Hello, my name is Pancham Wisenheimer. I'm Vice President of Research at Promega and I'm honored to be opening this for you. Let's start with a few thank yous. Thank you to the speakers for their transparency and new insights. Thank you for you that are present for your engagement in this conference and and thank you for the planners who really made all of this come together. I want to start by building a foundation on an often overlooked truth and that is the mechanisms that we are utilizing in our research are billions of years old. The ubiquitin proteomesome system is probably 1.5 billion years old and CRISPR system is probably a billion years older than that. And yet our understanding of these mechanisms is relatively recent and our ability to manipulate them to our therapeutic benefit is very recent. As community, we together have developed new editable genetic and protein reporters so that we can actually measure a quantitative endogenous level concentrations which are important in target protein degradation. We've also looked at live cell target engagement assays and we've even looked at live cell target engagement assays on complexes where we kind of unraveled the Ras RAF mechanism, even interpreting potential explanations for RAF a escape. In addition, we've even seen Med chemist shifting from just structure activity relationships into structure kinetic relationships, because now we have assays that look at both on rates and off rates. All these essays are not very meaningful unless you have the appropriate backgrounds. So it's not just about live cells and workhorse lines, but it's also about specialty backgrounds, Co cultures, even organoids along the way. With all these data sets coming in, we have many of us have focused on augmenting our interpretation with AI models. To do that, Permega is going to focus on two things over the next 10 years. One, new measures of cellular events in relevant models and two, assays that serve both humans and AI models. In the past we've always served humans twenty 30-40 years we've looked at linear relationships between an analyte of interest and a signal across 678 logs. But now that data seems a little thin for AI. And So what you'll start to see is 1/3 dimension of much more of a time domain. We'll have many more kinetic assays, which of course helps in our targeted protein degradation area. In addition, we'll see more controls and we'll see multi, more multi parameter outputs. Consider that the combination of all of these assays will allow us to reveal a little bit more of the dark proteome or a little bit more of the undruggables along the way, including even targets on RNA which at this point are difficult to actually see. So I leave you with an important question. What do you want to see tomorrow that you can't see today? Beyond this being a question that Promega wants to ask you, I think this isn't a question where I invite you to ask each other because with this question, you will be able to slowly reveal the next layer of biology, which is the basis of us moving forward. I wish you the very best conference. I hope you enjoy the talks. Take care. OK, Welcome everyone. Thank you, Poncho, and thanks to everyone for joining us today on our first day of our Cracking the Undruggable Code virtual symposium. Over the next three days we're going to be talking about some of the really important research areas that are helping the field to tackle tackle some of these, you know, classic undruggable targets and really advance therapeutic, make therapeutic advancements. Today we're going to focus on one of the probably the most historically well known undruggable target and this is the Ras signaling pathway. And we have some really exciting presentations lined up to to do that. Before we get started, I'm just highlighting a couple of housekeeping items. So on your screen you can see multiple windows, and these are all movable and resizable. So feel free to move them around however you prefer to get the most out of your desk desktop space. We also really want this webinars to be interactive, so please submit questions at any time during the event and we will address them either following each presentation, during our live Q&A panel discussion at the end, or via e-mail or through chat. There's also a resource library with a lot of helpful materials that you can use, so feel free to download any of these. And also noting that after the presentation, there will be a a short survey and we'd love to hear your feedback so we can continue to improve. And finally, there are going to be some polling questions throughout the event, so we'd love to have you participate in those so we can learn a little bit more about our audience and what you all are thinking. So here is our line up for today. And again, a lot of great talks today that we are looking forward to. But before we get started on the talks, I am going to start out with one polling question to learn a little bit more about about you, about our audience and kind of what your areas of research are. So, so this question is just asking you to describe your current research focus. So do you primary primarily study the Ras pathway? Do you study other protein, protein interaction actions outside of Ras signalling? Or do you focus on both Ras pathway as well as other PP is? So let us know, let us know which one of those best describes you and I'll give you a couple minutes there to put your answer in. OK. Looks like we got about half of you. So I'm going to go ahead and take a look at the results. So it looks like about half of you are studying other PP is outside of REST signalling. OK, that's great to know. So hopefully some of the learnings and and ideas that we, you know, share here with respect to the REST pathway will be applicable to you work as well. All right. And with that, I'm excited to introduce our first speaker is Doctor Arvindar. He is a professor in the Chemical Biology program at Memorial Sloan Kettering Cancer Center and Vice Director of the Center for Experimental Therapeutics with a PhD from the University of Toronto and postdoctoral training at UCSF. He's previously LED a research group at the ICANN School of Medicine at Mount Sinai. Currently, Doctor Darr's work focuses on kinase structural biology and small molecule design, leading to innovative therapeutic leads for both rats and wind driven cancers. He's also Co founder of Nested Therapeutics, which is focused on advancing innovative therapeutics aimed at challenging cancer targets with the goal of increasing the number of patients that can benefit from precision medicine. So welcome, Arvin. Great. Thank you, Amy, for the introduction. First of all, it's really honored to be here. I've always really enjoyed my interactions with the people and the scientists at Promega Poncho who did the introduction at the beginning, Matt Rovers, Jim Basta, various other individuals. So it's really great to be part of this symposium and it's cool to see in the survey that people are very interested in PP IS and parts of the Ras signaling pathway. That's certainly what I will mostly talk about, although at the end I'll transition to where we're building from our previously published work that you may or may not have seen before, but I'll describe it into a better understanding of interactions with the Ras directly. So this is my disclosure as you've heard before. So before I get into target engagement and what we're doing using tools that have been developed at Promega, I just wanted to put in context sort of the bigger picture problem that my lab is thinking about and I'm sure many of you are thinking about you're going to hear about from the other speakers today. And that's this general challenge for the field overall of developing broadly active well tolerated anti RASP pathway antagonist. I would say those don't currently exist. That might be changing in the near future, but we know it's a really important problem. We know it's important for a wide number of patients for which there are no current Ras or targeted therapies available. This includes non G12C mutant K Ras driven cancers, NRS driven cancers, NF1 mutant cancers. It's on the order of about 5 million newly diagnosed patients per year worldwide who have these rash driven cancers where there's no treatment options. So it's really important that we can address these particular cancers. It would also be important in the context where we do have targeted therapies, including breakthrough drugs like KRAS, G12C inhibitors like sotorasib or adagrasib or other molecules where unfortunately resistance is a major challenge. And the major form of resistance is usually reactivation of the Ras pathway, changes in the biochemistry, changes in the protein, protein interactions that can evade these targeted therapies. So we've been really trying to figure out and understand the protein, protein interactions in the pathway and trying to address this bigger picture problem. How do we develop well tolerated efficacious pan Ras pathway antagonist? The fact that drugs that would meet those criteria don't exist is of course, not for the lack of trying. This has been one of the most intensively investigated areas, I would say, in cancer biology, both in academia and in industry. And this is just really high level. Here's a beautiful review written earlier this year from colleagues here at MSK, Neil Rose and Dave Solid and others on various inhibitors in the RASP pathway, especially targeting RAF and downstream effectors. And on the right, it's just a high level summary of of various approaches targeting K Ras or Ras directly. So there's a lot of reason to be optimistic about the possibility of changing that dynamic where we don't have drugs for a lot of rash driven cancers and we'll see what happens. And I think there are a lot of people in the field who are quite optimistic about what's happening right now. Invariably, I would say there's two major challenges that any drug that is going to be advanced to this pathway has to overcome. The first being drug resistance. And what we've learned is that there's many different types of resistance that can develop either intrinsic resistance where a drug is just not efficacious for some reason about how the pathway might be wired in a in a various context. And there's acquired resistance, which is what many people think about. That's where you can have mutations that might not go to drug binding. But then there's this sort of in between resistance, adaptive resistance, but again, probably changes in protein, protein interactions, how the pathway is wired that can lead to loss in activity of a drug. The other major challenge for any drug targets pathways, therapeutic index sort of the bigger other side of the equation is how do you effectively block the pathway in the tumor cell but not have on target side effects in other tissues. And a lot of the pathways and targets that we focus on have overlapping an important roles both in cancer and outside of cancer. So again, another major challenge. So in my lab, we're we're thinking about targets and approaches that remain untested. And I think especially in the context of protein, protein interactions, what we've been learning about biochemistry of, of this pathway, especially in in vivo context where tools like Nano bread and various approaches that I'll talk about later become really important. What it's telling us is there's actually a lot of different ways, a lot of different targets, a lot of complexes that just have not been tested, not been exploited in specific fashion. And so we've been focused on chemical biology tool development in my lab to try and test these hypotheses. This is just very high level overview of different projects we have going in the lab. A lot are based on molecular glues, a lot are based on orthostero kinase inhibitors. We're also interested in that potential overlap between orthostero compounds and molecules that affect protein. Protein interactions might act like molecular glues themselves. So, but I thought for today's talk, I would start with this question and this may be is a little bit highlighted in Poncho's introduction about mechanism, these fundamental biological mechanisms that we know are important in cancer. What we started thinking about a number of years ago in my lab was we were thinking about compounds that were already existing in the field and we're asking ourselves, do these compounds actually operate as advertised? Do they operate by a mechanism? Does a compound X operate by a mechanism Y? Of course, that's a very important question even for early stage compounds through to clinically developed molecules. The unfortunate reality is most drugs are not going to be successful in clinic, but we can still learn from that if we really understand the underlying biology mechanism of how a molecule works. So my lab, we started thinking about this question in the context of small molecule MEC inhibitors. This is one of the effectors downstream of Ras. This is just a few different compounds that have been developed as as direct acting MEC inhibitors. One of the molecules that stood out to us for a variety of reasons with this compound, tremetinib, it seemed like it became a little bit of an outlier. You can see there's some structural similarity between all these compounds as highlighted in blue, but there's differences also in kind of highlighted in black. These, unlike most kinase inhibitors are not ATP competitive. They were known to be allosteric, they were known to be very selective, but they showed very different biochemical through the cellular, even patient activity. And we wanted to try and better understand that one of the reasons Trimetna was sort of an outlier was it wasn't actually developed as a Mac inhibitor. It came from phenotypic screens developed at a company called Japan Tobacco. And then that eventually molecule kind of progressed. It was required by GSK, the Novartis. And of course Novartis ushered this molecule through final clinical development and it was approved as the first single agent MEC inhibitor for cancer therapy for B RAF mutant Melanoma. But of course, it's mostly used in combination with RAF inhibitors. So we were thinking about this molecule. We wanted to better understand how it was working. There were some hypothesis in the field about it might affect protein, protein interactions in certain ways. But it was hard to really understand this, in part because there was no structure available of this compound bound to Mac, no crystal structure, which is extremely rare, of course, for a clinically approved drug. And having done my PhD in structural biology, I thought this was very odd as well. There was many structures of Mac down to other Mac inhibitors, but for some reason there was no structure bound to tremetinib. So what my lab did was we focused on trying to solve structures of Mech bound to Trimetna, but also in complex with various binding partners, proteins that form complexes with Mech. And we focused on the Rath family of kinases that include the active members A RAF, B RAF and C RAF, and then the pseudokinase relatives KSR 1 and KSR 2. As you can kind of see on the right here, KSR 1 and KSR 2 have mutations in an active site that characterize it as a pseudo kinase. And the KSRS also lack a RASP binding domain. So they're not direct sensors of Ras GTP like other RAF, sorry, Ras effectors. So it kind of it's been a bit of a mystery about how KSR might be working. Anyways, our strategy was to Co crystallize Mac with trimetnib and also with KSR and that worked actually very well and we were able to solve the first crystal structures of Trimetnib in complex with its target. And what we saw was Trimetnib bound sort of where we expected we knew was HP uncompetitive. So we could still see an HP analog bound and the Mac active side and also in the KSR active sites and Trimetnib was bound adjacent to that. But the part we didn't expect to see was that a lower portion of the molecule, this panel acetamide kind of extended away from the conventional Alistair pocket and made direct contacts with KSR. In the case of KSR one shown on the left here in the case of KSR 2 shown on on the right. The other thing we realized was from other crystal structures was that this ability of trimetinib to traverse the interaction interface was quite unique to this molecule. We subsequently found other compounds and other examples, but at the time this was the first one we found. And then in comparison to other well known clinical MEC inhibitors, columet of cellulumet or mirdemet nub, also known as PD 03259 O1, These molecules were much further removed from the interaction interface, in some cases up to 10 angstroms. And then finally the other part we realized from our structures was that the MEC inhibitor binding pocket was actually changing quite significantly between the isolated state shown on the left and the KSR bounce state shown on the right. And in the KSR bounce state, what we realized was the drug binding pocket was changing in shape and size in the complex. And so it was actually becoming slightly larger. And what Tremendous did really well was with that extended phenyl acetamine fill in that, that sort of reshaped reformed Alistair binding pocket. So we had this beautiful structural data. We started to think about this question, you know, could KSR actually be the physiological Co receptor of tremetinib? If so, how could this be missed? We we thought about that question as well. Nobody had really described KSR as being important in the mechanism of action of tremetinib or for in mech inhibitors in general. So we thought this was sort of an interesting for us to find that based on our structures they could have significance and how mech inhibitors might be binding to their target. And then another question we've thought about is nor are there other related targets that have similarly been overlooked. I would say now with the tools that we've been developing based on Trebetna, been exploring KSR, extending that to other complexes has been really illuminating for us. So one other piece of data that we had at the time, but so we first saw the structure, we had some biochemistry. We looked at CRISPR genetic screens that people had done in the presence of trimetinib or trimetinib in a RAFT inhibitor. And we could find a strong signal that KSR was probably important in the mechanism of action of trimetinib. If you knock out KSR, actually that could drive resistance that would be consistent with our crystal structure in our biochemical data. I told this story earlier when I was at Primeca. I thought I would tell it again. So at this time we had the structure, we had biochemistry, we had genetics to support the mechanism. We submitted the paper to top tier journal to Nature and unfortunately it didn't get accepted. And the big thing we were missing was a connection between the structure, you know, our in vitro data and what our insights were telling us and the physiological data, the CRISPR based data, the cellular data, we didn't have that in between that could connect the structure to the physiological mechanism. And this is just a few of the sentences from our unfortunate rejection letter. But we did get, you know, one of those classic lines that, you know, there was the very kind way of saying the paper was not accepted, but then also, you know, a bit of a teaser saying, you know, they would consider revision if we could address all the concerns. And the big concern was we didn't have this connection between, as I mentioned, between the in vitro biochemistry and the cellular function. So at this time, we had just started, you know, looking at the technologies that were being published by the Permega scientists, especially as I mentioned, Matt and Jim and Matt Rovers, Jim Vasta. And we saw these amazing tools where you could basically run live cell biochemistry in vivo, test drug receptor interactions, test the hypothesis that might come from the structure, it might come from a genetic screen. But add in that missing link between those two domains. And I was lucky to also have very talented post doc in my lab at the time, William Marsiglia. He was Co first author on our original 2020 paper. And then we had a subsequent paper published and the 2023 actually in journal in 2024 where he built and he started exploring banana bread tools to look at this question of in vivo target engagement. Could we get significance for the interactions we saw? And So what Billy did was he took Trametna and he added the linker and the Vodipi onto the molecule at the a sediment position. The idea was that this could be our tracer, this could be our acceptor energy acceptor for light produced by nano luciferase tag versions of MEC or KSR or other proteins in the complexes. This is how we originally envisioned the assay. The idea was that if you expressed MEC as a nano luciferase fusion, then the nail luciferase if given its appropriate substrate would produce light that can be absorbed by the Bo dippy on tremetna. If in close proximity, of course it would admit as second wavelength and then we could detect that. That's the, of course the fundamental basis of Brett. The other idea we had was that in the context of KSR bound back complex, we can put the luciferase. We thought we could put the luciferase on KSR and then similarly look at this complex, but in this case here we will be looking exclusively on the right side at the ternary complex. On the left, we could be looking at all states of mech and a cell either bound to KSR or bound to other partners or in the free form. So here's a crystal structure and now I'm going to try and start this animation, which I have to press this button. Where is that button? Oh I think you are seeing the animation on your screens. So the animation is showing the the tracer molecule bound in the KSR Mac interface. We very fortuitously had this right hand turn by the linker motif and place the boat dippy right at the side of the complex. So we knew is perfectly positioned to look at ternary complex formation. Basically the binding of TRAMBO or Mac inhibitors. Molecular glues found in the KSR Mac complex. So this is just one of the experiments we did, but I think to me it's one of my favorite experiments because it's very illustrative of why the protein, protein interactions in this particular context could be very important. So we actually measured steady state affinity of trimetna on Mac expressing an analytics fusion or the KSR Mac complex. And at steady state actually we look, we could measure very similar IC 50. So it it looked like actually the binding was very similar in both context. But what we realized was that if we did wash out experiments and we did more of a kinetic experiment, actually the binding activity of the drug was very different. So here we're looking at the binary Trambo MEC complex. And what we do is we express that in cells, then we pretreat with a free ligand, in this case trimetinum. We wash off trimetinum off the cells and then we add our tracer. And then what we're looking at is over time, what is the build up curve or the build up for a new Brett signal. And of course, this would be inversely proportional to the dissociation of the free drug. So we can see in the case of free Mac in the binary complex, we did a lot of mutation structure function to convince us that we were looking either binary complexes or ternary complex complexes. We can see that although the drug bound potently, it could wash off very, very quickly, suggesting that it had a fast off rate. Whereas in the context of the ternary complex, once that complex was formed, basically the drug would not dissociate. There was no buildup of of tracer onto the complex, suggesting that once KSR forms that complex that it's basically a barrier to dissociation. So this was a really interesting insight. It was the link I think that put the structure back to the genetics that helped us explain resistance is really a a really important finding for us. So what have we been doing since? So as I mentioned under standard conditions under A2 hour assay, which would be the normal way of running an interbred assay, we actually see a steady state affinity for tremendum on MEC alone or the kissarmic complex is very similar. But we realized if we do more extended experiment in 24 hours, we start to see actually a separation in the steady state affinity consistent with that slower off rate from the complex. You might say, you know, it's a 24 hour, I say even relevant. So we're still asking that question ourselves. But one thing I will mention is Trementa actually clinically has an extremely long half life, about 120 hours or five days. So we're starting to think, you know, is this a condition where we should be looking at interesting protein, protein complex effects, especially for micro glues, especially to find these, you know, unique molecules that have these very unusual mechanisms. I will say that we've done broader based screens and we've started to find other molecules. This is I'm just listening as compound X yet right now, because we haven't actually published this particular compound. It's in the literature, but it's not described as a molecular glue. And but we find in our assays that has very interesting molecular glue like properties, sort of like trimetinum. It's dose dependency is shifted in the presence of KSR, but on a very on actually a faster time scale and then at a greater depth that's highlighted by these arrows. Where it's also been getting interesting for us is looking at mutations, especially mutations that give rise to resistance to direct acting Ras inhibitors. So you may be aware that with inactive state switch to pocket binders, molecules such as that aggressive that have been put into patients, individuals will develop resistance. One of the common resistance mechanisms is downstream activation of proteins including Mac, the protein that I've been talking about, K57 and 256P. This D1 O2 one O 3 deletion is very common. This can give resistance to various Ras inhibitors, including newly developed Tri complex modulators. This has been shown in a publication at least in preclinical cell and models. So of course we've been exploring these mutations. One idea you might think is that, you know, you could use MEC inhibitor, but what we've realized is actually it's it really depends on the compound. So again, using trement just as an example molecule. So not only do these mutations give resistance to direct Eck and Ras inhibitors, they'll also give resistance to most allosteric MEC inhibitors, including tremetina. We can see a loss in potency with various mutations, more dramatic with the deletions. But again, we're looking for these interesting compounds and we can find molecules that actually break break that trend. They go in the other direction, higher sensitivity on the mutations, especially in various complexes. Going to speed up in the last few minutes. So I mostly talked about one specific complex, the KSR MEC complex, but this is an exemplary of many other related complexes that occur in cells. Eventually we will look at the pathways to draw it in a as a linear diagram, but there's a lot of different combinations that can be put together from similar components. As I mentioned, KS Rs relate to the rafts. There's multiple rafts, there's multiple Max. You can get a lot of different assemblies that can occur. These are actually experimental structures solved by my lab or other labs in the field. These are predicted structures for hetero American assemblies. And you know, just from a small number of components, if you count the Rasis, the rafts, the KS, Rs and the various flavors of each, just from a few components, you know, from 4 different components, let's say G protein, Roth, KSR, MEC, you can come up with probably several thousand different biochemical entities. So we've, the next question we've been thinking about is, you know, how do we better assess all these different complexes in vivo? This is an alpha full model depicting, you know, one of those types of complexes MEC bout to KSR bout to RAF, also in complex with KRSGTGTP. We think this is possibly a realistic depiction of how these proteins will actually assemble together. We have mutational data to support this complex and functional data as well. What we've been thinking about is, you know, how do different inhibitors actually engage their respective binding sites, especially in the context of these interactions. Of course of MECH allosteric blue site will be bound in close proximity to an orthosteric RAFT inhibitor binding site, which is actually is in relatively close proximity to various Ras direct acting Ras inhibitor sites. So this has been work of two people in my lab, a postdoc around Mystery and a student, Lucy Om. I won't go through exactly how we're doing the assay, happy to discuss it in the question. But what I will say is this is just exemplary. The type of data that we've been looking at this is K Ras detail C induced RAF RAF or RAF KSR complexes. These are four different RAF inhibitors. Three of these are type 1. They're of the similar class. We can see the very well known paradoxical activation induction of various raft dimers. We also looked at a paradox breaker and what was interesting here was that the paradox breaker is ability to dissociate various dimer complexes was strongest not on the complexes that are actually induced by type 1 inhibitors, but it was actually on a distinct set of complexes. This is an interesting insight for us. It sets up a lot of new hypotheses for us to think about. The other thing we've done is looked at other classes of compounds, but in the same assays. So we have sort of, you can almost think of it as a modality agnostic way of looking at Target engagement. And what's interesting here is you now can categorize compounds. This is just looking based on individual molecules. But if you do hierarchical clustering or some other processing of this data, you can see which compounds are most similar to others. And what's intriguing is a lot of MEC inhibitors look like certain rats inhibitors, but each rats inhibitor has a sort of its own profile. There's certain complexes that are intrinsically not sensitive to target engagement depending on which effector complexes are found. Again, this is a very interesting new hypotheses for us to think about. So I'll finish there and just a few conclusions. So as I mentioned, target engagement has been really important tool for my lab. It's really permeated every project. It's been a very enjoyable direction for the lab to engage on as well. We're using these approaches to look at molecular glues, to look at specific complexes in the RASP pathway, and we're also exploring some other projects as well. And I just want to make a quick shout out. So I'm very lucky to have some of the technologies from the lab licensed into a new startup company, Nested Therapeutics. A really talented group of drug hunters and scientists have built on a lot of these discoveries, built the first clinically active pan Ras, RAF Mac molecular glue with a very unique profile. It shows really strong preclinical data on these molecules. NSD 628 is actually just entered the clinic earlier this year. And finally, thanks to the members of my lab, I mentioned individuals working on the project, previous people working on the KSR, tremendous story, current individuals. Lucy, I'd mentioned her all as well. Two people specifically thinking about target engagement. These are some different collaborators that we have as some Kettering and then these are our funding institutions. So thanks very much for your attention. Great. Thanks so much, Arvind. Great presentation. And if anyone has questions that they would like to ask specifically to Arvind, please drop them in the chat and we will, I think we'll address them when we do the live Q&A at the end. I didn't see any comment yet, so, but you know, we're really would love to, yeah, address any questions you have. So please, please drop them in the chat and we will do that when we have our Q&A session following the last presentation. So with that, before we transition into our next speaker, we have another pool to learn a little bit about our audience. So this question is what innovative approach do you believe holds the most promise for targeting the protein, protein interactions in raft signaling? Arvind just gave us a lot to think about there. So A would be developing small molecules to disrupt the PPI. B would be a different approach of instead of small molecules, stapled peptides or peptidominetics to interfere with interactions. See applications of computational modeling and AI to identify novel bonding binding sites. I think Arvin did a really nice job of showing the value that can come from that or D exploiting allosteric sites to modulate confirmation and function. So yeah, I think, you know, what do you, what do you think when we think about what holds the promise here? Where should we really, where's, you know, the most benefit coming from? So I'll give you a minute or so to think about those and submit your response. It's taking a little longer. I'll go ahead and see the responses. So it looks like we've got really AI think a well well distributed set of opinions. So some of us thinking about small molecules to disrupt PPI's, some interest in computational modeling and AI to identify binding sites and quite a bit of interest in exploiting those allosteric sites to modulate confirmation and function. So thanks for sharing in that and and these are some things we could potentially talk further about in our Q&A discussion at the end. So next I am happy to introduce our next speaker, Hoyen Cho. Joe is has worked at Novartis for 15 years in the high high throughput screening group clone Co leading projects in oncology and cardiovascular disease and metabolism. She is also Co leads the advisory team at the Novartis Cambridge site supporting assay development, HTS and automation to optimize project screening outcomes. Recently she has focused on establishing platforms for chemical and genomic screens using HUP human IPSC derived cardiomyocytes. So welcome, Joel. Hi everyone, thanks for having me today. The focus of my talk today will be on introducing new essay technology for quantifying postulation on target by bench marketing P arc which is a crucial for studying the core cascade of the last laugh Mac arc signals transaction. I will be sharing the preliminary data of the P arc lumi cellular essay in 5th in 36 well title plus clinic format using the Novartis Chemogenetic Library as an alternative reader for the costly TR, FRET or STRF cellular asset. Lumiant immunoassay utilize none of its split luciferase technology. The principle of this assay is that in the presence of analyte the two antibodies come into close proximity align small bit and large bit. As you see here, the form are active enzyme, in this case nanovisciferase and the generate of prana. Since it's a homogeneous one plate essay without requiring washing steps, it can be a faster and more reliable alternative to Elizas and western block. Therefore, it's easy and no wash work floor makes it as automation friendly essay as to internal project. Our project team has evaluated both Lumid and the TR Fred SA to quantify the postulation over target protein in human IPS drive to cardiomycide. Traditionally, we developed essay for screening with the TR FRET or we can call it HTRF essay first. However, due to limited essay window less than two four in this case in 384 F format, we parked this approach and we look for alternative essay for the development immune essay. We made the CRISPR knocking cellular model here by knocking in small bit which is 11 amino acids at the C terminus of endogenous target. Once our target is a postpulated after additional primary antibody and secondary antibody are fused with the Rajiv, we measured increase nano reciprocal signal. This is a hit map in 1536 year old for men. As you see in the essay window was robust and as a six fold window. Therefore, we chose this Lumia essay as the economic and the sensitive alternative to tier Fred and H Terab or Alpha light assays to date. Cardiomycide produced at the scale of it were 9 billion cardiomyocyte to enable the high terpus screening and we screened 300,000 low molecular weight. So far before triggering the high terpus screening, we have faced stability issue of this lumen substrate. As you see in this left panel, the Lumi substrate diluted in traction buffer provider in this kit resulted in 50% signal loads at two hours and over 90% at 10 hours. I have reached out to Promega and the Met Kishan from Promega recommended alternative Gliser ethanol buffer to dilute the roomy substrate in the middle panel. Here, grooming substrate was stable up to 24 hours in new buffer condition in the right panel, although the overall low signal was lower with a new dilution buffer. The identical folder changing and IC50 were observed from two different subject evolution buffer here. Therefore, we proceed the height of the screening with a new buffer condition here. As the Arthalizer and HRF kits are quite expensive these days, especially for large scale screening, Our goal internally was to establish this new technology as a cost effective and highly sensitive alternative for. Title The screening. Promega offers a wide range of cellular essay kits, including sets of phosphos specific antibodies for several important signaling Not. After reviewing the list, we were particularly interested in phosphorus Step 3 and force PROC limit essays in the left panel. Based on the data provided by the Promega, the rumen assay sensitivity was comparable to iphaliza in the right panel. The. Additionally, as Promega has conducted extensive research with the PR readout by profiling oncogenic K rasmutant drugs, we have chosen this PR immunoassay readout for internal evaluation. We in the Novartis frequently use force for essay as the functional readout in many projects as it is a key regulator and a major signaling not in biology. For example, the core cascade Ross and the Mac work signal transduction leads to the regulation of a great variety of a cellular process AS2SA configuration. As you see here, we utilize the sandwich method with the two primary antibodies, so one for binding total ARC protein and one for binding phosphalated residue, and then two secondary antibody fused with each small bit and the large bit so that we don't need to have the engineer the sender the model for this evaluation in the right panel. Sharing the brief the essay protocol in 1536 your playful math 1st, we seed MCF 7 cells into 1536 well with the in low serum. Then we treat the compound and 20 micro mole at single dose and however, and and we use as actable control. We use the TriMet meat. This is important and selective and neck 1 to inhibitor and we incubated for 30 minute then we stimulate MCF 27 with apna and then incubator for the 30 minute and then we lyses a cell and add the detection antibodies and primary and secondary antibody all together. And after 2 hour incubation we read the plate in parasta for the screening we. For low molecular weight compound screening we have used Nobody's chemogenetic library composed of about 7000 compounds. It is more commonly known as MOA box internally by This is by date made by data mining and crowd sourcing institutional expertise across Nobody's biomedical research. This collection is a distinguished from other such libraries with respect to the wide brothers of coverage. As you see here in the right panel about more than 2200 mammalian gene targets with a less than one microbiochemical potency. These are well annotated compound with targets such as here a lot of GPCR agonist and antagonist and many kinase inhibitors and iron channels and so on. This collection of a well defined chemical probes provide tremendous failure to drug discovery such as probable concept or new technologies. So that's the exact reason we are using this chemical genetic library for this limit essay technology and also we can evaluate drug discovery floor chart and counter screen design and also facilitate understanding of a related pathways in the penotypic essays and off target effects. In addition or searches set include anti MOA box which is useful for improved MOA hypothesis generation with the inactive analog. So these different chemo types have a different off target and also active and inactive analogues have the same off target. So we can have a comprehensive data by screening this library. So so this MO box is celebrating its 10 year anniversary now. So in this time it has grown in size as well as impact and is the most common screen, a collection of a compound within nobody's biomedical research. Now here we perform the primary screening at 20 micromer single dose. On the top it is a heat map of a compound treated into cells in 1536. Well, the positive control, the traumatinib treated at the last two columns as you see here shown as a blue and the inhibitors were shown as a blue dots in the. So each box is 11536 web plate and the activator shown as red dots here and the neutral compound included DMSO shown as white. Average for the change between DMSO and the active control is about 18 fourth and the G prime was about point zero 2.62, which indicates the good assay quality and reproducibility in well activity histogram. Below it is heat distribution as DMSO as 0% in the middle and the active control through alternative as a -, 100% activity with a plus and -2 standard deviation cut off. Because we are interesting find the both activator and inhibitor. The hit rate for these inhibitors about 3.4% and hit rate for activator is about 1.3%. These hit rates are are relatively high compared to the traditional hydrogen screening as this library has enlisted for some targets such as a kinase. After primer screening, we ran the secondary screening which is a for those responsive validation in duplicate. We treated a compound top the 40 micromer and eight point health log dilution and tested about 229 CUP inhibitors and 52 available activators from our the compound herb. The right bottom dose of this fast curve are the example of a validated inhibitors with a compounds annotated as a ERC and Mac and Mac K1 inhibitors and some compound. Some compound is a sole proton as you see here in the top right, but the signal is a complete inhibitor at lowest concentration which is A10 animal. After those responsible validation of a highly validated human target the target pairs. We are showing that these scatter plot with the AC 50 from active and inactive pairs more than with more than five 4th selectivity difference. Activity from the active analogs and in in X axis and activity from inactive analog are in Y axis and lower scale. For your information, the inactive compounds in Y axis shown as a 40 micromole on the top here and which is the highest concentration we have tested. Wire potency from active analog shown here are below 6 micromole in the top panel. The numerous ARC tube and the Mac inhibitors with multiple chemotypes were captured validating this Lumi P arc Lumi immune assay for quantifying the postulation on the ARC protein. Interestingly, the compound targeting the GSK 3 was identified, which it seems crosstalk with the Earth pathway. In addition, in the bottom right eye in some selective activators were identified majority targeting PKC. And then I think we think of PMA in our the SA protocol PMA used to activate sensitize MCF 7 cells that's why and this PMA works through PKC pathway that's probably we are picking up a lot of PKC activator or inhibitor are here. And however, further in depth analysis and heat follow up activities are ongoing currently. As a summary and next step Lumia cellular essays to quantify the postulation of a target proteins have been successfully optimized in 1536 well to enable high double screening including P Urc Lumia immuno essay urc Lumia The cellular essays are showing good essay performance in 1536 well G prime above a .5 which indicate good SA reproducibility. Also, lumid substrate diluted in the trashing buffer provided in the kit was initially unstable resulted in the 50% signal loss after two hours and the 90% the signal loss efforts. Through collaboration with the Promega team, we addressed a substrate stability issue by switching to new buffer condition. We should provide to be suitable for height of screening over 24 hours screening. Nobody's chemogenetic library revealed the enriched targets from MOA box with a greater than 500 fourth over AC 50 compared to inactive analogs from active MOA compounds. As the next step, while we are doing in depth data analysis, we will do the counter screening to filter out any essay interference compounds such as nano reciprocalized inhibitors or activators. The configuration will be we will use the secondary antibody from the same species and those antibody will automatically come to the close proximity and become the fully active. But nano reserve rises. So we can filter out any compound directly inhibit or activate activate this nano reserve arrays. Also we will need to run orthogonal screening, but usually orthogonal screening we we do we measure the same signal and with the different without. In this case, we will use PRP HRF assay which is commercially available to confirm heat correlation of AC50 and the selectivity and again the further investigation of high value compound both inhibitors and activators with a full dose medicine active and inactive analogues. We will continue investigate and therefore hopefully we find the interesting target of which paper next year with the Pro Mega together. This is it for for my presentation and thank you so much for your attention and please feel free to share your question and thoughts. Many thanks to Doug and Carol, my site team and chemo genetic library creators and automation teams in Novartis. And also a special thanks to Matt and Hisham from Promega for the collaboration in terms of a scientific and Technical Support. Thank you very much. Thank you, Joe. That was a great presentation and really interesting to learn about how you've been applying these new Lumid assays in your screening efforts. Again, please drop your questions into the chat and we will collect them and discuss following our next two presentations. For now, I'm really happy to introduce our next speaker, Doctor Tommy Turbyville. Tommy earned his PhD in Cancer Biology from the University of Arizona and focuses on using imaging and innovative methods to understand cancer signaling and develop therapeutic strategies. For the past 10 years, he has been a team leader within the NCI supported Ras initiative, studying the molecular mobility, residence time and activation of Ras and its effector RAF on the plasma membrane of cancer cells using real time methods. His team utilizes advanced techniques including single molecule microscopy and nanobrat to investigate mutant wrath interactions with small molecules and really advancing strategies to target oncogenic wrath and improve cancer therapies. So welcome, Tommy. Thank you very much and thank you for organizing this the symposium. It's enjoyable to see the other talks and hopefully people will learn something from from my talk where I'm going to present some of the work that we've done here primarily with Nanopret, which is a a wonderful tool. So I don't want to go over this in too much detail, but we do work on Ras. And Ras is a signaling molecule that's primarily localized to the plasma membrane of cells where it switches between a GDP and GTP state, an on and off state. And when it's in the GTP state, it activates a lot of different pathways, including the MAP kinase pathway, which we've already been hearing about, but other pathways as well. So when Ras is mutated, it's stuck in the on state and it is going to be signaling constitutively downstream through all of these pathways contributing to phenotypes like cell growth, migration, invasion and all kinds of things that contribute to the the cancer phenotype. And it's a a well known fact of Ras that it is mutated in numerous cancers. K Ras is the major Ras oncogene and it is common in pancreatic cancer where over 90% of pancreatic cancers have mutations in in K Ras, but also in lung cancer and colon cancer. And you can see here on the right from this review article that the spectrum of different mutations in the in the in Ras occur mostly in codon 12 and codon 61. But you can see here that there is a, a number of different kinds of mutations that can happen in Ras. It can be AG12C mutation or AG12D mutation. And there may be subtle differences between those mutations in the way that they behave in cells. And that's one of the things that we, we are also interested in here at the Ras initiative. So I mentioned to you that we study Ras dynamically. We particular, my team, we, we really like to study Ras in the context of cells because we know that in order to be able to target Ras effectively, we're going to need to understand its Physiology in the cell. We do a lot of biophysical and biochemical and structural biology at the Ras initiative where we understand Ras behavior in at atomistic detail, but we also need to understand how Ras is behaving inside of a cell. And you can see from this video here looking at single molecules of Ras that they're extremely dynamic. They're moving around in the plasma membrane, They are changing their their mobility all the time. And this map, this diffusion map over here on the far right is showing you that the diffusion rate of Ras changes depending on what it's interacting with, either with lipids in the plasma membrane or its effectors. So we try to understand that and elaborate on that. And one of the ways that we're hoping to be able to do that further is by interrogating the small molecules as they interact with Ras inside of cells. And that's because the plasma membrane is where Ras is primarily active and signaling, and the plasma membrane is a very complex organelle. Basically it's a 2D organelle where you have a very hierarchical structure. It's supported by a mesh work of actin cytoskeleton. There are other proteins, RTKS, all kinds of GPCRS and you know many, many proteins and then over 200 different kinds of lipids all interacting in a very dynamic way to form a kind of hierarchical system in which to initiate signaling. So you can have freely diffusing molecules that get further confined in little areas where they can interact with other proteins and generate signals downstream. And that is exactly what Ras does with some of its major effectors. So we try to use structural, biophysical and biochemical assays to really target Ras here at the Ras initiative. And we have a number of of really wonderful clinical agents that are now going into patients. That work is incredible, but we're also here trying to understand what those interactions are in a cell context. And in order to do that, you need the disease Physiology is close to what it's like in the patient. Cells and tumors are very dynamic. They have metabolism, cell cycle. They are in different states of quiescence and stress and the proteins and macromolecules in the in the cell in the context of a tumor, very dynamic moving around. And so that's where assays like FRET, these protein, protein interaction assays can be very valuable because you can actually look at these interactions as they are occurring in the cell dynamically. So early on when we were talking about different ways of looking at these interactions, we came across this paper. We had tried a few different forms of Brett prior to that using our our Luke and Venus and other, other fluorophores and, and bioluminescent molecules in order to generate a Brett interaction. But when we found this nano Brett system, it was ideal because there's very, very high signal coming from the luciferase and a really nice separation between the acceptor and donor in terms of their fluorescence and emission that allows you to get very little background from the signal. And so it's the ideal platform for developing bread assays. And so when we develop bread assays, we follow a kind of a, a protocol where we first of all evaluate whether the C terminal and terminal fusions of the different probes are the best. RASP can only be N terminally labeled. You can't, you can't put a probe on the C terminus because that's the part that interacts with the plasma membrane. But the effectors that we're looking at interrogating, you have to actually evaluate which is better, the n-terminal or or C terminal end of of RAF for instance. And then optimize all kinds of things for the assay, number of cells to see which transfection reagents to use, which cell lines to use, the amount of donor to trans fact and and acceptor so on and so forth. We use this saturation curve as a really important way to evaluate these parameters. For one thing, if the signal is non specific, if it's just because of collision, you tend to get a very linear response. And so that tells you that it's not a specific interaction. If you get this hyperbolic curve that tells you that you have a specific interaction, you're holding the the acceptor, I mean the donor concentration constant at a very low level and then you're adding increasing amounts of the acceptor. And as it starts to saturate the acceptor signal, you get this hyperbolic curve that's telling you that there's a specific interaction happening. And the Brett values give you some information. High Brett Max values give you some information about the number of interactions that are happening in the cell. And then the steepness of this curve gives you some information about how much affinity there is. It's not a real KD sort of evaluation, but it does give you an idea of how much specificity there is in terms of the interaction. So we've done this for numerous acceptor donor combinations looking at Ras and effectors. And I'm going to show you some data of a screening essay that we did in collaboration with Eli Lilly looking at the interaction between RAF one, the major effector of Ras in the MAP kinase pathway, and the oncogene K, Ras 4BG12D. And this is just showing the saturation curve. So you can see that you get this very nice hyperbolic hyperbolic curve, very nice Mili Brett value as you increase the concentration of the acceptor in the cells. And then as a control, we have this arginine 89 leucine mutation in BRAF. This disrupts the interaction with Brass because it's in the RASS binding domain of wrath. And so you can see here that this protein does not interact with with the RASS molecule in in these cells. And then we look at these parameters formally to try to find the best amount of DNA to use for our transfections. We're looking for a high signal to background signal so we can distinguish hits from from non hits. And we also want AZ prime factor that is very high above .5 ideally. And in this case, we get a very nice Z prime. This is a, a, a calculation based on the standard deviations and the means of your, of your signals and gives you a nice readout for the quality of your assay. It's a quality control measure. It's used a lot in screening. We also evaluated that these interactions have some sort of biological meaning in cells because we're overexpressing these these proteins in cells. And so here we're just using the wild type version of K Ras in combination with the effector at A at a stable combination level of expression. And you can see that when we Co express untagged gap protein, so this NF one, this turns Ras off. You can see that you get a pretty dramatic decrease in the signal as you would expect, because all of the active Ras is being turned off in the cells by the overexpression of this gap. Or if you stimulate these cells with a growth factor, you can see an increase in the signal relative to the basal levels. So these combinations tell us that these are biologically meaningful interactions. We're not just, you know, combining things in a in a cell sort of randomly, but they're actually functional in the cell. And we evaluated this for different effectors including PI3 kinase and Rao GDS. So those are other effectors that are known to interact with Ras and we can see similar patterns where EGF stimulation increases the activity of the combination. So this is just an example from one day of screening. So we screened a library of about 80,000 compounds from from Eli Lilly was a diversity compound library and they sent us the library in plates and 384 well plates and we evaluated in that one dose in 384 well plates. We have a little bit of automation here, nothing like a big pharmaceutical company, but we do have some ability to to automate some of this. And you can see from here that the controls from different plates performed very, very well. So this is the R89L control here. So this is showing this is sort of akin to inhibition. It would be 100% inhibition if a compound was behaving this way. And you can see that on each plate, these, these controls behaved very, very well. They're almost identical. And then down here in green are the the positive control. So this is the high highest interaction. And then the different colors here represent the compound libraries from each plate. And you can see that we got a a variety of activities. Most compounds were inactive, so they they didn't inhibit the interaction at all. Some increased the interaction. And that's one of the advantages of the bread assay is not only do you see disruption of the protein, protein interaction, but you can also see enhancement of the protein, protein interaction. And then we saw some compounds that inhibited the interaction. So we did a lot of follow up experiments on this. But just to show you how the overall campaign performed, we had a relatively low hit rate, which is what we wanted. We didn't want to have all compounds hitting in the in the assay and and then when we look at day-to-day. So it took us about 15 days work days to kind of proceed through all of the 80,000 compounds that the signal to background was very stable in all of our controls and the Z prime factors were all about above .8 for the most part, which in cell based assays, I can tell you from my own experience, this is almost unheard of to get a cell based assay that performs this stably and this well where you get such a nice tightly controlled assay. And so that allowed us to really prosecute this assay in a way that we we could trust the data that the hits were actually hits. So this is just an example of a compound that came out of that screen just to show you that there we counter screen them against the Brett control, which is just a a luciferase and nano Luke fuse to a Halo with a linker in between. So it gives you a a very high Brett signal and any compound that showed activity in the primary screen, we screened against that Brett control and this particular compound had no activity on the Brett control. So that tells it's a specific to the interaction. And then in addition it inhibited the interaction of Ras with its other effectors. So not just raft one, but other effectors also. So this was telling us that this compound is active in the assay and in in addition, we did some experiments just looking at the behavior of the compound with different isoforms of Ras. So K Ras was our primary target because it's the most common oncogene, but there are other Ras isoforms, N Ras and H Ras and K Ras 4A and we saw that this particular compound seemed to have greater activity against K Ras compared to H Ras. So there is a big difference between the activity of this compound. It seems to be specific to K Ras, which was a nice surprise because the the differences between Atras and K Ras are very subtle. In the globular domain, there's hardly any differences in the amino acids and there's only in the hyper variable region that you actually have major differences in terms of the amino acid sequence. And this is just highlighting that there's a a big difference in the IC 50S between K Ras G12D and H Ras and red down below. So this LED us to an interesting hypothesis that maybe it has to do with the interaction of the HVR with the lipid environment and this particular compound in cells. And so we did an experiment where we swapped the HVRS, the hyper variable regions of the C terminal region that is involved in the interaction of rasp with the plasma membrane. So we took the HRASHVR and replaced that on to the K Ras 4 BHVR and vice versa. So now we have AK Ras 4 BG domain with an H Ras hyper variable region and A H Ras G domain with AK Ras 4B hyper variable region. And to our our astonishment and kind of delight, it turned out that the activity followed the HVR. So in the in the case of RAF 1K Ras 4B and light green you see the dose response and then you compare that to the H Ras. Now with the 4B HVR, it has the same kind of dose response activity and the HRSHBR is not active. So this tells us that there's something happening with this compounds interaction with the the plasma membrane and Ras and it's HBR and we don't fully understand that yet. Something we're actively trying to understand further this is just to show that this compound actually showed activity in cells. It's not a very potent compound. It's micromolar in terms of its activity. But just looking at these pancreatic tumor cell lines, you can see that phospho ERC is inhibited at higher concentrations, which is close to the IC50 that we see in cells, little bit higher than what we see in cells. Cause in cells it's 15 to 20 micromolar and here we're seeing the inhibition at 40 micromolar in these panc one cells and these ASPC one cells. We also see inhibition of phospho AKT downstream. And then in this this particular pancreatic cancer cell line, this is a wild type brass, it's rare pancreatic cancers, it has a mutation in wrath. And so in this particular cell line we don't see inhibition of MAP kinase. We continue to see some inhibition of of phospho AKT. So at least in in in these cell based assays, the activity in the bread assay is carried over into activity in these cancer cell lines in downstream signaling. So as I said, you know these HV Rs are distinct between the different isoforms of of Ras. So just to highlight a couple of features and K Ras 4B, there is a sequence of lysines that are right next to each other and these carry of a positive charge on them. And so it's been shown in numerous studies that these positive charges interact with negatively charged lipids in the plasma membrane including phosphatidylserine. A trash does not have this lysine tract. It has an addition to the C terminal palmitilation rather farnesylation. There is also some palmitilation sites in Atras and Nras. So it has additional lipid tail on it and and then none of these positively charged residues. So it's believed and that it's been shown in a few studies that these these differences kind of give these proteins a different ZIP code in the plasma membrane. So plasma membrane has these different domains where you have enrichment of particular kinds of lipids, cholesterol that are more that are more hydrophobic and then there are these negatively charged lipids as well. And so in order to kind of interrogate this further, we've been working on actually studying these interactions with lipids in live cells. And there are a number of fluorescently tagged lipids that we can add directly to cells. These are just showing a few of them and they localize in ways that you would expect. So for instance, phosphatidylcholine and single myelin are both mostly outer leaflet plasma membrane lipids. And you can see that they label the membrane very specifically, whereas cholesterol and phosphatidylstyrene are not just in the plasma membrane, they're in other parts of the cell as well. And you can see that their localization is is more complex. And we've taken advantage of this to develop a number of of protein, protein interaction assays. I'll talk to you a little bit about a nano bread assay we've developed. But first, I just want to show you that we have developed fluorescence lifetime imaging experiments. We have a a detector and a confocal microscope that allows us to do this. We can measure the lifetime of the donor. We don't need to look at the fluorescence of the acceptor because the lifetime of the donor is very sensitive to whether or not it's interacting in FRET. So if it is fretting the lifetime will decrease of the donor and, and, and so you can measure this very quantitatively in cells and from images. And so we've developed some assays to actually look at the lifetime of different fluorescent lipid, the fluorescently tagged lipids in combination with with Ras. And this is just showing the lifetime of the different isoforms of wrath wrasse in cells. This showing that there are differences which we expect because the local environment of each of these Ras isoforms is different in the cell based on their HVR. So we see these differences and then we can take this further and add in our fluorescently tagged lipids so they're tagged on the acyl chains and and we can add them directly to live cells and then use cell lines that we have that express GFP K Ras. 4B. We've looked at this also in combination with mutant forms of Ras and we can do the same sort of titration that we were doing in the bread assay and see a signal saturation as we increase the concentrations of lipid. And what you see here is as you might expect, because KRS 4B has these positively charged license in the hyper variable region that it interacts most with PS in cells. So the saturation curve goes higher, the bread Max is higher than when you compare on some of these other lipids. So this gives us confidence that we're actually able to study the interaction of rasp with lipids in in live cells. And we've gone further in this to develop a Brett assay because a Brett assay is more amenable to experiments in the multi well plate. We can do screening. So we've thought about maybe developing a, a nano Brett assay for interrogating brass and and PS or phosphatidyl serine interactions, taking advantage of this specific domain that comes from a, a protein that interacts with phosphatidyl serine. So it's the Lac C2 domain, well characterized in the literature. This is a structure showing the particular residues that interact with the phosphatidyl serine. And this is a confocal micrograph that shows that these this lack C2 tag with M Cherry gets to the plasma membrane where we would expect it to and that it interacts very specifically with K Ras G12D as measured by Flynn. And so we've gone on to do the same sort of optimization experiments in nano bread combining K Ras and we actually interrogated all these other Ras molecules with flax C2 as a nanobred assay. And we can see that we get a nice signal with K Ras and the black C2 domain and we get very little interaction or much less with H Ras as you might expect. And then we also have a really great control where we have mutated the farnesylation site in K Ras to a serine and so it no longer gets pumped. Are insulated and doesn't interact with the membrane and so you see almost zero or actually 0 interaction of this probe with the Lac C2 biosensor. So we've tested a small library of compounds. This is just, this is very preliminary work, but we're just optimizing this for, you know, as a possible high throughput screening kind of assay and, and, or also as a secondary assay to study molecules that have activity against Ras. And you can see that there's a variety of activities. And, and so we're developing this further. So to kind of summarize, as I said at the outset, it is very important to evaluate drugs, potential drugs or tool compounds in a environment that is as close to the tumor Physiology as possible. So cell lines have all kinds of limitations. I'm not trying to claim that they're like a patient, but they're they are very important for any sort of evaluation of compounds. They preserve the biochemistry and the interactions that are very dynamic of a cell, many of them non equilibrium as opposed to biochemical assays which are often at equilibrium. So it's very important to evaluate compounds and we've heard already the target engagement assay, how valuable that is. FRET based technologies are really good because you can look at these interactions and the perturbations in a, in a, in a complex environment of the cell and they're quantitative. But there are some limitations. These assays depend on tags. So Halo and Nano Luke, they can produce artifacts. We try to avoid those. We try to test to make sure that our our tagged proteins are still active, but there are artifacts that we can't ignore and we also have to overexpress proteins for the screening campaign. Although I know that the next talk is going to talk about some endogenous expression levels. So that's a really great possibility that we would like to explore also. And then one other limitation is that sometimes, well, not sometimes these reagents can be pretty expensive. And so to run a, a very large campaign, you know, if you're in a smaller lab or you know, it might be hard. So you're going to, you know, need to collaborate with people who, who have some resources. So that's my presentation. I'd like to acknowledge 2 primary people. John Columbus, who has developed numerous nano bread assays here in the lab, in addition to all of the work, including the screening campaign that I showed you here. And then Pedro Andrade Bonilla, who's doing the flame experiments in cells with fluorescently tag lipids. And then a whole bunch of other people across the Ras initiative. I'm just showing a few of the people here who developed cell lines and also the the constructs that we use, we make those all in house. So Vanessa and Dom design and develop those. So a lot of collaborative interactions here that make this work possible. So thank you very much and if you have questions, I'm happy to answer them now or later. Great. Thanks so much, Tommy. We are doing pretty good on time. So we could take a couple questions now. There was one that came into the chat from Madiha, kind of a technical question. Just wondering if you could elaborate a little bit more on how you calculate the signal to background ratio in your nano bread experiments. So we take our negative control, which is the R89L mutated construct and we compare it to the interaction with the with the positive control, which depending on the assay format is probably going to be a Brett 80. So we choose from that saturation curve and 80% of Max, the Brett Max because we want to have some headroom above to go to a higher Brett signal. And then we also want to have plenty of room below to inhibit the signal. And so we just take those values and we calculate the ratio. Great, thank you. And then the, the Z prime calculation is more complex, but if you just Google that, you can see the way that that's done it. It's basically looking at the mean values and standard deviations from your positive and negative controls from the plate, from the assay plate. So each one of those assay plates, we have positive and negative controls in the exact same position across all of the different plates that we screen. And we take those ratios from each plate. And if it doesn't perform, we we don't use any of the data from that plate. We rescreen the compounds so that that that Z prime factor is very important for evaluating the performance of the assay. Great. Thanks so much for giving a little bit more detail on that. So it looks like that's all the questions we got in the chat so far. So thanks again. And yeah, everyone, please feel free to keep dropping them in and we'll have some time at the end to further discuss. So now kind of wrapping before we wrap up with our last presenter, we have one more poll before we talked about opportunity and now we're wondering what your thoughts are on challenges. We know Ras is one of these notoriously very challenging signaling pathways. So what do you, what do you perceive as the most significant challenge in developing therapeutics against that target, the Ras pathway? So the high affinity of Ras for GTPGDP making it difficult to disrupt its active and inactive states. Oops, I didn't mean to advance, but I did. So there we go. Lack of suitable binding pockets for RASP proteins for that small molecule engagement. We know that's been a real challenge historically, the complexity of RASP isoforms and their redundant functions. Tommy talked a bit about that. Difficulties in effectively targeting Rasp's interactions with the plasma membrane. Also relevant to what Tommy just presented. Or all those downstream compensatory mechanisms and feedback loops. So out of those, what do you see as really the biggest, most significant challenge? All of them definitely challenges, but you know, interested to hear your thoughts on which one is really most significant as you think about this area. OK, Looks like we've got a good amount of people that have submitted their thoughts and quite a bit of consensus around the complexity of RAF isoforms and their redundant function not as true as well as the compensatory mechanisms and feedback loops. So very complex pathways biologically. And I think this does, you know, definitely contribute to the challenge that this field represents. So thanks for sharing your feedback. Now, I am happy to introduce our last presenter, Dr. Marie Schwinn. Marie joined Permega in 2010 and currently leads the Endogenous Biology team in the Advanced Technologies group. Here at Permega, she focuses on designing genome editing tools and workflows for creating engineered cell models to study native biology. She has also contributed to developing luminescence and fluorescence reporters for analyzing protein interactions, abundance, localization, and post translational modification. Marie earned her PhD in Biochemistry from the University of Wisconsin, Madison and a bachelor's degree in chemistry from Drake University. So welcome, Marie. Thank you for the introduction. My name is Marie Schwinn and today I'll be talking about our efforts with developing protein protein interaction assays that are fully endogenous that allow us to study complex biological systems. When I talk about endogenous biology, I like to start with this image by David Goodsell. What I like about this image is that shows how complex in compact and crowded the cellular environment is. So this particular image shows VEGF signaling, but it really could apply to any signaling pathway within the cell. In order for the cell to bring about a certain phenotype, this complex array of proteins and molecules has to connect and interact in a very, very specific way. In all of this, proteins are definitely the drivers and there's different aspects of proteins that come into play in determining what the phenotypic outcome of a cell is. So those factors include protein abundance, modifications, localization, confirmation activity, and important for the talk today, protein interactions. Scientists really struggle with understanding complex biology, mainly because we don't have the tools to predict how protein dynamics and drug actions are going to affect the cell biology. So at Promega we have developed a couple technologies that have been pretty important in studying proteins in the cell. The first technology is the Halo tag. So I'm going to quickly describe this before I go into protein interactions. The Halo tag itself is a 34 kill Delton protein. It's an engineered dehalogenase. It interacts with what we call the Halo tag ligands. And the Halo tag ligand is really 2 pieces. So there's a linker. This linker interacts with the Halo tag protein irreversibly. So once it's found, it's found. And then there's a functional group attached to the ligand. And it's this functional group that is really important for both the versatility and modularity of the Halo tag protein. So what's important about this functional group is that you can put really anything you want on this ligand. So that functional group, it could be a fluorescent dye, it could be a surface like a resin, or it could be a reactive ligand. So you could be creative and put anything you want on that ligand. Because of this versatility and modularity of the ligand, there are a huge array of applications for the Halotax system. It's pretty popular for cell imaging, including super resolution microscopy, but it also has been used in targeted degradation, protein purification, proximity labeling and complex isolation. And then central for the talk today, it's also involved in protein interactions. So the second technology that I'm going to describe today is the nano luck technology. So there's also two parts to this technology that are quite important. The 1st is we have the nano luck enzyme. So this is a 19 kilo delton luciferase. It was engineered for optimal performance in cells. In that end, luck enzyme uses a fermicine substrate to create a very bright stable glow type signal within a cell. I think 2 of the characteristics of this nano LEC technology that make it really vital for endogenous studies are that it is small and it is bright. So what I mean by small, it's only 19 kilo deltons. So if you think about Firefly or ranilla luciferases, 19K deltons is about 1/2 and 1/3 of the size of those two proteins. When you compare brightness of nano luck with the other enzymes too, it is a hundredfold brighter than either F luck or R luck. So again, brightness and size make it really valuable for studying endogenous proteins. There's also a wide range of applications for the nano luck enzyme and those range from measuring gene expression, viral tracking, target engagement, degradation as well as imaging applications. And for this talk today, it's important for our protein interaction assays. OK, so how do we use Halo tag and analog for protein, protein interaction assays? So first let's talk about nano Brett. So nano Brett is a luminescence resonance energy transfer assay. So that's a mouthful. So we abbreviate it as Brett. So what this is is you have nano luck and it's extremely bright, has a lot of energy and it's able to excite a ligand that has a fluorescent dye attached to it. And so that creates a Brett signal. When you're talking about how to use this as a protein, protein interaction assay, what you do is you fuse the Halo tag to one protein and you fuse the nano lock to another protein. And then you treat the cells with firmazine and your fluorescent ligand. And what happens is when the proteins of interest interact, the nano lock is in proximity to that Halo tag and it excites the Halo tag ligand and gives a signal. So the other technology that we have that measures PPI is our nano bit system. So this is a complementation based system. So what we did is we took our nano lock and we divided it and we divided it into two pieces such that one is a peptide that is one point 3K daltons and the other component is a polypeptide that is 18K daltons. We evolved each of those pieces for brightness and affinity. So the affinity of the small bit for the large bit is extremely low. It's in the range of 100 to 150 micromolar. So in other words, those two components are not going to come together on their own. So in a protein protein interaction essay, what we do is refuse the small bit to one protein, large bit to another protein, and when the proteins of interest come together, the small bit and the large bit are able to refold into an active enzyme. All right. At Promega, we have been integrating our reporters into the host genomes of cells using CRISPR. So we use a very basic workflow for integrating the tag sequences into the genome. We published this workflow in this paper that I cited here. What we do is we take a purified cast 9 and we combine it with a guide RNA and we form an RNP complex and then we electrocreate that RNP complex into the cell in the presence of donor template. So for our smaller tags such as high bid or small bid, we use a single stranded oligo. For our larger tags like Halo tag, Nanoluck, large bid, we use a plasmid donor. So once all those components are in the cell, the cast 9 cuts the DNA and then the cell repairs itself to integrate those tags into the genome. We then go about and we isolate clones so that all of our cells contain protein fusion. We have worked on improvements to this workflow. And what we have found is that you can add 2 inhibitors to your reaction when you're doing the electroporation, and they greatly improve knock and efficiency. So those two inhibitors are a DNAPK inhibitor as well as a whole Theta inhibitor, and those regulate DNA repair. By adding those, we see significant increase in knock and efficiency as well as an increase in homozygosity. At the same time, we see a decrease in unwanted mutations because we have now improved the knock in efficiency and decreased mutations during our CRISPR. We're much more efficient at being able to generate double knock insurance that are perfect for endogenous PPI assays. So here's what we're going to do. We're going to take our nano breads and we're going to take our nanobic components and we're going to knock them into the cell to generate double knock in cell lines. So we've done this two different ways. We've done it sequentially so that you knock in one tag, purify the cells, knock in your second tag, and we've also done it at the same time. And that is you knock in both of your tags at the same time and either works quite well. Why would you go through this process? We know that these technologies work very well in overexpressed formats. Well, there's several reasons that a researcher might want to do that. The main one for me right now is you don't have to do transfections. So what that means is you have assay to assay consistency. Your cells are always going to have that reporter sequence expressed, and that expression of that reporter fusion is always under control of endogenous promoters, not an artificial promoter. Further, that promoter is being regulated by cellular signals. There's a couple things you can also do with these endogenous cell lines that you can't do with overexpression. For starters, you can do extended time courses. So you can do a course, that time course that might be 4872 hours long, and you're always going to have your fusion present. Another interesting application for these endogenous lines is that you can now make 3D models. So normally the 3D models take several days to make and then you have to worry about the expression of your your fusion. But in this case you will always have your fusion presence. So this brings me to the main topic of today and that is the development of endogenous PPI assays specifically for Ras RAF interactions. So we initially selected Ras RAF as a model system to demonstrate proof of concept for our endogenous PPI assays. Because the Ras RAF system is super complicated, we know how important that these two proteins are in cell signaling. They control proliferation, growth, differentiation of the cells. We also know that these proteins as well as other proteins along that pathway are heavily mutated and implicated in cancers. And also we know that this pathway is extremely difficult to drug. There are many reasons why these proteins are considered undruggable. They range from maybe the binding pocket isn't suitable for a therapeutic, there may be inhibition going on, there may be compensatory mechanisms that are all preventing drug action. One thing that I'll talk about today that has played a role in the inability to develop effective therapeutics is the RAF paradox. So if you have a cell that perhaps has a mutated RAF protein and you come in with an inhibitor and your K Ras is wild type, the inhibitor does what it should, you get a decrease in signaling. However, if you have K Ras mutants, those same inhibitors are actually going to increase the signaling. So what we wanted to do is develop an assay where we could monitor RAF Ras interaction in response to inhibitors to see do we see that paradoxical interaction that some drugs can cause. So we created two different endogenous cell lines. We created a nanobret line that has nano luck fuse to C ref inhaler tag fuse to K Ras. And we created a nano bit cell line that has large bit fuse to C RAF and small bit fuse to K Ras. We did this in an HCT 116 cell background. So it's important to know about this cell background is that it has AK Ras that is well type on one allele and mutant on another allele. So what that means is we would likely be able to see paradoxical interactions in response to RAF inhibitors. Let's first start with the nano Brett PPI assay and let's look at how well that performs in cells. So for starters, we just took the cells and we plated them in a 96 well plate. We treated these cells with inhibitor and then we wanted to see can you see an increase in signal. And sure enough for the RAF inhibitors that we tested, you can see a very nice dose dependent increase in interaction. You can see potency as well as efficacy differences. And importantly here you can see that our control drug, which is PLX 8394 did not induce interactions of C RAF and K RAF. And that's important because this drug is a newer class that is shown to decrease signaling but not increase interaction of Ras RAF. And that is exactly what it did here. So you can also take these same cells and you can do kinetic assays. So in this particular example, we treated the cells and we monitored in real time the association of the C RAF and the K Ras. And you can see differences in association rates caused by different drugs. We also monitored the dissociation of the two proteins. So what you can see here is that there are clear differences in the dissociation rates of the two proteins in response to different drugs. One of the proteins LY 300-9120 shows almost no dissociation of the raft Ras, and that is great because that compound is actually shown to have an extremely slow off rate and that's exactly what we see here. So in addition to doing simple played assays, we can also image ourselves. So this is an image acquired by an LV200 luminescence microscope. The top panel is N luck C RAF. So this is our donor signal and then the bottom signal would be the Halo tag K res acceptor signal. In what you can see here. If you look at the top panel, as you can see a very uniform expression amongst all the cells of the nano luck, you can see that when you add inhibitor, you see now an increase in the acceptor signal, which is representing the induced interaction that is incurring. What's great about this is you can tell from the imaging that the interaction is occurring correctly at the membrane. So another powerful reason to use an endogenous protein interaction assay is that you can start doing 3D modeling. We took our cells and we simply plated them in ultra low attachment plates. We incubated the cells for four days and during that time period the cells formed spheroids. We were then able to assay the protein interaction in those spheroids. So this graph here represents the spheroid Brett ratio in the presence and absence of drug. Each of the dots represents the signal from 1 spheroid, and you can see here that the LY compound induces an interaction of eight folds over untreated cells, and that is exactly the same folds observed when the cells were in monolayers. So further, you can do imaging of your spheroids. This is a little more challenging because you're now trying to image a 3D spheroid as a pair as opposed to a monolayer, but you can still see the induction of signal in response to drug with the imaging. All right, so let's switch to nano bit and look how the nano bit assays perform in measuring the C RAFK Ras interactions. I think what you're going to see here is that it's very similar to the nano Brett. So this is the graph of the dose response of the interaction occurring once the cells have been treated with various RAF inhibitors. You can see the various compounds induced the paradoxical interaction, you can see differences in potency and efficacy of the drugs, and you can also see that our paradox Buster drug does not induce interaction, which is exactly what we are hoping for. So once again you can look at kinetic association. You can see differences in association rates with different compounds. And then in our kinetic dissociation experiments, you can see that after washing out the drugs, you have rapid decline of signal for the one drug, the GDC 0879. And you can see that our LY 300-9120 compound really does not come off and signal remains stable. So again, we thought to ourselves, OK, what else can we do with these cell lines? We took a while to make them. They took about two to three months to get a clone. Can we utilize them in a different way? So our cell line, if you can recall, is the HCT 116 cell line. This is heterozygous for the K Ras mutation. So it is wild type on one allele again and then it has the G13 activating mutation on the other allele. We were wondering what would happen if we can somehow change that wild type allele to the G13D mutation. So now you would have a homozygous mutant K rest line. At the same time, we wondered, OK, what would happen if we would get rid of that G13D mutation? So now you would have a homozygous wild type K rest line. And actually it turns out you can do this quite easily with CRISPR. It's just a single base mutation. So the rate of napkin efficiency or editing efficiency of that one nucleotide is very high. So it doesn't take that long at all. So we made our cell lines and then we went back and we said, OK, how do they perform in some of our assays? So this is similar to the graph I showed you on the previous slides in that we treated the cells in the 96 well plates. This is the heterozygous C13 wild type that I had shown you previously. And you can see that you get a nice dose response. When you do the same experiment with the homozygous mutants, you see that there is also a dose response, but the drug is actually much more potent at inducing the interaction of the two proteins. Interestingly, when you get rid of mutations altogether, you see a much greater decline in induction of the two proteins. So if you look at potency of the drugs, the G13 homozygous mutants in which the K Ras is always primed to interact with the RAF, that is the most potent combination, followed by the heterozygous and then lastly the wild type. Up until now, everything I have shown you has been taken advantage of the RAF inhibitors to induce the interaction. We were wondering what would happen if now we tried to induce the interaction from the native pathway of the cell. So that would be what would happen if we would take a growth factor like EGF and added that to the cells. Would we see an increase in the RAF Ras interaction? So this is the heterozygous cell line that we have been working with throughout most of this talk. You can see there's a mild increase in signal with EGF. When you get rid of that wild type allele and suddenly all the cells are expressing mutant K Ras, you see almost no induction of the interaction. Now when you get rid of the mutant all together, what you see is that you have a huge interaction increase. And so the potency is essentially switched from what it was before. And in this case, your wild type cells will have the biggest response to the native pathway stimulator EGF, whereas the mutant cells have the least stimulation response. All right, so I started out the talk today and I said cells are very complex. They're crowded environment, they're 3D, there's a lot going on. And we as scientists are really lacking in tools that help us predict how proteins are working in those cells. And further, we have no good tools to figure out how the drugs are acting on those proteins. So we took our Ras Ras model, which is again very complex and people are having trouble drugging it. We thought this would be the perfect test case to try out our endogenous protein interaction approach. So we took our two reporters that we have the Nano Brett and the nano bid system. We utilized CRISPR to knock those into the cells and create cells that are expressing both the nano Brett and Nano bit reporters used to wrath and wrath. What we were able to do with those cell lines was measure potency and efficacy. We were able to monitor kinetics and we were able to even monitor localization of the interaction. What's really significant that you cannot do with overexpression at all is that we were able to create spheroids to monitor the protein interaction in a more complex 3 dimensional environment. We're also able to generate Isagenix cell lines so that we could monitor the effect of different mutations on the protein interaction. So that concludes my talk today. Thank you for your attention and I will answer any questions that you have. OK, great. Thanks so much Marie. Great, great presentation and really appreciate all the work you do to keep advancing, advancing the sub models that can be used. So now we are going to transition to our live Q&A session. So thanks to everybody who submitted questions so far and feel free to continue to drop questions in as as we continue to discuss here. So I will start things out just with a couple questions that have come in already. So the first one is for Arvin. So this is from Dalton. The question is you touched on this near the end of your talk. Do you expect a combo partner of Trebetnib like the Brethenib to change the off rate of KSR? Can you hear me? Yeah, can hear you. OK. Yeah. So we've done those type of experiments. We've looked at different combinations of RAF inhibitors and MEC inhibitors. We haven't looked so much at the off phrase because those are a little bit more complicated to do, especially you can imagine a lot of permutations between the different paralogs, different forms of brass and then if you start adding different combinations of inhibitors. But we have done combinations looking at like steady state changes. And yeah, Dibrafnib, Trematnib specifically, we haven't looked at the off rate, but we've looked at affinity on various targets. But it I would say the answer actually really depends on which specific complex you're looking at. So in different complexes, we might shift seemingly to higher potency to other ones to lower potency. But at a high level, I guess if the question is can you like use those tools to look at combinations and is that an interesting direction? I think yes, definitely. OK, great. Thank you. And kind of a follow on question to that, Melania asks in which cell lines did you test, did you work with the tremat trematinib? What was the cell line background you were working in? Everything we've done is in 293 cells. We've played around a little bit with overexpressing in tumor cell lines. We haven't done anything as beautiful or as elegant as, but Marie just presented on endogenous. I think that would be very, very interesting. But everything we did was in 293. Great. Thanks. Yeah. And I think, you know, kind of thinking about like the different model systems that you all have used, you know, like maybe Marie, do you have any comments on sort of what you've learned here as you've gone into these more endogenous systems or these more complex models? You know, is there any like real key learnings that have popped out as you've kind of started to move in that direction? Yeah, it's been really interesting. I think one of the biggest take homes for me has been there's a lot of preconceived ideas about tags and tagging proteins and the fact of the tags and the proteins. And some people think, well, nanobits going to be a lot better because you're putting a small tag on your protein. But actually found in many cases both nano Brett and Nanobit work really well. The large tag actually doesn't interfere that much. The other learning I could say is that you're working endogenously, right. So you're going to take a loss in signal versus when we're used to overexpression and you have massive levels of protein there. So in some cases you are working almost at the limited detector, the limit of detection, the lower limit. And so to do a 384 larger screen might be challenging with some of those endogenous proteins if they're at the low levels. So it's great for things that are relatively abundant, medium, you know, medium, medium, low, but for some of those more rare proteins it becomes very challenging. Well. Thanks. Any other comments on on that area? Well, I just to kind of add to that, I just have noticed from some papers that I've looked at that the endogenous levels of these different proteins, like let's just take grass for instance, varies across cell lines and you know, the relative proportions of 4A to 4B can be different in one cell line versus another. So I think it's a good idea to kind of see, I mean to actually look at what the impact of those different expression copy number effects are in, in different cell line backgrounds. Yeah, I, I think it, it, it's kind of an important thing to to do. And so I'm glad that you guys have taken that on 'cause. If I had a one more comment I think it'd be in the biggest benefit. Those things split down the risk price we can study endogenous protein and we the very small the like a not small bit or high bit that is only 11. I mean assays respect the least interference with those small tech. So it makes it easier engineer our cell, especially when we work with our very precious stem cell models. You know, it's crucial we have we have intact biological effect. So I think that's the most, the biggest benefit for this technology. Yeah. And just one more thing to add about the size. If you are going to engineer cells, that small tag is actually a lot easier to get in than trying to get in a larger tag. Technologies are advancing, the methods are advancing to make those larger tags easy to integrate, but small tags really fast, low off target integration and you can buy everything and have it within two weeks. Right, Madea, Madea just submitted a question with respect to the 3D model. So do you see the signal intensity increase in the 3D model based on like all the cells are structured? Right. So this is something we're still trying to figure out. Obviously right now we're limited with the equipment that we have to analyze these spheroids. So we're just taking an overall all the luminescence that's in that. Well, we don't know like if you start even cutting into that spheroid, how deep that signal is, how deep our reagents penetrate. So that's a work in progress and I I don't have much more information beyond that. Thanks. Yeah, definitely a new a new area, so much to learn. And yeah, I think like you said, all the combination of instrumentations with the the models, you know, they really have to grow together. Just one more question about cell backgrounds on this is virtual the questions from Katrina, did you were the, was the data that you showed was that assay run in cardiomyocytes? And as a follow on how long did that type of screen take from sort of beginning of your project and of the project inception as you go through the the validation steps to the end? So kind of a two-part question. Yeah. In the beginning, we tried this lunar technology with the internal project, which is with the cardiomyocyte. However, I couldn't reveal any information of ongoing internal project. That's why with the collaboration of Proneca, we picked up lunar essay to evaluate in there. For those evaluation we used the MCF and MCF stem cells in that sense we didn't have engineer those a cell. We switched it to the sandwich method. We have to add more antibody but in terms sensitivity it we didn't lose more sensitivity compare engineers aligned in with our on my side. Thanks. Thanks, Joe. So there was a question that came in about just tag location. How does the tag location re in I guess influence the assay? You know, I think with like Tommy said, sometimes you don't have a lot of choice, right? You have to tag a certain location to Other times you do have choice of where you put the tag. And any comments on from any of you? As you've all used these tagging technologies, you know how how what to consider there and what you might want to look out for. I can start this one out. Tag position does matter in a lot of cases. As Tommy pointed out, sometimes there's just a biological reason why 1 tag can't be put on one terminus versus the other. There's also modifications, you know, localization, things that A tag could interfere with. In other cases, it becomes almost a geometry issue. So for nano bits, obviously those two components need to be in proximity so they can refold into an active luciferase. If they're too far apart, they may not efficiently come together. You may see a lower signal or decline in kinetics. And then in Brett you have that's the energy transfer. Obviously you want them to be relatively close in proximity. So orientation does matter. Thank you. Any other comments? Yeah. Go ahead. I was just going to add, yeah, I think the orientation matters. And as Tommy brought up too in his talk, the specificity of signal is really important to consider. I like using, you know, of course, testing empirically multiple locations of tags and seeing where you get maximal signal, but then using mutations that are specific to certain interfaces. The only thing I would say that is a little bit of a challenge is sometimes our assumption for how a mutation might work based on literature might not pan out always in the assays. So sometimes I'm not sure if that's the actual assay telling us something incorrect or if the literature and the dogma in the field is actually incorrect. So that's sort of been, I think as people start using these technologies, one of the things been interesting is like we start to learn new things about the basic, even biochemical interactions. Like we've used the RBD mutations that Tommy has mentioned, and those are good loss of function alleles. But actually we found scenarios where they're not loss of function depending on which RAF complex you're looking at. So yeah, I think tag matters, controls matter, and then fairly accepting literature really matters, and then trying to figure out if you have something new or not. I'll be curious. Oh, go ahead. I have a similar answer. Yes, target the the direction is matter For cardiomyocyte we test force and terminus and the C terminus. We had the greater robust window at the C terminus. Again, it's based on how your target board and the interface of other protein. Yeah, it does matter. Did you have another comment, Arvind? Oh my other comment actually was more of a question for the other panelists. I'm just curious about this is how many times have they learnt or been surprised from a result from either a target engagement or bread based assay in, I don't know, looking at an interaction or or a specific mutation? And what are their follow-ups to try and figure out if, you know, like I said, is it something biologically meaningful or is it a maybe some sort of technical artifact good. Question. Well, I sort of along those lines, Arvind is you know, the the Brett or Nano Brit assays, it's a binary interaction. You can set it up, you know, you could combine a small bit, large bit and then a Halo to make a ternary interaction. But you're, you're really only looking at, you know, 1 to one interactions and there's all of these other things going on at the same time, other interactions that could be competing for the same binding site. You know, so like when you think of the effectors of Ras, and one thing I've really wondered about a lot is how do these different assemblies, how do they, these complexes assemble? Is it, are they, are they waiting around for their turn? Or are there different complexes that are forming simultaneously? And that has to do with localization of the different factors. You know, there's a lot of complexity that it's hard to measure with these binary interactions. So, you know, not that they're not good. I'm they're great, but it's just sometimes I, I wonder about that, what we're missing. Yeah, no, I 100% agree. And and I think on top of that beyond just like what we're looking at either binary or you can look at a lot of different combinations of binary interactions or journey interactions. There's also the time scale of the effect, which I didn't think about it so deeply before. I just took it for granted about, you know this typical ways that most assays are run. But now I'm getting a better appreciation for that. I showed a little bit of data where even in with reversible binders like the profile of a compound can change quite substantially over very long periods of time. Of course, there's a lot of things that can happen in a cell, but I wonder if, you know, that is a reflection of what you've mentioned is like how are the complexes actually assembled? What are the time scale on which a lot of these interactions are modulated in a meaningful way? And then like how are compounds like affecting all those equilibrium that and there's so many different things happening all the same time. So like kind of with all that complexity, are you finding that some of these like AI modeling tools are helping you to better understand what this might look like or you know how they can be addressed or are they still not useful to at that level? Well, I can comment a little bit. We have a, a collaboration with the Department of Energy, Lawrence Livermore lab to do these very, very large simulations. And I think we are learning things, but it's, you know, to, to set these simulations up and to run them and to learn from them requires I think, you know, the team of people. There are a lot of computational people, machine learning people for the AI sort of component. And it's, it's, it's way more than than what ChatGPT leads you to believe. You know, like I, I just, so I, I worry a little bit that that some of these, these technologies are, they're new and they're exciting, but they haven't really fully been developed yet and, and road tested to really be able to say, OK, is this, is this the process that is happening in the cell? But you know, I'm all in favor of them. I'm just saying it. I don't think there's an app for it yet. Very good. Any other comments on that? In addition to the Doctor Dara's question about what could be surprising. So in the sense of besides the the compound is a mechanism of action either focus on the content on the essay interference. So the bodies have internal data comparing essay interference compound of focusing on activate or inhibitor on those luciferase itself. Surprising there was a very high hit rate for inhibition for Firefly, however very low inhibitor and from screening our full deck in the nobodies with nano receipt rates. However, when I screwed the 3000 thousand compound with this nano receipt rates, we found a decent amount of inhibitor and also surprisingly some activator too which is our non target related activity. So I think recommend to pay attention what you got and you need a good orthogonal essay to filter in and out your the compound of interest. Yeah, very good comments. Yeah, I can second that, that a lot of the things that look like they were inducing stronger interactions between Rath and Ras turned out to be false positives. They were, they were interacting with the assay. So having a counter screen, you know, to to look for those luciferase and other, you know, Brett inhibitor, whatever the mechanism is important. Yeah, and maybe controls too, I guess, right? Having the right controls like to your point Tommy, and setting your plates upright so you know that you think they're working how you expect and when you can trust the data and when you need to to scrap it because something didn't work quite right. OK, So kind of finishing up here, maybe just a couple of sort of future looking questions. So you know, we've talked a lot about kind of a lot of the research that that you've worked on. What do you think about obstacles in translating the laboratory findings on on the RAFT pathway and compounds that impact the REST pathway into clinical applications? So, you know, how, what can the research community do to, you know, address these challenges and help to bring therapies to patients Sooners, do you have any, any thoughts on that? It's a very, that's a very big picture question. That is a big picture question. Yeah. I mean, I would just say that there's still a lot we don't know. That was sort of what I was looking to about the surprises. So like a lot of inhibitors have taught us a lot of surprises about the pathway. The paradoxical effect of RAF inhibitors that RAF inhibitors switch to off state inhibitors can still be active on K Ras, even though the original dogma was that the mutations would be locked in the GTP bound state. So we can see that they're continuously cycling. There's cross talk in interactions, there's like gaps that shouldn't work on K Ras that apparently work in the context of the mutant. So I think there's just a lot of black box on the biology in the mechanism and we still are trying to figure out all the ways to like understand that complexity. The tools that we talked about today I think are very, very helpful. They fill in a lot of gaps, but I still think there's probably a lot we don't know and just bringing out ways to like appropriately assay, you know, drug binding on target. It seems like a simple question, but it's always, it's not always so simple. Context really, I think can matter quite a bit. So I, I mean, to me the, the big, the big challenge is just a lot of unknowns on the biology side still, even for something that's as intensively as brass. Yeah, yeah. Any other thoughts there? Just to kind of follow up on that a bit, I'm really a proponent for endogenous assays, but it's all like context dependent. And the cells I'm using have different mutational backgrounds. They have a lot of them are immortalized cancer cell lines. They have multiple alleles. And sometimes you can get a different result working with one cancer cell line versus another cancer cell line. So it becomes really complicated to go from those cells into like a person or an animal because there's so much context in those cells that we don't always appreciate or we don't even understand. So especially with Ras, Ras, I think it's very, very challenging right now to make that that jump. Yeah, definitely. OK. And then closing out, I'm going to do one kind of future direction slide kind of as you look forward and in your crystal balls, what kind of emerging technologies or methods do you believe are going to have the biggest impact on the field of rest biology going forward? Anything that you're super excited about or that you wish was there, that's not there as as a tool. Well, one thing I can say is that the fact that there are these compounds now offers a lot of tools that didn't exist previously for studying the biology kind of getting to what Marie's talking about. And so I, I think that's going to help a lot in the future. And they're, you know, they're more of these coming. I mean, now there's like lots of them coming down the pipeline. And I think another sort of valuable thing that's happening is that there are a number these endogenous expression tools and then the different families of fluorophores and different ways of multiplexing so that you can look at more than one thing at, at once in, in a 3D cellular context, you know, on, on a, on a plate, you know, in microscopy or in other, in other modalities. You know, I think like there's, you know, now mass spec imaging and things like that, we're really starting to understand how these interactions are happening dynamically in, in the cell or in a tumor model. I think that that's an important, yeah. I don't know what other people think, but I think those are those are definitely going to help a lot. Any other thoughts? Closing comments. Maybe just add to what Tommy said, I think those tools are so important because they're like especially the clinical ones, they're basically experiments and then that cycle going back from patient because you'll learn something often times related either to efficacy or resistance. But then going back and cycling it to the lab iterating like that seems like it's getting faster and and more information we get from early clinical trials. I think it's will always continue to be super important. Yeah. All right. Well, very good. Thank you all for spending time with us today, all of our presenters for sharing all this great data and their thoughts, and all of the attendees who joined us today and participated. We really appreciate it. Hope you all got something useful out of it. We do have two more days of our virtual events. So tomorrow we'll be talking about NextGen Therapeutics, the power of induced proximity and target protein degradation. And on Thursday, diving into unlocking the potential of RNA and new paths for targeting a so targeting RNA. So a couple of great topics coming up. I'm happy to have you all join us again. And with that, thank you so much. And there's a couple of contacts here if you want additional. I know there's a few questions we didn't get to that are a little technical. If you still have those, feel free to reach out to the per Megatechserve team. They're always happy to help. So thanks again. Bye. Bye. Thank you. Take care everybody. Bye everyone. _1734241501156

Time: 10am CDT / 8am PST / 4pm London / 5pm Amsterdam-Paris

"Illuminating New Frontiers: Cracking the Undruggable Code" invites you on a groundbreaking journey across the evolving landscape of drug discovery. This three-day virtual event unites trailblazers in science and industry to explore how innovative approaches are reshaping our ability to tackle "undruggable" targets. From the complexities of the RAS pathway and the intricacies of protein-protein interactions to the pioneering fields of targeted protein degradation/induced proximity and RNA targeting, each day illuminates a different frontier of therapeutic potential.

We stand at the cusp of a new era, where challenges become opportunities for breakthroughs that extend beyond conventional boundaries. This event is a testament to the power of collaboration and innovation in unveiling new therapeutic strategies, offering hope for untreatable conditions. Join us as we navigate the cutting-edge of science, inspiring a future where every target is within our reach, and no one is left without options.

Day 1: Rethinking Undruggable Targets: New Insights into RAS and Protein Interactions

In the realm of drug discovery, certain targets have long been deemed "undruggable" due to their complex nature and the technical challenges associated with modulating them therapeutically. Among these, the RAS pathway stands out as an important element in cell signaling implicated in numerous cancers and diseases. Historically deemed undruggable due to its complex biochemistry and lack of suitable binding sites, RAS has recently witnessed a paradigm shift. Recent innovations in molecular biology - coupled with breakthroughs in drug design and technology - have challenged the notion that the RAS pathway is undruggable, leading to the development and market approval of RAS-targeting therapeutics. Similarly, protein-protein interactions (PPI) - key elements of RAS signaling - represent untapped therapeutic potential, offering avenues to intervene in disease processes in ways previously deemed infeasible.

In this session, we will explore the latest research targeting the RAS/MEK/ERK signaling pathway and learn about novel approaches being developed to manipulate and measure PPIs within RAS and other signaling networks. Join the conversation and engage directly with the experts in a live round-table discussion following the presentations.

_1734241501318