need a full effector function, you can use a different technology with maybe higher affinity. The same for nucleic acid. For example, if you want to remove beta amyloid with an anti-ibeta antibody, you need to recruit the immune system because the anti-ibeta antibody just tags the site. You need the excavator. You need the immune system to really dislodge the. You need to recruit the immune system. If you recruit the immune system to ibeta plaques, you also recruit the immune system to red blood cells.
There is really a very delicate balance that I think we can really optimize better than most other technologies.
Okay. How do you think about the context where transferrin and CD98 are the most appropriate target?
So far, the sort of transferrin is more validated. It's been in the clinic, both in the context of enzyme replacement therapy and in the context of antibodies. Although there are cases of anemia, especially again, if you have full effector function, as in the case of Trontinemab with Roche, it sort of seemed to be manageable. So transferrin, the limitations are known, and they are really mainly safety associated with hematologic side effects. With CD98, there is still less understanding of the safety profile. We are working very extensively to really establish the safety profile of CD98. I mean, there are differences like in the kinetics of entry to the brain, in the cell types that express the different shuttles that may, again, impact brain distribution. Ultimately, it will be good to have more than one shuttle. So far, everyone is using transferrin.
It's a lot more validated.
Right. How did you think about selecting an Aβ antibody for your first ABC candidate?
Yeah, we are sort of aiming to really develop the best-in-class anti-amyloid beta antibodies with brain shuttles. We are optimizing each of the components. The lead program now is the Trontinemab from Roche. Their naked antibody actually was inactive. The Gantenerumab, which was the precursor for Trontinemab, was inactive as a naked antibody. We are choosing an epitope that is very potent as a naked antibody, more similar to the real antibody, the Trontinemab. Also, Roche saw incidence of anemia. They claimed that it is manageable, but so far, patients were not treated for that long. I think anemia is a cumulative effect. It will become worse with time. We have a different transferrin technology, which we think will have much less anemia risk. Compared to other technologies, people are trying to solve the anemia problem in different ways.
For example, Alector and AbbVie are inactivated. They have an inactive effector function. They hope that the transferrin receptor will somehow recruit immune cells. We think that this approach is less likely to fully recruit immune cells. I think that we will have an optimized antibody that has the best epitope as a naked antibody, the best transferrin technology to minimize anemia.
The other program that you have using the ABC technology has a GCase enzyme. I guess, why did you select that? And sort of could you give us the high level on that molecule?
GCase is a lysosomal enzyme that's mutated in over 10% of Parkinson's patients. It's causing like 10% of Parkinson's patients. It causes up to 30% of Lewy body dementia. It's a validated enzyme replacement modality because if complete loss of function of GCase leads to Gaucher disease, and there are good enzyme replacement therapies for Gaucher disease. It's a relatively validated enzyme replacement target. The only issue until now for Parkinson's disease and Lewy body dementia was that it didn't enter the brain. We basically now engineered the enzyme to become significantly more stable, more active, and we integrated the blood-brain barrier shuttles. We think it would be a good drug for Parkinson's disease and eventually Lewy body dementia.
What are the next steps for those programs in terms of getting into the clinic? To what extent does one program validate the technology for multiple?
Yes. We are targeting both the Aβ and GCase to be in the clinic next year. Yes, each one of them will validate the technology. As I mentioned, there are actually different variations. We are also developing tau siRNA with brain shuttles, which will also have a tailored technology. I think each of these drugs, like the antibody, Aβ antibody, the GCase enzymes, and the tau siRNA, will validate maybe the drug modality. They will validate the technology in general if there are little safety issues and good brain penetration.
Okay. Maybe in our last minute, you can just talk about kind of your strategic and overall approach to thinking about capital allocation across the early-stage portfolio versus the later-stage programs.
Yeah. So we have sort of resources. Like our runway is through the second half of 2027. We have over $350 million. Our clinical programs are already fully funded. Like our phase three drug is fully funded. Our phase two drug is fully funded. We have resources to develop A beta, GCase, and tau sort of toward the clinic. We think we will have, like in the next two years, we will have a really good portfolio of clinical programs and earlier stages of clinical programs.
Maybe since I have 30 seconds, I'll ask you one more question, which is, you guys had partnered a lot of your early programs. As you look to the ABC candidates, will you take a similar approach, or do you want to wholly own them?
Our preclinical programs are currently fully owned. Yeah, if there is a proposal that we can't refuse, we will explore it, absolutely. At this point, we are developing the preclinical programs as fully owned.
Perfect timing. Thank you so much for joining us. Really appreciated the discussion this morning.
Thank you very much.