Welcome to the live discussion of the symposium, A-beta Targeting Therapies in AD number one. My name is Yona Levites. I'm from Center for Translational Research in Neurodegenerative Disease in University of Florida. And my co-moderator is Dr. Luka Kulic, and he will introduce himself, and then all the members of our discussion will introduce themselves. Thank you very much, Yona. Yes, I'm Luka Kulic. I'm a neurologist by training and I'm a medical director working at Roche Pharma Research and Early Development in Basel. I'm leading several Alzheimer's disease programs at Roche. I would hand over to Laura. Thank you. My name is Laura Nisenbaum. It's a pleasure for me to be part of this session. I really appreciate the invitation. I lead the Diagnostic Pathway Group at Biogen and am also responsible for the biomarker strategy for aducanumab. Thanks. Michelle? I'm- Sorry. Okay. I'm Michelle Gee, and I'm in clinical development in the neurology business group at Eisai, and the clinical lead for the elenbecestat program, which was included the MISSION AD phase III trials. Last but not least, Konstantinos. Hello, everyone. My name is Konstantinos Avgerinos, I am an MD, and I have a master's in medical research methodology. I'm currently a visiting fellow at the lab of Dimitrios Kapogiannis at the National Institute on Aging, located in Baltimore, U.S. Thank you. I hope we will get in a lot of questions from people about our talks. For now, I don't see any yet. I think we can use this time to ask among ourselves. Just a little ask for the audience. Please, place your questions in the Q&A and add also the name of the presenter that you would like to ask the question. Thanks. Yona. I have one question for now to you, Luka. Very nice presentation. I was wondering, with the levels of antibody that you detect in the brain, did you see any ARIA or any other severe side effects that we would expect from that much antibody in the brain, in animal and human studies? Thanks very much for this very important and also very good question. The answer is short, no, we have not seen ARIA so far. However, we need to take into account that the models that we used for the clinical molecule, so the non-human primates that we used, they do not have overt amyloid deposition, at least at the ages when we tested them. In the absence of amyloid deposits and also in the absence of good murine ARIA models, we were not able to assess the risk of ARIA in our preclinical models. We did not observe any ARIA or ARIA-like events in the single ascending dose study in healthy volunteers. However, this study, as you might recall from the presentation, was done in young healthy volunteers. The age range was 18 years- 40 years. That on purpose because we wanted to exclude this potential risk or confounder in this very first study of RG6102. No CAA or microhemorrhages in the mouse model? No. There are not really good ARIA models. Increased microhemorrhages have been reported in the literature with antibody treatments, right, monoclonal antibody treatments in mice as well, but we did not really observe anything like that. No ARIA- E, especially ARIA- E is a big problem for preclinical models. No. Thanks very much. Maybe I can then proceed with the next question, and maybe I will get the question back to you, actually, Yona. I really liked your presentation on the single-chain variable fragment construct with the collagen domain. I was wondering about the immune activity of this collagen domain. From what I understood from your presentation is that the presence of this collagen domain prolongs the half-life of the molecule in central nervous system and in the systemic circulation, and this is kind of the main rationale. What happens then? In terms of the therapeutic activity. Is this domain recognized by glial cells? Does it facilitate phagocytosis? What is your thoughts on that? We actually looked pretty extensively at mice injected with collagen domain by itself. And we looked at the RNA-Seq from those mice, and we looked at behavior, and we did not see any changes or any kind of effects from that. On top of that, our main problem with single chains is its short half-life and instability. Prolonging the half-life would be, we feel, more beneficial than problematic. I see we have a lot of questions coming in. We'll start with the first one for Luka. Have you seen any anemia or changes to reticulocytes or red blood cells in the patients? Curious about any potential target tox. Right. Thank you very much for this question. It's an important question because our RG6102 construct is a transferrin receptor 1 binder, so this is a potential safety concern, anemia. What I can say is that in the single ascending dose study, we did not observe anemia and also no hematology-related dose-limiting adverse events. What we did observe, and we will show a little more at upcoming conferences, so I cannot comment today on the entire safety results of this study because this is a topic for upcoming conferences. What we did observe, and also in line with our preclinical observations, are transient effects on reticulocytes that were obviously transient in nature, and also based on our modeling predictions, are not expected to result in anemia upon chronic repeated dosing, especially not in the ongoing Brainshuttle AD study in AD patients. This is, of course, something that we will be monitoring very closely in the Brainshuttle AD study. Thanks for this question. Maybe I can proceed with lots of questions from myself, actually, but we should also include the other presenters. I would have actually a question for Michelle regarding elenbecestat. You presented the results from this full dataset analysis on elenbecestat from the MISSION AD study, I found interesting that you observed this cognitive worsening in six months. This was the finding that really remained also in the full dataset analysis. I was wondering, so how does that fit or relate to the other findings from other BACE inhibitors? Also an additional question on that is, with the other BACE inhibitors, there was this interesting finding that there was worsening on cognitive tests that were memory-related, like ADAS-Cog, but at the same time, there were improvements that were reported with BACE inhibitors on, I think, verbal fluency tests. I was wondering, did you observe something similar with elenbecestat? Luka, thank you for the questions. Starting with the MISSION AD dataset and the cognitive worsening. There was subtle and transient worsening seen on the ADAS-Cog 11 at the six-month time point. It was not present at the 12-month time point or the later time points, including the main 24-month time point, and not seen in other scales either. In terms of how that compares to other BACE inhibitors, I think, as we know, several BACE inhibitors have now seen some aspects of cognitive worsening, but there are slightly different attributes between them. I think they're not necessarily looking the same amongst the different BACE inhibitors. In terms of what perhaps was driving the ADAS-Cog worsening that was seen at the six-month time point in the MISSION AD dataset, we have looked into that. In terms of specific items, it was really being driven by word recall and by word recognition in the full MISSION AD dataset. Thank you very much. Here's another question for you, Dr Kulic. Does any data exist on expression of TFR1 in elderly versus young? In which system? Hematopoietic, brain, or general? This is not specified, I guess. I guess so. I think it's a very good question because the target population is elderly. I am personally not aware of any studies that looked at changes of transferrin receptor 1 during aging in, for example, at brain vasculature. This is something that I'm not aware of. An important question, yeah, because our patient population is elderly, AD patients. A question from Fadi Rofo of Uppsala University. He is asking if you've measured different species of A-beta protofibrils, fibrils clearance in your studies. This is a question for me? Yes. In which studies exactly? Is this the human studies? I think what they mean, your clearance of A-beta, if you've confirmed protofibrils, oligomers fibrils clearance. I guess this might- Almost all A-beta. This relates to preclinical experiments. Right. in mouse models. Okay. We did not specifically look at oligomeric species. The preclinical experiments I'm aware of and that I presented in this meeting, there we mainly looked at essentially ELISA and also histology. These were the readouts. We did not, to my knowledge, look specifically at certain oligomeric species like protofibrils or other oligomeric species. All right. We have a question for Laura. The CSF Lumipulse assay, has this been approved for screening or diagnosis, and will it be reimbursed by insurance? Thanks, Luka, and thanks for the question to the audience. In the U.S., this in vitro diagnostic test has been submitted to the U.S. FDA for 510 clearance and is currently under review. These assays also are CE marked in other regions globally. That's the current status. Maybe I can follow up on that question and as a take-home message from your presentation. It sounds like that the CSF assay and the amyloid PET can be used essentially in an exchangeable manner. Do you envision this also for the future if, for example, aducanumab gets approved, that we could treat patients who only had CSF analysis instead of an amyloid PET imaging? In the current status, obviously the amyloid PET tracers are approved in the U.S., and the CSF tests have undergone review or are undergoing review by the FDA. Certainly once those are approved, this would provide a mechanism for being able to use CSF as part of the confirmation of amyloid pathology. Certainly in the future, that would be a path by which the amyloid confirmation could occur. Maybe just an additional question on that. Do you envision that also the therapeutic effect will, to some extent, have to be monitored by either amyloid PET or any other markers, maybe CSF biomarkers? The pharmacodynamic effect of the treatment or what are the discussions there? Do we need to show that aducanumab really reduces amyloid burden also once it gets approved? Luka, great question. Thanks so much for that. Really what we focused with this data set was looking at whether the CSF Lumipulse assay would provide concordance with screening samples. Clearly, any work looking at monitoring, we would need to look at subsequently. Thanks. Thank you. Here a question probably to Michelle. Does BACE inhibitor have a future in Alzheimer's disease, and what would be primary aspects of improvement? Okay. Thank you for that question. I think it's fair to say that BACE inhibitors have a very high hurdle if they are going to progress as a treatment in AD. Looking across the BACE inhibitors, one of the commonalities is that we haven't yet seen demonstration of clinical effectiveness in early AD. That is an incredibly high hurdle ahead for this class. Okay. Thank you very much. A question maybe for Konstantinos. Thanks very much for this very nice meta-analysis of different anti-amyloid antibodies and the randomized control trials of these antibodies. Clearly there are differences, obviously, between the antibodies which were assessed in this meta-analysis. The question which I have for you is this a reasonable approach to combine all these different antibodies in one meta-analysis and assume essentially a class effect, right? Given the fact that the MOA, so the mechanism of action, may be really different between the antibodies. We know, for example, solanezumab mainly binds monomeric beta, while aducanumab, gantenerumab, BAN2401, bind more, aggregated A-beta. What is your thoughts on that? That is a great question, and thank you for asking this question. That was exactly the reason that, in addition to the main meta-analysis, we performed the subgroup analysis. First of all, we performed the subgroup analysis based on the individual drug. Then we performed additional subgroup analysis based on the mechanism of action. We also performed the subgroup analysis based on the possibility for Amyloid-Related Imaging Abnormalities. In general, this subgroup analysis did not reveal any substantial differences between the subgroups. What we found was for the drugs that were a little non-specific to the target, these drugs were bapineuzumab and crenezumab, and these drugs did not have any clinical effect. The other group was, of course, aducanumab and gantenerumab. These two drugs were targeting oligomers and fibrils, and they had some positive effect. The other subgroup was solanezumab, which was the only drug to target monomers only and also had some positive effect. The conclusion is that from the subgroup analysis, based on the target, we found that drugs that have non-specific targets do not tend to produce effects. We have still to verify that with future data and future analysis, but this is what we found in our subgroup analysis. That was a great question. Thank you. Thank you very much. Question to Laura, if you measured any other biomarkers in the CSF related to inflammation, synaptic plasticity, and. Thanks, Yona. Thank you again to the audience for that question. It's a very interesting question. Obviously, there are many exploratory biomarkers that have been identified recently with recent studies. What we did with the aducanumab studies at this point is really prioritize the established biomarkers related to Alzheimer's disease so we could get an assessment of what was happening within that ATN framework. We will certainly in the future explore other biomarkers to look to see whether there might be impact in some of these other areas. Thanks. Laura, from my side, maybe a question. I think there is a great interest actually in the CSF biomarkers also upon treatment, right? Especially with such an antibody like aducanumab that really reduces significantly the amyloid burden. My question for you is, do you already have any data on that? What happens to the CSF biomarkers? You looked at screening upon treatment, so how do they change? Also, in particular to A-beta 42 in CSF. Does it normalize over time once plaques really get cleared from the brain? Do we have anything or are you able to share any information on that? Sure. Happy to address that question. It's a very important question. In fact, some of this data has previously been shared at CTAD last year, and I also direct those interested in further understanding of the efficacy and safety data to the presentations that will be shared on Saturday. In the previous disclosures, we showed that there were some changes longitudinally, as you might expect to see for some of these CSF biomarkers. In particular, there was a reduction in CSF p-tau and t-tau. Again, I would encourage you to tune in for the presentations on Saturday for further information. Thank you. Looking forward to the presentation. I have a quick question to you, Luka, and then there is one from the audience. In your impressive preclinical data, is this the only one antibody that you've tested with the conjugation of TFR or there are many others? Were they similarly improved? In terms of exposure, I guess this is the question. We are exploring, this is a Brain Shuttle platform. We are exploring certainly also other mechanisms or other large molecules that could be used as cargos for the treatment of different indications, actually. This is essentially the most advanced program that is now in the clinic. I meant the other way around. Did you test other antibodies with the shuttle? You mean, anti-amyloid? Right. No. The answer is no. From the very beginning, this was clear. It's also a big opportunity for us because we really have a head-to-head comparison with our clinical molecule, gantenerumab, and then can really test the shuttle version of it. This is really a big advantage here. A question from Tom Coleman. Based on preclinical data and existing clinical data of donanemab, do you have sense of how plasma concentration levels translate into levels of amyloid clearance? Maybe very low dose will be sufficient for clearance. I think this is a very good and very valid question. In our preclinical experiments, for example, the chronic dose study in APP London mice, which I presented, clearly indicates that we can achieve a similar pharmacodynamic effect or similar efficacy with regards to amyloid plaque clearance at a significantly lower dose. This is something that we have shown preclinically. Whether this will be also the case in humans, in AD patients with amyloid deposits, is something that will be now assessed in the Brainshuttle AD study, which has just started essentially. There we simply do not have the data to share. There is some modeling work that has been done. However, this has to be taken with a certain amount of caution. Every model depends, of course, on the input that you provide into the model, and there are certain uncertainties with regards to the actual effect. Clearly, RG6102 has the potential to result in a faster and also increased and more homogeneous penetration of brain tissue. Our hope is that this will translate into a faster amyloid plaque clearance and also ultimately, the ultimate goal, of course, better clinical efficacy. There are different options, and another option could be really that you can achieve essentially a similar efficacy at a significantly lower dose, which is, of course, associated with potentially better safety or patient convenience. This is also something that could be beneficial. Thanks for the question. All right. We have a few minutes left. I would have a question actually for Michelle regarding elenbecestat. What I found interesting, and to me at least this was new information, was that the cognitive worsening, coming back to this finding at six months with elenbecestat, was also associated with an increase in NfL in plasma, from what I recall from your presentation. What is your take on that? How do you interpret these findings? Does this suggest that the cognitive worsening is also indeed associated with some neuronal damage, or what's your interpretation? Yeah, you're right that the nominally significant worsening on the ADAS-Cog at the six-month time point, there was also changes in plasma NfL at that same six-month time point. We actually looked at that relationship post hoc, there was no association between those changes in plasma NfL and the ADAS-Cog. Okay. Do we have more clarity? There is, of course, lots of speculation what could be behind this effect, right? Any learnings from your program there or any speculation? Not in terms of speculation, but in terms of actually analyzing that data, there was no relationship between the changes in plasma NfL and the ADAS-Cog 11 data at that six-month time point. Thank you. I see another question here again to Luka. Your presentation was popular. Infusion time, what biomarkers will you be monitoring for target engagement and efficacy? I'm trying to interpret the infusion time part of the question. What biomarkers will you be monitoring for target engagement and clinical efficacy? Clearly in the ongoing study that has just started, the Brainshuttle AD study in patients at 12 months and in the two higher dose groups, which will be evaluated also at three months. We would like also to assess this fast amyloid clearance aspect with this molecule. This will be the biomarker to assess target engagement and pharmacodynamics, essentially. I'm not so sure about what the infusion time means in this case. How long? Maybe the treatment period. This is 28 weeks for the Brainshuttle AD study. This is a study which will essentially, the primary objective is to establish safety and tolerability of RG6102 in patients with AD. As I said, there are a couple of secondary objectives and amyloid PET, PK, CSF concentration, and various other biomarkers in CSF and plasma we will be looking at. Thanks. I think it is time to wrap up the discussion. I'd like to thank all the participants and the audience for the questions. Thank you very much. See you next year in person. Yes. Thank you very much. Thank you.