PTC Therapeutics, Inc. (PTCT) Earnings Call Transcript & Summary

Ladies and gentlemen, thank you for standing by, and welcome to the PTC Splicing Deep Dive Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded. [Operator Instructions] I would now like to hand the conference over to your speaker today, Alex Kane, Executive Director of Investor Relations. Thank you. Please go ahead, sir.

Thank you, and good morning, everyone. Happy Friday. Thank you for joining us today for this second webinar in our deep dive series focused on the splicing platform. With me on today's call is our Chief Executive Officer, Stuart Peltz; and our Chief Scientific Officer, Mark Pykett. We're also pleased to have on the call 3 key members of the research team who have 30-plus years of combined experience at PTC, largely focused on RNA biology. With us on the call is Nikolai Naryshkin, our VP of External Innovation; Chris Trotta, our VP of Biology; and Matt Woll, our VP of Chemistry. Before we start, let me remind you that today's presentation will include forward-looking statements based on current expectations. Please take a moment to review the slide containing our forward-looking statements. I will now pass the call over to our Chief Scientific Officer, Mark Pykett, who will review the agenda and provide an overview of the splicing platform. Mark?

Thanks, Alex. Good morning, everybody, and thank you for attending today's session on the PTC Splicing Platform. First and foremost, we wish everyone -- we hope everyone is safe and healthy. We're really excited to do a deep dive into our long-standing splicing platform today, it's something we wanted to do for quite a while actually. And I'm pleased to be joined today by several colleagues from research, as Alex noted. I'll first do a brief overview of the splicing platform before turning it over to Nikolai, who will review risdiplam as the seminal proof statement on the splicing platform, then Matt will cover the novel chemistry and libraries that support our efforts, and Chris will talk about the mechanisms of biology and targets and our databases in the splicing space. And then we'll have Stu wrap things up. So we hope this will add to everyone's understanding of the breadth and the depth of our splicing platform and PTC's leadership in the field and how fertile this ground may be. So with that, I'd like to remind you that our splicing platform is just 1 of 3 core PTC platforms. We also have our Bio-e and gene therapy platforms. But of course, today's session will focus on our long-standing work in RNA and, in particular, on our splicing platform. The RNA space is the field in which we have been pioneers for over 15 years, where our experience and expertise have positioned us as a leader. And it's a space for experience where the learning curve, what works and what doesn't work, which chemistries are useful and which are not, which targets can be pursued and which cannot is really invaluable. Of course, everyone is familiar with our nonsense mutation read-through product, Translarna, the first drug developed from our RNA platform and a product which generated over $250 million in ex-U.S. sales last year. Translarna is really a great example of how we have deployed our deep understanding of RNA transcription and processing to bring products out of the overall RNA platform, and it is now being sold in over 50 countries. Now there's an ongoing dystrophin study for Translarna being conducted at UCLA. Due to COVID-19 related issues, the final muscle biopsies have not yet been collected for the last 8 boys in the study. In accordance with the analysis plan, we remain blinded to all study data. We continue to monitor the situation in Los Angeles and are assessing all potential options in order to have a data readout by year-end. So for today's webinar, we'll provide a deep dive into our splicing platform and risdiplam is really the flagship program from this pricing platform, which Nikolai will talk about. So with that, let me first review some of the basics of RNA in splicing to highlight the fundamental biology of our core splicing platform. In this diagram, you see that the DNA, which carries the genetic code, gets transcribed into a basic pre-mRNA molecule. That pre-mRNA molecule has coding sequences called exons that are interrupted by noncoding sequences called introns. And these noncoding introns need to be spliced out to form the final mature mRNA, which then gets translated into protein. This processing of the RNA has many components and steps, which actually represent areas where we can intervene in disease states to develop new therapeutics. This is something that PTC has been studying for over decade and a half and has developed incredible insights into. So as I said, splicing is a complex process and involves many participants. Many of the steps involve various RNA and protein interactions that splice the pre-mRNA into the final mRNA. It's more complicated than it looks. There are enhancers, inhibitors and modifiers, which affects splicing in a sequent specific manner. And what we now know is these diverse elements could actually be targeted. For example, the U1 snRNP is a prime target for intervention. You might find it surprising but about 30% of the genome encodes splicing-related proteins, which speaks to the universe of potential targets. I won't go into all the details, but the abundance of targets really provides a rich set of opportunities, and PTC has excellent insights into these targets, which Nikolai, Matt and Chris will talk about. As mentioned, we started our work in splicing a long time ago. And initially, people didn't think you could target splicing with small molecules. But it seemed gradual to us to target splicing because the machinery involves highly specific binding and enzymatic activities, including mRNA itself, that can, in fact, be targeted like other enzymes inviting proteins. So there's a lot on this slide. But over the last 15-plus years, we've built proprietary databases, libraries and high throughput screening systems. We have novel chemical matter with remarkable properties that could influence how this splicing process is carried out. And these capabilities built over years are now aligned to create strong differentiation and to enable us to accelerate new targets and compounds, putting us in a unique position to continue to lead the field. These capabilities have led to the proof statement in risdiplam. It was really on the back of risdiplam that we learned a lot about the platform. And the platform has already led to 2 additional development compounds for familial dysautonomia and our Huntington's disease program. PTC518 for HD is the next program from our splicing platform to go into the clinic and is on plan to enter the first in-human study later this year, so it remains on track. So we're very excited about this program. And again, the deep understanding developed over years is invaluable in terms of the insight into what works, and we think positions us well to continue to bring more splice and compounds forward. All of this unifies into a strong value proposition that we believe will continue to bear fruit. There are hundreds of splice sites that can potentially be targeted to develop new drugs. These cut across the key therapeutic areas where PTC operates today, including CNS, neuromuscular, metabolic and ophthalmology. We have a large number of splicing assets in discovery today and several moving forward through proof-of-concept studies. And we're finding that the hits are relevant not only to our core rare disease space, but also to larger indications, which may provide new avenues to grow our business. So what you're going to hear today is how we're deploying this robust platform to push new programs forward. As Nikolai, Chris and Matt will lay out, we're utilizing the systems to advance a host of new programs through discovery and proof of concept. And Chris will discuss 2 new assets, in particular, which we're very excited about, one for SCA3, one for MAP tau. So you get the sense that what you see today and where we are now is really just the tip of the iceberg. So with that overview, I'll turn it over to Nikolai now to discuss risdiplam and how it represents a good success story from this splicing platform. Nikolai?

Thank you, Mark, for a very nice introduction of the PTC splicing platform and the exciting opportunities that it enables and opens up for us. PTC has been the leader in therapeutic targeting of splicing for the past 15 years plus. And our efforts in this area started with SM -- spinal muscular atrophy as an indication that culminated in the discovery and development of risdiplam, a molecule to treat spinal muscular atrophy. Risdiplam validates and makes a very clear point that, yes, it is possible to target pre-mRNA splicing, act with high potency and selectivity and generate clinical compounds. It provides a very strong translational foundation where we can track, target engagement from cells in vitro to animal models in vivo and all the way to humans. We demonstrated the selective modulation of splicing that, yes, there are unique structures and there are unique elements that can be addressed with small molecule drugs as the rest of the presentation will showcase. We also have built out and enabled screening tiers sets of tests to enable the discovery and development of such molecules. We have worked out and elaborated on a defined mechanism of action of risdiplam, and these principles that we discovered in targeting pre-mRNA splicing are expandable to other targets carrying noncanonical sequences. We have also, in the context of this and in combination of all of these inputs, we have laid the foundation of the PTC's splicing platform that provides unique and critical insights into the drug discovery and development process in this area. So we started with SMA, which is a genetic disorder and about 1 in 11,000 newborn children are born with SMA. PTC is collaborating with the SMA Foundation and Roche to advance treatment in SMA in this program. And underlying this is really deep insights and understanding of the unique molecular genetics that is driving spinal muscular atrophy and that was elaborated starting in 1995 and all the way to today. In population -- in general population, there are 2 very related genes, SMN1 and SMN2. And at least a single copy of SMN1 is sufficient to provide each cell in the body with sufficient quantities of the SMN protein, such that the physiology is functional. In SMA patients, SMN1 gene is deleted, then reactivated, and the cell has to rely on highly related, highly similar gene called SMN2 to provision -- for provision of SMN protein to this cell. The challenge with SMN2 is that due to a single point mutation and one nucleotide change, its splicing enhancer is turned into a splicing silencer depicted by violet oval here. That in combination with a noncanonical weakened interaction at the critical site, we call 5 prime splice site results in the events that the majority of the splicing the production is shifted toward the Delta7 messenger RNA, which generates and produces highly unstable and rapidly integrated SMN Delta7 protein. And only a small fraction of this splicing reaction is going toward the full-length message, which generates normal wild-type SMN protein. The challenge that is a sufficient quantity of SMN protein in different cells and different tissues of the body that results in spinal muscular atrophy. And the important element here that is instrumental and essential to understanding of our splicing platform and how we think about targeting splicing for therapeutic purposes is depicted on this slide. Each exon is defined by interactions at the 3 and the 5 prime splice sites that are respectively targeted by U2 and U1 snRNPs. These interactions are highly regulated and integrate a multitude of inputs from various sequences and other interacting proteins that in combination enable the splicing reaction to proceed. We call this exon definition. And marking up each of those elements in DNA and pre-mRNA that Mark presented in an overview that need to be processed during the splicing reaction. And it turns out that while about half of these U1 5 prime splice site interactions are canonical and that means that they form a perfect complementary structure in the duplex between the 2 RNAs and proceed very efficiently in interim removal, about 45% -- as Chris will elaborate in greater detail, about 45% of these interactions are noncanonical. And that means that U1 5 prime splice site complex has imperfections, has elements of structures that are not completely formed or are formed in such a way that productive splicing cannot be fully enabled. And it is these interactions that provide both the ability to alter the splicing outcome as well as the selectivity to such molecules. And this is the case in SMN2, in which as we spoke just now, the 5 prime splice site here with a noncanonical sequence and in isolation or by itself, there is insufficient positive input to enable splicing. We have identified several series of chemical compounds that act as molecular glues effectively and correct or improve the interaction of U1 and that noncanonical 5 prime splice site such that the majority, if not all of this splicing outcome is directed toward the full-length mature messenger RNA that generate the full-length protein, thus increasing the cellular protein to the levels that are sufficient to maintain cell health and cell physiology. In order to identify such compounds, we build a series of tests and assays starting in vitro and going all the way to animal models and larger animals to qualify, characterize and enable the discovery and development of molecules with high activity and selectivity to the SMN2 splicing reaction. I will also note the knowledge and the expertise that we have developed in doing so in SMA, the principles of this approach are applicable more generally. And they have become the part -- one critical part of our splicing platform. In identifying such a compound, here is an example of its behavior and target engagement, in a mouse model that carries the human SMN2 transgene. As you can see, we have a very nice dose-dependent increase in SMN protein in both the brain and the peripheral tissues, such as blood. And that provides an important avenue for us to monitor the physiological outcome and the target engagement in vivo because it's much easier to assess this in tissues such as blood than in other organs. And we have seen this because SMA is a systemic disorder, and it was important to have a molecule that distributes to all tissues in the body, we saw it in all tissues and organs we have looked in the mouse. Consistent with it, once we started our human clinical trial program, we saw that SMN protein levels are increased in the presence of risdiplam and they are at or above SMN protein levels observed in healthy volunteers in one of the studies. We have also demonstrated, and I know you're all familiar with these data that in 2 pivotal studies in infantile-onset and later-onset SMA known as type 1, 2 and 3, and then the broadest population that has been assessed in placebo-controlled pivotal studies -- or pivotal study, such as FIREFISH, that there is efficacy. There is unequivocal efficacy and activity of this compound as well as very appropriate safety profile. As such, risdiplam is becoming the most competitive molecule to address the spinal muscular atrophy as a whole body disease. It is a molecule with systemic mode of action. It engages the right target in a selective manner. It has been tested and has experienced in greater than 450 patients starting from newborns all the way to 60 years of age. And it really validates the approach that we have taken in this program, and it also exemplifies how one can get a small molecule drug that is active and selective with respect to target. So we can generalize this, we can apply this knowledge and expertise to a variety of other targets in the core areas of our interest. And the next 2 sections presented by Matt Woll and Chris Trotta will document this in greater detail, starting with Matt who will present an overview of our chemistry and our thinking in how we discover such compound. Matt?

All right. Thanks, Nikolai. I'll take it from here. So Nikolai nicely outlined how splicing modifiers for SMA stabilize a complex formed between U1 and a noncanonical 5 prime supply site in the SMN3 pre-messenger RNA. If we blow this up at the molecular level, we can glean some interesting insights that I think is very relevant to our strategy to discover novel splicing modifiers. So shown here is an NMR structure of an SMN splicing modifier that's a close structural analog to risdiplam. And it interacts with duplex RNA at the site of a mismatch, making contacts that stabilize this complex. You can think of these U1 pre-mRNA duplexes as molecular surfaces so on the left is canonical duplex, it presents perfectly matched base pairs. On the right is noncanonical duplex, has pre-messenger RNA that does not perfectly match with the U1 sequence. The duplex may still form, but with imperfections that are due to these mismatches. The structure of the surface will be unique to the pre-messenger RNA sequence. A noncanonical duplex is also less stable than the canonical duplex. In fact, sometimes, they're destabilized to the point that the weakened interaction does not even initiate splicing. If you could imagine finding a perfectly matched small molecule that will bind the surface of the noncanonical splice site and stabilize the duplex, you could initiate that splicing event. Each imperfect sequence will have a slightly different interface, and thus, will require different molecules for activation. Risdiplam only influences a set of noncanonical duplexes with a very similar sequence to that found at the 5 prime splice site of SMN exon 7. We've also found a completely unique set of molecules that selectively influences a splice site that's different but relevant to familial dysautonomia. The really cool thing is that there's no cross-reactivity between the SMA molecules and the FD molecules. So in other words, different sequence requires different molecules. So what's different about the molecules from a molecular perspective? It's really about the shape. Both the SMA and the FD molecules have similar structural motifs, but they have different 3-dimensional geometries. And it's this concept that forms basis for our library design. Most screening libraries are composed of a combination of internally synthesized compounds and commercially acquired compounds, often in the form of diversity sets or so-called diversity sets. To demonstrate the diversity of a library, companies will often plot molecules in a 3D format, using molecular property descriptors to stratify the compounds. I've got a cartoon on the left here of one of these pictures. And although these produce a nice picture, these models give no guarantee that the molecules will cover a wide array of small molecule accessible biological activity. We desired a better way of describing the chemical space of splicing molecules and really have the knowledge and experience to do it. So as I described a few minutes ago, we have found that splicing modulators have common structural motifs and that the spatial arrangement of these motifs gives rise to activity and selectivity. Using this knowledge, we've derived a shape-based splicing-focused chemical space. And as a way to classify molecules within this space, we've subdivided it into 28 unique shape categories. Interestingly, if we search some large commercially available diversity sets within our own library, we find very few molecules that fit the descriptions for these shape categories. It is no surprise, however, that molecules that we have synthesized for our internal splicing programs populate several of these categories. And this has been a great enhancement to this specialized chemical space. We founded a fertile ground for additional splicing discoveries. We're now invested in a very deliberate effort to fill these diverse categories by mining the commercial space, not by buying vendor-provided diversity sets, but by handpicking molecules that provide value to our set. Additionally, to really achieve the diversity necessary to influence the widest array of splicing events, we've engaged in a very purposeful synthetic effort to produce molecules for our screening library. This isn't a resource-intensive effort, but what it provides our unique molecules that are exclusive to our library and cover this diversity space. This is an entirely new approach to reclassify chemical space using methods focused on the key RNA-centric structural motifs and molecular shape. This makes our library -- our splicing library one of a kind. And we believe, ultimately, it will lead to a much higher productivity rate in our screening efforts to identify new splicing modifiers. So in summary, PTC has a library built for success in splicing modulation. Some of the keys that make our library unique to any other: we own the largest collection of molecules from successful splicing programs. We have the know-how to handpick commercially available molecules that fit our library description. We're engaged in a bold strategy to synthesize novel screening molecules within our own uniquely defined chemical space. And lastly, this is all built on the foundation of principles learned from selective slicing modifiers. So I hope over the past several slides, I've done a good job of convincing you that in order to find the right molecules, you have to have the right library. But I do want to be clear, however, with one final point. The exact molecule discovered in a high throughput screen is very, very rarely suitable for clinical development. There's a refining process you have to go through from hit to development candidate. PTC has built over the years a world-class organization for lead optimization. And Nikolai showed a similar slide earlier for the development of risdiplam. A molecule undergoes a transformation. You improve potency, you improve selectivity, pharmaceutical properties. This is an iterative process that requires tremendous chemistry resources that is if you want to refine molecules in an expedited way. But ultimately, your development candidate must be efficacious and meet all the safety standards required for clinical development. We, at PTC, are experienced in navigating this funnel, and we do it well. So Nikolai has laid out the foundational framework and slicing regulation with the risdiplam story, I've described the application of these principles to the design of an entirely new type of screening library focused on splicing, and now Chris is going to present how we put these molecules to work with our powerful screening platforms. Chris, take it away.

Thanks, Matt. Great description of building an RNA-targeted library, really, the crux of any platform is the library and the molecules. In the end, those become the drugs. And from Nikolai, we learned quite a bit about how risdiplam taught us something about splicing. So in my section here, I'd like to go through how we built around that to build a platform. We knew SMN couldn't -- can't be the only splicing target out there. And very early on, we wanted to make sure we built the infrastructure and the ability to exploit this mechanism to develop other programs, other targets and other drugs. And so I'm going to go through this, and this puzzle gives us a representation, gives us the idea of the proprietary knowledge that has gone into the basis for our drug discovery platform targeting splicing. The information that we know from mechanism really enables our ability to design a platform. And I think Matt showed very clearly how that insight has led to an enhanced and unique compound library. And on the other side is the screening part and what are the targets that are amenable to this approach and how do we access, how do we bring together the targets and the library to find new potential drugs. And so I'll spend a few minutes going through this notion of canonical versus noncanonical splicing because this is really the basis. And it's not something that was appreciated before the discovery of small molecule that can actually target a noncanonical splice event. It wasn't appreciated that this would be a drug target, but it turns out it is. And the reason for that is pretty clear. So a canonical 5 prime splice site, as Nikolai clearly described, leads to a perfect interaction between U1 and pre-mRNA. And U1 is involved in the recognition of 5 prime splice site to really begin the splicing process. And so it's at this step that the cell decides to splice or not to splice. And if U1 is present and is present in a fashion that creates a stable structure to several splice. So 50 -- it's no surprise that 55% of all exons in the human genome are canonical. Efficiency is a must when you're trying to splice a messenger RNA with dozens up to 100 different exons, you really need to be able to efficiently splice. However, there are constraints around the 5 prime splice site. As you see in this picture, the 5 prime splice site contains 3 nucleotides or more in the exon and the exon codes for protein. And so therefore, these constraints are not acceptable to a cell that is trying to create a diversity of proteins with various functions. And so there must be diversity at this 5 prime splice site amongst these 3 nucleotides. And that's really how the cell has evolved U1. So within the intronic part of the 5 prime splice site, there's almost perfect complementarity for most splice sites. It's much greater than 55%. The diversity comes in the exonic part, where the cell is trying to achieve diversity to allow for different splice events. And so this causes the production of noncanonical 5 prime splice sites. These are splice sites where you have imperfect interaction at the splice site. And as Matt and Nikolai both described, this really leads to structures that can be targeted by small molecules. And so going into the notion that we're trying to develop a platform, this is where we felt we needed to focus our attention on these noncanonical 5 prime splice sites, really the sequence variability at the end of the exon. And it turns out, it's really the minus 1 and the minus 2 position that distinguishes the -- this canonical from noncanonical. And I'll show you a little bit of what I'm talking about there. So Matt laid out very nicely how small molecule acts as a molecular glue. Oops, sorry, let me go back. I knew I was going to do that. Small molecule acts as molecular glue to stabilize a noncanonical 5 prime splice site. And so how do we access that 45% of noncanonical exons? Well, you look at what they are. So basically, there are -- since it's 2 nucleotides at the exonic boundary to perfect matches an AG at that site, but there are 15 other combinations that make up noncanonical 5 prime splice sites. So in the case of SMA and SMN2, it's a GA 5 prime splice site, and that's one class of molecules that we have talked about so far. And so the job of the platform was to really understand of those 45% of other splice sites, which ones could be targeted. And additionally, if we could target them, which ones are involved in human disease. And so that's really -- it's a difficult task to do this without data. So we thought we would -- we designed an experiment where we could ask. First, how many of these are affected by a mechanism similar to small molecule. And so the small molecule stabilizes the interaction. And so therefore, we genetically reprogram U1 to do the same thing. And so what you can do is you can take U1 and you can change the nucleotides in U1 at that particular minus 1, minus 2 position to now perfectly base pair with any 1 of the 15 noncanonical 5 prime splice sites. And in doing so, we created a separate U1 variance. In this case, I'm showing a GC variance that will potentially bind to and stabilize all GC 5 prime -- noncanonical 5 prime splice sites in target exons and the like for all the rest. And so we created these 15 cell lines and then asked what exon of the 45% were amenable to inclusion. So these would represent targets where if you can stabilize the interaction, these exons, the cell could decide whether to include or not to include these exons. And so the results of that experiment are shown here, and I'm showing one sliver of the results, but it's a really interesting sliver. And so what I have here are across the top are the different cell lines. And so what we do is we treat this cell line with that variance, we allow it to then reprogram splicing with that variant. It still has wild type, so the cell survives. All normal 5 prime splice sites are utilized. But in addition, you now start to see the ability of these U1 variance to affect splicing on a target. And so the colors represent the different noncanonical splice sites. We then take all the events. And in this case, we're looking at analysis is of exon inclusion. So how many exons were further included by these U1 variants? And so we take all of those events, and each line here represents one of those events. And as it gets more red, that's indicating that it's spliced more. So it's a measurement of the amount of splicing of that exon. And so taking the different cell lines, we now match them to the splice site and we sort all of the events by their 5 prime splice site. And what we find is that, for instance, in this case, we had a GC variance expressed in the cell. And when we sorted and look at where the GC exons that were effective lie, they directly lie. You can see a diagonal here. So for each variance, you get a set of targets out of that 45% of noncanonical exons that is now affected by U1. This is exactly in line with our thinking that if you stabilize that interaction, you will affect slicing at that 5 prime splice site. And so what we found were several thousand events out of that 45%, that could potentially be drug targets, could potentially be affected by something that can stabilize them. I mean, in this case, it's U1. It's a direct complementary U1. In the case we're looking forward to develop the platform, we would be looking for small molecules that can afford the same effect. And so out of those 1,000 splicing events, we find inclusion of cassette exons, similar in respect to SMN. This is a cassette exon that is normally splice. And in the case of SMN2, the mutation drives splicing to is much weaker and therefore, require some help to splice. In addition, we also found -- which was interesting is exclusion of cassette exons. And I'm not showing that data here. But basically, in a similar fashion, we see exclusion and the thought process behind that is if U1 is a perfect interaction with that 5 prime splice site, perhaps it's so stable that now the splice skips over it, it doesn't have time at kinetics to allow for splicing. But nevertheless, these are exons that tell us that perhaps we could skip exons by using this kind of a mechanism, if you could find a small molecule analog at this kind of a U1 variant. And surprisingly, we found a new class of exons called psiExons or exon. And I'm going to talk in more detail about those as they form the basis of our Huntington's disease program, and what is a pseudoexon? A pseudoexon is -- it's like a cassette exon, only it's pseudo because it's unrecognizable as it doesn't have a canonical 5 prime splice site, and it's located within an intron. And so these are previously unknown exons that can be promoted by a strong interaction with U1. So while they have a strong interaction with U1, it's clear that these type of exons also have the elements of exon definition that Nikolai talked about, and that's really crucial. So it's not just the 5 prime splice site, but it's the 5 prime splice site in combination with the exon definition that leads to this large lift of potentially regulatable 5 prime noncanonical splice sites. And so we -- with that information, we developed a proprietary bioinformatic pipeline to add together all of these elements to discover further potential exons that could be affected by this kind of strengthening of the U1 interaction with the 5 prime splice site. And so this gets us to target. So what are the targets? So using both the experimental data that I just presented as well as the bioinformatic analysis we had a set of several thousand of those 45%. In fact, somewhat around 10,000 or more exons that we knew could be affected by strengthening the interaction between U1 and that 5 prime splice site. And so how do we get to the term -- how do we identify targets out of this? Well, what we did was we cross-compared all of those exon events to dominant genetic diseases, haploinsufficiency diseases and splicing-driven diseases. And I haven't mentioned yet, but oftentimes, splicing mutations will affect splicing -- sorry, mutations will affect splicing. And in particular, these mutations may happen at the 5 prime splice site creating a noncanonical exon. So we added all of this information together to build our target list. And so when you cross-reference these 2, we determine that there's potentially, out of those 1,000s to 10,000s, there's potentially hundreds of targets that exists, where we know that strengthening the 5 prime splice site will lead to an effect on splicing that will help -- will alleviate a potential disease phenotype. And so from that, we built a database, and that's shown here in high detail for 5 targets. And so on this list is SMN2, IKBKAP. I'll be talking in a little bit more detail about that. Our Huntington program, I'll also highlight this program; and 2 additional programs, which I'll talk about today, and these are interesting because ATXN3 and MAP tau responsible for SCA3, spinocerebellar ataxia and tauopathies caused by MAP tau aggregation. These are exon-skipping events. And so this was one of the discoveries that we made from running our experiment with our genetic reprogramming of U1 that you can actually skip by strengthening -- we hypothesize strengthening too much the interaction between U1 and the 5 prime splice sites. And so we're really excited to pursue the targets that are on this list, and you can see there are many, and we're really working our way through this list as we speak. But how -- so how are we working our way through the list? So how do we find of those potentially hundreds, and as I pointed out, there are thousands of others, which may not be involved in disease, but will give us information about the kinds of splicing modifier as we can find. How do we screen so many targets? If we do it the way we did for SMN or for FD or Huntington, it's one at a time target. It's not -- it's efficient and it's focused but it doesn't get us access to the sheer number of targets, which, in this case, with our knowledge, we know we want to see. We want to see activity on multiple targets to understand how the small molecules are functioning, which gives us great starting points. And as Matt alluded to, these starting points can then be further optimized for their selectivity, for their pharmaceutical properties to lead us to development candidates and ultimately drugs. And so this is the final piece of the puzzle. So as Matt alluded to, we have a proprietary and strong RNA-targeted library or I should say our library has been enhanced by these efforts. We now understand the targets. We have a very nice solid target database with many examples of disease, relevant splice events. And so we needed to develop platform technologies that could allow us to screen multiple targets against multiple small molecules in an efficient manner. And so we developed 3 general platforms here. And so the first one I've already alluded to. It's the bioinformatic pipeline. But in this case, we're not applying it to a U1 variance. We're applying it to small molecules, and small molecules that can induce splicing changes. And so this has been very fruitful to look at all kinds of molecules, as you can imagine, published molecules and proprietary molecules in our library. And it really has given us incredible insight into the other kinds of splicing event beyond U1 and the 5 prime splice site that might be targetable by small molecule intervention. For example, U2 has -- also has an interaction in a sequence dependent manner and -- at the 3 prime splice site. And in the case of the 3 prime splice site, there's even further diversity. And so you can imagine lots of potential structures there. We also developed more directed platforms. We call our Isoform plex platform and our HTSpliceseq platform. And these have -- these are essentially trying to answer the same question. Can we identify small mice molecule splicing modifiers? However, they're slightly different and for reasons based on what we're trying to achieve. And so for Isoform plex assay, we're really trying, in this case, to measure endogenous isoforms that are created when you treat cells with potential splicing modifiers. And so this is where our experiment with the genetic recombination U1 gives us a great heads up to what type of targets to put into this kind of a screen. We know for the -- all the different noncanonical 5 prime splice sites, we understand a set of targets for each that can allow us to then pull a large number of molecules. And this plex could have a potentially 100 or more targets in it. And so that gives us the opportunity and the ability to panel very deeply into the sequence diversity that's at the 5 prime splice site. That extends beyond the minus 1 and 2 position. And then our second platform is HTSpliceseq. And so this is based on the fact that a large proportion of our understanding of splicing and disease and how they're connected comes from many gene analysis. So when a researcher discovers some mutation that affects their protein of interest, and it's at the splicing level, almost the first thing they do every time is it's splicing, they'll still generate a minigene. And I'm showing an example with SMN here. You take the proximal exons and the exon of interest and for SMN, it's not necessarily a mutant, but you're looking to see -- you're looking to create a minimal system where you can measure the splicing decision. So when the cell splices this, it either includes or does not include the exon. And so the goal is to minimize that to the point where you have a model that you can then screen. And so we've built hundreds of disease causing splice-defect minigenes and now we can screen those in a high-throughput precision and against a number of molecules. And so -- oops, I -- yes, so let's put it all together. So I've shown you that we understand the targets, we have a library, and we've built the screens to look for this, and now I'll talk about the results of that screen. And so when we do something like -- when we run a screen, in this case, it's a plex screen, it's a complicated set of data, but there's a couple of interesting points I want to highlight. First, you see red and blue. What does that mean? So the red and blue are whether you have exon inclusion or exon skipping, being blue. Exon inclusion being red. There are 24 targets we're looking at. It's a range of targets. I mentioned druggable splice events, controls, noncanonical probes. These are some of the other potential active events. And across, you're looking at various compounds. And so what you see, the take-home message here is we have a variety of degrees of splicing activity and selectivity across our library. And just to go back to a point Matt made, the typical hit rate in HTS is 1% or less. Here we see 4%. So we really have enriched our library with molecules that can affect splicing. And so it's now our job to develop programs out of this. And so in the next few minutes, I'll take you through some of our programs that have emerged from these efforts and others. And so there's really 2 main things we're trying to do by targeting splicing. We're trying to either restore protein or reduce protein. And we do that in multiple ways by targeting this 5 prime splice site and noncanonical splicing. SMN2 is an example of an endogenous exon, where we're trying to treat SMA by modulating the splicing, mutations that create a weak 5 prime splice site found in -- such as those found in familial dysautonomia, another potential target for splicing. To reduce protein, we can promote inclusion of pseudoexons, and I'll talk about that in a moment. Oftentimes, since their introns arrive, pseudoexons will have stop codons, and that leads to loss of a messenger RNA when they're present. And so that's highly a Huntington's program. And we can also promote exon skipping, which would potentially lead to frame shifting or, again, a stop codon -- premature stop codon that could cause decay. And that's really the basis for our spinal cerebellar ataxia and tauopathies. So FD is shown here, a very similar set of slides that you saw to Nikolai. It's a genetically driven mutation at plus 6 position where you weaken the 5 prime splice site. We sought to developed molecules that could strengthen that 5 prime splice site and therefore, shift back to inclusion of exon 20 and creation of a functional IKBKAP, IKAP protein. We, indeed, working with Susan Slaugenhaupt at MGH, did identify such molecules, and I'm showing here the results of an animal study, where in a mouse model for FD, we can show upon treatments here are 6 animals in each case, each dose, low dose, high dose, you can see that. By endpoint PCR, we get exon 20 inclusion. You noted that's full length, and it is dose responsive and that's quantitative down here as well protein also is increased when you increase the RNA. And so it's a great example. It was actually chronologically our second example of the spicing program that targets a noncanonical 5 prime splice site. And so that takes me to HD. So HD is interesting because it arises from that novel class of pseudoexons. And so the disease is a neurodegenerative toxic gain of function. It's a dominant disease caused by a repeat expansion in the Huntington's gene. And this slide sums up in a nutshell what our DC -- PTC518 does. And upon treatment with this molecule, it will cause the inclusion of pseudoexon within the Huntington mRNA. That pseudoexon carries with it a stop codon, and that causes degradation of the mRNA. And in dominant diseases, such as the repeat disease,-- nucleotide repeat diseases, you want to get rid of the RNA and the protein, and this accomplishes both. So you're targeting splicing to grade the RNA. And obviously, the RNA doesn't exist, there will be no protein. And we show here on the right, in vitro, in cells, you first begin to see the emergence of the inclusion events with low concentrations of compound, which increases. But over time, you begin to decay the RNA. So now the steady-state level is reached, and you lose the RNA, and that's what's depicted in the bottom with the graph. And so we've taken those molecules, obviously, all the way to the DC stage. So here's some of the convincing evidence that these molecules function in ways that would be beneficial to the treatment of HD in humans. And so this is a back HD mouse model. And you can see upon treatment with various doses of the molecule increasing doses, we get a dose-dependent reduction of the mRNA in the brain. And that's really the target organ for the disease. And so since it's a small molecule, it's oral, it crosses the blood-brain barrier. And in addition, it crosses into all regions of the brain, and that's really important for Huntington, where the striatum, a very deep part of the brain is involved in disease pathology. And so in addition, in the cortex and the cerebellum, we see equal lowering throughout, and we're hitting our target of 50% lowering, which is what we feel is necessary, will be important for disease mitigation. And so a key point, and Nikolai highlighted this beautifully with SMA and how it accelerated that program, we're able to also measure peripherally in tissue -- easily attainable tissues or blood this lowering. And in the case of our PTC518, we see an equal one-to-one brain to blood lowering. So this gives us confidence that when we measure lowering peripherally, we can -- we know well what's going on in the target or in the brain. And so the drug is oral across the blood-brain barriers, in IND tox and this ability to measure mRNA and protein in healthy volunteers is essential because it gives us an early read on the target engagement and proof of concept. And so Phase I trial is planned for the -- before the end of this year. There was no lag there. That was me. Not clicking early enough. Okay. So back to our database, and I previously described to you, these were all inclusion events. And we found this the possibility that we can also skip exons with small molecules that target 5 prime splice sites, which are noncanonical. So the next 2 programs really demonstrate this. And so spinocerebellar ataxia, it's a progressive cerebellar ataxia. And to date, there's no disease modifying therapy, and it's a nucleotide repeat expansion disease, similar to Huntington, and it occurs in the ATXN3 gene. And so in the case, we have biology on our side. So it turns out that ATXN3 is an alternatively spliced messenger RNA as -- or probably more than half of the genes in the genome. But what's interesting is that this alternative splice event, which naturally occurs, actually skips exon 4. That creates a premature stop code on and degrades the mRNA. And so this is how the cell regulate the level of an RNA. It will create some by-product to control how much of the RNA is present at steady state level. And so with that in mind, we know it's a targetable event. So therefore, we search for small molecule compounds that could do the same thing, to actually push further the alternative splicing event. And indeed, we found such a small molecule in a screen, and that's shown here. Similar set of data where we have these endpoints that pictorially demonstrate the splicing change. Only this time, we're looking for a larger band, the inclusion band, to go to a shorter exon skip version. And you see dose dependently, exon skipping, it's quantified here to about 75% or 80%. And what's very nice is this western over here showing the protein and the protein is concomitantly lowered. And so we're hitting the target we expected, and we're seeing the expected lowering of the protein. And so with that in mind, we looked at these early molecules and realized they were -- and we tested them in mice, we realized they were pharmaceutically suitable for -- to go in the mice only there wasn't a humanized version of this mouse model available to us at the time. So we basically set up a xenograft where we created in the cell, a xenograft, the tumor to be able to look upon treatment of these mice, would we see lowering of the human ataxin mRNA in this xenograft in this tumor. And indeed, we do, and there's a low dose and a high dose. So again, we're seeing dose-dependent lowering of ATXN3 mRNA and on the right, the protein. And again, it's a similar lowering and it's also dose responses. So another example where we get a clear connection between splicing and protein lowering and an identification of a small molecule that can do that. And finally, I'll sum up with MAP tau. This is to correct tauopathies. And I see I'm running a little late on time. I'll go relatively quickly. Basically, for MAP tau, it's a phosphorylated microtubule-associated protein. And when there's over phosphorylation, that creates aggregates, and that leads to neural cell death, and that's kind of shown in the picture here. And in particular, this FTDP-17 is a type of a disease where it's MAP tau driven. So mutations in MAP tau lead to this over phosphorylation of tau in these aggregates. And it's depicted here as a balance between a protein isoform, which is called 3R and 4R. So 3R and 4R are in balance. If 4 -- if you have more 4R, you get an imbalance and that leads to this -- these aggregates. And that's because this particular section of protein in the isoform is hyperphosphorylated. That event is directly correlated to an alternative splicing event in exon 10. This is a noncanonical exon. And its inclusion or exclusion really governs how much 4R and 3R occurs. So our goal here was to cause skipping of this exon, driving from 4R to 3R reestablishing a more healthy balance and we see, indeed, our compounds, again, by end point, show a very nice skipping and loss. We're increasing 3R, and that's depicted here. And so again, we've identified a small molecule that can skip in exon. And so to sum up my section and the platform, Mark showed this slide earlier. Early on, one at a time, we would go through, run screens on targets on -- in diseases where splicing was the leading cause of the disease. And so accelerate -- taking that knowledge and accelerating our compound library diversity into areas that were enhanced by our splicing programs and then adding on top of that, our knowledge and building out even further now additional molecules so that we don't only access the targets that informed us, but we want to access all targets, all 15 potential noncanonical flavors of the 5 prime splice site. And I hope that I've convinced you that in addition, we've built up several screening platforms that really allow us to leverage the information and knowledge coming from our early experiments and some of the data still emerging from molecules like those that are targeting SCA3 and MAP tau that all feeds back into this to allow us to take our library to new places and to pick additional targets to pursue. And so this really represents, to me, the tip of the iceberg. And so we're really -- this is where we go next. And so we're really beginning -- this is the point of a deep dive. We're going down to pull up the rest of these targets. And so with that, I close, and I'm going to turn it over to Stu, my boss, and take it away, Stu.

Thanks, Chris. And thank the team for putting this together. I thought it was a really a nice overview of what we're trying to do here. And in closing, what I really wanted to say a little bit is just to summarize and say, look, we've built a very strong engine and the platform for splicing, that's actually quite novel, quite unique. And I think -- and you can see the building blocks of that is not only the understanding of these assays and technology as well as the databases and the integration of all of this, with the chemistry that allowed us to move forward on this. When I sit back on, when we started this with SMA, this was considered a loonshot. And in parts because people thought, okay, proteins are the only thing that 1 could target. And we always thought that wasn't true. We've been doing now RNA biology for well over 2 decades now, and it's very clear that the initial enzymes in the primordial soup weren't proteins, they were RNA. And even to this day, there are RNAs that form unique structures that are catalytic that are quite important to this day. So we knew that there's an ability to do this, and we've done this now not only with Translarna, but also now with the splicing platform, which is really quite impressive. And we put together all the pieces and learnings of the biology, the assays, the chemistry and the experience. And that there's not another company that has that whole breadth of experience that allows us to be able to identify such molecules that modulate splicing. And I think a trick to the platform is always can you do it more than once. And you can see that we were talking about with of familial dysautonomia and Huntington's disease. So I think it's really quite clear that we very systematically over the decade and have put us selves in a very strong position of building the platform of technologies, a strong chemistry library in understanding how to optimize that. And starting with SMA going through and showing that one can move forward with multiple compounds, and we're really quite excited about Huntington's disease, you heard Chris talk about that and that will be first in humans this year. We also think, I guess, really the notion of an oral bioavailable small molecule, especially diseases like Huntington, where it's a consequence of making a toxic protein once the cells die, and you have to hit every cell. An orally bioavailable molecule that gets to every cell in the brain, we think, is just a way to go. And most of these are systemic, and they have other defects. So I think there's -- we think just like SMA, where we think this is, I think, the results are quite stunning with our risdiplam and also the fact that it's systemic ultimately puts itself in a unique position, not to mention that it's orally bioavailable. We'll do that with Huntington. But more importantly, the foundation is there, and we've just shown you a couple of other examples with SCA3 and MAP tau. But the database and the knowledge of defining what are the other targets that are proprietary, and to us, really puts PTC in the unique perspective of being able to do this time and time again. And so as Chris had showed you is that this database of targets and how we can screen and more our chemistry from one target to other is we think quite exciting. And so part of what we've done is really to build this. And I think part of what -- part of the capital that we have run on is really to expand this and get a large number of programs such that within the next several years, try and get 3 to 5 development candidates before we move forward. So we're really quite excited about this. And I think you could see the power of this technology. So let me stop there, and thank, everyone, for listening. And if there's questions, we'd be happy to take them. Thank you.

[Operator Instructions] Our first question comes from the line of Robyn Karnauskas from SunTrust Robinson.

I know there's a lot of work went into this. So just on Huntington's, let's talk a little bit more about what are your risks around the safety and threading that needle between getting that protein or the wrong protein down? And if you have a tighter binder, does that mean you actually have to have less or more protein? How does it affect the PK/PD of the molecule? I'm just trying to understand the risk of the drug that you're putting into the clinic and maybe some of the risks around what we might see initially from the program?

Yes. I think -- thanks, Robyn. It's a good question. I think the point really there, the risks around the safety, it's always a question of safety when you have new molecules and you test them. And we spend a fair amount of time looking at safety within our program. One of the big things we do quite early is not only look for efficacy and potency, but look for safety effects as well. So we measure that very early on and so that's what we do quite often. Because it's really no difference, I think, from any other molecules, what are the off target, what are the effects of that. We're always looking for in the case of Huntington's product and the nice aspect, I think, of what Chris has shown you with the Huntington's disease is we're very titratable. And so the level of the protein that goes down is dependent on the exposure, which we can measure within the blood. And as Chris had showed you that the level within the blood really reflects what is seen in the brain. So we have a good estimation of that, and you could see in the animal model test, the level of reduction of Huntington was equivalent both in the blood as well as what we saw in the brain. And by the way, within all aspects, wherever we look within the various tissues within the brain. So I think from that point of view, when we think about what the molecule does, it's really about an exposure of what level do we reduce that RNA and therefore, what level does the protein measure. And that's what we're measuring when we look at the exposure. And so what we're shooting for somewhere between 40% or 50% reduction and so we could define an exposure that does that, and that's -- in the sense in the clinic, we'll be doing in healthy subjects so we can look at that and see the effect. And one nice aspect of doing Huntington's in healthy subjects is when you'll start with single accelerating doses that will get a PD effect in relation to both RNA and protein. And so -- and I should point out that was very similar to what we did in SMA. So we knew very early on that we were altering the level of SMN transcript going to the one that included exon 7, similarly here, we'll be able to measure the reduction of the RNA level and reduction of protein as a consequence of that as well. And so we'll be measuring that. So we'll get a pretty good handle in healthy volunteers, both single ascending dose as well as multiple ascending dose out what it looks like in the effect of that.

And just a quick follow up. From a time line perspective, I know everyone's focused on time lines for this program, is there anything that could extend those time lines, but maybe you have to look at more dose groups? Like what is your concern around being able to actually get data in a timely fashion and meet and get data next year from the program?

Yes. We feel pretty good about this. It's in healthy volunteers. The reason we're -- our plan is to go into healthy volunteers is actually for speed, right, because you take multiple cohorts at a time, you test them in the sense and you measure both PK and PD. You get the results, you then look for safety following the single dose and move quickly. So we don't anticipate issues within the healthy volunteer studies, then you get up to a dose and then you do a multi-dose and follow that over time. So I think and the beauty of that is that you could do this quickly in Phase I units. So we think this is -- should be -- we should be able to do this in the normal time that you can do Phase I healthy volunteer studies. So we don't see anything in the way of doing this now.

Our next question comes from the line of Brian Abrahams from RBC Capital Markets.

This is David Szeto on for Brian. Can you hear me?

Yes.

Great. I was just wondering if you could perhaps elaborate on how you might use your recent cash inclusion from the monetization of risdiplam royalties to potentially accelerate the development of these next-gen splicing modulators? Is it about interrogating the library and high throughput screening perhaps so that identifies more targets in the business or broadening the pipeline and bolstering maybe the lead assets in Huntington's and obviously the disease the states that you mentioned? Could you just elaborate on that a little bit?

Right. You were sort of going in and out a little bit, but I think your question is in terms of the proceeds we brought in, what else will we be doing? Yes, we'll be -- in terms of the Huntington's disease program, we'll be doing some additional work in there in terms of biomarkers as well as natural history studies and just to -- and do some work, obviously, to continue to bolster the Huntington's program. You saw a couple of examples of some other targets that we were doing. And so we plan on expanding of those groups to be able to rapidly move forward of the splicing programs. Our goal is to get in the next 3 to 5 years, 3 to 5 development candidates that are splicing candidates so that we're excited about that. That will require additional resources in terms of the number, in particularly, both chemistry, biology and pharmacology to expand on that because as Chris and Nikolai showed you, we're now capable of rapidly identifying these compounds. It's now a matter of having the resources in terms of chemistry to rapidly in a sense really to pull the crank in terms of the internal process of finding the compounds that have the best efficacy, potency and pharmaceutical properties that we want. And we're a well-oiled machine that can do this. It's now just about expanding on that in terms of the number of programs that we're going to be expanding those groups to have -- to continue to crank on a significant number of additional targets.

Got it. And then if I could add just a quick follow up. I was just wondering if the process with risdiplam, if there were any synergies that you've learned that might help further the discovery and development efforts and processes around your next lead splice modulator assets. If there's anything that might speed the process up?

Well, it's a total learning experience. That's the -- I think was one of the major point that you saw is that we've built up not only the platform technologies, this green platform technologies, but the library itself has the shapes and forms and that we've been able to actually use our library to get more higher hit rates. That's the -- I think Chris had talked about, you saw the 4% increase. So we get better hit rates, and I think there's a morphing of compounds that can go. So we -- I think there's a more rapid in terms of efficacy and potency. And so I think from -- this is now a decade and a half of work that's gone into really the understanding of the biology, the chemistry, the building of the platform, the focus on making a better library, the identification of the type. This is really now an integrated process and all the learnings that we've gotten from these programs are now all integrated into that. So we think there are certainly a huge number of synergies that occur as a consequence of this. And now it's really getting the compounds and then building out the pharmaceutical properties and the efficacy and potency and that just becomes a sort of almost an engineering, the chemistry effort to go through that screening test so you can identify the molecules that have the characteristics that you want to move forward on. So yes, I think there is a lot synergy and now it's really just about us getting these targets [indiscernible] through them, building the larger group, to be able handle more of the different compounds under different programs that will be moving through. But we're excited about this and that is really full speed ahead, both with the Huntington's. And we think that this is going to ultimately be the best therapy for these patients as a consequence of the fact that is already bioavailable, that it is systemic, that it hits every cells and that's going through for all of the compounds that we move in terms of splicing. So we think there is a real importance to this platform and technology that has no other company better than we're positioned to be able to continue to do this.

Our next question comes from the link of Alethia Young from Cantor.

I just want to talk a little bit more about -- I mean, I know we all talked about risdiplam and antisense and how that stories might play out or look like it's going to play out. Just how do you think about some of the different approaches to Huntington's whether the antisense or sRNA? It would be helpful to see your color there what's your sweet spot?

Sure. Sort of what I was alluding to just before is that for bioavailability of the molecule, right, it's simply to take, it distributes throughout the body. It goes into the blood, and therefore, every cell is treated with the molecule. And that's, I think, even more critical like in Huntington's disease, right? Because at the end of the day, it is a whole brain disease as well as other aspects within the body. And I -- and I think the same thing is -- what's true with both gene therapy, it's true with antisense technologies. And I think it's always been somewhat true is the delivery process. You just can't get to every subtype. And so the notion of having orally bioavailable molecule that passes the blood-brain barrier in the case Huntington's disease. I guess, every call, that's quite efficient in terms of reducing both the RNA and protein. And by the way, it's both the RNA protein that probably has toxic -- a consequence of the toxicity that we think, again, that put small molecules at the end of the day to have the best chance of actually helping patients in the most complete way because of just being able to treat the whole brain. When you look at HD patients, it's not just the cortex and striatum that's affected, it's a whole range of disease. And so by getting -- by being able to get it to every portion of the brain, that's really a major advantage. Not to mention that it's orally bioavailable, easy to take. There's just some advantages to this that we think it ultimately should be the winner.

Our next question comes from the line of Gena Wang from Barclays.