Perfect. I think we'll get started with the next session. I'm Matthew Harrison, I guess, deemed recovering research analyst at this point. I work in investment banking at, at Morgan Stanley. Very pleased to have, Prothena with us here. Just quickly before we get started, I need to read a disclosure statement. For important disclosures, please see the Morgan Stanley Research Disclosure website at morganstanley.com/research, disclosure. So, really pleased to have, Gene and Wagner here. Maybe I'll just turn it over to them to introduce themselves, and then we can jump in the Q&A.
Yeah, thank you very much. First, thank you for having us. It's really nice to be here. Appreciate the opportunity to speak. I'm Gene Kinney, I'm the CEO of Prothena.
I'm Wagner Zago, the CSO of Prothena, also glad to be here.
Super, super. So, I thought we'd start with Alzheimer's, probably big surprise. I think maybe just help for everybody's benefit, just to get us in the right spot, right? There's been a lot of drugs, we've seen a decent amount of progress in the field over the last year or so. You know, your asset, PRX012, you had the opportunity to sort of see some of those agents while they were progressing through the clinic and design it. So maybe you could just talk about how did you design that asset, and what do you see are the limitations of the current drugs that you've tried to overcome, and how you designed that?
Yeah, no, I appreciate the question, and, you know, Wagner can certainly speak at length about the steps we've taken to really optimize the development of PRX012. And I think, you know, it really goes back to 2017. So for folks that don't know Prothena, we spun out from Eli Lilly Pharmaceuticals back in December of 2012. So we've had a long history, understanding the development of particularly anti-amyloid-based approaches in the context of Alzheimer's disease. And what we were seeing, you know, even, you know, in the 2017 timeframe, was a real evolution around clinical trial design and execution. We were seeing, you know, better examples of patient selection in the context of clinical trials, better use of biomarkers, better endpoint usage, that were more sensitive to the early phases of Alzheimer's disease.
and even a better understanding of observation periods, what ARIA was biologically, and how we dosed around and through it, really to give us the exposure levels that we needed with these anti-Abeta agents, in order to see activity. And we saw a class of anti-Abeta agents at that time that targeted the N-terminus of Abeta, that we felt very likely would be the first class of disease-modifying agents in this disease. And of course, you know, badly needed. Alzheimer's disease, 6.5 million patients affected by that in the United States, fatal disease. And so anything that can move us from purely symptomatic treatments to disease-modifying treatments, of course, is a huge step forward as we think about the treatment armamentarium.
Now, as we looked at that first class of anti-Abeta agents, we felt that really to fully address the burden that patients would be having, not just because of the diagnosis, right? You're early in the disease, you get a diagnosis of Alzheimer's disease, it's a fatal disease, you have to cope with the potential pending loss of independence in the course of the disease. So a lot of burden being put on you from a patient perspective early. We don't want to add to that, with the treatment approaches. And so we knew these first generation antibodies would likely require very regular treatment. So in the case, for example, of the approved agent like aducanumab, twice-monthly IV treatments.
They are testing a subcutaneous formulation, which we'll hear more about at CTAD this year, where we're thinking somewhere around two injections on a weekly basis, so up to eight injections a month. And, you know, as we think about the patient burden, we really wanted something that could remove that patient burden in the treatment paradigm. Something that would increase the total addressable market, something that would provide for a more convenient approach for patients. And as we thought about removing that patient burden, we really set our sights on a treatment that would be once-monthly subcutaneous, relatively simple to administer, a low volume approach. And so to do that, we had to design in certain biological characteristics into the antibody.
Maybe Wagner, you can tell us a little bit about how, how we achieved that goal.
Yeah, but just to expand a little bit and to emphasize, PRX012 was born to be subcu, that was the objective of the program. It was not an antibody that we had IV, and then we checked if it could be subcuable. All the qualities that you would need in an antibody that deliver subcu were there in the screening process, in the design process of PRX012 from the very beginning. So back in 2017, very excited with the first generation antibodies, particularly at that time, aducanumab. And the confidence at that time that if you clear amyloid rapidly and substantially, it would reflect in clinical benefit. And over time, that just solidified with the other antibodies.
But then to reach a level of target engagement in the brain, and then biological response that we would need with a sub-Q-able and small volume antibody, there were a number of properties that we need to create that were new. There was no precedent. One, that is very clear, a higher affinity antibody. So with the smaller amounts of antibody that you deliver in the small volume sub-Q, they have to be sufficient to enter the brain, opsonize the plaques, and induce the same level of biological activity as the first generation antibodies. So affinity was a very important component. I'm gonna park for a moment here. There are a number of other things that are very important to have in an antibody when you're talking about sub-Q. So bioavailability, stability, high concentration, formulation, all of those things, we were experts in those things.
So it were kind of easy for us to de-risk the program, or to accentuate the properties of X12 in those round. But in the affinity side, we were asking a level of affinity for plaques that was unprecedented. And it was a level that we knew we could not reach with the classical methods of affinity maturation of antibodies. We had to take a different route. From the beginning, we knew that. And we were just in the lab and looking at how our chemists do their own work in a small molecule with SAR, or Structure-Activity Relationship, where they know very well how is the structure of their compounds and the structure of the target, and trying to improve the interactions.
In the small molecule world, they talk about pockets of interaction so that they can increase the affinity and the residence time. When we ask ourselves: Why can't we do the same thing with antibodies? We've been generating antibodies based on biological variability, injecting animals, and hoping that we get the best of it, or with random mutations, which is very ineffective, very lengthy to get to a point. We said, "Let's try, now knowing everything that we know about A-beta," and at that time, we already had a very good understanding about the structure of A-beta, particularly in a fibrillar form, like the plaques, how these antibodies bind, including the antibodies that we made.
But at that time, aducanumab, lecanemab, we knew exactly what are the points of contact that they have with, with the plaques or with the fibrillar beta, and we improved upon. So we started with the highest affinity antibody we could find, and now by using computational, three-dimensional conformation studies, we looked for pockets that could be accentuating that binding, could be increasing that binding in, in the plaque. Particularly when you're talking about an antibody, the ability to bind with the two arms simultaneously, it's a very, very, efficient way of increasing the affinity or avidity of the antibody to, to the target. So finding the pockets in the Aβ that were simultaneously exposed to an antibody and corresponding amino acids that we, starting with the modeling, but then generating hundreds of antibodies to, empirically determine, we found PRX012.
So what is particular about PRX012 is that once it binds to the plaque, it stays for a very long time. There is very low, very small level of dissociation, very slow off rate, and that is because the two arms of the antibody bind to the plaques where you have epitopes presented in an optimal way simultaneously. So the chances of having the two arms dissociated at the same time is much decreased. What does that reflect to you? A same level of occupancy as the other antibodies can be reached with a much lower concentration of the antibody, and that lower concentration of the antibody allows us to now deliver subcutaneously the same amount and reach the same efficacy.
I guess two follow-on questions to that before we talk about the data obviously that people are focused on at the end of the year. First is, you know, if you have such high affinity and things like that, should people think, or is there a risk about tolerability and safety, especially related to the characteristic AEs with your compound versus some of the compounds that we've seen?
Yeah, no, it's an important question, and I think, you know, specifically you're talking about ARIA or amyloid-related imaging abnormalities. You know, which are related to targeting A-beta, and so it's an on-target effect. You know, as we understand the science, and I think, you know, Eisai has published similar data, we believe that the maximum occupancy of A-beta at any given point in the variation of exposure is probably primarily associated with those ARIA events, whereas the average exposure over time is probably more relevant to amyloid reduction and ultimately positive clinical outcomes. We've seen a number of proof points that support that notion. In terms of, you know, hard data, we've seen modeling, for example, from Eisai with their own molecule that supports that as well.
We think that's something that is important in the context of PRX012. So what you said, you know, is a more potent antibody likely to produce more substitute ARIA for safety events. That would be true if you used the same dose levels, but I think the whole idea here is that you're moving to a lower dosing requirement to get to the same levels of occupancy, which affords you the ability to move to this subcutaneous approach, particularly on a once-monthly basis. So I think, you know, there we would expect certainly not worse tolerability effects.
And in fact, there are some arguments if, in fact, the variance or the delta between the maximum occupancy and the average occupancy is smaller due to a more convenient approach, like a subcutaneous administration with a high-potency antibody, that in fact, biologically, you could test whether in fact you're advantaged there. I don't think you need to win on that piece. I, I think at the end of the day, from a commercial perspective, if you are equivalent or in the neighborhood of equivalent around safety and efficacy, but you move the field to a once-monthly subq presentation, I, I think that is sufficient to really take a, a dominant place in the marketplace.
But, I do think that the characteristics that allow you to move a molecule to a once-monthly subcutaneous presentation are the same characteristics that would allow you to ask the question of whether that lower variance between maximum occupancy and average occupancy affords you a bigger therapeutic window. And that's something obviously that can be empirically be looked at in the context of clinical trials.
Okay, perfect. And then, the second question is just around IP. You know, given that there are a lot of compounds targeting these pockets, et cetera, just talk about that and how you see about your IP position.
Yeah, I think we—I mean, these are obviously novel molecular entities. We have composition of matter patent on PRX012, which, you know, is relatively fresh. And we'll be continuing to file as we learn about additional properties on PRX012 through the development path. We feel like it has a fulsome portfolio of intellectual property protection.
Okay, super. So, why don't we move on to data this year? And I guess I think, and you can correct me, but I think what you said, you obviously have a phase I program ongoing. You have a set of single ascending dose data and a set of multiple ascending dose data, and I think you've talked about the SAD data, plus the first cohort of the MAD data. So maybe just tell people what that encompasses in terms of patient numbers and dose levels and things like that?
Yeah. So maybe just starting at the top and working down, I think, you know, again, to, to reiterate, you know, we think at the commercial level, you know, showing comparable efficacy and safety profile with a once monthly subcutaneous presentation is a winner in this space. You know, so coming out of our phase I program and the full results of the phase I program for all dose levels, we would expect next year. You know, we would be looking to what we'd call an optimal dose level, and an optimal dose level in our mind is a reduction of amyloid plaque that has shown to actually translate into clinical benefit.
So there, you know, at a six-month time point, you'd be looking at somewhere around a 44% plaque reduction over a six-month time period, and a safety profile that again, is somewhat equivalent to where our first-generation antibodies are out there in the field. So you're looking somewhere around 24% or so ARIA rates or ARIA-E rates all baked in. That's from an optimal perspective. Obviously, the way that the phase I is set up is that we have overlapping single dose and multiple dose cohorts ongoing. What we've committed to share this year are two dose level cohorts from the single dose portion of the study, and one dose level cohort from the multiple dose portion of the study. What we're expecting from that first read is obviously information about safety, tolerability, PK, and immunogenicity.
From the single dose cohort, that's important, as we set up. You know, if you agree with me that, you know, from a commercial perspective, moving to once monthly subcutaneous presentation with relatively equal potency, or I should say equal efficacy and equal safety, is a winning hand. Really, we're gonna learn whether we have a likely ability to meet that target profile, even from the single dose data, so understanding something about PK and exposure and brain penetration. As we move to the multiple dose data, the multiple dose portion of the study is a little more extensive than one would usually think about in a multiple dose study for monoclonal antibody. It is six months of treatment.
And obviously, we haven't said exactly how many patients we have per dose level cohort, but it's larger than a typical phase I multiple dose study, such that we can get good inference into what the levels of plaque reduction and ARIA rates are. So what we'd be looking for as a company is really understanding that we have a therapeutic index, that we see, you know, an appreciable level of amyloid plaque reduction with a window to the ARIA rates, that if, in fact, we need to further optimize the dose level, we have the opportunity to do that, right?
that it's a pretty straight picture on how to get from here to there, from a dose response perspective, and obviously, getting us to our goal of selecting an optimized dose level so that we can move then beyond our phase I program in 2024 into registration-directed studies.
Okay, helpful. So why don't we just touch on a couple of things? Because you said a lot of stuff there, and I think also for people figuring out how to interpret one MAD dose cohort and what that means to the other dose cohorts, you know, is probably helpful. So I guess the first question is, you know, what have you said, or what should people think about the dose exposure response curve, or what you're gonna be able to tell people about that curve, such that if we see one dose level, how easy it's gonna be to interpret what that may mean at other dose levels?
Yeah, I think it's an important question. I mean, you know, I think the traditional way you think about, for example, small molecule development programs, you might go in, you know, at some fold lower than an anticipated biological effect level, and then start dosing up. Here, you know, there's well over 8,000 patients' worth of data in the field that tells us something about amyloid occupancy, which can really be derived from understanding the potency of an antibody, the exposure in the periphery, and ultimately the brain penetration. And what that actually does in terms of impact to plaque reduction and how then that plaque reduction translates to clinical benefit. And so we actually go into these studies with really a pretty sophisticated level of knowledge about the levels of occupancy that we're targeting.
So we do expect this first dose level to be active, and by active, we don't mean 1 or 2% reduction of amyloid plaque. We mean an appreciable level of amyloid plaque on which we can really get a good assessment of a therapeutic index. And I think the way that we'll think about this is based on that therapeutic index, what our ARIA rates look like, what our amyloid plaque reduction rate looks like across the six-month period. How does that actually track back to our understanding of the biology and the assumptions that we made moving from the preclinical to the clinical space? Replace those assumptions with real data, of course. And then obviously, that should give us a very clear picture on where to move, if in fact, we need to move from there.
If, you know, if, for example, that first dose level is not already the optimized dose level. But maybe let me pause and see if Wladimir wants to add.
No, we've seen... It’s incredible. In fact, the consistency, once you normalize the data for plaque occupancy versus clearance from any antibody that binds to the plaques and clear, of course, there is a very high level of consistency, and we build models around it. And all the data so far indicates that the limiting factor for clearance is the occupancy of the target. So if you dose higher, you're gonna see a dose response in terms of clearance. And with that understanding of how these molecules work, the first-generation molecules work, of course, we picked a dose that was expected to be active. But there is an uncertainty that we carry from preclinical to clinical, and the biggest one is exposure.
We have, just in terms of CSF exposure, for typical IgGs, we are talking about between 0.1% and 0.3%. That's a wide range... We have to determine for PRX-012 what is, what is the actual concentration in the brain, and then maybe that first dose that we selected is the dose that we're gonna move forward. Maybe it's higher, maybe it's lower than the, the target. But the confidence that if we need to go higher to achieve higher levels of clearance, that we would do so, it comes really from PRX-012 and other preclinical, but all the other antibodies. There is a dose response, the limiting factor, there is occupancy of the target. It's not, for example, availability of microglia.
Maybe just I wanna move on to some other topics, but I guess last question is: once you identify what you think is the optimal dose level, what do you do from there? Do you move right into a phase III program? How should people think about your ability to execute and move quickly?
Yeah, no, I think I think the clear path, obviously, to a traditional approval is a standalone phase III study. What we're learning, and of course, the field continues to evolve in this space, is, you know, through careful patient selection, good use of quality endpoints, that in many cases now we're seeing phase III's that are reading out essentially overpowered for the effect size that you need in order to get the safety database size that you need. So we think that, you know, there's really become almost a democratization of the clinical trial approach, both in terms of cost and efficiency. You know, our background, again, as a company, we really were the originators and initiators in this space.
We feel that we know this space quite well, both from a preclinical and clinical side, and we feel we certainly have both the knowledge as well as the scope and bandwidth to execute in that space. Along the way, obviously, we'll be assessing whether an accelerated pathway continues to exist. If so, you know, we'll clearly look to take advantage of that. You need a couple things to think about accelerated pathways, serious disease. Your program has to reasonably address an unmet medical need. And I think as we talk about burden and accessibility, you know, there's a case to be made there around the potential for X-twelve to address an unmet medical need.
Okay, super. Why don't we flip to tau? Most of the time you hear tau antibody, and you think hasn't gone well. So maybe that's a good place to start to say why, why you think it might go better for you.
Yeah, and I think in part, you know, if we think Abeta is a complicated protein to target in its pathological form, which is 38-42 amino acids, let's talk about tau, which is up to 440 amino acids, has multiple phosphorylation sites, truncation sites, 6 splice variants, a couple major isoforms. And so I think one of the big challenges in the field has been how do we actually target this protein in a way that actually matters? And, you know, as we were working on this some years ago, and Wagner can go into detail, you know, we really didn't have a lot of information to go on, so we took a very empirical approach.
We simply made antibodies to just about every different form and part of tau that we could think to make antibodies to. It's an approach we call empirical epitope mapping, which is just a way to muscle through this and make a lot of antibodies and then ask what the characteristics of those antibodies are through a series of in vitro assays, all the way through to in vivo model systems. And what we're looking for in terms of readout is consistency and robustness of biological effects. So, you know, having a 30% blockade of cell-to-cell transmission in an in vitro system or some you know partial recovery of some toxic cell stress model, or some small effect or inconsistent effect in animal isn't gonna get it done for us.
We wanna see robust effects. We wanna see consistent effects. And it wasn't until we zeroed in on a region of the protein within the microtubule binding region, the mTBR region, that we found a place where we were satisfied with the level of consistency. I would tell you of all the tau antibodies where it hasn't gone well so far, none of them target that area. So, you know, this is one of the major differences, and maybe Wagner, if you wanna share some insights.
Yeah. So the first generation of the antibodies that have not worked so far, they have two characteristics. They bind to the N-terminal portion of tau, and as Gene said, when we saw efficacy with antibodies targeting those regions, it was small and variable, so we could not replicate, for example, in two consecutive studies. So we quickly abandoned that region. And the other characteristics that they were also what we call the effector function, reduced. So they would not trigger clearance of the target once the antibody binds to the target. What we learned from our screening, basically, empirically determining the qualities of the antibodies, that you have to target the mTBR portion of tau in order to block the cell-to-cell transmission completely.
We saw that reflecting in consistent and robust efficacy also in animal models. But also, the effector function was important. So this idea that if the antibody binds to a small molecular weight pathogenic forms, it's not gonna induce clearance via FcR gamma. I think it's lacking. I think there is a process of clearance that's not phagocytosis, but it's called pinocytosis, that also helps in the brain, the cells get rid of that pathogenic form via a clearance mechanism. So we put the two together. So PRX005 is an mTBR antibody with effector function. While we were finding empirically these results, other labs were starting to describe that tau, certain forms of tau, bind to neurons via the mTBR domain with...
By interacting with receptors on the side of the cell, like heparan sulfate proteoglycan domains. So all of that made sense. So what we think X-five is doing is binding to tau and preventing that tau from now interacting with a receptor or one of the receptors on the postsynaptic side and preventing that uptake, that if it happens, if that seed of tau gets inside the neuron, then you would expect to have a seeding process of an increased probability as you have aggregates being formed, and also, of course, blockade of the transmission, but also very quickly getting rid of that pathogenic form via an Fc mediated clearance. We think that that is what makes X five with a higher potential than the first generation.
So maybe just give us next steps there, and obviously, given partnering, you know, how and when we might hear something about that.
Yeah. So we, as you kind of allude to, we enjoy a partnership with Bristol Myers Squibb, that incorporates PRX005, our mTBR targeting tau antibody, among some other targets as well. Recently, in fact, you know, we announced this, just recently, they have opted into the worldwide rights for that program, and started to assume control of that phase I study. So they'll be taking that on at this point, communication will fall back to them. So we're excited to see them continue to move that forward. And we were excited to see them come in, you know, at a time point before they were required to do so, to take the option on that program. We have two additional molecules that remain preclinical at this point, in that collaboration.
The way that that collaboration is set up is there was an upfront payment, and then on a molecule-by-molecule basis, there are two option periods. The first is at the time of IND filing or right around the time of IND filing, where they have a, an option period where they can come in for U.S. rights. That's an $80 million payment for that option. And then we would typically conduct the phase I study, and then on the back of that, would start their option period, where they could avail themselves of rest of world rights for an additional $55 million option.
Okay, super. And maybe just since we touched on BD, what's the view on BD broadly, right? You have a mix of partnered and unpartnered-
Yes
... assets. How should investors think about, you know, what you're focused on?
Yeah, I think we've been very purposeful in establishing partnerships. First, we enjoy partnerships with Roche on our Parkinson's disease asset, and with Novo Nordisk around our ATTR asset, and of course, we just talked about Bristol Myers Squibb. We enjoy those partnerships. We feel like they're very productive. Those organizations bring a lot of value to those programs. We've been pretty strategic. Obviously, as you think about diverse portfolios, such as we have, we have a vaccine that we plan to file an IND on this year. That'll bring us to six molecules in the clinic. So three will be partnered, three will be unpartnered. That's very much by design. It's about thinking about capital allocation, making sure that we have the capital to deploy to our wholly owned assets.
Our wholly owned assets are really places where we feel like we have unique expertise. We have the ability to advantage patients by moving those molecules faster through just our expertise, our understanding of the science, our understanding of both the preclinical and clinical space and also, you know, our ability to ultimately commercialize. So our lead asset is a phase III asset in a peripheral amyloid disease space, which is rare by nature, very efficient from a commercialization perspective. We think we're the right organization to commercialize that asset. In other areas, we may be very advantaged by additional skills, by additional bandwidth, and in those areas, we actively seek partnerships. So I think it's been a good mix.
It's been very purposeful, but I think it's been a responsible way to make sure that we both bring in capital and deploy that capital in an appropriate way within our portfolio.
Okay, super. Maybe in the last minute or two, just to make sure we touch on amyloidosis and your phase III study and Parkinson's. Maybe just give us an update on those studies, and then also for the phase III in amyloidosis, just remind us if there's an interim there and how to think about data timing.
Yeah. So let's start with the AL amyloidosis program. There's a couple forms of peripheral amyloid disease. Most folks are very familiar with ATTR amyloidosis. AL amyloidosis, of course, is where our lead program is. That molecule, the name is birtamimab. Based on a prior study, which ended early, due to futility, when we did an analysis of that study, what we found was in the more advanced patients, patients that were more prone to a mortality event. In fact, by literature, you would anticipate these folks, surviving somewhere between four-six months. On our best standard of care arm without birtamimab, it was, just over 8 months. We found a survival benefit, a hazard ratio of 0.413 over a nine-month period of time.
So almost, not quite a 60% relative risk benefit on all-cause mortality. And that's a pretty sizable benefit. We, of course, did a significant amount of analysis of that to ensure we weren't leading ourselves down the wrong path, interacted with the FDA, and ended with a special protocol assessment agreement, or a SPA agreement, with the FDA, where we would conduct an additional study, which is ongoing. This is the AFFIRM-AL study, and at an alpha value or a P value of 0.10, so typically it's 0.05. So at 0.10, we would be eligible to register that therapeutic based on the strength of the prior data and the totality of our safety database. That study is enrolling, it's ongoing. We expect top-line data from there next year.
That's a wholly owned asset, and we do have plans to commercialize that ourselves. I think the other asset you asked about was prasinezumab, which is our, our Parkinson's disease drug. That Parkinson's disease drug is with Roche. Roche is conducting a phase IIb study called the PADOVA Study. This is a study with, I believe, 575 patients across two arms. They'll be looking at a time to event analysis for progression on something called the MDS-UPDRS Part III, which is a scale of motor dysfunction, so things like tremor, bradykinesia, typical symptoms of Parkinson's disease. The reason five points is relevant is that's something a patient can feel and appreciate, in terms of worsening.
So what they'll be looking for is for patients that maybe are already on symptomatic therapies, how do they do on this time to event analysis? Roche indicated on their first quarter earnings call that they were fully enrolled in that trial. It is time to event, so it's a little hard to know exactly when that will read out, but they're currently suggesting that will be a 2024 event. So we're excited to see those data as well.
Well, super. Thanks for being here-
Thank you.
Gene Kinney, appreciate it very much.
Thank you.
Appreciate it. Thank you.