Annexon, Inc. (ANNX) Earnings Call Transcript & Summary

Good day, everyone. I'm Doug Love, President and CEO of Annexon Biosciences. We're thrilled that you're able to join us today on the second of three series that we are conducting on the role of C1q and the classical pathway in complement-mediated diseases. Our first series involved the role of C1q in autoimmune disorders, which we conducted in August of this year, and we invite you to our website to find materials as it relates to that. Today, we will be discussing the role of C1q in the classical pathway in complement mediated neurodegenetive diseases. And our third series on the role of C1q in ophthalmic disorders will be conducted later in the year. With that, we shall dive in. We will be making forward-looking statements as part of our discussion today, and we invite you to our website as it relates to that. We are quite excited by the lineup of topics that we have before you today as well as the speakers who will be bringing them to you. In addition to Ted Yednock, our Chief Scientific Officer, and Sanjay Keswani, our Chief Medical Officer, we're joined by four outstanding key opinion leaders in their respective fields. Ted will get us started with an overview of C1q in complement mediated neurodegeneration and then turn it over to Dr. Henrik Zetterberg, who will review neurofilament light chain, a key measurement of neuronal damage and a biomarker that we are using as part of our neurodegenetive programs. He will then turn it over to Dr. Beth Stevens, a former post-doc of our scientific Co-Founder, Dr. Ben Barres, who will review the role of complement-mediated neurogeneration, specifically in Huntington's disease. From there, Dr. Ed Wild will review, providing clinical perspective on Huntington's disease and the use of NfL there. And finally, Dr. Angela Genge will review from a clinical perspective, the role of C1q in ALS as well as neurofilament light chain. And finally, Dr. Sanjay Keswani, our Chief Medical Officer, will close us out with an overview of Annexon's clinical programs and complement-mediated neurodegenerative disease with a focus on Huntington's disease and ALS. Many of you are familiar with Annexon, and we're quite excited to provide just a brief update of where we are as a company. You will recall that our aim is central in focusing on bringing game-changing therapies to patients suffering from serious complement-mediated diseases, and we're quite pleased by the progress we've made to date. As you will recall, by targeting C1q at the start of the complement cast date, we are targeting two distinct disease processes: One, antibody-mediated autoimmune disorders; and the second, which we're talking about today, complement-mediated neurodegeneration. That lends itself to prosecuting three therapeutic franchises in autoimmune neurodegeneration in ophthalmology with three clinical stage drug candidates currently in five Phase II plus three clinical trials that are underway. And finally, we anticipate producing clinical data over the next 2 years in seven distinct programs. We're well capitalized to do that with cash on hand through early 2024 to really deliver on this game-changing mission to bring important therapies to patients suffering from serious complement-mediated diseases. We have quite an audacious pipeline. As you are familiar with, that is both diverse and wholly owned. In the autoimmune section, we are making quite great headway with our programs there for which there are four, two currently dosing, -- syndrome and warm autoimmune hemolytic anemia, and two that are on the horizon, lupus nephritis -- and motor neuropathy. As it relates to ophthalmology, we have a Phase II program that is underway in geographic atrophy. And today, we are focused on the neurogenerative franchise, where our lead indications are Huntington's disease in ALS. These indications are diversified in that Huntington's disease is a classic CNS indication, and ALS is an indication that involves both the peripheral and central nervous systems. These -- both of these programs are biomarker-led programs where we are looking to, in addition to establish a safety profile, also ensure that we have effective target engagement across the blood-brain barrier as well as a meaningful reduction of neurofilament light chain. Now we're quite encouraged by these programs and that we have been able to demonstrate both full target engagement as well as a statistical reduction of neurofilament light chain preclinically, both in Huntington's disease as well as in ALS models. And we've also been able to do so in patients in our Guillain-Barré study. So we're encouraged as we continue to advance these programs and look forward to bringing new data in the near to midterm. Finally, as we think about Annexon and our three therapeutic areas, we're quite excited by the vast potential across all of them and super excited to bring to you today really the first overview of complement-mediated neurodegeneration and the role of C1q there. There were a pioneering approach as it relates to CMD and by seeking to preserve functioning synapses, neuronal health and behavioral and cognitive function in these patients. Now we are targeting what we believe to be a very common pathway for neurodegeneration, that is the loss of functioning synapses that leads to neuronal death and ultimately decline function, both cognitively and motor -- in both the brain and eye diseases. As a result, we are able to prosecute a wide range of diseases, and we will show you examples of that today. We are using a very objective biomarker that is well-established neurofilament light chain in these early programs to show our drug's effect before advancing into a Phase III pivotal program. Importantly, as you will hear over the course of today, -- correlates quite nicely with patient disability and reduction of NfL has shown nice correlation with clinical benefit in multiple diseases in the clinic. Finally, we have made good headway towards this approach, as I alluded to on the prior slide in that we have demonstrated full target engagement across the blood brain barrier, both preclinically and clinically and we have demonstrated a reduction of NfL in now three distinct therapeutic areas, Guillain-Barré syndrome in patients and Huntington's disease in ALS and preclinically. So with that, I'm going to now turn it over to Dr. Ted, our Chief Scientific Officer, who will provide an overview of the role of C1q in complement-mediated neurodegenerative diseases. Ted, over to you.

Thank you, Doug. So what I'm wanting to talk about today is cell complement-mediated neurodegeneration. And this is a discovery by Dr. Ben Barres and Beth Stevens. Ben was our scientific founder at Annexon, and Beth Stevens was a post-doc in his lab at the time, and she now has -- she's a professor at Harvard. And what they found is that C1q mediates synapse loss in a process and development that occurs normally, but then also, again, in the adult as a process called complement-mediated neurodegeneration. And this appears to be a common pathway in many neurodegenerative diseases that leads to neurodegeneration and disability. And what they found is that this process is independent of the inciting etiology, so whether it's A-beta or Tau or increases in intraocular pressure, that it precedes the loss of neurons and it correlates with functional decline. And what we and many other labs have found is that inhibiting C1q protects against synapse loss and protects cognitive behavior as well as motor function. So it's important to be targeting C1q in the classical pathway, and I'll talk more about that coming slides. So going back to their discovery. Again, normally in development, C1q recognizes synapses. And it does -- it is independent of antibodies. It's a unique function of C1q as a recognition molecule. It can distinguish weak synapses from strong synapses. And it binds to the surface of the weak synapses, triggers classical complement activation, microglial cell recruitment that then prune these weaker synapses away from the neuron, leaving the stronger ones in place for appropriate circuits. Now in the adults, that's development. So in the adults, when neurons are stressed, and again, this could be an Alzheimer's or Huntington's, independent of ideology, C1q will revert to this developmental pathway and begin to recognize these stressed synapses and neurons again. And it will trigger synaptic removal to the same pathway, microglial cells will come in, it recognizes the complement coded synapses and prunes them away from the neuron. But this is depriving the neurons of trophic support and is causing neuroinflammation, and it really becomes a driving part of the neurodegenerative process itself. Now in support of this, we and many other labs have found that inhibiting C1q is protective in a wide range of animal models of neurodegenerative disease, including chronic indications, such as Huntington's or Alzheimer's or spinal muscular atrophy, diseases of the eye, glaucoma and AMD as well as acute brain injury and traumatic brain injury. So it's a broad-ranging mechanism. Now just showing some data behind these statements. This is showing Ben and Beth's original discovery in development, looking at the visual pathway where we can see C1q in green is tagging synapses in red. These are synapses that are adjusting to be eliminated. And they found that by genetically deleting C1q, the animals would go through this developmental period and end up with about twofold more synapses. And these are functioning synapses. So C1q is recognizing functional synapses and triggering their elimination. So again, a developmental pathway. What they then found was that after this developmental period, C1q will begin to accumulate on synapses with normal aging. So off of the top, we're looking at the mouse brain from 1 month of age up to 2 years, 2 years is quite old for a mouse. And you can see this very broad staining of C1q, particularly in the hippocampus and the striatum, which are regions relevant to brain diseases, but also in the cortex and throughout the brain. So this is largely synaptic staining. And they found that this increases about up to 300-fold compared to young animals. And this is a process that also occurs in the human brain. So this was exciting because C1q accumulation on synapses occurring with age, age is one of the biggest risk factors for neurodegenerative disease. And so it implies that this mechanism could be an important component of the risks associated with neurodegeneration. So based on this foundational work, truly spawned an entire scientific field. So there have been labs throughout the world who have reproduced and extended this work looking at all these different animal models of disease that I've discussed. And the key finding from all of this work is that synapse loss is a major driver of neurological disability and blindness in the case of the eye. It precedes the loss of neurons, and it really correlates with functional loss and cognitive decline. As an example of this, this is a very unique study looking at Alzheimer's patients where they did a performance task. And then postmortem, there was an analysis of synapses. And what they found was that there was a strong correlation between synapse number and performance on this cognitive task. Now similar results have been shown in patients with Huntington's disease, and this is the work that was done with -- in Beth Steven's lab at Harvard and she may be talking about that later today. Now looking at these animal models, I'm just going to give some very high-level results showing how synapse protection matters. And so in this case, again, we're coming from Beth Steven's lab, showing that exposure to A beta as a model of Huntington's disease causes synaptic loss very rapidly over a 2- to 3-day period. And then in the presence of an antibody -- our antibody against C1q, these synapses are protected both the number of synapses as well as their function as measured here by electrophysiology. So that's synaptic function. Now in Eric Wayne's lab at UCSF, we get a very different model. This is a model of frontal temporal dementia. And he found that over an 8-month period, these animals developed an obsessive compulsive behavior, which is typical and patients with FTD. So this is looking at the red bar versus the black bar. So these mice developed this behavior. And if he genetically deleted C1q over this 8-month period, the animals did not develop this behavior. So it's showing a protection of synapses as well as then a behavioral function. And then finally, looking at motor function. This is work coming out of George Mentis' lab at Columbia University and a model of spinal muscular atrophy. And what he did with -- this is a very aggressive model of SMA. The animals basically only lived for about 2 weeks. And in the red line, it shows that the animals really never gain an ability to ride themselves and to walk. And so it's -- they're stuck at the 60-minute time limit for the task. Compare that with the black line, which are animals that were treated with our antibody against C1q, this is systemic administration, the antibody entered the spinal cord, protect the synapses and importantly, allowed the animals to game motor function and, in some cases, even walk over this 14-day period of time. So here I've shown how synapse protection can protect cognitive behavior, motor function and so now I'm going to talk a bit more about the classical cascade itself and how C1q is involved in initiating this cascade. So C1q is there at the top. When it binds to synapses, it will activate the components downstream in the pathway. So a single molecule C1q can activate a couple of molecules of C4, which then goes on to activate more molecules of the downstream components. So this is an enzymatic amplification cascade. Each of these different components is playing an important role. C1q, C4 and C3 will all associate with the cell surface and they function as receptors for microglial cells or macrophages and can trigger them the direct attack of these immune cells on the nerve surface. The downstream components, C5 to 9 are involved in poking holes in membranes, so they cause membrane damage and could contribute to the disease progression in that fashion. So we think that it's very important to inhibit the classical pathway up at the top by blocking C1q because we're blocking the activity of all of these downstream components. So we think that's important for efficacy. From a safety perspective, inhibiting C1q will specifically block the classical pathway, showing there in the graph on the lower right, in the red line, but it does not impact the activity of the lectin or the alternative pathway. So those are left intact for their normal immune function. Now just giving a bit more detail to this process or just showing C1q binding to the synaptic surface and how this leads to the deposition of C1q, C3 and C4 and C5 on the synapse surface. These first three components are recognized by microglial cells largely through the receptor CR3. So all of these molecules are important in the pruning of synapses away from the neuron, whereas C5b to 9, as I said before, will be causing -- can cause membrane damage. But what we found by blocking C1q at the top is that we're blocking all of these components. So we're blocking C1q C4, C3, and C5 deposition on the cell surface. Now this is in contrast to other approaches of blocking the complement pathways. So inhibiting C5, for example, all of these upstream components that I've mentioned will still be in place. And microglial cells will still be able to prune the synapse. Inhibition of C3 will block this microglial recruitment, but it will still leave C1q and C4 in place, which can drive further complement activity and membrane damage directly. So -- and then by blocking C1q obviously, all of these components are gone and will allow the synapse to remain and continue functioning. So that's the background of CMND and the complement cascade, and I'm going to talk a bit about some of our lead indications -- two of our lead indications, Huntington's disease and ALS. So in Huntington's disease, shown on the left side, just looking at the histology, there's a lot of deposition of C1q and C3b in this -- item. This is both on neurons and on synapses. On the right side, we're looking at an MRI image of the entire brain. And this is looking at microglial cell activation. So in the Huntington's patients, they're sort of in the middle of the brain. There's a lot of microglial cell activation. And this is important because microglial cells are the major producers of C1q in the CNS. And they're also are responsible for them pruning synapses once there's been a complement deposition. Moving to an animal model of Huntington's disease. This is the R62 mouse model, a very aggressive model of Huntington’s. We treated these animals with our drug and measured -- we're able to measure levels of drug in the brain. This is important because we're trying to effect the central process. So we know that we have good levels of drug in the brain and that we're inhibiting C1q as shown there on the right. Now importantly, we found that treating these animals reduced the levels of neurofilament light chain, which is an important marker of neuronal damage. NfL is made within neurons, it's specific to neurons and as their damaged, is released into the brain and into the circulation. And so with 7 weeks of treatment with our drug, we found that there was significant lowering of neurofilament light chain, reflective of neuronal protection. Now looking at ALS, again, there's a lot of complement deposition in the spinal cord, in this case, of ALS patients. So that's looking at the control versus the ALS patients, a lot of staining. There, we see C1q in brown, and it is staining both synapses and neurons. And so it's quite a general stain. Then looking at -- so that's within the spinal cord, the central nervous system. Looking at the right side, we're looking to neuromuscular junction. This is where the peripheral nerve joins the muscle and innovates it. And so this is the dark structure that you can see in control patients. And then on the right side is ALS patients, again, the dark structure is still there, but it's now surrounded by C1q deposition. And so what this paper found was that C1q is deposited arm on the neuromuscular junction prior to its destruction, so implying that it could be playing an important role in complement-mediated destruction of synapse. Now looking at our animal model of the ALS, this is one of the SOD1 models. We treated these animals with our drug and found that it was able to occupy or inhibit C1q in the plasma as shown on the left side. On the right side, it's entering the spinal cord and inhibiting C1q there. And an important thing to note is how much C1q is accumulating in the spinal cord compared to control mice. And then in the muscle, we were also able to -- on the drug is able to enter the muscle and block C1q here. So it's really blocking C1q and all of the key compartments. Again, in this animal model, what we found is the treatment with our drug block neurofilament light chain. So first, note the difference in NfL accumulation within the CNS -- within the CSF of control mice in the black versus the maroon squares. So there's about a 600-fold increase in neurofilament light chain. So reflecting a great deal of ongoing damage in these animals as it was actually in the -- from R62 model of Huntington's. And again, treatment with our drug, in this case, for 9 weeks, allowed a reduction of neurofilament light chain consistent with protection of neurons. Importantly, this reduction in NfL translated to a function. So here, we're looking at our compound muscle action, potential electrophysiological measure of muscle force for muscle strength. And again, this is reduced in the SOD animals compared to controlled mice and is increased then by treating the animals with the drug in this case for 9 weeks. So in summary, C1q is mediating -- is complement-mediated neurodegeneration via synapse loss. It really becomes part of the neurodegenerative process independent of the insight in etiology. And we found, and many other labs have found, that inhibiting C1q protects these synapses as well as cognitive, behavioral and motor function. So with that, I'm just going to give a teaser for the next time, and this is looking at ophthalmological diseases. So here, we found that CMND is playing an important role in diseases such as glaucoma and also in geographic atrophy. Showing here is a model of photoreceptor damage caused by bright light as a model of GA. And we're seeing synapse labeling very strongly with C1q, seeing microglial cell recruitment and loss of the synapse. So there's a very strong correlation between synapse density as well as C1q labeling, but we'll talk about this more next time. So thank you for your attention. And with that, I'm going to turn it over to Dr. Henrik Zetterberg. Thank you.

Thank you very much, Dr. Yednock for the kind introduction. So my name is Henrik Zetterberg, I'm a professor of neurochemistry at University of Gothenburg and University College London. And I will talk about my favorite biomarker, neurofilament light. So I will give a perspective on neurofilament light as a biomarker. And I will try to address how it can be used to detect onset and intensity of neurodegeneration. So this is a bit of a busy slide, but I often show it because this slide shows that biomarkers we have at hand at the moment, that measure different compartments in the brain and subcellular compartment of relevance to neurodegeneration. So we have -- if you look in the upper right, synaptic markers, we have towers and axonal marker. Neurofilament light is a marker that is highly expressed in large caliber bilateral axons. And when those fall apart because of neurodegeneration or that you have acute neural injury on the basis of a stroke or traumatic brain injury or something like that, then neurofilament light gets out in the extracellular fluid that communicate freely with service bind fluid. And it turns out that the communication with blood is also quite direct. We also have assays to measure Alzheimer's pathologies and astrocytic and microglial markers and amyloid and some other proteins of interest. But this talk will be focused on neurofilament light. This marker, if you look in the Olsson biomarker database, which is a biomarker database you can look into at Olsson forum, we see that neurofilament light is the second most Alzheimer associated biomarker. It is in studies of mild cognitive impairment, we see that high neurofilament light levels correlate with a loss of cognitive function. So if you are in the upper range of the NfL levels, then you lose more points on cognitive tests. You also lose more brain volume if you see the NfL level is high. So ventricles get larger and the whole brain volume shrinks more in the higher end of these values. But the marker is positive across neurodegenerative diseases. So in this slide, we have looked at our -- at Swedish memory clinics and the diagnosis that were made and then compared with the CSF, NfL concentrations. And in the right most part of this slide, you see frontotemporal dementia. Frontotemporal dementia and ALS are the two neurodegenerative diseases with the highest levels. But also vascular dementia typically has high NfL concentration, but also Alzheimer disease and other neurodegenerative dementias go with increased CSF, NfL. We can also measure NfL in blood, and this is because we have developed sensitive analytical tools. So traditionally, we have measured CSF, NfL concentration by enzyme-linked immunosorbent assays. And these can typically measure down to concentrations of around 50- to 60-nanogram per liter if the assays are good. Lower concentrations are impossible to measure with ELISA. So we worked quite a bit in other groups to develop -- discovery-based assays that were tenfold as sensitive as ELISA. But that was not good enough for blood until we started to use single molecule array technology, which is around 100- to almost 500-fold sensitive to ELISA. And with this similar type of assay, using the same antibodies and the same calibrator protein, we can get this type of nice correlation that you see here between CSF, NfL concentration on the x-axis and the blood NfL concentration on the y-axis. And this correlation is -- it's a quite good correlation. So we think that we can use blood NfL almost like a surrogate of CSF, NfL in many conditions. We have also in traumatic brain injury, been able to compare repeated CSF and blood samples from the same individuals. So in the A panel, you see CSF, NfL changes following traumatic brain injury. And in the B panel, you see the corresponding blood levels. and you see that they follow the same pattern, which, to us, means that the dynamics of CSF, NfL turnover is similar in blood. So CSF and blood have similar kinetics of NfL, which I think is an important result, especially when we now try to use NfL as a biomarker in clinical trials. So this is a study on NfL concentration in plasma in Huntington’s disease. And in the E panel, you see how -- plus NfL concentration correlates with progression of Huntington's disease. And there is a quite clear increase. This is a logarithmic scale on the y-axis, and you see how as Huntington starts and gets progressively worse. You get higher NfL concentrations with this type of progression. NfL is also a prognostic indicator in ALS. So here, you see a longitudinal study on ALS and you see that a high NfL concentration correlates with disease progression and the impact on survival is you have almost a twofold increase in death rates if you are high in plasma NfL at baseline. The markers have also been used in treatment trials. So in the left part of this busy slide, you see how NfL levels change from baseline in response to treatment with ocrelizumab in multiple sclerosis. And you can also in the upper part here, compare with interferon beta treatment. You see in the upper left here, the blue is treatment with ocrelizumab, which is a more effective treatment and that leads to a stronger reduction in NfL concentration. We have also done studies in the middle part on spinal muscular atrophy, which is a childhood neurodegenerative disease where you lose spinal motor neurons. And this is a disease which is quite severe, especially the type 1 form of SMA. It's caused by a mutation in the gene encoding survival of motor neurons and a [indiscernible] treatment has been developed by you injecting the lumbar sac of these kids, an antisense oligonucleotide that stabilizes the transcript of the missing genes, so that the protein can be expressed a little bit more and then these kids survive, and they can be -- they are clinically improved. And NfL concentration in CSF is lowered in a very nice way after -- already after three doses, you see a lowering with which -- actually, you can see this lowering in 4 to 8 months -- weeks, sorry, post first dose. Then we also have another -- in the rightmost part of this panel, you see NfL levels in response to treatment with antisense-mediated downregulation of mutated SOD1. And then you see that the plasma and CSF is levels go down in response to this treatment, which is a very promising finding in ALS in this case. This is a case series which was important to me because I've always thought that, well, the apparent half-life of NfL in CSF from blood is 2 to 3 months following acute injury. So I have been thinking that in 6 months, in clinical trials, one should be able to see a reduction in NfL if the drug is effective at reducing the intensity of the neurodegenerative process. But this is the case here that in the start we did together with Institute of Child Health in London. And the treatment here was enzyme replacement treatment in neuronal ceroid lipofuscinosis. And this is an established treatment. It's not a wonder drug, but it works. But in the A panel here, you see how noisy the CSF/NfL concentrations are 300 to 500 days post the first infusion. But then if you follow these kids longer, up to say 750 to 1,000 days post the first infusion, then you see that eventually the NfL levels go down. This has made me a little bit humble in regards to how quickly one should expect changes in NfL because this treatment does something to the disease process, but we had a hard time seeing it during the first years of the study. But I must admit that this is a K-series and the numbers are low, but I still wanted to show this because this could stimulate some discussions on how to interpret NfL. Now I got two slides. Here, we have the conclusions on NfL. CSF and plasma NfL are biomarkers of neurodegeneration/neuroaxonal injury. It is a dynamic market, and this works in both CSF and blood. High level suggests active neurodegeneration, ongoing injury and drugs slowing neurodegeneration should reduce or, at least, stabilize NfL levels, but it is a slow marker and one should be careful with over-interpreting a lack of reduction. And perhaps if there are other signs of drug benefits, one should continue the trial as long as possible, basically. And here is my thank you slide. And I'm happy to discuss this presentation with you. But now, I would like to hand over to Dr. Beth Stevens, who will continue presenting in this exciting symposium. Thank you very much.

Thank you, Professor Zetterberg. My name is Beth Stevens, I'm a faculty member at the Boston Children's Hospital and the Broad Institute. And I'm happy to tell you today about some of our work involving the complement cascade and microglia in mediating synapse loss, not only in normal health and development, but also in the context of neurodegenerative diseases like Huntington's disease. Now as you heard from Ted earlier today, we have shown previously when I was a post doc at Mentis' lab, it's a classical complement cascade plays many roles in the brain, but in particular, our work is focused on the role of this new pathway in synapse loss or synapse elimination, which is a normal developmental process by which neural circuits are sculpted. It is necessary for precise brain wiring and what we showed a number of years ago was that complement, much like an immune system, these immune proteins are essentially tagging synapses during development. And that these complement proteins, including C1q, C4 and C3 are recognized by resident microglia, resident macrophages. And this is an instructive signal that tells them to prune or engulf specific synapsis during development. And what we went on to show is that microglia actually phagocytosed parts of these synapses are axons during development. And if you block either C1q, C3 or the receptor on microglia, that can lead to defects in pruning and synaptic connectivity that has behavioral and functional consequences. Now over the last many years, work for many groups, including our own, have gone to show that the same or very similar developmental pruning mechanism mediated by the complement cascade in microglia can become aberrantly reactivated to drive or mediate pathological synapse loss in a number of different disease models ranging from Alzheimer's disease, glaucoma, FTD and others. Today, I'm going to tell you about some unpublished and new works that's focused specifically on Huntington's disease. Now Huntington's disease, as you know, is the most common autosomal dominant neurodegenerative disease characterized by involuntary movement, dementia and psychiatric symptoms. We know this is caused by a CAG repeat expansion that translates into this polyQ repeat of mutant Huntington protein. Although we certainly know the gene, and we also know this selective impairment of the basal ganglia and certain circuits in the brain, that surprisingly, we still don't know much about the mechanisms, in particular relative to this early pathological synapse loss and cognitive impairment, which is what we're focusing on. And one of the reasons why we decided to look at Huntington's disease. So despite the fact that these are monogenetic disorders, there is this really selective vulnerability, both with respect to the circuit, the brain region. And also, as you'll hear in a moment, the specific synopsis that are lost. Now here is the circuit on a larger level. This is the connections between the cortex and the basal ganglia, or specifically the projections to the medium spiny neurons in the striatum. And if we zoom in on that circuits, we can actually label these synapses with different antibodies, for example, selectively. So the inputs coming from the cortex to these medium spiny neurons can be labeled with an antibody against presynaptic, a marker called VGluT1 whereas other inputs from other brain regions like the thalamus can be labeled with VGluT2. And so using these tools, we decided to go and ask a number of questions. We want to better understand, one, what are the mechanisms that regulate this process and in particular, in the context of today's talk is the complement cascade in microglia actually driving or mediating some of this function? So how early does the synapse loss occur? Could it be modeled in a mouse? And I'm going to convince you, I hope that it can be, thanks to some very good genetic mouse models. Two, what are the molecular mechanisms? Does it involve some of the same pathways that I told you about earlier? And most importantly, can early intervention of the complement cascade at the level of C1q and down to C3 and CR3, can we intervene and then protect not only synapses, but different aspects of disease progression, in particular, with a focus on some of the cognition and cognitive aspects? And so these are experiments I'm telling you. It's unpublished work for long-standing collaboration with my lab, William Yang's lab at UCLA and a number of other collaborators, including William Fall and others that involve human samples in this tissue. And this is Dan Wilson, who's post-doc in my lab, who's really been meeting these studies. Now fortunately, there are a number of animal models of genetic models. Many of them have been developed by William Yang's lab that enable you to really model some of these mechanisms. And in particular, early stage disease, which is what we're going to be focused on. We're also pairing some of the preclinical animal model data with validation in both human brain tissue from Huntington’s disease patients at different stages as well as CSF, which I'll tell you about at the end. And let me kind of run through some of the main findings. First of all, using a number of tools that my lab has developed and the antibodies and approaches to visualize these synapses, we've been able to, for the first time, really hone in on quantifying synaptic loss in these various animal models of Huntington's disease. And the take home is if we stay in the brain sections with these different antibodies, in particular, that recognize the cortical striatal synapses versus other synapses, what Dan was able to show was that there's a progressive loss of these particular synapses over disease progression in these animal models. This is a q175 Huntington model, but we also see some of the other models, including BACHD. And most interestingly, and relevant to the human, there's also a selective loss in the striatum versus other brain meetings like the hippocampus, for example. So now again, using those antibodies, 1 and 2, Dan was now able to take more quantitative approach and not only look in the animal models, but also in human. I also wanted to mention that the synapse loss, I think this is one of the first real observations that this is happening very early in the context of disease progression, which is also relevant to human. In this case, as early as 3 months before any overt clinical and other pathological symptoms in these mice. Then again, we see the specificity in the cortical stratial VGluT2 synapse loss versus the -- sorry, but VGluT1 versus the VGluT2 synapse loss. Moving on, we got samples of brain from the New Zealand Human Brain Bank from Richard Faull, and Dan was able to take a similar approach and quantify synapse loss of these brain samples in Huntington's versus control. I'm just showing you a representative image, but he's quantified this and similarly shown that in the HD2, this [indiscernible] grade 2 tissue, which correlates with mild and moderate clinical score, we already can observe decrease in this cortical striatal synopsis that gets progressively worse at the later stages of disease. And interestingly, and this is very rare tissue, we were able to do some biochemistry and also show that there seems again to be a selective loss of those VGluT levels in these presymptomatics, so very early HD1 stage versus control. So again, this is just data and from the human that's again supporting the idea that the synapse loss that's selectively lost in Huntington's patients much like CMND animal models. So what about complement? Is it upregulated? Is it localized to these synapses? And if so, is there specificity in these cortical striatal synapses? So again, using very similar approaches, Dan was able to show a significant up-regulation of C1q in these cortical -- synapses even as early as 3 months compared to other brain regions or other synapses. And again, looking not just with C1q but also C3, we were also able to show a very similar upregulation and synaptic localization in the caudate nucleus brain autopsy samples from Huntington's patients. You can see an elevation in both C1q and C3, which is, again, being modeled in the animal models. So we're constantly kind of going back between the preclinical models and the human tissue to try to do validation. So what about the microglia? Now these are cells, of course, my lab has been studying in many contexts. And they do a lot of things in the brain, ranging from neuroinflammation to synaptic loss and pathological synapse loss. So again, using the tools that we've developed for other studies in the lab, what Dan was able to do was to label the cortical striatal synapse in those mouse models and then measure the engulfment or phagocytosis of synapses that are labeled by microglia. So you can, for example, inject dye or tracers into the motor cortex, they project to the striatum. And using the methods we've developed, what Dan was able to show and quantify, is microglia engulfment of those cortical striatal inputs. And then you can look across disease progression and then you can do that after perturbing the different pathways. And this is just showing a significant increase in the engulfment of synapses, again, suggesting these mechanisms are at play, pretty early in the context of Huntington's disease. So what you're all probably wondering is can we actually target this pathway, not just with genetic models, but also its therapeutic blocking antibody like the one I'll tell you about from Annexon? And then does that actually reduce the synapses dysfunction? So we've used a number of approaches. I'll take you through both the antibody blocking data as well as the genetic studies we've done. And this is just telling you that we can target both C1q, the initiating protein of the cascade that then activates the pathway that leads to deposition of C3, so we can also block C3 and we can block CR3. And I'm just going to summarize some of the main findings, beginning with the C1q blocking antibody. So here, using Annexon's M1 ANX005 C1q blocking antibody, what Dan was able to do is inject in 4-month-old q175 HD mice and wild-type mice IP. In addition, another cohort of mice, all done blind with a control IgG. And then over the course, every 48 hours, continued these dosing and then did weight measurements. And then after about 1 month, so at 5 months, performed a number of assays. One, we quantified this number of cortical striatal synapses. And two, we've done some slice electrophysiology as well. And I'll show you both of those data that are quite exciting. So we assessed again after 1 month of regular treatment with 40 mgs per kg of this ANX005 C1q blocking antibody versus control IgG. So here, again, we're blocking C1q and they're reading out synapse loss and functional synapse loss by electrophysiology. Here is showing evidence that the antibody gets in and reduces C3. This is the C3 tag that's recognized by microglia. And when you block with C1q, we see a significant reduction in the C3 in the brain, which is good news. And then quite excitingly, when Dan quantified synapse loss using the same antibody markers, VGluT1 and Homer, again, its cortical striatal synapse markers, showed a significant increase or more synapsis after 1 month treatment with a C1q blocking antibody. So it's protecting some of those synapses. We recently went on to ask whether their synpases are functional, are there functional benefits, so that required some electrophysiology. So Kevin Mastro in collaboration with Dan Wilson and my lab, did some slice physiology from the same paradigm after 1 month of treatment and quite excitingly showed also increase in the frequency of firing in a number of other electrophysiological readouts I don't have time to tell you about. But basically, the data tells us that this is also improving or protecting some of the functional synapses in this mouse model in the cortical striatal circuit. So what about behavior and disease progression? Again, using a combination of tools and approaches, we started to look at behavioral readouts like some of the motor defects like rota-rod, all the way up into very exciting data that I'll tell you at the very end that suggests that protecting complement or blocking complement cascade leads to a protection in cognitive and other behavioral phenotypes. So here, we're taking the approach where we're going to block C3. This is a genetic cross C3 knockout with the BACHD mouse. This is a different HD-mouse model that William Yang developed. So he did all these crosses in his lab, and then they sent the tissue to us to quantify synapse loss. And what Dan found, this is again at couple of different stages. I'm going to show you, in this case, the 12-month old data, which is quite exciting, is that mice that BACHD have a significant decrease in synapses, but when you knock out genetically C3, we see protection of that synapses loss compared to the BACHD. So this is important because it's also showing we see sustained improvement of synapse numbers. What about behavior? This is still work ongoing, but at least initially, the rota-rod looked quite promising in the same cross as we see a benefit or prevents the development of some of these motor defects at the 12-month age as well as some differences in the open field. And we're now looking at some of the earlier time points that's work in progress. In my lab, we're taking another strategy in addition to looking at C1q blocking and C3 blocking, we're also wanting to just manipulate the microglia, CR3. So in this case, Dan crossed the CR3 knockout mice with another HD mouse model, in this case, the q175, and much like we saw with the previous work, this is also showing that blocking the receptor on the microglia that recognizes complement can also lead to significantly more synapses even at those earlier time points. And as I mentioned, caught of the press data from Dan, where he's been developing more cognitive tests, cognitive flexibility tasks, and what these data show is a genetically ablating complement receptor 3, against CR3 knockouts cross to the q175 prevents the development of early visual learning and cognitive flexibility deficits that have been observed in these HD mice. So we recapitulated some of the published work that the CR3 knockouts have these defects. But as you can see on the far right, the CR3 deficient HD, which is purple, looks very similar than the wild type in this cognitive flexibility task. So this is quite exciting because it's really targeting a cognitive function that we think is more relevant to some of the synaptic mechanisms. And this is also relevant beyond Huntington's, of course, because the strongest correlative cognitive decline is synapse loss. And so trying to really pair behaviors with the synaptic phenotypes is what we're trying to do. And I think this is data to support that here. And in the very last minute, I just wanted to tell you that we're very interested, of course, in understanding and trying to translate these findings not only by trying to validate some of these mechanisms, but also to see whether we see changes in the context of these -- sorry, I'm having some problems here. Sorry about that. So what we're actually now doing is measuring some of the changes in CSF with these complement levels. Analysis of complement levels in HD clarity cohort of HD is work ongoing in the lab. But quite excitingly, using this -- these samples collected from the HD Clarity cohort, Dan has been using ELISA-based methods, much like you may have heard already from Annexon, to look at and measure and C3 fragments in the early stage of this cohort in their CSF samples. And the data thus far are quite promising and show that the levels of C3 and their activated cleavage component called iC3b are, in fact, elevated and altered in the context of disease progression and disease stage. We see a significant increase in the late pre-manifest Huntington's and early manifest HD compared to the earlier stages and of course, compared to controls. And quite excitingly, this work also shows that it also is correlating very nicely with the CAT score, which is a measure of disease burden. So on that last note, just wanted to acknowledge a lot of collaborators in this work, which made it all possible and for all of you for listening. And with that, I think I'll turn it over to Professor Wild. Thanks very much.

Thank you, Beth. My name's Ed. And today, I'll be talking to you more about Huntington's disease and the unmet need in the field where neurofilament light might fit into that need and some thoughts on therapeutic development. So let me introduce you to one of my patients. This is Kim, and she has Huntington's disease. You can see why it used to be called Huntington's chorea because she has chorea. These involuntary jerk-knee movements. But the other thing you can see from looking at Kim is that she has lots of voluntary motor function, loss of balance and that is more disabling than the chorea. There are treatments -- medical treatments for chorea, but we can't treat that lots of balance and lots of voluntary motor function, nor can we treat the cognitive dysfunction that's progressive, which is essentially a form of dementia. And many of the behavioral features of Huntington's disease such as personality change, aggressive behavior, depression and anxiety are very difficult or impossible to treat. So a large unmet need and no treatments for most of the features of HD. Even in the symptomatic realm, most of the treatments we use don't have a strong evidence base, although they probably work, but it hasn't been shown in the disease-modifying sphere. There's nothing at all that's been shown to slow or reverse the progression of HD. So it's a pretty bleak landscape. And when I think in therapeutics, tend to divide them according to this conception I have of HD as a river. With a single stream at its source, the gene and the mutant Huntington protein, a very sort of palpable single thing to blame for everything that happens. But as soon as that protein is produced, the stream branches into dozens of dividing, interacting, reforming pathogenic pathways, each of which is a potential therapeutic target. But we need to choose them wisely and know how to modulate them if we're going to have any hope of making a reasonable impact on what happens next, which is that the streams reunite into neuronal dysfunction and death. Things then get complex and we get clinical progression and things are really simple again at the point where the patient passes away. But you can broadly divide the therapeutic approaches to slow HD into these very upstream events targeting the gene and protein and then the more pathway targeting approaches. The best and most famous example of the former is perhaps the Roche tominersen program depicted here as Ikaros because I think that, that program failed earlier this year. It was trying to lower the concentration of even Huntington. I think it failed because the dose used was too high. I don't think it's because lowering Huntington is a bad idea. But what it does show is that it's an extremely difficult thing to do therapeutically without accidentally causing harm along the way. This is my current sort of sketch of the Huntington lowering sphere in Huntington's disease. As you can see, although rush is at least temporarily sidelined, there are a number of entities pursuing these Huntington lowering approaches quite aggressively, including oral Huntington splicing modulators. So this is a busy sphere. It's possible that all of these will work or that none of these will work, but I suspect we'll probably end up slowly somewhere in between the two, where it's shown that Huntington's disease has some effect. But because we will never either want to or be able ablate Huntington completely, there'll always be some of that Huntington protein in that stream, causing dysfunction in those pathogenic pathways. And so here, in the downstream therapeutics target area, there's, I think, always going to be scope for something that can make a real difference. Well, how does that sphere still look right now? I mean we've had over 100 therapeutic trials of things like coenzyme Q and creating nutraceuticals, substances thought to be largely good for the brain in a broad sense. Things are a little bit more focused now and that there's a sort of claimed mechanism for most of the things that have been tried. But in all honesty, even the current generation of pathway targeting therapeutics really were medications in search of an indication. And the evidence that those indications might be relevant to HD largely was generated after it was decided to apply the therapeutic to Huntington's disease. And hopefully, we can move beyond that. But I think what this calls for is pathway targeting interventions that target solid well-described arrangements, things that we've known about for a long time, where if you asked 100 scientists what are the top 10 things you would like to target in Huntington's disease, they'd be on that list. And I don't necessarily think that's the case of the things I showed on the previous slide. The other thing I think we need in order to get really solid evidence for moving into patients and this biomarker support from the preclinical stage to translate into clinical and then from early stage to later-stage development. I really think that biofluid biomarkers need to be and have every right to be a key decision-making tool along the way. And this is where neurofilament light comes in. We've heard already what it is and how it may have been helpful or may continue to be helpful in other diseases. What about in Huntington's disease? So NfL and CSF was first described as being elevated in 2009, quite striking separation between controls and HD patients here. The first large detailed study was done by my group in 2017 using samples of data from the large track HD longitudinal study. And what we showed in those blood plasma samples was the NfL, top left, was elevated with every successive disease stage compared to controls and compared to the previous stage. Bottom left, you can see that, that elevation is CAG repeat dependent. The more CAGs you have in your Huntington gene, the earlier and more steeply the NfL rises. We showed that baseline NfL predicted subsequent brain atrophy. And in the bottom right, you can see that if we divide the pre-manifest, i.e., healthy mutation carriers who haven't been diagnosed with HD, we divide them in half according to the median NfL, the ones with the higher NfL were much more likely to progress in the subsequent 3 years. So the bulk of what we know about NfL in HD longitudinally comes from the HD-CSF study, which I led at UCL Institute Neurology. This is a 24-month study of 80 participants with brain imaging and multiple lumber punctures. So these are data from the baseline. And in all the slides that follow, you can kind of ignore the green, which is studies of mutant Huntington, although obviously, you can look if you're interested. Blue is NfL and CSF. Red is NfL in blood. And top left, we can show really striking associations with clinical scores, motor score here and really striking associations with both regional and local brain volumes, so regional and global brain volumes. And using a bioinformatics method, we were able to show that these changes were among the earliest detectable changes in Huntington's disease, preceding any of the other clinical or HD imaging changes. In 2-year longitudinal analysis, which was published slightly less than a year ago in Science Translational Medicine, we showed that NfL rises really consistently throughout the spectrum of HD in a way that's entirely distinct from its profile and controls. We were able to derive estimates of how quickly we would expect NfL to rise in CSF and in blood. And these can be used -- if we assume linearity, these can be used to estimate how much NfL would be expected to rise over any smaller or any larger interval with the caveat that over intervals smaller than this 2-year one that we actually measured, those estimates may not be so reliable and may, of course, have less statistical power, but useful rules of thumb perhaps. We showed that the NfL level was a very strong predictor of clinical progression here shown with the composite UHDRS score and using a random Forest analysis, we were able to compare all of these markers head-to-head, not only the biofluid markers but also more traditional measures like predictors of progression like age and CAG repeat. And what we showed was that the NfL at baseline was the strongest predictor, but that NfL and CSF and plasma, the rate of change of NfL was an independent predictor of progression. So if you have high NfL at the start, you progress quickly. But if your NfL then changes more during that period, you're likely to progress even further. And NfL is really clearing all of the hurdles when it comes to validation as a biomarker, and it's now being looked at in a huge HD clarity data set. So how does that translate to utility in an early phase clinical program? I think there are 3 potential cases where NfL can declare or can allow a drug developer to declare victory. Clearly, the gold medal comes from if your placebo patients, NfL increases during your trial and your active treatment NfL falls either towards or to the control level. I think in that situation, it's a pretty open and shut case that you have protected or rescued neurons and that's likely to be highly significant. And I think a silver metal comes if you don't perhaps lower NfL but you slow the increase in NfL. So in this case, haven't gone down towards controls. But what you have done is gone up less, so that the active treatment arm is now distinct from your placebo arm. And I think that would also be thought of as being a potentially important result. But even in this case scenario, this is -- I still think that bronze medal because in this case, you haven't slowed the increase in NfL, but crucially, you haven't caused a dangerous increase in NfL either because NfL, I think, is also a potentially really important safety biomarker. If NfL goes up in your trial, it's safe to assume that, that is because your drug has caused some kind of unexpected harm. And how we -- and so if you don't deflect NfL down, all is not lost. It may just be that the trial wasn't short enough, but crucially, you haven't done harm, which I think turns out to be quite easy to detect because it's now long been known that the Roche drug tominersen did cause an unexpected spike in NfL. And at the time, we hoped that, that might somehow be an occult sign that the drug was doing something therapeutic like releasing NfL from aggregates. In retrospect, this was a fairly loud canary in the coal mine, which was a sign that the drug was causing some sort of early harm, and that, that was then not something that could be retrieved during the later course of the trial, even though the NfL fell. So NfL going up, potential safety signal and a very important signal, but 3 potential scenarios where NfL gives you strong reason to proceed with the program. That's all I have time for. Happy to take questions, including questions on the Roche program to whatever extent I can talk about them. And with that, thank you for your attention, and I'll hand over to Dr. Genge.

Thank you, Professor Wild. I am Dr. Angela Genge, Director of the Clinical Research Unit at the Montreal Neurological Institute at McGill. And today, I'm sharing with you some information really on the clinical aspects of ALS as it pertains to this program. As you may be aware, I run a clinical research unit at the Montreal Neurological Institute and an ALS Center of Excellence, which provides care for patients with ALS. I've been heavily involved in trial design and clinical trial design in ALS programs and have focused most recently on early drug development for ALS. Today, I'm going to talk about ALS, which is amyotrophic lateral sclerosis or Charcot's disease. It is a fatal neurodegenerative disorder, which is caused by progressive loss of upper and lower motor neurons. These motor neurons exist throughout the upper and lower extremities and in the bulbar region and about 75% of patients initially present with a weakness or difficulty using 1 of their 4 limbs. About 1/4 of the patients actually present with a change in their speech, a slurring of speech similar to when you had too much alcohol to drink or swallowing. It is a devastating disease in that it affects patients with no warning, and patients die within 5 years from disease onset. About 90% of cases are sporadic. And at any one point, 225,000 people are affected globally. However, if this disease became a chronic disease, the number of patients would increase significantly as the actual incidents is similar to multiple sclerosis. Disease progression in ALS is extremely important to understand. There's a progressive decline in motor function and in activities of daily living, which is the reason that the -- a primary outcome measure that is most frequently used in ALS clinical trial programs is one called the ALS Functional Rate Scale or ALSFRS, for short, which looks at changes in gross motor, fine motor, bulbar and respiratory function and shows a decline on average of 1 point per month in the clinical trials. There are fast progressors and slow progressors, but the average is a change of 1 point per month. Media and survival, as I earlier mentioned, is 3 years. And that contrast with the other big neurodegenerative diseases, I think of them as the big 3, ALS, Parkinson's and Alzheimer's. And this is the one that has the most rapid progression and ultimately is fatal in a relatively short period of time. So survival probabilities currently from the onset of symptoms is 12 months in 92% of patients, only 5 -- only 28% of patients live 5 years and approximately 13% live 10 years. If patients have no intervention to replace the respiratory function, such as a noninvasive ventilator or an actual ventilator, death usually occurs between 3 and 5 years from onset. ALS affects motor neurons, wherever they exist. So they exist both centrally and peripherally. The central nervous system has upper motor neurons that extend through the corticospinal and cortical bulbar tracks and originate in brain primary motor cortex. These neurons are responsible for carrying impulses from the voluntary motor region of the brain through the corticospinal -- cortical bulbar tracks to other motor neurons that exist in the bulbar region and the spinal cord. The lower motor neurons, or the anterior horn cells, carry impulses from the spinal cord or from regions of the brain stem to the muscles that they innovate. These motor neurons are what die off in ALS. And we have discovered recently over the last couple of years that a very good marker for the loss of motor neurons in patients with ALS is, in fact, neurofilaments. These neurofilaments are measured in both the serum and the cerebrospinal fluid. And natural history studies suggest that elevation of neurofilaments correlates with symptom onset in ALS patients, can separate ALS patients from patients with diseases that mimic ALS and correlate fairly well with the rate of clinical progression. What we mean by that is the higher the neurofilament levels are above normal or level of detection, the more rapid we can expect that patient's ALS to progress. Simoa assays, or fourth-generation single-molecule arrays, have enabled these neurofilaments to be detected even more reliably and sensitively particularly in the serum. Prior to the development of the Simoa technology, the neurofilaments levels were being obtained solely through this free grow spinal fluid and were found to be elevated in the CSF, particularly the phosphorylated heavy chain neurofilament. Neurofilament levels have now been shown to rise up to 1 year prior to symptom onset. Although before these studies done by Michael Benatar following patients with genetic forms of ALS, we were noting that elevated neurofilament levels could be used as supportive indicators of the presence of ALS in patients presenting with typical signs and symptoms of ALS. So we have gone from recognizing elevated neurofilaments in the CSF as a supportive diagnostic test validated around the world to now being able to look at serum and CSF of patients with ALS and with the genetic mutations and be able to follow these patients prior to symptom onset and identify when they are going to develop ALS prior to the onset of symptoms, a really important recent development in our understanding of ALS and the use of neurofilaments. Neurofilament light serum, neurofilament light, has now been clinically validated as a prognostic biomarker, again by Michael Benatar and his group. And it's been demonstrated that you can observe it over time and that you can also observe that serum neurofilament levels are stable over time. Once they increase in the serum and CSF using Simoa technology to measure the increase. There is no evidence that potentially serum and CSF, NfL may be a stronger correlation than even phosphorylated heavy chain neurofilaments for the rate of progression and the onset of ALS in a particular individual. Serum neurofilament light is prognostic of future ALSFRS decline or slow book decline and survival as mentioned previously. Reductions in neurofilament light documented in SOD1 ALS patients. SOD1 ALS patients are patients who carry a genetic mutation in the gene SOD1 and are the group of patients that are being targeted with an antisense oligonucleotide therapy, tofersen, in the Biogen program. In the original paper presented by Dr. Tim Miller in 2020 in the New England Journal of Medicine, he presented a number of rapidly progressive patients who responded to tofersen by having a decrease in their levels of neurofilament light in their -- in the -- after being treated with tofersen and this decrease in the levels of neurofilament correlated with a decrease in the rate of decline of the ALSFRSR. So the patients who showed a decrease in their neurofilament levels also showed a sustained function and lack of deterioration in function as measured by their ALSFRSR. So in summary, ALS is a fatal neurodegenerative disorder and this fatal nature has a time course of between 3 and 5 years. The key endpoints are still functional rating scale and survival for registrational studies. However, the FDA called for a biomarker development program in ALS in their guideline paper in 2019. And the neurofilament light appears to be a viable surrogate biomarker in ALS. Large increase in neurofilament light occurs at the symptom onset and then stabilizes at elevated level. It appears that a therapeutic agent that can potentially improve function by slowing the decline of ALS measured by the ALSFRSR, can also have a reduction in the neurofilament light and potentially herald a target effect by a therapeutic agent by decreasing the neurofilament light while the patients are being followed for clinical change. And with this, I wish to thank you, and I'd like to welcome Sanjay Keswani, who is the Chief Medical Officer for Annexon. Thank you.

Great. Well, thank you, Dr. Genge for that really informative presentation. So I'm happy to give an overview of Annexon's clinical programs targeting complemediated neurogeneration. So as Doug alluded to in his presentation, we have three therapeutic areas: Autoimmune diseases, ophthalmology and neurodegeneration. We have clinical programs in all of these three areas but today, we are focusing on neurodegeneration. So we feel we have a really groundbreaking approach targeting complement-mediated neurodegeneration, specifically C1q media synapse loss is a common pathway in mediating neurodegeneration and disability across a wide variety of diseases of both the brain and the eye. And with an anti-C1q approach, we therefore feel that we can protect functioning synapses as well as neuronal health and function. In addition, we're utilizing a biomarker that's really emerged recently. It's a very promising biomarker of neurodegeneration, specifically neurofilament light chain or NfL. And interestingly, elevated levels of NfL correlate with patient disability across a range of neurodegenerative diseases and also reduction of NfL has been shown to correlate with clinical benefit in multiple diseases. And lastly, we have some promise from our C1q platform with respect to NfL reduction being observed in our GBS Phase Ib proof-of-concept study and also preclinically in Huntington’s disease and ALS preclinically. So this slide highlights C1q firstly, normal developmental role, as Ted mentioned, with respect to sculpting the brain and retina. And then the system being quiescent. And then in neurodegenerative diseases gets aberrantly triggered or reactivated in response to a variety of neuronal stresses that occur in different neurodegenerative diseases. And this results in removal of functioning synapses. And as mentioned, we believe that this mechanism is applicable to a whole range of neurodegenerative diseases regardless of insight and etiology. And indeed, as listed on the lower half of this slide, C1q blockade has resulted in functional benefits in a whole variety of disease models including Huntington’s disease and Alzheimer's disease, spinal muscular atrophy, glaucoma, geographic atrophy, frontotemporal dementia, traumatic brain injury ALS as well as progressive multiple sclerosis. Of note, our anti-C1q approach is differentiated from other complement approaches, and Ted highlighted this in his presentation. As with our approach, we offer the only way to block all the early classic complement components that are involved in macrophage-mediated clearance of synapses. Now as Dr. Zetterberg highlighted, neurofilament light chain is a promising biomarker of neurodegeneration and does underline our proof-of-concept approach in neurodegenerative diseases. It's a sensitive measure of neuronal damage in degeneration. It correlates with patient disability and predicts outcomes in a number of different diseases. And Dr. Wild, Dr. Genge mentioned this in the context of HD and ALS. Interestingly, NfL is reduced by effective therapies for multiple sclerosis, spinal muscular atrophy as well as ALS within 3 months of treatment. And we, ourselves, have observed NfL reduction clinically as well as our preclinically, and can highlight that on the next slide. And we're utilizing this as a key measure in ongoing Phase II trials in Huntington’s disease and ALS. So this slide illustrates the 3 examples we've seen so far of NfL reduction by an anti-C1q mechanism. On the left, we have data on significant NfL reduction in GBS patients has occurred acutely in our Phase Ib study of note in GBS for high levels of NfL, which correlates with patient disability. In the middle, Ted highlighted the data on our HD preclinical model, specifically R62 transgenic mouse model, which is an aggressive model of Huntington’s disease. And here, after about 2 months of treatment, we saw an NfL reduction. And on the right, we also saw NfL reduction in another aggressive mouse model, this of ALS, in one mouse model, where we also saw reduction of NfL after about 2 months of treatment. And of note, C1q inhibition has also resulted in functional benefit in patients as well as in animal models. On the left, we have our Phase Ib GBS data, where when we did a responder analysis looking for large improvements in function on the GBS DS, which is a primary approval endpoint for acute GBS patients. And these large improvements were akin to patients going from being bed bound to being able to walk independently or run. And nearly 1/3 of patients treated with 005 have this large improvement versus zero patients in a placebo group. So encouraging data indeed we'are in the midst of the registrational program in GBS. In the middle, we have an example of spinal muscular atrophy, which again Ted highlighted with respect to gain of motor function in these animals being associated with preservation of functioning synapses and also behavioral improvements in animals with a model of front-end temporal dementia. These are just illustrations are other examples, but these just highlight the functional benefit that we've been seeing. This slide illustrates our pipeline in neurodegenerative diseases involving the brain. Huntington’s disease is primarily a brain disease. So clearly, CNS target engagement is important. And we've been as careful, as we can, with respect to choosing dosing paradigm, so 005 intravenous antibody in ensuring CNS target engagement. ALS is a mixed disease involving both CNS and the peripheral nervous system. The target engagement, therefore, in both compartments is important. We also have ANX005, an intravenous antibody with superior dosing characteristics, which is coming into the clinic early next year. And this will offer another opportunity to treat a whole host of neurodegenerative diseases in the future. With respect to our rationale for Huntington disease, we believe we have robust scientific data, underlining our C1q approach in Huntington's disease. As Ted and Dr. Beth Stephens mentioned, aberrant C1q activation has been noted in synapses for Huntington’s disease patients. And then C1q is efficacious in preclinical models of Huntington’s disease. In addition, NfL can be utilized as an objective measure of neurodegeneration in Phase II studies in Huntington’s disease. We know it's primarily from Dr. Wild's work that elevated NfL levels correlate with disability. And that NfL longitudinally shows an increase in levels over time in Huntington’s disease patients. In addition, the dosing paradigm that we're utilizing with 005 should result in full CNS target engagement. Of note, we saw full CNS target engagement in GBS patients and also in preclinical studies as well with a peripherally administered antibody. This slide represents an overview of Huntington’s disease Phase II trial. It comprises a 6-month treatment duration period and 3 months off treatment. Sample size here is 24 patients. The study population are adults with or at risk for manifest Huntington’s disease with a total CAP score of more than 400. And also UHDR's independent score of more or equal to 80%. So these patients are largely early manifest HD patients and should have baseline elevated NfL levels of about 5 to 7-fold above normal. With respect to primary endpoints in this study, this includes safety tolerability of 0.05, PK and PD of 005 in both serum and CSF, and also NfL reduction in both plasma as well as CSF. Exploratory endpoints in this open-label study include quant EEG. Of note, it's been observed that there is decreased power in the alpha band on quant EEG of Huntington’s patients. So exploratory manner, we'll be seeing, we have any impact on that. But also we'll be collecting clinical data with respect to the UHDRS as well as its composite components as well. Again, this will be done in an exploratory fashion. With respect to our data readout, the initial readout is anticipated in the fourth quarter of this year. The patients were all enrolled in May 2021 and will complete 6 months of treatment in the last quarter of this year. Now with respect to the 3-month follow-up period, all patients will complete that in the second quarter of next year. So with respect to the data we'll actually have at the end of this year, we'll have data on a subset of patients, about 16 patients that completed a 6-month treatment period. And this relates to the time taken to collect the samples and to analyze the data. With respect to what kind of data, this will include safety and tolerability of ANX005, target engagement in both serum as well as CSF as well as NfL reduction in CSF and plasma. We'll also be looking at our final data set that's been the second quarter of 2022 with respect to patients on the off-treatment phase, and we'll be looking at clinical exploratory data as well as the primary endpoint that I just mentioned. Now moving on to ALS, which is our second neurodegenerative disease currently in a Phase II study. There is also a strong biological rationale for targeting C1q in ALS. Again, as Ted yet not mentioned, average C1q activation has been noted in the CNS and PNS from patients with ALS. Anti-C1q is efficacious in the preclinical model of ALS, both with respect to NfL reduction, but also a functional benefit as well in the SOD1 mice. In addition, NfL can also be used as an objective measure of neurodegeneration in Phase II studies as we know that NfL levels are elevated in ALS with chronic disability, and Dr. Genge very nicely articulated that, but also published NfL longitudinal data shows stability over time. And that allows us to interpret a drug-induced effect on NfL. And lastly, we anticipate full CNS and PNS target engagement with 005, again, really important to target C1q in both these compartments. And we believe we've chosen our dosing paradigm to achieve this robust target engagement. So this slide highlights our ongoing Phase II ALS trial that we expect to design. So it's a 3-month treatment duration study with 3 months off-treatment follow-up, sample size being 24. Of note, ALS is a rapidly progressive disease, and this dictated the choice of treatment duration for this study. With respect to study population, we're targeting all forms of ALS rather than a specific genetic subset. Of note, we're agnostic with respect to the exciting etiology in our neurogenerative diseases. We believe that C1q synapse loss is a common pathway across all these different diseases regardless of insight and etiology. And of note, sporadic ALS, which would be largely the ALS that we'll be targeting here occurs in more than 90% of ALS patients. ALS will be diagnosed according to the E1 score criteria. The onset weakness would be within 3 years prior to enrollment. Slow biocapacity is more equal to 60% of predicted normal, and the ALSFRS-R would more equal to 30. So these are patients primarily with early ALS in this study. With respect to primary endpoints as we include 005, PK and PD in serum as well as NfL reduction in plasma. And we know, and Dr. Zetterberg highlighted this as well in his talk, that there's close correlation between plasma NfL and CSF, NfL. With respect to exploratory endpoints, this includes electrical impedence and myography, which relates to innovation of motor endpoints in neuromuscular junctions and also clinical scores, such as the ALSFRS-R, ALSAQ-40, which assesses our quality of life for ALS patients as well as handheld dynamometry and slow vital capacity. So these are all exploratory endpoints in this open-label study. So with respect to data readouts for the ALS Phase II study, the initial readouts anticipated in the first half of next year with respect to trial metrics, enrollment is ongoing in North America, and enrollment should be completed in the first half of next year. And that again, is when we would anticipate our initial readout. In terms of the type of data that we will show, this will include taking tolerability of 005 chronic dosing, 005 PK and C1q target engagement, NfL reduction in plasma. Lastly, we'll have some exploratory clinical data, particularly after the 3-month off-treatment period. So in summary, C1q mediated synapse loss is a common pathway of neurogeneration and disability in multiple diseases regardless of insight and etiology. Anti-C1q is protective in several disease models of neurogeneration, including Huntington’s disease and ALS. Elevation of neurofilament light chain is a common biomarker of neuronal damage across a whole host of neurogenerative diseases and does correlate with patient disability. And our objective in Phase II studies of Huntington’s disease and ALS is to show a reduction in NfL as a demonstration of impact of anti-C1q on the neurodegenerative disease process. So with that, I'll hand it over to Doug Love, for some closing remarks and for our Q&A session.

Dr. Stevens and Dr. Angela Genge, we appreciate the good work you've been doing in your respective fields and the work you're doing with Annexon and thank you for being with us today. As you all can hear, we're quite excited about this opportunity with treating complement-mediated neurodegeneration with our anti-C1q classical pathway approach. It's quite well supported by the literature as well as what strikes me a range of studies with multiple investigators across the globe in diverse animal models, all very consistently showing inhibition of C1q protects functioning synapses, neuronal health and behavioral deficit. Because we are targeting a common pathway for neurodegeneration, that is the preservation of functioning synapses, there's broad utility to this approach. And so we're excited to bring it forward across multiple therapeutic areas, including Alzheimer's and Huntington's disease -- or I'm sorry, including ALS and Huntington's disease. And we're, of course, leveraging, which you've heard quite a bit about neurofilament light chain, an established objective measurement of neurodegeneration. So we will continue to build on this approach. We look forward to producing data over the course of Q4 and into the first half of 2022, which we think will go a long way towards validating this approach and bringing it forward to patients in need. And there is a significant need in this approach. Indeed, our classic platform as a whole has the potential to provide benefit to millions of patients across autoimmune, neurodegeneration in the eye and neurodegeneration in the brain. And when you think of just about neurodegeneration in the brain alone, our lead indications, Huntington's disease and ALS, have substantial need and both represent blockbuster opportunities in and of themselves. And finally, I'm going to close and open it up for Q&A for all of you. As I've said, we're encouraged by this C1q classical pathway approach. We think it has broad applicability across a range of diseases in the autoimmune and neurodegenerative space. We're pleased thus far with the data we've been able to produce both preclinically and clinically to validate this platform approach. And we're well positioned to deliver on 7 distinct, well-supported clinical studies over the next 2 years. Importantly, we have cash into early 2024 to do so. And so we look forward to driving these catalysts with intentionality with a winning team, a team that has deep drug development experience. When you're playing with their hearts and minds as you've heard me say before, this is mission-led work for us. Many of the folks in our organization have been impacted with family members and others with the diseases for which we are pursuing. And so we're excited working with intentionality to bring this forward to all of you, and we like the early promise that's been shown thus far. So with that, I want to thank all of you for your attention, and I want to open up the line for any questions that you may have.

[Operator Instructions] Our first question comes from Joseph Stringer with Needham & Company.

Got a question on NfL here in terms of the some -- one of the KOL slides sort of showed the rate of NfL in HD specifically looks like it's sort of plateaued in sort of later-stage disease. And so I guess, relative to the Phase II trial here, you mentioned that patients are going to have their earlier-stage patients, they'll have elevated NfL, but is the thinking that they're still on the up slope in terms of the trajectory of NfL levels such that a 6-month treatment period, you could see a difference in NfL? And I suppose what gives you sort of some level of confidence that 6 months is going to be long enough to see NfL reduction for HD? And then I have a follow-up question.

Thanks, Jay. Good questions, and thanks for participating this morning. I'm going to turn that over to Dr. Sanjay Keswani, but just briefly, the short answer is, yes, with regard to increase of NfL in the patients we selected. As Sanjay alluded to, we have selected early manifest patients who are relatively early in their disease where NfL is back to still be increasing. But Sanjay, if you can elaborate on that, that would be great.

Yes. Really good questions. I mean, so firstly, and I will ask Dr. Wild to actually comment further because primarily it's his work. So we know, particularly in manifest HD patients that there is an upswing in terms of NfL over time. And so some very nice 2-year longitudinal data, which shows an increase of about 20% over that 2-year time period. As you mentioned, the NfL change is sigmoidal in patterns. So at late stage of the disease, it tends to plateau. But we're still targeting patients whose NfL is increasing over time, and we expect it to increase over the treatment duration of study. And with respect to the duration of intervention, that's kind of the million-dollar question here. There hasn't been any examples of NfL reduction clinically in HD. We've looked at our R60 transgenic mouse model, which is an aggressive model of Huntington’s disease obviously, in the mouse. And there, we saw NfL reduction after 2 months. In our GBS Phase Ib study, we saw NfL reduction within a month of treatment. In our other models, such as the SOD1, we saw a reduction in about 2 months as well. And clinically, we know that in multiple sclerosis, SMA as well as the SOD1 subset of ALS, sponsors have demonstrated a reduction in NfL within a few months. But that's something that we still are not 100% sure in terms of the ideal treatment duration. Of note, with respect to the treatment duration in 6 months, and now there is a 3-month off-treatment period, we're expecting some target engagement from that 6-month treatment duration to bleed into the 3 months. So it will realistically be about 7 months or so of target engagement.

Okay. Great. I guess kind of a follow-on to that is a similar question for ALS and that again, correct me if I'm wrong here, but on one of the KOL slides, showed the NfL for -- in ALS patients sort of with a larger increase initially at symptom onset and then sort of plateauing or stabilizing at elevated levels. So I understand that with the Phase II ALS trial, you're a little bit constrained in terms of the treatment duration readout 3 months by the more rapidly progressing ALS. But is the thinking on NfL specifically for ALS is that you'll have enough sort of increase of NfL in your patient population at baseline, again, to kind of see the reduction -- you'll see a meaningful reduction or a reduction and you won't have kind of reached that plateau level?

Yes. So happy to address that. Yes. So what we're trying to do in these diseases is actually look for a reduction in NfL, particularly from baseline ALS patients. And so we know that longitudinal data has shown some stability and the patients we're targeting should have stable NfL levels over time. The increase that was shown was really prior to symptom onset in patients with familial ALS. Many of our -- most of our patients who have sporadic ALS already have symptoms. And so that stable level of NfL is elevated multiple fold above normal levels. So that indicates a high degree of neurodegeneration. So what we're trying to do here is reduce NfL from those high levels, which are, as I said, stably elevated. And that will give us some confidence that we're impacting neurodegenerative disease process in ALS.

Thanks, Joey. Any questions from anyone else on the line?

We have a question from Philip Nadeau with Cowen and Company.

A couple from us as well. A follow-up question on the Huntington's disease data that we're going to get in Q4. It looked like from Dr. Wild's slides that there's a fair amount of variability in both the rate of increase in NfL as well as the absolute levels of NfL in Huntington's patients. I guess first question is, is that impression correct? And then second, given that the Phase II initial readout is only going to have 16 patients, I guess how likely is it that there's a consistent signal of NfL decline in such a small patient population given what seems to be heterogeneity in NfL measures?

Yes. Good question, Phil. Maybe we'll turn to Dr. Wild first to answer the first part of that and Sanjay, you can answer the second part.

Thanks. It is a great question. And I think it's -- there is a good degree of variability between patients. But that, I think, is biological variability. In other words, I think that, that is a genuine reflection of differential rates of neuronal dysfunction and death from one patient to another. The -- for instance, the early stage 1 HD cohort is still pretty wide in terms of the range of disability that it encompasses. And these clinical measures are relatively crude in terms of how they categorize people. What's crucial I think about NfL, though, is that if you look at the stability within participants, it is really quite striking. It's not my slides, but it's in our science translation medicine papers. You can measure NfL at baseline and then just a few weeks later, and it's really strikingly stable in CSF and in plasma. So that, I think, tells you that that's not noise, but what you're looking at when you see that variability is actually a more biological way of dissecting the heterogeneity that we know exists even within relatively narrow clinical balance of HD. And essentially, what we're measuring is the thing we want to slow, which is neural dysfunction. Reliably, though, from baseline to follow up over longer periods of several months, almost everyone's NfL rises. The rise is, again, there's variability there, but that is biological, I think. And so the variability is what we want -- is something that we want to capture because we want to -- we're hoping to deflect each person from their own trajectory. And of course, when you look at whether your drug is deflecting a biomarker endpoint, what you're going to do is take each person's endpoint and subtract it from their starting point. And that, I think, will put everyone on a level baseline, but you can look at the percentage change or absolute change in NfL within each participant.

Dr. Wild, maybe if I could just follow up, while we have you, is there a cohort of patients who have spontaneous reductions in NfL from baseline over 6 months? And if so, what proportion of patients would that be? I'm kind of curious as to what a false positive rate could be using NfL as a biomarker?

That's a great question. And unfortunately, the answer is we haven't looked. In our cohort that the -- by design, it was baseline to 2 years. And there's virtually done looking at the raw data here and the there's virtually no one who fell from baseline to 2 years. I think that when that might happen is if they have an artifactual elevation of NfL at baseline. So for instance, in the context of recent head injury or some other recent undetected secondary brain thing going on that could increase the NfL at baseline. If that then fades away at follow-up, you would expect to fall. But as I say, it's certainly over 2 years, that's extremely rare. It could happen over 6 months. But I'd be surprised if that's a significant contributor to overall signal.

Perfect. That's very helpful.

Sanjay, did you want to add anything on that? Or shall we go to the next question?

I thought Dr. Wild answered that really well. Thank you.

Actually, I do have -- sorry, one more follow-up on ALS. We recently saw the program from AstraZeneca be discontinued. In your prepared slides, you showed some reasons why C5 may not be the best target in neurodegenerative diseases. I'm curious to get your perspectives on the discontinuation of the SOLIRIS study. Why you think that did? And whether there's any learnings that you'll apply to your ALS program?

Yes. I mean I think you know, Phil, we do think the target matters. C5 being downstream really is not blocking upstream macrophage microglia activity associated with the disease. But maybe, Sanjay, you and Ted can elaborate further on our thoughts with regard to that program.

Well, I'll start and then, Ted, you could jump in. So as you mentioned, obviously, we're targeting C1q, which is upstream with respect to the classical pathway. We think that's important in the context of ALS as there is a fair amount of cellular activation, particularly in macrophage activation and macrophage phagocytosis of synapses. And so we think it's important, therefore, to target upstream versus at the level of C5, which really is only targeting the memory and attack complex in these patients. And so therefore, we feel we're differentiated from the efficacy point of view. The second point I'd like to make is CNS target engagement is as well as peripheral nervous system engagement. As one can appreciate, ALS targets the brain, spinal cord, the motor nerves, they leave this spinal cord as well as the neuromuscular junctions. And so we've utilized a dosing regimen that should give us CNS target engagement as well as engagement peripherally. And we think that potentially might be an important area of differentiation as well. But Ted, I don't know if you have any further thoughts on that?

No, I think you've nailed the two main points. That's good.

Our next question comes from Tazeen Ahmad with Bank of America.

It's super helpful. A couple for me. Just starting with NfL. So with the data that you're going to present in the fourth quarter for Huntington's, is any impact on NfL going to be considered meaningful by your definition of what needs to be seen to move forward? And how can we get a better sense about durability of response? So we'll take a look at the data at the 6-month time frame, but what, if any tools you have available that can give us confidence that the effect that you'll see will not just be a point-in-time estimate and can be relied on to be durable post that time period? And then I have another follow-up.

The question is, Tazeen, and thanks for participating this morning. Maybe I'll start and kick it over to Sanjay and Dr. Wild. Yes, we think borrowing Dr. Wild's Olympic analogy, we frankly think any reduction of NfL in Huntington's disease would be important. It's never been done before. And certainly, it indicates that patients aren't getting worse. It would suggest that you are having an impact on the neurodegenerative disease process. And then with regard to your second question, it is why we built on that 3-month off-treatment phase of the study so that we could assess the impact on NfL once we've withdrawn treatment with ANX005. But Sanjay, Dr. Wild, please elaborate further.

Yes. Maybe I'll just say a few things, and Dr. Wild, if you have follow-up comments, that would be great. So yes, just to Doug's point, we think that any level of reduction at all could be clinically meaningful in these patients as an indicative of reduced nerve injury in these individuals. With respect to durability, we are collecting a number of time points all the way through the study. So for example, for plasma NfL, we're collecting that Q2 weeks during the entire study. And we also have regular lumbar punctures as well throughout the study. And so we'll be looking at temporal reduction of NfL as well. As Doug mentioned, there's also an off-treatment period as well. So this pattern should be elucidated by the number of time points we have in the study, hopefully sustain NfL reduction. And then potentially, as drug wears off, we might see an increase back to baseline levels. Dr. Wild, I don't know if you have any further comments on that?

No. I mean, I agree. I think we've never seen NfL reduce in any human trial of Huntington's disease. I'm not aware of any other preclinical animal study in which a therapeutic compound has lowered NfL. It's increasingly a focus of assessment in the field of Huntington's disease. Last month, when it was earlier this month, the European Huntington's Disease Network had several talks in which reductions in sort of soft clinical endpoints or supposedly favorable deflections of brain atrophy measurements were presented. And in every case, the first question from the HD scientists and clinician audience was what happened to NfL, because it's really become a benchmark, I think, in terms of the -- potentially what it tells you, you're physically measuring the thing you want to rescue and you're doing so in a way that's much, I think, less prone to artifactual influence than something like brain atrophy, which is just such a much more complex signal. And so having said all of that, I do still think it's a win if NfL doesn't change. In other words, over a short study with a small number of participants. NfL, if your drug is unexpectedly causing harm, you will certainly show an increase in NFL and that's bad. If your drug is producing benefit, you might show a reduction or you might show a slowing of increase. But it may be that the power of that study over a relatively short time means that you may have to wait a little bit longer -- a few months longer for the NfL benefit to emerge. And so there are many ways to consider a potential success, and NfL will tell you if your drug is doing harm.

Okay. Now I think earlier in the presentation, you had talked about like a gold medal versus a silver result. Would you be in a position to interpret based on the NfL reduction, which of the scenarios is most likely at this data readout?

I mean for the mice that's a gold medal result, right? You've reduced NfL from baseline back towards what we presume is the wild-type NfL level. I think conceptually it's important to remind ourselves that the level of NfL from a single measurement, is it a reflection of the current rate of neuronal damage, okay? So in a sense, it's like the speedometer of a car, whereas the odometer tells you how far you've come. The NfL is a single measurement that tells you sort of is -- what the current rate of damage is in the brain at that time. So you, I think, have a very reasonable prospect of reducing it from over reasonably short periods of time. And it's not something where I think the absolute level is potentially much more important than the rate of change. If the level doesn't fall, then that's when you can look at the rate of change and say, well, we didn't slow the rate of damage, but did we stop it -- stopped the rate increasing. So anyway, from the mice, certainly, that's very much a gold medal result. Remains seen what we've seen here. I think there's many potential ways NfL could be helpful.

Okay. And then maybe if I could squeeze in one question on ALS. Can you give us a sense about why you're measuring the reduction in the CRM only of NfL and not in the CSF initially?

Sanjay, you want to grab?

Another really good question. So we know -- and this has been demonstrated in ALS specifically that there's a really nice correlation between plasma NfL and CSF NfL. And with respect to the pragmatics of the study, not having repeat LPs does remove a hurdle with respect to patient enrollment. So those are the two reasons why relying on plasma NfL for ALS. I should just add, and this relates to what Dr. Zetterberg said is that the plasma NfL measures have really improved over time, specifically with the utility of the Simoa assay, which allows one to have really sensitive measures of NfL in plasma. So there's a lot more confidence in terms of relying on plasma as an index of NfL measures and change in these patients.

We have a question from Anupam Rama with JPMorgan.

I just had a quick clarification point. Have all 24 patients in the HD study been enrolled? And when we think about that, I think it was 2Q '22 readout. Will it be data from all 24 patients on and off treatment periods?

Yes. Good question, Anupam. Yes, and thanks for joining us this morning. Yes, all 24 patients that have been enrolled and the Q2 readout will include on- and off-treatment data. Frankly, we were hoping to do that by the end of this year, but you have to have a cutoff, right, before analyzing the data, and so we rolled it into next year.

Yes. So specific -- great. So specifically, we'll have about 16 patients worth of data at the end of this year. And as Doug mentioned, that relates to the time it takes to actually collect the samples and analyze the data. So there's a bit of a time lag between when they're dosed and obviously, data collection and data analysis.

Operator, are there any other questions in the queue online?

There are no other questions on the phone.

We do have a few written questions. We'll go through a few of those, one of which is will the slides be made available? Yes, they will be. You'll be able to find them on our website. We have a question with regard to -- [indiscernible] regarding inhibition of C5 or C9 via inhibition of C1q. And maybe, Ted, I can turn this over to you to respond to this question.

Sure. So the question has to do with the efficiency of inhibiting upfront on amplification cascade versus inhibiting in the back. And what we found is we've made antibodies against the early components of pathways, C1q, CR, CS, C4 and what we've found is that by blocking C1q from binding to the cell surface in the first place, it really shuts down the entire cascade. For C2, there are bypass mechanisms. In a chronic or strongly a disease with a strong activation, if there's any ongoing complement activation, be it through a weak C1 inhibitor or C2 inhibitor then the entire complement cascade will propagate. And likewise, if you inhibit downstream's C3 or C5, the upstream components will build up. And these will themselves induce their own bypass pathway. So high levels of C4 can continue to cause membrane damage. And there's some papers recently published on this in particular. So in our experience, inhibiting C1q is really the best way to stop the entire cascade.

Good. Okay. Another question we have here is, have we measured C1q levels in HD or ALS patients in the CFS or serum?

Yes, yes, we have. C1q levels, by and large, don't change a whole lot. They're relatively high in the circulation, and there are about 1,000-fold lower in the CSF, but they stay maybe one or twofold higher, sometimes with disease. But what does really change is the activating conditions. So C1q will begin recognizing synapses and trigger the complement activation. So you can measure elevated levels of activation products such as C4a or C3a. So C1q levels themselves is pretty much are constant and ready to respond if there's something to make them respond.

Great. Another question, in the HD mouse model, is there a window after where C1q inhibition no longer reverses NfL? In other words, zero time frame.

That's a good question. What we've done in our first studies was really go toward the end of the model. So we dosed for about 2 months. And at that point, it was a stage of the animals' development that they were beginning to die of disease. And so we've just kind of pushed the limits to that end, and we're going to push it further and then upstream as well, but that's where we are at this point. So we are treating once the disease has already started, and it's really toward the end state, and that's where we're seeing our effect.

Good Question. We have a question on the predictive value of inhibition or adoption of NfL preclinically in Huntington's disease in ALS as well as clinically in GBS to our ongoing clinical studies. Obviously, it's a -- I don't know if you can answer it definitively, but I don't know, Sanjay, maybe you want to opine on this?

Yes, there is some paucity of data with respect to translation preclinically to clinical. What was interesting for the ALS SOD1 clinical trial is there was NfL reduction seen in that patient subset. And that was mirrored to some extent with respect to a preclinical model of SOD1 where Biogen also showed an NfL reduction. There are models of multiple sclerosis, where in the EAE model, which is a preclinical model of MS, NfL has been reduced by clinically effective treatment, and that has also been mirrored in the clinical scenario with respect to relapse remitting MS patients. For HD, clearly, we have some preclinical data, which we think is compelling with respect to NfL reduction. But that translation hasn't been proven yet.

Thanks. And last question, Dr. Wild, maybe a bit provocative and so you choose if you'd like to answer it, but whether you have an interpretation of why the Roche study failed in HD?

So my answer here is just my answer. I think that the milligram dose was too high. And 220-milligram doses of -- ASO, given what 28 days apart to all active participants, whichever group that they were subsequently rolled over into, is too much for the HD nervous system to cope with. And I think it probably caused a degree of inflammation early on in the trial. So to continue my Olympic analogy, I think that the -- you may have the first just sprinter in the race. But if they start the race by running 10 meters backwards, it's very unlikely that they're going to beat the competition across the finish line. So even if the drug -- the lowering of Huntington had some virtue that took a while to emerge, and we know that the NfL spike preceded the Huntington reaching its floor level, it would have been impossible for it to show that benefit within the trial. But again, and so two things about that the relevant to this call. First is there certainly is inflammation happening. The white cells and protein level were up in the CSF, which is more or less a neurologist definition of chronic --. And that probably played havoc with the complement pathway in ways sort of highly undesirable and uncontrolled and is therefore, more or less the opposite of what we're trying to achieve with the Annexon program. And the other is, of course, that it's really highlighted the value of NfL is something to listen to in the preclinical and translational development of any therapeutic Huntington's disease. And up is bad, down is good.

That's very well said, Dr. Wild, and that concludes the questions that we have here this morning, folks. Really do appreciate your attention. We're happy to engage with you directly if you have further questions for the company. As we said before, we're excited about our approach to tackling complement mediated neurodegeneration. We are tackling it at the most proximate end of the complement system to provide perhaps a more complete complement inhibition throughout the entire chain. And so we look forward to turning over the data cards later this year and into next year on both of these very important programs. We thank you for your attention, and wish you all very well.