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

All right. Thanks, everyone, for joining us here at the Goldman Sachs Global Healthcare Conference, Day 2. Really decided to be joined by Arnon Rosenthal from Alector. Really excited to kind of catch up on the story and where things are and where things are headed, obviously, a pre-advanced company in CNS. So as ever, we're required to make certain disclosures and public appearances by Goldman Sachs' relationships with companies that we discuss. The disclosures relate to investment banking relationships, compensation received or 1% or more ownership. We've prepared to read aloud disclosures for any issuer during the sessions upon your request. However, these disclosures are available in our most recent reports available to U.S. clients on our firm's portal. In addition, updates to those disclosures are available by ticker on the firm's public website. Goldman Sachs agrees to host this conference on the basis that no third-party seek or provide confidential and material nonpublic information. In addition, by attending this conference, you provide Goldman Sachs the right to record and redistribute the conference information. The views of third-party speakers do not necessarily reflect those of Goldman Sachs. So with that, let's begin.

So only to start at a very high level. So overall, how would you describe Alector's approach, particularly for neurodegenerative disease.

Sure. Thank you for the invitation, and thank you, everyone, for coming. As you know, for the last 2 decades, the dominant approach to treat neurodegenerative disorder was to target misfolded proteins. So there's been more than 100 clinical trials targeting different misfolded proteins that typify different neurodegenerative disorders. So people are targeting A-beta and Tau in Alzheimer's disease. They target alpha-synuclein in Parkinson's disease. They target the Huntington disease -- in Huntington's disease, they target for C9orf72 in ALS and FTD. And so far, with some modest exceptions, this approach has been largely a failure. We, at Alector, about 9 years ago, decided to embark on a different therapeutic strategy. And instead of going after the misfolded proteins, we are targeting the brain immune system to counteract multiple disease pathologies. So conceptually, what we do is similar to immuno-oncology. As you know, what revolutionized cancer treatment in the last, I don't know, 5 years is that instead of trying to destroy cancer directly with chemotherapy or radiation or toxin conjugate antibodies, we now recruit the brain -- the immune system to counteract cancer. We are doing the exact same thing for brain disorders. So all our targets are practically immune checkpoints for the brain specific immune system. And unlike cancer where there is no genetic validation for immune checkpoints. In neurodegeneration, there is genetic validation for targeting specific immune checkpoints. So all our targets are immune checkpoints, which are genetically validated as risk gene for neurodegeneration. So we are really expanding the immune recruitment concept to brain disorder and in this, we are really pioneering. Nobody else have done it until we started this approach.

I want to follow up on something you mentioned in your opening remarks, which is I think the kind of a general rule that targeting things like beta amyloid was a bit of a 4-letter word for a long time, but obviously, that landscape has changed pretty considerably in the last -- say, last year or so. So how do you look at how things like Leqembi and donanemab have changed. How we think about kind of beta amyloid. And where do you think the white space still remains for immune modulation in neurodegenerative disease.

Yes. I think that the beta amyloid therapeutics. From one end, they rejuvenate the field, there is a lot of new interest in the field because they show, for the first time, that you can develop disease-modifying drugs for Alzheimer's disease. The white space is the fact that all the anti-a-beta therapeutics that have tested so far show very modest effect. They show approximately 30% slowdown in cognitive decline over 18 months. And there seems to be a ceiling effect, the 3 anti-a-beta antibodies that were reported shows very similar clinical benefit. It is a statistically significant clinical benefit. It's probably clinically meaningful, but it's at best model. So there is a lot of room for improving the treatment. And our immunomodulatory drugs could do that, and they could do that both a stand-alone therapeutics, we know that the immune system in the brain, not only is the garbage collector, this is the cell type that controls or that is responsible for removal of all the misfolded proteins, but this is also the cell type that replace damaged synaptic connections, that control the replacement of damage myelin that patch leaky blood vessels and damage brain tissues that stimulate the supporters in the brain to enhance brain functionality. So we know that the immune system in the brain have very broad effect on brain health, both in disease and in health situation. So enhancing this immune system would have broad therapeutic benefit that, in our view, would lead to better efficacy. So this is as a stand-alone therapy, but in addition, we think that immune-modulatory drugs like ours like that could act in combination with anti-a-beta drugs. What the anti-a-beta antibodies do, they just tag [indiscernible] for the proteins and then they have to recruit the immune system to dislodge the misfolded protein and destroy it. So we are basically bringing the excavator to the field or basically, we are complementing the anti-a-beta antibodies, they just tag the misfolded protein. So we see a natural combination between antibodies against any misfolded proteins and other drug that enhance the immune system that really excavate the mis-folded protein. So in this respect, we see synergy and sort of just a lot of room for improvement of the existing anti-a-beta drugs.

So let me start then that context with your lead program, the sortilin monoclonal antibody, AL001 for granular mutant FTD. So can you walk us through the rationale for increasing progranulin levels, specifically by sortilin blockade as a treatment for this subset of FTD?

Sure. So frontotemporal dementia is a rapidly progressing early onset form of dementia. It's the most common form of dementia under the age of 60 and a significant component of this disease is caused by familiar genetic mutation. And one of the genetic mutation is a secreted immune regulatory protein called progranulin. So in subset of frontotemporal dementia patients that have the progranulin mutation, they have less than 50% of the normal level of progranulin. So one way to treat this subpopulation is to find ways to elevate progranulin linked back to normal level. And this is, in a way, an enzyme replacement [indiscernible] and you have a familiar disease that caused by -- sort of under-representation of a single protein. And if you find a way to stop this progranulin back to normal level, you may have a therapy. So we found a way to restore progranulin back to normal level by blocking the gradation cascade for progranulin. And again, we did it through human genetics. Human genetics told us that there is a single receptor called sortilin, that down regulator basically induce the degradation of progranulin. So people that have mutations in sortilin, loss of function mutation in sortilin have 2 to 3x higher than normal level of progranulin. And we developed a drug that basically mimic the human genetic and block sortilin and basically prevented a degradation of progranulin and enable elevation of progranulin, both in the brain and in the serum by 2 to 3-fold. And we have shown in animal models that this elevation of progranulin is safe. It's chronic and it's therapeutically beneficial basically. Animal models of frontotemporal dementia that have one good and one bad copy of progranulin show behavioral deficits and our drug was able to reverse this pathology. So with this data, we went to the clinic in healthy volunteers, we were able to show that we can elevate progranulin to 2 to 3 fold chronically in a safe way, we did a Phase II open-label study. We showed that we can slow down cognitive decline by almost 50% over 12 months, that we can normalize lysosomal and inflammatory biomarkers that we can normalize GFAP, which is a neurodegeneration biomarker comparing to age-match control, we also were able to show slowdown in brain volume loss, something that no other neurodegeneration drugs were able to show. And again, with this strong package, we went to a Phase III, and we are now at a very advanced stage of placebo-controlled, double-blinded Phase III study that would be -- would enable us to file for a BLA when it's completed.

Before we get into thinking about the Phase III trial, kind of dynamic range questions earlier at the conference, you spoke to another company working in the progranulin space. And they raised the supposition that there's a case for a super physiologic expression of progranulin in people who are deficient in progranulin. It's called haploinsufficiency. How do you all think about that in terms of what does the dynamic range for progranulin restoration needs to be to translate into kind of neurological benefit?

Yes. We are restoring progranulin back to normal level, not too little and not too much. We are like in the Goldilock range. Progranulin is immune regulatory molecule. It's a mitogen and like any human hormone or protein, if you over-express it, like if you think about insulin, if you think about human growth hormone, if you think about any secreted protein, if you express it or you make too much of it, you lead to a disease. So we think that there is a bell curve of efficacy and safety. And I think too little is not going to be effective and too much is going to be unsafe, and we think that you need to elevate the drug at the level of protein just appropriately to restore it back to normal. There is a reason why progranulin is in a certain level in the body and it's very tightly regulated by positive and negative mediators. And we are just following the physiology of progranulin. We are restoring progranulin exactly the normal level. And because of our mechanism of action, we have a safety feature that does not allow us to overproduce progranulin in the body. We exactly restore it back to normal level. We think this is the right thing to do. This is what the physiology tells us. And we think that overexpression could lead to adverse effect.

Okay. Great. So kind of in that context of both thinking about the dynamic range of progranulin and the Phase II data that you walked through earlier, can you please walk us through the existing design for the Phase III INFRONT-3 trial for 001 in progranulin mutant FTD.

So the current design is a placebo-controlled, double-blinded study that's going to last 2 years long. In our view, it would enable us sort of BLA filing and approval. The study is -- the primary endpoint of the study is the sort of the CDR variant of frontotemporal dementia that was agreed on with the FDA. We have 4 additional clinical readouts that are the secondary biomarkers. We are looking at volumetric MRI because it's a rapidly progressing disease that's associated with rapid brain tissue loss. We are looking at multiple biomarkers. So we'll have a very sort of complete package of 5 different clinical readouts, biomarkers like GFAP and neurofilament and MRI imaging data that we think will allow us to really file for a full approval, not accelerated approval. The current study was designed to include both symptomatic patients and at-risk patients. It was a combined study that was intended to test both prevention, [indiscernible] and treatment [indiscernible]. More recently, we have decided to focus on the symptomatic patients, which are much less variable in their disease progression and will allow us -- so a smaller study and possibly a shorter study.

So let's get to that part. So can you walk us through some of the recently published observational data about progranulin mutant FCD that's led you to potentially change the design of INFRONT-3?

Yes, we did not change the design that the endpoints are the same, the clinical -- again, the clinical readouts are the same. The patient populations are largely the same. The only thing that we decided to do is to separate the symptomatic from at-risk population. And again, the reason is that both the published data that you referred to as well as our own sort of blinded analysis suggests that symptomatic patients, the progression is much more consistent and less variable than the at-risk patients, less than half of the variability. And basically, that's -- induced us to really separate the symptomatic and at risk to 2 different clinical trials in our end to focus on the symptomatic patients. This is something that is not controversial in the community. It's not controversial with the regulatory authorities. They are happy to see a more stratified consistent patient population. And it will allow us to, again, have a trial at the same statistical robustness as before, but with a sort of smaller patient population, and we think even shorter trial duration.

So let me come in to that, given a focus on symptomatic patients on the forwarding INFRONT-3. How should we think about the powering assumptions? What kind of effect size are you looking for? I guess before you were essentially looking for a 40% difference between drug and placebo on CDR sum of boxes, is there a reason to expect that should change given kind of your focus on symptomatic versus symptomatic plus at risk?

We have not changed that like the effect size that we are looking for is still 40%. Because in the past, we included at risk, which had high variability, we had to compensate for the variability by a much larger patient population, a much larger trial. Once you remove the at-risk patients, you can reach the 40% effect size with a smaller population of much more homogeneous and much less variable disease progression.

Can you remind me in INFRONT-2, how many patients were symptomatic versus at risk?

So if we had about we had 12 symptomatic patients and 5 at risk. But in the Phase II, the at-risk patients were recruited just due to their genetic mutations. In the Phase III, we had we added an additional criteria. The patients had to have the genetic mutations in progranulin and they also to have high level of front of neurofilament and high level of neurofilament predict conversion like we -- we thought within 1 to 2 years. Now with newer data, it seems that if the conversion is between 3 and 5 years. So the conversion rate is much more variable than initially the academic community does.

Okay. So with kind of following up and thinking about the -- INFRONT-3, I guess any variable question that comes in all neurodegenerative disease trials is, how do you think about the potential for placebo effects in a disease of this nature? This study, you given a readout like CDR sum of boxes?

Yes. The only way to really see placebo effect is to have placebo-controlled, double-blinded clinical trial that failed and compare it to historical control. And there is 1 case of this in 2020, there is an international group that run a blinded clinical trial in frontotemporal dementia. It was a year-long trial with 220 patients. The drug failed. It sort of had 2 groups with 2 different doses of drug that was supposed to be disengaging or disintegrating misfolded proteins. The clinical trial failed, but there were data on clinical disease progression and the clinical disease progression was not different than historical control like that the consortia -- that the FTD consortia reported. So there is no evidence that there is placebo effect in the CDR at least over a year duration. In Alzheimer's disease also, the evidence suggests that if there is a placebo effect, it's short-lasting, it's only a few weeks. It doesn't last for, again, for a year or more. So we think based on the published data and our discussion with the KOLs, there is really no long-lasting placebo effect with this readout. And in addition to the cognitive readout, we are having multiple imaging and liquid biomarkers, which are also completely objective means we don't think that placebo effect would lead to changes in the rate of brain tissue loss or to the level of GFAP or the level of progranulin. So we think that, again, we will have -- for the Phase III, we will have a solid package. And also for the Phase II, we saw, again, normalization of both of multiple biomarkers, and these were tested before and after treatment in the same patient. So it's almost impossible to have a placebo effect over the year period that we looked at these patients, the data suggests that there is really no meaningful placebo effect. So we are confident about the data. But again, the Phase III, which is double-blinded placebo-controlled will hopefully confirm that.

All right. So maybe shifting gears from progranulin into C9orf72 mutant FTD. So can you walk through the rationale of increasing progranulin there where there isn't kind of an intrinsic genetic deficiency in progranulin. Where kind of progranulin accumulation might have benefit in that context?

Sure. So C9orf mutations or [indiscernible] is another cause, genetic, familial, genetic cause of frontotemporal dementia and -- or ALS. And in this case, the level of progranulin is largely normal. And we still tested our drug in this patient population to see elevation of progranulin, 2 to 3 fold above normal is still therapeutically beneficial. And the scientific rationale for this trial was twofold. One is human genetics, People that have C9orf mutation and a regulatory mutation in progranulin that reduces progranulin just by 10 to 15 fold -- 10% to 15% decline much faster and the age of onset is earlier. In addition, people that have C9orf mutation as well as people who have the progranulin mutation that typifying misfolded protein in this disease -- diseases is the protein TDP-43. And in animal models, it was shown that elevating of progranulin beyond normal can counteract TDP-43. TDP-43 aggregates in the size of progranulin in these patients or in the animal models and high level of progranulin stimulate the disaggregation of TDP-43, it can protect nerve ending. It extends lifespan. So there are multiple evidence in animal models that elevated level of progranulin is therapeutically beneficial even if you start from a normal level. And mechanistically, it could be because among the many functions of progranulin, it's an enhancer of lysosomal function and immune cells and neurons with better lysosomes can counteract and degrade misfolded protein better. So progranulin enables the lysosomes. Again, which is the trash compactor in the cells to degrade misfolded proteins like TDP-43 more effectively, and this could be part of the mechanism. So based on the human genetics that show that even modest reduction in progranulin leads to more -- sort of more severe disease progression and earlier onset and the animal model showing that elevation of progranulin above normal is therapeutically beneficial. We tested this hypothesis in human, in C9orf mutation carriers with FTD symptoms. And indeed, it was an open-label study, but compared to historical controls, we saw over 50% slowdown in cognitive decline and also normalization of biomarkers like -- or reduction in biomarkers like GFAP. So we think that this is consistent with what we saw in the progranulin mutation carriers and we think that elevation of progranulin beyond normal level is still therapeutically beneficial. And this really opens up the progranulin therapy for other neurodegenerative disorders where progranulin level is largely normal and where small reduction in progranulin is detrimental. And these are diseases like Parkinson's disease, ALS, Alzheimer's disease, LATE. And in all these cases, progranulin is largely normal, but even modest reduction in progranulin is a risk for the disease. Progranulin is a formal risk gene for Alzheimer's disease, Parkinson's disease, ALS, LATE. So we think that elevating progranulin would be beneficial in all these diseases and to test that, in addition to the C9, we, together with GSK, our partner are planning on starting a Phase II trial in Alzheimer's disease late this year or early next year.

So maybe following for them all, let's come back to the dynamic range question because obviously, that comes up. So when you're haploinsufficient for progranulin, you give sortilin antibody to restore you back to 100%. I guess kind of you mentioned here the idea that you're essentially at 100% progranulin and you go up, say, 2 to 3 fold. What do you think is the effective range where you're getting clinical benefit in C9orf72 FTD, other kind of sporadic Alzheimer's, Parkinson's and so on, but not so much that you're getting the kind of negative sequelae associated with progranulin overexpression.

So it's a great question. I mean the animal model data, at least in ALS model than TDP-43 pathology suggests that elevation of progranulin by twofold leads to therapeutic benefit in other animal models, people didn't really quantify this question. So it's still open. But there are multiple data showing that high level of progranulin sort of induced proliferation of 2 more cells of primarily breast cancer cells, but also other tumors. So it is a mitogen that at high level could stimulate cancer growth. And again, overall, it's a pleiotropic growth factor that high level means it could lead to unintended consequence. So we think that, again, the restoring to normal level and twofold elevation is still within the Goldilock range, but above that, it could be sort of risky. And this is -- we compare what we do to SSRI, like to serotonin and norepinephrine reuptake inhibitors, they're also like the therapeutic benefit is if you increase the level of neurotransmitter by 2 to 3 fold. If you go above that, you start getting adverse effects. So we think that it's similar to what you see with using progranulin. But again, the clinical data will test that.

Right. Maybe let's move to -- on your other lead clinical program, AL002 for their TREM2 program for Alzheimer's disease. So again, why do you walk through the mechanistic rationale for pursuing TREM2 as a target in Alzheimer's?

Great. So TREM2 is, in a way, an activating immune checkpoint for the brain immune system, a cell type called microglia. It's a strong genetic risk for neurodegeneration. If you don't have TREM2 at all, you develop neurodegeneration by the age of 40, you die soon after, and it's a 100% penetrant. If you have one of about half a dozen point mutations in TREM2, that reduce the ability of TREM2 to interact with its natural ligands, you triple the risk of developing Alzheimer's disease, which is similar to the risk you'd have if you have 1 copy of APOE4. So loss of function of TREM2 invariably leads to neurodegeneration, whereas multiple animal models show that overexpression of TREM2 are protective in multiple types of dementia. So based on the human genetics and animal model studies with overexpression of TREM2, we decided to develop a drug that activate TREM2, that basically binds the receptor and activate it and basically compensate for decreased activity in the aging brain. And with this drug, we tested this drug in multiple animal models and in 2 types of Alzheimer's animal models, we show clinical efficacy and multiple sclerosis models, we showed efficacy and much faster regeneration of myelin. And with this combination of strong human genetic and animal model data, we took the drug to a Phase II in Alzheimer's disease. And we think that this is really the beginning of second-generation drug for Alzheimer's disease that will go beyond the anti-a-beta therapeutics.

Maybe with that in mind, so you mentioned this Phase II trial, INVOKE-2. Can you walk us through the design of INVOKE-2?

So INVOKE-2 is a 4-arm study. We have 3 different doses of the TREM2 activating antibody AL002 and placebo-controlled. It's a double-blinded study that recruits early-stage Alzheimer's disease, it's -- what's called the common close design over 12 months. We have a broad range of readouts. Again, we'll have 4 different clinical cognitive readout for this trial. We are going to have PET imaging for Tau. We are going to have PET imaging for a-beta. We have MRI imaging for brain volumes. We have multiple liquid biomarkers in the serum and CSF including a-beta 40 and 42, different forms of Tau, GFAP, neurofilament and multiple other biomarkers. So we think that -- at the end of this study, we will have a pretty comprehensive readout to continue to Phase III and possibly to apply for accelerated approval. The interesting thing is that we are practically completed recruitment for the Phase II, it's 324 patients. And what we see is evidence of ARIA, which is imaging abnormality that so far was observed with all the anti-a-beta antibodies that remove beta amyloid. The ARIA that we see based on the radiographic imaging, the clinical presentation, the early onset, the reversibility, the dependency of APOE genotype, the responsiveness to steroids is indistinguishable from anti-a-beta antibodies. So this -- again, we -- the study is still blinded. We don't know the actual effect on a-beta, but we think that we can extrapolate that our drug may remove a-beta similar to anti-a-beta antibodies. And this is an exciting signal to us that we have an active drug. And again, the study is practically completed recruitment. We are going to have data by the end of 2024 that we'll share top line data.

Okay. So I guess kind of you mentioned before the safety signals for agents in the TREM2 space. And so how do you think you can manage those specifically in INVOKE-2 to kind of get the balance between demonstration of efficacy relative to kind of safety overhang for the study.

Yes. Generally, the ARIA is a managed side effect now with our anti-a-beta antibody. Our ARIA is right there with all the other anti-a-beta antibodies we really in regard to prevalence means. The issue -- ARIA is just an MRI observation like the issue is the prevalence of area with adverse effects -- or severe adverse effects, and this is in very few cases, less than 1% of our ARIA are with adverse effects. And in the cases that we have like the ARIA is reversible and it's treated with steroids. And I think the field, in general, is learning how to manage ARIA and how to reduce ARIA by dose escalation -- by gradual dose escalation. So we think -- and the field now think that ARIA is a manageable side effect now that it's a package deal with sort of efficacy in Alzheimer's disease. And again, Alzheimer's is a lethal disease. People die within 7 to 10 years after diagnosis, and it's a horrible life before that so some level of adverse effect which I think people will learn even better to manage with time is tolerable.

Okay. So finally, the question we're asking every company at the conference. What is the reason to own Alector stock in the next 12 months?

So globally, we are really pioneering a new therapeutic strategy for dementia and neurodegeneration that, if successful, could be as impactful for brain disorder as immuno-oncology is for cancer. So we are doing something. We are pioneering something that no other companies have pioneered. We are way ahead of everyone else in this approach. And in the next 2 years, we will have 2 major clinical readouts, 1 for Alzheimer's disease, and you know what happened in the Alzheimer's field, and a very positive readout as a major financial impact on the stocks of the respective company. And we also will have a pivotal study readout in another major sort of neurodegenerative disease that will again open up the drug to many additional indications. So in the next less than 2 years, we will have 2 major clinical readouts in unmet medical needs in very large sort of clinical indications like Alzheimer's disease. And I think it could -- both -- either of them would transform the therapeutic field for neurodegeneration.

Okay. Well, thank you so much, Arnon, for joining us, and thank you all for joining us here at the Goldman Sachs Global Healthcare Conference.