Welcome everyone to our day two of our Cracking the Undruggable Code virtual events. We're happy to have you join us here today. Yesterday we had some really excellent presentations and discussions around one of probably historically one of the most notorious and robable targets, the KRAS signaling pathway. And had a lot of really nice discussion about the the complexities of this of the biology of the signaling pathway and a lot of the new assays and tools that can be used to better understand this biology so that we continue can continue to move forward with advancing therapeutics. Today we're going to transition a bit to talk about one of the really exciting areas used to be used to tackling undruggable targets, which is induced proximity and targeted protein degradation. We've got a really great lineup of speakers that are going to provide some some new science and some nice discussion on this topic as well. So before we get started, I just wanted to cover a couple of housekeeping items. So you'll see on your screen a number of different modules and these are all customizable and movable. So feel free to adjust these and move them however however you see fit. There's a number of ways that you can interact with us during this presentation. So you'll see you can, you can raise your hand, you can turn on your camera, and also you can turn on your microphone to ask questions. We would ask that you hold on doing this until we're in the Q&A sessions, but then we'd be happy to have you talk with us in this way. There's also a chat box that you can use to type in questions if you'd like to present questions in that way, and we'll answer them either through in the Q&A or through the chat as well. There's also a few resources that are available to download that you might find useful, so feel free to check those out and there will be a survey at the end. We always love to hear from you and get your feedback so we can continue to make these valuable to you. So please take some time to complete the survey once we're done. So as I said, we do have a really excellent agenda, a lot of great topics and discussions coming, coming up today. But before we get to that good stuff, I wanted to start out with a quick poll really does just help us better understand you as a group and kind of where you are and working with these types of degrader compounds. So select the one that best describes you. Are you new to this field and really just still learning? Do you have some experience but you're really building up that expertise? Or do you consider yourself very experienced in this area and work with the greater compounds quite regularly? So we'll just take a minute or so to let you put in here what how you feel the best describes your current state in the field. I think we've got a good number of respondents, so I'm going to go ahead and take a look at the results. So we got a really, really good distribution here. So we've got some pretty experienced folks. So we'd love to hear from you and hear from your thoughts. But we've also got about half of us are new to the field and still learning. So that's great. And we hope that you know you find this information educational and useful to you as you continue to grow your expertise in this field. So thanks for sharing that information. And now I'm going to introduce our first speaker, Doctor Fleur Ferguson. Fleur earned her PhD in chemistry from the University of Cambridge and completed her postdoctoral research at the Dana Farber Cancer Institute at Harvard Medical School. She is currently an assistant professor in the Department of Chemistry and Biochemistry and the Skagg School of Pharmacy and Pharmaceutical Sciences at the UC San Diego. Her research focuses on developing proximity pharmacology technologies for disease areas where traditional therapies have struggled, such as degenerative diseases. Dr. Ferguson's work has garnered numerous prestigious awards, including the NIH Directors New Innovator Award and the NSF CAREER Award. So welcome, Flora, and thanks for joining us. Thank you so much for the invitation, the opportunity. I'm really excited to present 2 short stories from my lab. We're really trying to expand how we can develop targeted protein degraders and molecular glues against tricky targets. So these are my disclosures before we get started and as many as the title of this symposia suggests. Many of us are interested in the undruggable proteome. And really when we're thinking about molecular glues, many of us are thinking about how we can target proteins via protein, protein interactions, think about disordered proteins and think about transcription factors which have fallen into this category where they have very clear links to disease, but we don't yet know how to drug them. And I think these come in multiple flavours. And I'll talk about two of these flavours today. Pull these from a nice review at the bottom here. And these are the bifunctional molecules such as Protax or toxic protein degraders, as well as molecular glues. And in particular, we're interested in molecular glue degraders for degrading transcription factors and these can be fantastic. You can use your bifunctional molecules to degrade undruggable domains of ligandable proteins or leverage protein degradation to enhance selectivity if you only have a multi targeted ligand. And I'll talk about that in our first story. The second component for molecular glues is really the strength of these is where you have undruggable proteins that lacks more molecule binding sites. And so you want to leverage that second protein to recruit your protein of interest. And one thing that I have been really involved in and excited about and continue to work on is the use of proteomics and chemoproteomics to map out degradable target space and start to learn some of the rules of these new modalities. And so I first started working on this back in the kind of 2020. We had this nice paper from my postdoc where we made a map of degradable kinases. And others such as Eric Fisher and Ben Ebert have done the same for IMID NEO substrates. And there's a whole exciting range of new preprints and papers out on this area in the past few weeks. And so we now have these nice maps. And what this enables is not only for us to take those maps and pick out the low hanging fruit, but it also allows our colleagues in the machine learning and protein modeling areas to develop new machine learning or chemical docking approaches using these large and uniform data sets. And that's really crucial because as a new modality, we tend to have few examples. And so creating these large scale data sets can really advance the field as in general. And something my lab has also done is developed, you know, these types of methods for doing the same thing for non degrading glues. And so we're now creating those maps for non degraded glues. And so I hope I show you in the introduction and these are widely available. I talked about these in the last Promega symposium. And so they've been adopted recently and widely picked up. And so I hope I kind of convince you that these large scale chemical proteomic approaches can really advance the field and help us to, as medicinal chemists, pick tractable targets and figure out what are the druggable proteins and also help work with our colleagues in the molecular modeling and machine learning fields. And so in this talk, I wanted to talk about really 2 efforts from my lab to build on this work. The first one is how do we create or how do we design chemical proteomic workflows and assays that allow us not only to figure out the low hanging fruit, but to better figure out how to optimize those molecules once we have an active compound. Because what you get from these maps are starting points and druggable, but then usually it's up to the chemist to figure out how to make those kind of multi targeted or weak compounds into excellent degraders. And the second one is really leveraging those degradable protein maps to kind of build a potent and selective tool for molecular. And this is a cerebone recruiting glue for the transcription factor ZBTB 11. So this first project was a collaboration between my lab, Eric Fisher's lab and Bill Seller's lab, led by the talented scientists you see on screen. And it was really motivated by the fact that as a field we have done a really great job of understanding the biological variables that contribute to target tractability for Protax. These bifunctional molecules. Some of this from these degradable kind of maps and machine learning approaches I introduced in the introduction as well as protein docking approaches for these proteins which are well modelled in by either the PDB or by alpha fold. And so we know which biological variables can influence our target selection, but what about the chemical variables that influence the potency and the kind of kinetics of our targeted protein degraders? This is something where we have these key chemical variables that we can modulate as chemists. So we don't have too much control over the sequence of our protein, but we can modulate our small molecules. And most of the biology of these targeted protein degraders is driven by this ternary complex. When you have your ternary complex, which is in the right confirmation to allow ubiquitination, it's really the residence time of the two ligands as well as the cooperativity of the interface that are determining how stable that complex is. You need enough stability to promote Poly ubiquitination, but not too much stability that you gum up the UPS machinery and slow down your Catholic rate of degradation. And a huge amount of the field has really been focused on developing these models using either pure mathematics or mathematics plus machine learning to figure out how cooperativity and the dissociation kinetics of the ligand in fact impact the protein degradation outcomes. And so there are many of these papers. I highlighted just a few here. But what you'll see is that that there's this really tight relationship between the K all for the KD and the cooperativity, where if we have a very potent molecule, we probably need weak cooperativity, otherwise we're going to have poor degradation. Whereas if we have a very weak molecule, we need a high amount of cooperativity. And our field has really focused on measuring and determining the impacts of cooperativity, which is achieved by linker modification and modification of the small molecule, but also really depends on the protein, protein surface and interface of the productive complex. And so it's really hard to predict which modifications to make to access the complete range of cooperativity and allow us to survey it in an efficient way when we're doing structure activity relationships of degraders. And for this reason, it's really hard to know when you've tried enough molecules. But there have been relatively few systematic studies of modulating residence time as an optimization strategy. Those which have been performed have focused primarily on BTK using reversible covalent and covalent chemistry, cysteine. And whilst they found that the residents time convincingly impacted the degradation outcomes, they're limited to just one target, one model system. And cysteine is relatively uncommon in binding pockets. And so it may not be an approach you can apply to all of your projects. So what we wanted to do was understand how widely the strategy can be applied to TP programs. And so our objectives were to create a data set and large scale data set to evaluate the scope and magnitude of the effects of modulating the residence time of the target binding ligand, which is something easy to do chemically. To generate mechanistic understandings of the processes that are influenced and to evaluate this as a generalizable approach for accelerating degrader drug optimization in the early stages. What we leveraged here was really nice chemistry from Jack Taunton's lab where they have taken this multi targeted pan kinase inhibitor and appended different functional groups to make it reversible covalent with a fast off rate, reversible covalent with a slow off rate or completely covalent. And so we now have pan kinase binders that access the full range of residence times from reversible to irreversible. We built a library of targeted protein degraders varying the linker length, the regiochemistry and the covalent chemistry and profiled them in cellular target engagement assays for VHL to ensure that they are cell permeable. We then evaluated degradation dependent pharmacology using viability assays where we have a cell line that has a wild type or the VHL knockout E3 ligase and so we can see here a difference in cell viability indicating degradation dependent pharmacology. This allowed us to narrow down to a set of four matched compounds that have the same E3 ligase ligand, E3 ligase attachment chemistry, linker and kinase binder attachment chemistry, varying only at the site that interacts with the catalytic lysine. And we profiled these in global proteomics analysis at two concentrations to give us a kind of idea of what the target scope is of these degraders and how residence time influences the target scope. What we can see, I'm just summarizing all of that data in this Venn diagram here, is that residence time really strongly influences the targets. And you'll see that there is only overlap between compounds with the closest matched residence times and no overlap, for example, between the covalent and reversible analogs. So this to us was really striking and we decided to 1st validate that these effects were real. So here's an example of our validation for CDK 6, where we've taken our degradation results from the proteomics, which are plotted on the left. And we performed a cellular target engagement assay using the Nano Brett system to make sure that our compounds were comparably engaging CDK 6 and this wasn't driven by binary affinity. We also performed the washout assay. So we use this nano bread assay in washout mode to confirm that covalent inhibition or reversible covalent inhibition was occurring on a time scale relevant to degradation, which we can see here that our covalent compound has very little washout. Our reversible covalent compounds have a sort of intermediate washout on the time scale of two hours, whereas our fully reversible compounds are completely washed out and the signal matches the DMSF. We next confirmed degradation by Westonblot, which again confirmed our results from proteomics showing this wasn't just a artifact, but we do indeed see the same trends using this orthogonal approach. And so finally, we wanted to look kind of wide to see if our selectivity profile across the entire kind of a change. And so we use this K 192 NICE Nano Brett Todd engagement assay to do so, where you can see that we have relatively low target engagement, which is typical of toxic protein degraders and we haven't significantly changed the target binding profiles. We next wanted to look at if it was the ternary complex formation that was really limiting our degraders or changing our degraders ability to to degrade their target proteins. To do this, we teamed up with Eric Fisher's lab to do VHLIPMS in the presence and absence of our degraders. And what we can see is that our degraders are able to form ternary complexes and sell lysates with almost the entire kinome. And so all of these compounds are able to recruit way more proteins to VHL then they're actually able to degrade. And I should say this assay is done in the presence of excess VHL and doesn't account for that intracellular competition. However, as predicted by our mathematical models, we see much larger accumulation and much greater occupancy of these ternary complexes with our covalent compound versus with our reversible to create it. So we wanted to see how these changes in ternary complex dynamics and ternary complex stability impacted ubiquitination. And so to do this, we teamed up with Bill Seller's lab to use their east of assay, which measures the VHL dependent biotinilation of differently of different kinases. And we're able to see protein wide that generally we have nice ubiquitination where we have high degradation that's confirming what we expect. But we also see some targets that have low ubiquitination but are not yet degraded, indicating they may be targets that undergo slower degradation. So to put all this together, we took a look at each of these different components to categorize our kinases into different different categories. We have kinases where fully reversible binding is favored and these are exemplified by AK1. And we anticipate this is because AK1 forms a very productive ternary complex and so you need to be able to let go of that kinase to allow fast catalytic turnover. We have a number of kinases such as CDK 6 which are favored by the reversible covalent molecules achieving this kind of Goldilocks zone of enough stability to promote Poly ubiquitination, but also to allow for release of the kinase. And then finally, we have somewhere the covalent, in fact, we have many where the fully covalent molecule gave us our best results. And we think that this is because we need a very high level of ternary complex stability in order to promote ubiquitination, indicating potentially unfavorable interactions there. And with this data set, we're now working with Cherry Chowdhury at J&J to build mathematical models to refine predictive studies. But I think the real conclusion of this work is that, you know, our ligand residence time really strongly influences our degrader activity and selectivity. And so this is a, this allows you to tune the residence time if you're just using this reversible covalent chemistry at lysine really quickly. It's rational. We know how to do it as chemists, we can build it in rationally. And so when you're doing protein degrader studies, this is something you could consider including in your early libraries along with Lincoln variations. We've been working with a number of different labs to apply this successfully. So we do think that it's really helpful in the degradation development process. And so in the last few minutes of the talk, I want to really give a very brief overview of how we're using some of these degrader libraries, in particular molecular glue degrader proteomic data sets to advance towards targets that we think are important but undrugged in cancer. And this project really is motivated by the fact that K Ras drives most of pancreatic cancers. But once pancreatic cancers are established, it has been shown that they can survive upon K Ras ablation. This is an old paper using genetics. Or they can survive by developing resistance to clinical kinase, sorry, clinical K Ras inhibitors and they do so by up regulating oxygen phosphorylation. However, inhibiting OXFOS in the clinic is a little tricky because you get neurotoxicity. So it tends to look great in mouse studies. But once you get into humans, you see this is example from a phase one clinical trial in AML and solid tumors, dose limiting peripheral neuropathy. So we were looking for OXFOS modulated targets that can overcome neurotoxicity because they really serve as a rheostat to up regulate OXFOS in cancer. They're not required in healthy cells. So we came up with this list and one of the targets that popped out was the C2H2 zinc finger ZBTB 11 which contains a nice CXCG motif and in convincing biological studies modulates oxidative phosphorylation upon knockout in mouse cells. Importantly, if we look in IPSC derived post myotic neurons using CRISPR eye, and this is a nice data set from Martin Kamman's lab, we see ZBTB 11 is not essential. And so our hypothesis was that that was really going to be a synthetic lethal interaction between K Ras inhibition and ZBTB 11 depletion. But CBTB 11 depletion would be well tolerated by neurons, healthy neurons that don't have that K Ras mutation. So of course we created a hybrid knock in cell assay, performed focus cerebron binder library screening and follow up medicinal chemistry, which allowed us to come up with a set of active degraders and negative controls that have an N methyl on the glutaramide to prevent cerebrum binding or a closely related analogue that binds cerebron but no longer degrade CVTV 11. You can see that in the dose response on the right. We performed of course all the rescue experiments to show that this is working via Cerablon and via the ubiquitous proteasome system and perform global proteomics to show that we have two off targets. And so this is a moderately selective chemical probe. For CBTB 11, we developed cell lines that were resistant to clinical K Res inhibitors targeting K Res G12C and G12D and validated that those cell lines do indeed have increased oxygen phosphorylation using a Seahorse assay. And I think this is really the killer result, which is that in the parent cell lines, which are of course sensitive to our K Res inhibitor shown here for example, such aggressive, we see very minimal effects of our CBTB 11 degraders alone. However, in our resistance cell lines where you see in red very little impact of K Ras inhibitors because they've developed resistance and very little impact of our K Ras degrade, sorry, our ZBTB 11 degraders alone, we combine the two in blue at the bottom. There we see very strongly cytostatic effects indicative of a synthetic lethal interaction. So we did a variety of omic studies to understand this. What we see, you know, as a summary is that across the proteome, we see that these mitochondrial proteins are increased in the presence of such a rasib and decreased by our ZV11 degraders, indicating we're normalizing the proteome. We can see that in metabolomics where we see reduced flux through the TCA cycle and we can also see that in our Seahorse assays where excitingly we can see reduction in oxygen phosphorylation within 24 hours and a change in the metabolic index of our cells. So that's exciting, but what about that neurotoxicity? If we do neurotoxicity assays and human IPSE derived neurons, we see that our mitochondrial membrane potential and Ross are strongly impacted by that complex 1 inhibitor I showed you in the beginning that failed in clinical trials, but not by our CBTV 11 degraders. If we go out to three days to look at viability again, we see profound impacts on neurite length and viability from Complex 1 inhibition, but not CBTV 11 degradation. We're now working with Yuri Manor and many others to understand why this is. And you can see from these preliminary results using competitive phase imaging that our IOX molecule is causing profound fragmentation of the mitochondrial network, whereas our ZBTB 11 degrader seems to be causing slight elongation of the mitochondrial networks, indicating different mechanisms of action. And so we've developed this toolkit and it's really been enabled by these large chemoproteomic studies to identify the right types of molecules to screen and the right degradable targets. We found that combining K Ras inhibitors with CBTB 11 degradation gives us anti proliferative responses in K Ras resistant K Ras inhibitor resistant P DAC cells, which represents a very important therapeutic kind of unmet need and that we have reduced acute neurotoxicity in our IPSE derived neurons. So with that, I'll wrap up. I'd really like to thank everyone in the lab who did all of the hard work as well as our many collaborators and our funders. Thank you to Permega for the invitation and to everyone here for your attention, and I'm happy to take any questions. Great. Thanks so much for that. Was a fantastic talk. Really exciting work you're doing. We do have time to take a couple of questions now. So if anyone in the audience has a question, feel free to raise your hand or type them in the chat and we'd be happy to happy to take a couple now if if anybody has any. I, I thought it was really interesting. I haven't really thought of before about sort of the synthetic lethality approach with degraders. And you know, how you're using the, the second story you told about, you know, kind of taking that approach. Have have you seen this used elsewhere floor with, you know, kind of identifying a target that would cause a synthetic lethal effect in a in a cancer situation? For molecular glue degraders, I haven't seen that yet, but I think that's possibly because we have still relatively few examples published, although there are probably many out there in different labs and companies of creating, you know, selective degraders for new kind of transcription factors. I think a lot of our efforts have been really on understanding the target space and the mechanism of action, which is super exciting, is now enabling these studies. But I think it's something that we're probably going to see a lot more of as the field advances as some of those studies start getting published. So in terms of synthetic lethality, I haven't, You're right, it's not an incredibly common approach for these molecular glues. But I think that's just because perhaps we're just starting to understand what they're capable of and looking forward to hopefully seeing many more of that kind of work done in the future. Great. Thanks. Yeah, we did have one question come in. Dipika asks. Perhaps she missed it, but could you elaborate on how you tackled selectivity assays and the site where the glue binds? Yeah. So for selectivity, we used global proteomics analysis and for the site where the molecular glue binds, we actually identified the Dagron. There are 12 zinc fingers in ZBTB 11 and I believe seven or eight of them have a CXCG motif. And so we identified the Dagron using a combination of mutagenesis studies and domain deletion studies, which I didn't have a chance to show, but are not preprint. So you can check those out to see how we did that. And then to understand the structural biology, we're using our Rosetta docking of the ternary complex. Right, thank you. But looks like we don't have any more questions right now, but keep thinking of them and we can definitely ask who are some more questions when we do our final Q&A. So thanks again. Who are next? I'm happy to announce our next speaker, speaker, Doctor Jorg Winter. Jorg is a chemical biologist and Principal Investigator at the CEMM, which is the Research Center for Molecular Medicine of the Australian Academy of Sciences. His lab focuses on innovative pharmacologic strategies to probe and disrupt aberrant transcriptional circuits in cancer utilizing high throughput technologies like quantitative proteomics and functional genomics. Doctor Winter. Doctor Winter's current research centers on proximity inducing small molecules that can rewire cellular circuits and advanced targeted protein degradation. His work has been recognized with multiple awards, including the Tetrahedron Young Investigator Award and the EFMC Prize for Young Chemical Biologists in Academia. Welcome, Jorg. Thank you, Amy. So I trust you hear me. Well, let me just see that I can advance my slides. All right, These are just my disclosures, but they are not really relevant for today's presentation. So it's in this some it's always great to go second because then you can trust that whoever was speaking ahead of you can has introduced the fields already sufficiently, which for has done a great job of. But very quickly currently in the field we're mostly interested in these two modalities either hetero bifunctional protects or monovalia molecular degraders that by and large function very similar principle which is inducing proximity between the target of interest and an E3 ubiquitin ligase. What we have gotten interested though over the years is increasing number of reports of inhibitors acting as degraders and there are several examples like that. I think one of the most prominent and most well elucidated one from the Ebert lab on small markets that induce that, that induce the degradation of BCL 6. There are others focusing on the estrogen receptor, but most are really reports on kinase inhibitors that induce kinase degradation. And this is something that caught our interest. And we were particularly interested based on the fact that for most of those cases, whenever people report an inhibitor that is designed for a particular kinase also induces degradation of that kinase. By and large, the mechanism that is described for this is that these inhibitors block the communication and interactions between the kinase and the HSPNID chaperone system, which is typically kind of scanning the kinase surface in order to ensure structural integrity. So we were interested by and large, looking at kinase inhibitors of how often is it really that inhibitors induce degradation? Is this extendable? And are there general mechanistic frameworks that go beyond this, this famous example of chaperone deprivation? So I'm going to tell you two stories over the next couple of minutes, one that focuses on 98 kinases and one that focuses on one, the oxygenase. And I hope in the end I can connect them in a way that it makes sense to you all the work that I'm going to present today is work that in my lab has been conducted by a very talented postdoc net to this calls. And so she basically said that we can or thought that we can approximate kinase inhibitor effects on kinase stability by over expressing a kinase antivirally with a nano luciferase tag. And then after treating the cell line with a particular inhibitor of interest, monitor and nano luciferous levels or bioluminescent levels over time as a means to approximate how strong the degradation of that inhibitor is on that kinase. So we didn't do that for one inhibitor and one kinase. But Natalie generated a saline panel of 98 different kinase nano luciferous fusions alongside 2 controls, which is the GFP nano Luc fusion and the DGFP, which is the destabilized GFP nano Luc fusion. As as as controls that allow us to frame our data because the GFP nanologue fusion is a very stable protein and the DGFP nanologue fusion is a very instable protein. And we then screened this panel of 100 cell lines against 1000 library of 1570 different kind of inhibitors in this time ranging Ases. And by and large, we collected more than 1.5 million data points that we could then dive into for our analysis. Very briefly, we collaborated with in terms of the analysis with Patrick Alloy from the RB in Barcelona. And what we did by and large is other than removing outliers, we looked at the whether or not we found the effect of this compound is much more pronounced on one kinase as opposed to the 97 other kinases. We compare this effect also to several chemical controls that we've put in place. So we don't only have the protein controls GFP and D GFP, but we also have controls where we just block transcription elongation. This is with this inhibitor NDP 2 or when we block translation by cyclohexamide. And So what we want to see is and this is data from the screen that you see now in this in this course we want to see the our compound of interest here AV 4/24/12 outperforming both the cyclohexamide and the NDP two treatment. So with these parameters in place, we have identified more than 160 selective inhibitor induced degradation events and more than 300 events that we refer to as same as selective. So what do I mean by that? You would see a selective example on your right where there's only you. What you see plotted here is the effect of this particular compound on these 98 different kinases. You see that from 97 of them in greater is no effect and you see this one kinase DDR1 being selectively destabilized by that particular inhibitor. An example for a semi selective inhibitor or degrader you would see on on the left side were again the vast majority of all kinases are not affected, but there are three kinases that are reduced in abundance in this time dependent trajectory. What we can do now is to analyze that data on a high level and the first thing that we can do is to test this HSP 90 chaperone deprivation hypothesis. We can do that by binning all the kinases in in different classes, whether or not they are strong interactors, weak interactors, or no interactive of HSP 90. What we do clearly observe is that we do see an enrichment of our hits among those kindnesses that are known to be strong or at least weak interactors, indirectly affirming chaperon deprivation as a relevant mechanism under pin, under pinning this inhibitor induced degradation. A concrete example so that the kinase that we see most frequently destabilized is her 2. And so here you see with this inhibitor neurotinib, again, data that is just pulled out of this large screening effort. Again, you see neurotinib has no effect on the stability of 97 kinases with the exception of her two. And we can now take all the inhibitors that have this destabilization effect on on her two and map them on a target space. And what was very nice to see here is then all the heats also overlap with a target space where we know that her two inhibitor spine. So here we have a strong evidence that the inhibitor induced kindness, destabilization or degradation is also via a direct mechanism. This is also something that we can test. In this particular case, we found a suit of different covalent inhibitors to induce her two destabilization. So when we now go in and mutate the relevant cysteine on her two, we can also rescue the degradation that is conferred by these degraders. By and large, if we look at the data set, it's however not that we see that covalent or covalency is, is a general feature of these degrading inhibitors. Again, remaining on the high level, we do see that that half life of the kinase is not correlated with how often we call hits against a particular kinase. And this is just something that gave us a lot of confidence that we're not just measuring transcriptional or translational interference or vice versa, that the controls we put in place allow us to nicely correct for that. We also benchmarked the data set with the the paper that Fleur illustrated earlier, her prostate work work with Nathaniel Graver. We could see that the ability of a kinase to be destabilised by an inhibitor is not correlated to its amenability to heterobiofunctional protect design. In other words, if if a kinase is frequently destabilized or degraded by an inhibitor, that does not mean that this is a kinase that is also easily degradable via protic approach. So if you have 160 examples or 320 examples, you cannot make a need to give follow up with all of them. So what we decided is that we're going to focus our validation efforts on three concrete examples. We've picked 3 examples where we have where we have observed selective and acute degradation. And we've also picked 3 examples that are different in terms of the HSP 90 client status, where BLK represents a weak client, then represents a strong client, and RIB K2 is not a client of HSP 90. So I'm going to focus most of my time today on Ribkinis too. But we have also published this study as a preprint. So if you're interested in the other examples that I'm still going to mention, you'll find more data online. So the first experiment that was important for us is to really go away from this time trajectory abundance measurements to protein stability measurements. So here we want to really ensure that the compound induced effect on in this case RIP K2 happens on a protein stability level. To this end, we use this protein stability reporters where we express our protein of interest with, in this case, RIP K2 as a BFP fusion, which is separated by AP2A set from M cherry. So we can use BFP over M Cherry ratios as a measure to approximate protein stability. And what we nicely see here is indeed when we treat cells that express the stability reporter with the inhibitor that we have identified here called rib K in four, that we do see this shift in rib K2 stability. This is something again we can record in time ranging essays. And here importantly also when we treat these cells with an HSP 9 inhibitor, the effect is clearly not phenocopied. We've also data that I don't show here that reports that this inhibitor rib gain 4 is directly binding to rib K2. So the next step was really to try to understand on a mechanistic level how this inhibitor induce ribkinase to degradation. And what we typically do are in order to answer these questions, what we typically undertake our genome wide or focused CRISPR CAS 9 gene disruption essays. So what we basically do is we take the stability reporter. We transduced them with a gathering, a library that covers in this case, the entire genome. We knockout thereby every gene in the genome and ask the question knockout of which genes affects the degradation that is elicited by our inhibitor, the creator of interest. And typically if we do these experiments with a molecular glue or a product, what we what we usually see is a very strong enrichment in this high fraction that you see on the right of your on this of these volcano plot, a very strong enrichment of one particular E3 ligase. And this is clearly not what we see here. In fact, what we see here is an enrichment of genes that are involved in in membrane trafficking and it lysosomal degradation. So based on these, based on these results, we surmised that probably the clearance of rib K2 after treatment with this inhibitor is is based on autophagy or lisosomal degradation. And in line with that hypothesis, we indeed saw that we can rescue the degradation of rib K2 by core treatment with bifidomycin. We could also show that when we knock out FIB 200, which is a master regulator of macro autophagy, again we can mitigate the effect of that inhibitor. What was very curious for us to see is that when we looked at degradation of FIB Kit 2 with high conduct microscopy is that prior to degradation, what we do see when we treat cells with these inhibitors is that there's a formation of these aggregates, condensate like structures that are later on eliminated. And again you see when we compare core treatment with vehicle control or core treatment with bacillomycin that the removal of these condensates is again dependent on the lysosomal machinery. And this is particularly interesting because it's known that ribkind is 2 forms this condensate or aggregate like structure that are called ribosomes. Also after physiologically ribkind is activation via the not one receptor structure. What we do see also is that both in the ribosome formation but also in the formation of this inhibitor and used aggregates that this is dependent on the CAR domain. I don't have, I don't show the data here, but this is again something that you can find at the preprint. So this is intriguing for us because it suggests that however that inhibitor is inducing degradation of this particular kinase is this cause via a phenomenon that is very closely related to its physiological turnover. So we were also then interested in this protein T-Mobile 1 and we could nicely show that again one we treat cells with this inhibitor, T-Mobile, one Co localizes with these REAP K2 aggregates. So it seems to be a player involved in that process. And likewise when we knockout T-Mobile, what we see is that the formation of these aggregates and their clearance are delayed. Again, this is something that we can not only see in a high contact microscopy, but we can also observe via Western blot. So as a model for this particular inhibitor induced degradation, what happens is that rib K2 typically is available as monomers. Upon treatment with this inhibitor, there is this higher order assembly that is formed. These assemblies are reminiscent of physiologically relevant ribosomes and they are then after they are formed, recognized and turned over by the laser storm. And So what we believe is that this particular inhibitor really co-ops the physiological route of K2 degradation and this Co opting physiological route of degradation is is the common denominator that you'll see in the other two examples that I'm going to share with you over the next couple of slides. Second example that I want to introduce here is this very profound destabilization of the kinase Lin that we see after cellular treatment with inhibitor SI 3. So you see here again data from the screen, we could show that also this particular inhibitor binds in recombinant assays to link kindness. So we assumed again this is going to be a direct event. Again, we could show that the effect is not only on abundance but really on lean stability using the same stability reporter that I've introduced before. And what was very interesting for us to see is that when we do this very typical chemical rescue experiments, while it was clear that the induced destabilization is dependent on ubiquitination, this is what you see with attack 243 core treatment. We could not rescue degradation when we blocked the protosomal loan or when we blocked the last Somal degradation alone. Only if we block both of these major degradation pathway, we could see a rescue of SF3 induced degradation of Lin. So again we turn to a functional genomics in order to identify what are the E3 ligases that are involved in that process. And we clearly identified here one hit which in this case is an E3 ligase called Sybil. In addition to asking the question via these functional genomics, what are the genes that are required? We also use proximity biotumination coupled to proteomics to ask the question what are the ligases that are recruited in physical proximity to the lean kindness upon inhibitor treatment. And here again, very nice and very congruent to the CRISPR screen, we did see recruitment of the E3 Ligi Sybil to Lynn, but we also saw recruitment of a second closely related E3 Ligis called Sybil B. When we turn to validation studies, again, we see these redundancies. So when we knockout Sybil alone or when we knockout Sybil B alone, we see no or only a very, very minor rescue effect. But we really need to knockout both of those lidases in order to completely rescue the degradation effect that we observed for SI 3. And So what was again interesting here for us to observe is that this dual and redundant involvement of Sybil and Sybil B is again involved in the physiological degradation of hyperactive Lin. So this is Lin for instance, can get hyperactive activated after Ige stimulation. One of the ways of how LEAN gets activated is via phosphorylation of CSK or CSK activity maintains LEAN in an inactive state. And so that prompted us to test in cells not in recombinant assay. And this is something we did together with Stefank Knapsleb, whether or not our inhibitor binds to CSK or LEAN. And to our surprise, even though in recombinant assays the binding to LEAN was very clear in cells, that inhibitor was not able to engage Lynn, but it was very well and very importantly able to engage CSK. And so here the model that we come up with is a really the way of how this inhibitor is inducing Lynn degradation is by inhibiting CSK. And Lynn lives in these two worlds where if it's active, it's very instable because it's frequently or it's acutely turned over by Sybil and Sybil B And when it's inactive, it's in a in a more stable state. So by blocking CSK, we push Lynn into the hyperactive state and that then allows this acute to turn over by both of those E3 live cases. The third example that I'm going to present to you is probably the most intriguing data. So here we're observed degradation of another SAC family kind is called BLK after cellular treatment with inhibitor TAC285. Again, this is an effect that happens on stability reporter in a nice time ranging manner. So again, in order to understand the mechanism of action, we're trying to do this genome wide CRISPR screen. And here we don't find any 3 ligase, but the effector that we find here is actually the gamma secretase membrane bound protease complex where very, I would say intriguingly, we identify 4 out of the four gamma secretase subunits is very strongly enriched meaning and this is something that we could validate pharmacologically that you need the gamma secretase activity in order for this inhibitor to induce build K destabilization. So here, and this is again one more work that we outlined in the preprint, what we could show is that again, BLK lives in two different states, one that is membrane anchored where it's stable and one that is cytosolic where it's instable. And again what the inhibitor TAC 285 does here that by a gamma secretase she pushes the membrane bound stable state into the cytosolic in stable state. And so in summary for for this part, I hope I could convince you that inhibitor induced degradation of kindness. This is something that occurs frequently. We did a firm Chevron deprivation as a mechanism that underpins many of these inhibitor induced degradation events. However, in the three cases where we had a closer look, this was not really the case. Instead, what a picture that a generalizable picture that emerges is that inhibitor induced degradation frequently superchargers physiological turnover mechanisms. So then basically all these kindnesses live in two different states, a stable and an instable state. And what the inhibitors do is that they're tilting the balance towards the less stable state. And on the mechanistic level, this can occur via different phenomena. It can occur via cellular localization, it can occur via modifying kinase activation level, and it can also occur via modifying the aggregation state of a particular kinase. So we were also interested whether this supercharging of physiological turnover mechanisms is a mechanism that can go beyond kindness targets. And this is a collaboration with a very cherished collaboration partner of ours have a development from the Max Bank Institute in Dortmund. And what Herbert's lab did is that they set up a phenotypic discovery drug discovery essay aiming to identify chemical matter that would act as an inhibitor of this protein idol, one which is a target of interest in immuno oncology applications. And this is involved in the catabolism of tryptophan to canurinase. And by developing this canurinin phenotypic reporter, they identified a pseudo natural product that acted as an item 1 inhibitor. But also as you see here in Western blot, induced degradation of item one in A in a time and dose dependent manner. Very interestingly, the degradation of item 1 was relatively selective. So here you see again quantitative proteomics and in addition to item one, there's only two other targets that are significantly destabilized. So we could show that that inhibitor or they're called here as IDEX also engage either one in cellular assays. And we could also solve the crystal structure that shows that Ida one binds into the heme binding pocket of Ida 1. So Ida 1 is a heme binding protein. And what was particularly interesting is that if we compare the binding mode or the structural state that is induced in IR One after Co treatment with this particular compound Irec 2, and we've compared this to a standard IR one inhibitor that reached clinical investigation. You see that at the C terminus of IR one, there's this bundle of helices and these helices all undergo a certain structural rearrangement which is most pronounced with a with a very last Helix that completely uses structure in that particular crystal structure. So binding of idle one causes this profound structural rearrangement at the C terminus. So again we wanted to understand what are the cellular effectors that mediate this inhibitor induced degradation and kind of in line with a relevance of the C terminus of idle one, we could observe destabilization of idle one with our stability reporter only if we put the fluorescent proteins on the north terminus. As soon as we touch the C terminus, those proteins or these protein fusions were inert to this itic chemical series. So again we turn to FAX based CRISPR screens and identified very clearly an E3 ligase called KLHDC three that is relevant for the inhibitor induced degradation of idle one. Now interestingly again here KLHDC 3 also appeared to be regulating IO1 level in steady state. So this is also the physiologic or the endogenous E3 ligase controlling idle 1 turnover which is again something that you see in the back graph here. But if we remove or if we knockout pulled S s here by Cal HTC Three, we clearly abrogate the activity of the IDEXX in inducing idle 1 degradation. So working with Brenda Schulman's lab, we now wanted to understand how that relates to basically ubiquitination of Idle 1. And what you see here are ubiquitination essays using different versions of Idle 1. And what is I think very interesting to see is that the APO structure again idle 1 is typically in cells present in in its heme bound form. So what you see is that APO idle 1. So where heme was displaced, meaning here the DMSO treatment is very actively ubiquitinated. The same occurs when we treat these recombinant essay or when we incubate idle one with one of the the chemical series, here IDX 6. But it is not occurring when we treat it with one of these clinically idle 1 inhibitors or when idle 1 remains heme bound. So this suggests that really the state of heme engagement or the state of how our inhibitors bind the heme binding pocket determines how good of a substrate for KLHDC 3 idle one becomes. This is something that we can also recapitulate when we just block him synthesis in intact cells. And this really leaves us with a model that IPO Idle 1 is very efficiently turned over, recognized and turned over by this. Like SKLHTC 3, this is a staple state is stabilized or mimicked after IDEXX bind to idle one, in contrast to when idle 1 is bound by him. And this is probably the vast majority of all idle 1 copies in a cell. Idle 1 is less, is stabilized and is less prone to degradation via KLHTC 3. So as a summary, I hope I could convince you that in addition to this classical proximity inducing modalities of how target protein degradation can occur, there is this world of inhibitor based supercharging of of native degradation events. So we have seen this with kinases, we have seen this with this protein idle 1 were in all of those cases the proteins that we had a look at live in these two different state, a stable state and an instable state. And in all of the cases, the inhibitor just tilted that balance via different cellular phenomenon. Again, this was activity oligomerization, cofactor binding or cellular localization. And this is something that I think is very interesting to observe further and should hopefully provide a blueprint for how inhibitors can induce target protein degradation frequently independent of a traditional proximity inducing event. So that leaves me with acknowledgments. So again, all that work has been conducted by Natalie Scolds, a postdoc in the lab. I want to thank all our collaboration partners and in funding sources. And if there's enough time, I'm happy to take questions. Thanks so much, York. That was great. And yeah, we definitely have some time for questions. So please feel free to put them in the chat or I'll raise your hand and and ask them verbally. I guess just a general maybe question from me. I thought your inhibitor work was really interesting. Do you see this as these inhibitors that are driving degradation as sort of chemical probes to help us understand just these mechanisms that could be further exploited? Or are you thinking like you could maybe take those those inhibitors and then further, you know, exploit these mechanisms to make more traditional degraders? Any comments on on that? I think for now it was mostly that study was mostly born out of curiosity in terms of mechanistically understanding, you know, how how these inhibitors typically induce degradation, whether or not they will really be able to compete with in terms of efficacy and potency with traditional degrader modalities. I have my doubts. But I think it's, it's just if you look throughout these examples, with exceptions, they, they don't reach the potency and the maximal degradation efficiency of you know, well optimized project or well optimized molecular glue degrader. Nevertheless, some of these degradation events that we do observe to happen at a concentration that is low enough that you can imagine that, you know, this will be relevant in a clinically set. I do think that it's very interesting to see that, you know, in in four different cases where we went into completely unbiased, we always ended up with the same phenomenon of kind of supercharging physiological degradation routes. And I would not exclude that this is a blueprint of making new degraders. But I do think that you need a use case why you would want to do that as opposed to a regular, you know, molecular glue or protect play. Great. Thanks. We did have a few questions coming in the chat. So some of these a little more technical. So Loretta is wondering the concentration of the compounds you used in the large kinase screen and how you set the threshold for what you considered a hit. Yeah. So the concentrations we did were always ensuring that we're not in a toxic range. If this was feasible, we screened at 10 micromolar. If not, I think I would believe we we screened at one or two micromolar. The thresholds for calling a hit had to do with statistical measures. But by and large it was really driven by selectivity of for one particular kinase as opposed to all the other kinases and being more potent than our positive controls. I can't tell you a concrete P value from the top of my head, but I think we we, we try to be relatively conservative. I think what we did do is we took a lot of advantage from this time dependent trajectories because that allowed us to not only base all of our adjustments or judgment on one particular time point, but we could take this entire time ranging interval into consideration. Great. Thanks. So another question, is it clear what happens in cells treated with these inhibitors if they lack the E3 ligase is responsible for degradation? Do you think these inhibitors would have vastly different effects if used in tissues composed of different cell types? That's a good question. I, I don't know whether if I'm interpreted correctly it is whether or not availability of that particular endogenous degradation circuit would dictate how these inhibitors function. I think typically right if the inhibitor will only have an effect where the kinase of interest is expressed and if a hard time anticipating a scenario where a kinase is expressed without the appropriate physiological degradation circuit in place. So we haven't really thought about that all too hard. I don't. I don't know how realistic of a scenario that will be. I don't know whether I answered that question or if I interpreted that question correctly. Feel free to drop me an e-mail or a note if I didn't. Sounds great. Thanks. And then Fredericks is asking, have you also observed stabilization of kinases with inhibitor treatment? Yeah, yeah. We have observed that actually quite frequently. It's something that I think is very interesting, but not something that we have at this point followed up systematically. And then one more question is about just, I'm not the, do these superchargers inhibit enzyme activity? So I'm wondering if I mean they are inhibitors of of enzymatic activity or I assume unless they're they're just binders, but maybe you could just comment on on the, the enzymatic inhibition and maybe how that correlates with some of the responses you're seeing. Yeah, that's a great point. By and large, all of them also inhibit the target for the IDEXX. So these inhibitors and degrades of idle 1, I don't think we saw a linear correlation with how good of an inhibitor as opposed to how strong of a degrader of those molecules are. So I think here in that particular case it really had to do with how accessible these the C terminals is for degradation Vitis physiologically like this KLHTC 3. So in a if we would optimize this chemical series for degradation, this is likely the parameter that we would need to optimize, not necessarily based on how potent of an enzymatic inhibitor these compounds really are great. Thanks. It looks like that's all the questions we had in SO. Thanks again, Jorg for this great presentation and answering some questions. Really appreciate learning about your research. Thank you for the invitation. All right. And so before we move on, we're going to ask you another poll question, this time thinking about challenges. So as we think about these degrader compounds and all that we're learning, you know, what are you as an audience think is really the biggest challenge that we have right now in developing greater compounds? Is it really around that targeted identification and identifying suitable target proteins? Is it a, is it the challenge around achieving selectivity and minimizing off target effects? Is it about the stability in the pharmacokinetics of the compounds as we're, you know, taking them into the clinic? Or is it really understanding some of what Fleur talked about really understanding and predicting the the biology and productivity of that ternary complex. So I'll give you just a minute or so to share your thoughts and where you think the biggest challenge lies. OK, It looks like we've got a good number of responses, so see what we ended up with. So the majority are thinking that achieving that selectivity and minimizing off target effects is still a real challenge to overcome. But also, you know, definitely a spread of of the other topics as well. So I think these are all very relevant challenges, but you know, it's interesting to hear what you all are thinking. So thanks for sharing. All right, and now we will move on to our next presenter, Dr. Kristen Reaching. Kristen received her PhD in Biomedical engineering from the University of Wisconsin, Madison and joined Promega in 2014 where she's established approaches to characterize the cellular kinetics, potency and mechanism of action of small molecule degraders. Kristen is currently a group leader focused on developing technologies that enable further insights into mechanisms underlying to greater efficacy and the investigation of new induced proximity models for TPD. So thanks, Kristen. I'll turn it over to you. Thank you, Amy, for that introduction and thank you for joining everyone. I'm really excited to share some of our recent work investigating ternary complex formation as well as degradation of Cerebrano substrates. And also just want to say that was a very timely poll question because we are also very interested in using some of our tools to enable characterization of selectivity of these molecules. So I'm going to talk a little bit about some of the insights that we have learned in this process. And as both Fleur and and Gay org really nicely introduced and, and highlighted it, I think you know what is the the huge potential here for targeted protein degradation in general. And that's being able now to essentially target proteins that were previously considered on druggable. And this has been best exemplified by the IMID molecular glue class of degraders, which function by binding to Seroblon and essentially modifying the surface of Seroblon to to essentially create a new protein, protein interaction interface and induce binding of a variety of different proteins called neo substrates. And upon binding, these neo substrates then become ubiquitinated and degraded. And many of these neo substrates, sorry, many of these neo substrates are proteins that are not ligandible and they consist of as flora also discussing finger transcription factors, but they also span multiple different target classes. But one of the similarities across these neo substrates is that they all or the majority of them contain this conserved structural Dagron, where there's a critical glycine residue in the beta hairpin of the structural Dagron. And many groups have attempted to mine the proteome to try to identify just how many proteins within the proteome might contain this Dagron. That could potentially be neo substrates of Cerrobond and it's estimated to be over 2500 proteins. And on the chemical side, looking at the structure of these imid molecular glues that they all seem to share this core structure, this glutaramide ring that many people are also now trying to explore functionalizing in different ways, branching out with different functional groups to try to induce recruitment and specificity for specific neo substrates. But some challenges remain in how we can identify and optimize molecular glues for specific targets because this isn't a rational approach and relies on understanding of that protein, protein interaction interface and how we can translate that into chemical design. But also how how can we characterize the selectivity both in terms of the target as well as the cell type. And of course, this this floor discussed very elegantly. Some approaches that are really critical in this space are are proteomics, quantitative proteomics and chemo proteomics. But the challenge with some of these approaches is that they're limited in terms of the time point and also the expression of the proteins that you want to be able to monitor in, in that particular cell line. And So what we have enabled is the ability to monitor at quantitatively and in a sensitive fashion degradation of endogenous proteins with high bit. And so high bit is a very small 11 amino acid peptide that has high affinity. It complements spontaneously with the large bit protein to produce a very bright luciferase enzyme. And we have used this extensively over the years to study Protac mediated degradation to character characterize the kinetics of degradation of endogenous proteins and understand mechanism of action. And we've done this by knocking in the hybrid tag to the endogenous locus where we can precisely tag a protein of interest and track the degradation in real time following treatment with a degrader. And more recently, we've started to apply this approach in characterizing and understanding molecular glue mediated mediated degradation kinetics as well as selectivity. And so just showing here I'm showing a few examples of IKZ F1 and IKZ F3 degradation that these are both high bit knock insurance to the respective proteins and degradation induced with aberdamide. We can see very rapid and and potent and complete degradation of both of these proteins with aberdamide. And then more recently Novartis published this degrader which is an IKZ F2 degrader DKY seven O 9. And we can also see with high bit that we can capture nice dose dependent and partial degradation of IKZ F2. But what about other neo substrate? So this is just looking at kind of the known targets of these compounds, but we want to be able to assess the selectivity. And so of course we're not looking at the proteome scale here, but we wanted to build kind of a a panel of new substrates that are all tagged with high bit. And these represent sort of the most common new substrates that have been observed with a classical molecular glue degraders. So we generated this panel of high bit knock insurance to to these various targets and assessed degradation of aberdamide as well as DKY 709 across this panel. So in addition to IKZ F3 and IKZ F1 degraded by aberdamide, we also see both of these compounds in fact degrade Cell 4. So DKY 709 degrades both IKZ F2 as well as cell 4. But we're also interested to see what the selectivity profile looks like if we were to compare the ternary complex formation induced by both of these compounds against the same NEO substrate. So what does that profile look like? So to do this, we turn to our nanobrat ternary complex technology. So this relies on a nanoluck fusion to the neo substrate coupled with the Halo tag fusion to Seroblon. So nanoluck serves as an energy donor for for energy transfer and then Halo tag is the energy acceptor. So upon treatment with a molecular glue, we induce turnary complex formation and we can see an increase in the breath signal. So looking first at Ibert amide with showing the same degradation profile that I showed you in the previous slide on top and then ternary complex, the selectivity profile on the bottom, we can actually see gratifyingly, that's the most potent ternary complex induced by Ibert amide is induced between IKZ F1 and Sarah Bond as well as I, KZ F3 and Sarah Bond sort of matching the degradation profile. And if we look at DKY seven O 9, again, degradation on top and ternary complex on the bottom, we also see that this compound induces the most potent ternary complex with IKZ F2 and Sarah Bond. So this is great. We can see potent ternary complexes forming in agreement with the degradation, but of course we also see that ternary complex can still be formed between neo substrates and Sarah bond induced by both of these compounds that do not result in significant degradation. And just to call out a few examples here, Aberdamide does induce quite potent ternary complex with IKZ F2 and then DKY 709 induces potent ternary complex with IKZ F3 despite a lack of degradation. So as I mentioned, one of the questions that we're really interested in, in trying to gain some better understanding is, is how can this information be leveraged into greater design and, and can we use these tools in a structure activity relationship campaign to try to dial in selectivity of a particular NEO substrate or alter the selectivity profile to then induce degradation of particular NEO substrates. So to that end, we were very fortunate to embark on a collaboration with Saint Jude and Zoran Rankovic's team, who now as as we know has transitioned to the ICR and we're going to hear some exciting updates from him very shortly here. But we were really excited about this collaboration with his team as well as Giselle at Saint Jude to try to identify selective CK 1A degraders that could potentially be used in treatment of various cancers. And so the team began to try to screen and phenotypic assays across a library of molecular glue, Seroblon molecular grooves to try to identify anti proliferative hits across 9 pediatric cancer cell lines. And so this this library consisted of traditional IKZF one and three degraders that did not show broadside of toxicity. Of course, GSPT 1 degraders did show broadside of toxicity across this panel of cell lines. But strikingly there was a set of compounds that seem to show cytotoxicity only in mold 13 cells. And from the cancer depth map, we know that mold 13 cells are actually highly dependent on CK 1A. And so the team hypothesized that these compounds could could actually be CK 1A specific to graders. So these initial compounds form the basis of series 1 compounds and we then took those and profiled them and our high that degradation assays looking at degradation of CQ and Alpha as well as IKZ F1. And just showing here is kind of a summary. We, we actually profiled all these compounds in kinetic fashion, but we quantitated from those kinetics the degradation potency, the D Max and the degradation rate. And so you can see sort of a broad range of responses, insignificant improvements in CK 1A degradation over Lenalidomide, which is currently only FDA approved drug to induce even partial degradation of CK 1A. But we also see degradation of IKZ F1 with all of these compounds, but less broader, more narrow range of degradation potency and degradation rate across IKZ F1. We also confirmed through proteomics degradation of CK 1A and IKZ F1 as well as IKZ F3. But then we also just wanted to make sure that we weren't touching GSPT one in in potential for potential safety liability of these compounds and hitting GSPT 1. And you can see here that these top two compounds from series one do not degrade really at all GSPT 1. And we confirm that they don't form ternary complex with Seroblon, but we do see robust ternary complex induced by both of these compounds with CK 1A. But of course, the intention here was to really improve CK 1A selectivity, and so could we then improve the selectivity and dial out the IKC F1 degradation? And so the Saint Jude team overlaid this series 1 degrader onto the crystal structure of Sarabon and CK 1A induced by Lenalidomide. And with molecular dynamic simulations, they really saw that by branching off of the C4 position on the molecule, there really were no opportunities to engage with CK 1A. But by perhaps switching this position now to the C5, this would present some opportunities to directly engage CK 1A and inform some contacts that could then stabilize the ternary complex. And so a set of series 2 compounds were made that now branched off of the C5 position. But just to kind of show you how we transition from series 1 to series 2. So this is series 1, which the top compound from series 1, which of course is a dual degrader. And we can see that degrader forms potent turnery complex with CK 1A as well as IKZ F1. If we switch to the C5 position now we can see in some cases we actually saw a select selectivity where Ikz F1 degradation was preserved was apparent, but then CK 1A was not being appreciably degraded. And we can see this mirrored in the ternary complex as well, where we saw robust and potent ternary complex formation with Ikz F1, but really very little ternary complex with CK 1A. But if we structure this and make a more rigid benzamide configuration in this molecule, we can actually switch that selectivity. So now we can drive even more potent and complete degradation of CK 1A, completely dial out the IKZ F1 degradation. And we can again see the that mirrored in the ternary complex, where we see more potent ternary complex formed with CK 1A and very weak ternary complex with IKZ F1. But Zorn didn't want to stop there. So he wanted to drive even more profound CK 1A degradation while still maintaining that selectivity, but also seeing if we could improve the PK properties of this molecule. And to do so, they generated a set of Series 3 compounds which were made by employing the spring formation strategy to attempt to approve the PK. And when we profiled these in our Kinetics, we saw really exquisite potent and fast and incomplete degradation of CK 1A down to single digit nanomolar with really while maintaining that selectivity for CK 1A over IKZ F1. So even with concentrations up to 10 micromolar, we saw no degradation of IKZ F1 and then they were also able to generate with this top series 3 degrader. Now this SJ 3149, a high resolution crystal structure showing the ternary complex forms between Cerubon and CK 1A. So now we can see not only are there new contacts formed with Cerubon, but now this molecule also directly forms contacts with CK 1A, which we think really helped to stabilize this ternary complex leading to more efficient degradation. And when we looked at this in our nano bread assay, so this graph on the right, the the top series 3 compound is in purple. And just for comparison, here I was showing you the Turnery complex previously with the top series 2 compound, which is in green. We actually saw a really massive increase in the the signal that we observed in the nano bread assay, which we think is is really strong indication of this increased stability of the Turnery complex and then leads to very exquisitely potent degradation of CK 1A. But looking back across the trends across the entire series, we were also very pleased to see such a strong correlation of degradation. So this is plotting CK 1A degradation rate against CK 1A degradation extent at 4 hours. And not only do we see a strong correlation there, but we also see very strong correlation between degradation of CK 1A and the phenotypic response. So the cytotoxicity that we observed in the mold 13 cells. So both faster and efficacious degradation of CK 1A seems to correlate very strongly with potent cytotoxicity in the mold 13 cells, despite being a very different cell background from that which we monitored degradation. And this is often the case that we are not always able to monitor very effectively degradation as well as the downstream phenotypic consequences in the same cell background. And So what we want to be able to enable is, is being able to profile degradation in those relevant cell models, but also to enable profiling degradation in many different cell lines very rapidly. So how can we do this And to 1st just to discuss why we might want to do this and why we might expect degradation to be different in different cell lines. There are many different reasons why a degradation could be different and in different cellular contexts that dependent on the target expression of different isoforms or mutational status. The turnover cell permeability could also be different and availability of the molecule inside the cells, expression of the degradation machinery as well as the de ubiquitinase machinery. But ultimately, we want to be able to characterize degradation as well and disease versus healthy cell lines where we want to target the protein in a disease cell context, but potentially want to spare the protein in a wild type or healthy cell context. So how can we do this? So we are developing some new technologies to enable a detection of degradation and unmodified cells using LUMET. And this is a relatively new technology that is based on an immunoassay format where we use primary antibodies against a target protein combined with secondary antibodies that are conjugated to the nano bit protein. So in this case, they are conjugated to either small bit or large bit, which are the low affinity nano bit proteins and rely on assisted complementation when brought into proximity by binding of the antibodies to the target. And so the workflow here is very, very simple. We simply treat the cells with a degrader and then after a period of time, we lice the cells, add in our antibodies and detection and we just simply read out the luminescence. So it's very rapid assessment to target protein levels and there are no wash steps required. So initially to start out, we wanted to screen for degradation of GSPT one. And the reason we wanted to specifically look at GSPT 1 is because it has essential cellular function and degradation of compound or with compounds that hit GSPT 1 is often broadly cytotoxic in many cells and it's a potential safety liability with degraders that function via Sarah Bond. So we wanted to start here to develop tools to specifically monitor GSPT 1 degradation in any cell context. So we developed a LUMA assay for detection of GSPT 1 degradation. And the first thing we wanted to do was compare the degradation we observed with that, that we measured in in the high bit cell line that we've used extensively. And so this is just showing a comparison with another fantastic Saint Jude degrader. So this is the GS PT1 degrader. We compared early time point as well as a 24 hour time point. And you can see that there is a difference at early time point. So we're detecting more potent degradation with Lumet at short treatment time of three hours, but we have a similar potency and D Max observed between both hybrid and lumens at 24 hours. So we were initially quite perplexed by this difference and and we wanted to ask whether we were detecting the same isoforms of GSPT one in both assays. And so looking more closely at our hybrid CRISPR cell line, we can actually see that because we tagged the N terminus and there are three different isoforms of GSPT 1, isoform 1, which actually has a different start site, that we are only tagging isoform 2 and isoform 3. If we however, switch and tag the C terminus of GSPT one, now we would tag all isoforms provided if they're all expressed in the cell. So we generated a hybrid CRISPR knock in to the C terminus of GSPT 1. And we compared the degradation kinetics of the C term knock in on the top compared to the N term knock in on the bottom. And in this case, I'm showing a comparison of two different compounds, the St. Saint Jude GSPT 1 degrader, but also CC-885. And while we can see very potent and complete degradation in both cases, there are some subtle differences and you can maybe see a little bit on the top with both compounds, the C term high bit tagging, we're actually seeing a little bit faster degradation kinetics. But importantly, in terms of the specificity, they both do not respond to a Bertamide as a Bertamide should not degrade. GSPT 1. So we went ahead and we quantitated these degradation kinetics in terms of the rate and the potency. And indeed we can see that there is a difference in the degradation rate. The C term knock in does seem to degrade faster than the N term and there's also a difference albeit smaller in terms of the degradation potency. And so by including isoform one in our hybrid tagging, we're we're actually able to observe faster degradation kinetics. And this suggests that isoform one may actually be degraded first or more efficiently compared to the other isoforms. So what does this look like now when we compare back to LUMIT? So now we can see that if if we compare the degradation potency at 3 hours in direct comparison to the time that we measured at LUMIT or for in the LUMIT, we can see that this better aligns now with the Lumit at these early time points. But it's also something to to keep in mind that we don't actually know what epitopes these antibodies, these GSPT 1 antibodies recognize and are they selective for one isoform versus another. But also they might because there's significant homology between GSPT 1 and it's related family member GSPT 2. It's possible that they could also be recognizing GSPT 2 in addition to GSPT 1. So it's important to keep in mind that what we're detecting with these different formats, these different assays, whether hybrid, hybrid is very specific to what it is being tagged, but that this could also be different, you know, depending on the antibodies that are used in various immunoassay formats including Western blots. But having validated this assay now we went ahead and wanted to monitor degradation in different cell lines. So I showed you hack 293. We also looked at a more disease relevant cell line on B 411 and we also wanted to characterize the potential for these antibodies to exhibit cross reactivity in different species. And so we looked at a rhesus macaque line as this is something that might be important for preclinical species selection. And we could also see degradation of GSPT one in all of these cell lines with similar rank order of these compounds. We've since expanded this now to a number of different cell lines as you can see here. So this is showing a three hour profiling and 24 hour profiling on the bottom across six different cell lines that span a range of GSP 21 expression as you can see from the signal to background signal that we measure in the LUMIT assay. And we can see upon treatment with the SJ compound about two log orders of degradation potency in these different cell lines that is not necessarily correlated back to expression level. And so this is another thing that we're interested in better understanding is what what is driving these really dramatic differences in degradation potency. This is sort of the first time that we've been able to quantitate with high precision degradation potency in different cell lines without having to do so on a Western blot. So we're actively trying to explore how we can understand what is driving these differences to better understand the degradation potency and how this information could be leveraged in the context of degrader design. But just to end kind of with a pretty picture, I also wanted to show you that we are able to now incorporate bioluminescence imaging. We have an instrument, a very new instrument called the Glomax Galaxy that is specifically designed to enable bioluminescence imaging. And we just recently were able to image our lumen assay on this instrument using the same lumen antibodies in a modified immunocytochemistry protocol. So all we do is add the degrader, fix and permeabilize, and then we can detect GSPT 1, endogenous GSPT 1 in these cells. So we can see it's primarily localized to the cytosol and upon treatment with degrader, how that signal is reduced. But with that, I would just like to acknowledge everyone at Promega who contributed to this work. We have a fantastic team here that I'm really privileged to be able to work with and also want to thank our wonderful collaborators, particularly Zoran and Giselle, and the rest of the team at Saint Jude that contributed to the CK 1A story that I presented. And then I want to thank all of you for your attention and open it up for any questions. Great. Thanks Kristen, really excellent updates. Yes, please feel free to raise your hand if you have a question and and ask it verbally or type it in the chat. There was one question that came into the chat kind of going back to the beginning part of your talk when you were looking at ternary complex formation and Bacion is wondering if there are ways, any ways to tell it the ternary complex formation is coming from direct interaction or from like a bridging protein? Yeah. That's, that's an interesting question. I mean, I in this case, I wouldn't expect because these new substrates are not always, I mean, I guess it's not really expected that they would be interacting together like, you know, in a complex themselves. I guess maybe some of them could, but I, I think that's unlikely. What we are detecting is primarily a direct interaction that is mediated by the small molecule. It's also something that we know from experience that, you know, if we're trying to detect and of course it all depends on the geometry of a complex, right? But if we're trying to apply nanobrat to study sort of a proximal interaction or you know maybe not maybe another protein that is involved in a complex but is not a direct binder of another protein that we don't often have the ability to detect that because the proteins really need to be sufficiently close to detect energy transfer. So I would say that the the ternary complex that we are detecting, those are. All direct interactions. Great, thanks. Another question about you showed sort of the impact of tag location and I Loretta has a question about the tag itself, right. So the natural high bit tag has lysings in it and but we've also worked with a lysingless arginine based high bit version. And Loretta is just wondering about if you're have plans to expand the knock and sell lines to include the the lysingless version or if you have any comments on that? Yes. So it is possible to use the lysine less variant and we have we have tested this and a few cases and you know in all the cases that we've studied all the week, we can't obviously test every every situation. And it may be possible in some limited circumstances for those license to contribute to degradation. It's it's certainly possible. We've never observed that to be the case, but, but certainly, yeah, it is a concern and and for that reason it is possible to use the lysine list variant. So yeah, that's something that is available if you'd like to to do that. I think technically we, we just ask that you let us know that you intend to do so. But but yeah. Great. Thanks. Yeah, I think it kind of lends, you know, to your point, regardless of the method of detection we're using, right, all assays have some some possibility for artifact whether it's antibody and how the antibodies binding or if you're tagging and how you're inserting the tags. So, you know, always I think important to to think about control, then, you know, really making sure you're doing everything you can do to properly interpret your biology. So that is all the questions we have for now. So thanks again, Kristen, really appreciate your presentation and thank you. Before we move on to our final presentation, just one more quick poll question. So before we asked about challenges and you know here kind of on the flip side, what is really the most exciting potential that you see with degraders in drug discovery? And you know, I think we can all see there's tons of really exciting potential. So it might be hard to pick one, but you know, from your perspective, what do you see as the most exciting potential? It's just expanding the druggable proteome and tackling some of these targets that we just haven't been able to tackle before, improving selectivity and specificity and reducing off targets compared to some of the traditional small molecule approaches. Addressing or overcoming resistance mechanisms that can occur often occur with traditional small molecules or using degraders in combination therapies with other modalities. So I will give you a couple minutes here to share your feedback on where you think the most exciting potential exists. All right. Looks like we've got done coming in. So pretty strong consensus around this one really expanding the druggable proteom and yeah, really tackling these targets that just we haven't been able to address with the classic inhibitors. So thanks for sharing that. I think all, all really exciting potentials, but you know, I think that certainly is an exciting one, especially as we're kind of talking about on druggables here today. So thanks for sharing. And now I'm going to introduce our last speaker, Doctor Zoran Renkovic. Zoran is a professor of chemical biology and the director of the Center for Protein Degradation at the Institute for Cancer Research in London. Previously, who was the Director of Chemistry at Saint Jude at Saint Jude's Children Research Hospital Hospital where he has established successful targeted protein degradation program that developed novel Seroblon warheads, Protax and molecular glue candidates. Zoran has held leadership roles at Eli Lilly, Merck, Schering, Plough and Organon directing teams that delivered multiple clinical candidates across various therapeutic areas. His current research focuses on advancing targeted protein degradation strategies to explore cancer biology and develop next generation treatments. So thank you, Zoran. I will turn it over to you. Thank you, Amy. Really appreciate. Actually, I'd like to start by congratulating you and your colleagues at Pramita for putting together such a wonderful webinar. And of course, thank you for the invitation. And it's really great to be here, whatever that is, in this virtual world. Today. I'd like to share some of the learnings from our product optimization projects, specifically focusing on the oral bioavailability. These are my disclosures now. This is just a brief reminder of the reasons behind such an incredible excitement about targeted protein degradation exists. From completely new and modality and catalytic mechanism of action to drugging and draggable. TPD is definitely a ground breaking paradigm in drug discovery. So my group is generally interested in both products and molecular group degraders developing and utilizing them to identify translation opportunities for cancer. We are currently focused mostly but not exclusively on Cereblon for the reasons highlighted here. Now of course the products are heterby functional conjugates, really great in terms of the rational design, but not drug like and consequently difficult to optimize. And that is precisely pretty much the focus on my talk today. But first just to I wanted to contrast the products with molecular glues and which are on the other side of our TPD spectrum. They present phenomenal opportunity for dragon and draggable, but they also present the design challenge for molecular glues you have to screen. This is why we built a library around 5000 compounds Ceramon binders that a wonderful really to talk to just a minute ago presenting some of our work. The design principles around this library were very much to what I refer is bio diversity. It's shown here All compounds contain the conserved cerebral binding glutaramide motif that presented broad chemical diversity at the cerebral surface. Importantly, the library is designed with the physical chemical properties in mind so broadly fitting the drug like space and the screening. This library ultimately led to design discovery of potent and selective graders such as GSPT one again, Christine really still introduce this compound. It's as you can see a small molecule shows a wonderful PK in mouse dose dependent pharmro kinetics and it's already orderly bioavailable and it's been out licensed for a clinical development straight from screening to hopefully clinical development for Ind early next year. This is not an isolated case. We just earlier this year we published the CK 1A and other potent and selective degrader Christine mentioned released. So again some in depth in vitro properties of absolutely potent potency and selectivity properties that characterize this compound as a CK 1A degrader. It's also one of those compounds here is amplified, is also already bioavailable. Pretty decent grassy law clearance and auto viability. But what about projects? Well, what to expect when you bought together 2 ligands with a linker in between. So no surprise we end up with the molecular monsters like this one, an LCK project that we published a couple of years ago now. This, this, this was really good. The grader in retro. But show in vivo activity for for to show in vivo activity, we needed AIP administration, which obviously is suboptimal when you think about clinical enablement and translation. So what to do in such cases? So what to do with molecular monsters like this? How do we approach the optimization? This was a steep learning curve for the whole community, and I'll share with you some of the findings from our own journey. I'll come back to this molecule at in in a minute, but I'll start first with, I'll go to one of our first product projects which started with our interest in medulloblastoma, which is the most common malignant pediatric brain tumor. There are at least four groups characterized by genotyping patient tissue and Group 3 specifically characterized by amplified MEC. For that reason we wanted to attract JQ one against the patient derived cells. JQ one is beer know well known well described beer before inhibitor that ultimately results in a lot of reduction in Mac expression and indeed in this medial bustoma cell line as the MBO 3 shows gradually reason reasonable activity, but less than what we hope for. So it was relatively recent during our work when when Bradner published this, the first one of the first products repeat that one. And we were interested if this product would show any better activity in the cell line. But to our surprise, same cell line, very poor activity, significantly less than inhibitor itself, which is unusual for products. So we're we went back and done Western blotch and this same cell line and then it did be shown lots of beauty four and beauty two. You can see here awesome corresponding CMA as well. And you can see even a hookah fact really at higher concentrations. So debit one works as a product in this airline. But this was at 8 hours when we tested the compounds at 24 hours, a completely different picture, there was a protein rebound. So you can see really, really no effect on Beauty 2 and and very little on Beauty 4. Now really we banged our head really. So what was going on really still here and cut the Long story short really still. So we ultimately realize that there is a chemical stability issue with with the D BET 1. So you can see here when we monitor the levels of compound and this is just the cell media and not no cells present, you can see really very rapid loss. So something's happening with the compound is degrading and with a half life for about two hours, when we look back into original paper from Brother, you can see really it's the same thing indeed at 24 hours the protein is coming back, so no effect. And when tested again, our compound in this was ML cell line M 411, We saw very similar effects of loss of a compound with the health life of about four hours. Now the assay in vitro mandoblastoma cell assays around for 72 hours. So you can see really still this rapid loss of the compound really explains the pretty low activity in this particular cell line. Now when we look back into the reasons, probably not surprising, certainly not for chemists really in the audience that that we found that what what's happening is actually hydrolysis or thalamide and and thalamide rings. And to address that we decided to look for potential replacements of the thalidomide as as a cerebral directing warhead in our projects. So we we tried several, but this was a quite successful really stir example where we replaced the thalamite we spent in Gutaramide. So the compound retains similar activity to to thalidomide and importantly shows rather improved chemical stability. So with that, we decided to make a direct analog of D BET one, which is we refer to it as a PG product. And when we test this compound in medioblastoma cell media in terms of chemical stability, again shows considerably better stability. And consequently we can see more prolonged in deep effect on BRD 4 protein levels even after several days compared to D BET one where the BRD 4 protein comes back very quickly, only about 24 hours or so incubation in in cells. So that's well reflected in terms of anti proliferative activity. And so you can see here that in respect to DIBET one, our PG project shows about 3 orders of magnitude higher activity and this was in HDMPO 3, the medevastoma cell lines, same thing we see in other cell lines including MV 411 like ML cell line here. So this was great. So our PG product performs really very well in vitro, but in terms of pharm kinetic properties is really no different to Tibet one really still. So basically no a difference in terms of pharm kinetic properties, very poor, high clearance poor or availability. We made a range of analogues focusing on the link region and we were able to compare to clearance data and look into physical chemical properties. And what we realized was that a couple of most important predictors of improved clearance or rather reduced clearance, our topological policy of this area and the number of attachable bonds. So with that in mind, we designed analogs with improved with reduced a number of rotatable bonds and positive this area as you can see here compared to the BET one and PG product, which also resulted in very important reduction of hydrogen bond donors and consequently compounds like this and showed improved pharmacotic properties including reduced clearance and improved or by availability. So this was a focusing on what mostly really is been reporting our literature in terms of optimizing overall in we training with properties, the focus is very much on on the linkers. So people how refer refer to that as link erology. But we posed the question, what if we keep the rest of the molecule the same, but just choose and synthesize compounds that would have a different warheads? This was a wonderful collaboration with talkers. They synthesize a bunch of analogs where commonly used. Several warheads were incorporated in our JQ 1 based project and and then these compounds were screened in vitro by all these wonderful assays that that Christine was mentioning describing her stall in her talk. So this was done by Elizabeth Kane released about lots of in vitro BRD 4 or BET protein degradation kinetics. Just some of the example some of the data really is to showing here is this is an alpha screen. This is biochemical ternary complex formation and surprisingly really most of the compounds this warhead analogs shows relatively similar ternary complex formation. So AUC is very similar but very different degradation potency as you can see on this graph on top. Really it was surprisingly or satisfying satisfyingly for us it was this Compound 8 or our PG project. But interestingly, what we haven't seen really still very often we do see correlation between AUC indicating stability or formation of turning complex in our screen. I say often correlates with the building for protein degradation. But in this case we haven't seen that correlation at all. So no, not much difference in terms of AUC or turning complex formation, but quite considerable difference in terms of protein regulation. So, so the next again, wonderful work from Liz Kane from Mega looking into several an engagement by our products. So here you can see a comparison between a live cell and lysate IC 50S. So this is a chemical pro displacement nano bread assay and you can see that most of the compounds fell on a line here where the Lenalidomide is suggesting really there is no there are no issues really in terms of cell permeability. So the only outlier is the bit one which obviously it seems not so permeable as as well as the other projects here importantly we see the actually interestingly we see that live cells IC 50s correlate rather well with beautiful Dmax 50s. So the protein degradation is higher for those compounds that have a higher cerebone affinity. So and then when tested compounds in this couple of cell lines and 411 and HDMP 3 cells, as I mentioned early on, most of them show rather potency. They're all more potent the JQ one or DBT one. And then again, we found that for this rather still small data set, but there is a reasonable correlation for both cell lines between the IC activity IC 15 in terms of viability and with the BRD 4D Max 50. So again, very interesting observation. This is more that said as I say, but we're going to look out more details in the future potentially into this previously unobserved release of a better correlation between cerebro and binding affinity with with protein relation. Next, we profile the compound well this set of compounds in but proteomics unbiased proteomics in MB 411 cell cell line. So we can see our lead product. The PG based product shows by Western blot and hybrid rather high degradation potency. So the concentration we selected for this experiment was 100 animals and early time point and and conclusion really is that most of the compounds show as expected relatively similar profiles. So degrading bats family of proteins mostly The only exception. So maybe it's easier to see on this heat map very similar profiles just for this oxyzolon analogue we picked up ZBTB 21. This is a novel near substrate not previously reported, just suggesting really still how this cerebral on landscape continues to be increasing really as we are generating more proteomics data. So not much difference in terms of proteomics profile, but a considerable difference in terms of pharmaconetic profile between these compounds. Again, these 9 compounds be profiled by Pharm Kinetics. This is really only PO leg at four time points, 3 animals per time point. But you can see really that maybe that's on this right hand side graph where Auc's have been compared. So this is AUC for the levels of compound in the blood after four time points and say and 10 care administration, you can see basically that out of nine compounds only about four or five of them show oral exposure. Some of the most best performing in vitro compounds were not orally bioavailable, but we kind of like really slow that our PG warhead is performing very well. And we generally really still now in all our work focusing and using PG and for, for for just one example really stored how we use that really. I, I'll decide to take this most one of the more recent work on LC LCK product development and the reason for our interest in LCK really came up from an earlier discovery by one of my colleagues at Saint Jude that around 40% of pediatric TLL display in vitro and vivo sensitivity to the satanib. This was rather surprising because the Satanib obviously was developed as a VCR able inhibitor, but as many kinase inhibitors, it's rather promiscuous and in fact shows LCK rather high KD, but also inhibit inhibition of LCK. And my colleague June, he threw some really wonderful detective work figured out that this, this activity, LCK inhibition is what makes TLL cells sensitive to the satin. So you can see that it prolongs the overall survival in vivo. And and just to say briefly, LCK is cytosolic kinase, a small subfamily of of said solid kinase is associated with the T cell receptors. So T cell active receptor activation results in LCK plus relation. So hyperactivation of of LCK and T cells leads to ultimately probably proliferation and and results in the TLL which is the background to to this this disease really. So this discovery is now being tested in clinic now. So this afternoon. So there's a clinical trial they still running. We still don't know the results really, but we thought just saying if, if this afternoon based product would show any better activity than the something itself. So we designed a bunch of products and one of them like the best of those initial round was this project to which we refer as AJ1646. Of course it's a PG based product. We tested them LCK affinity for this compound, it shows pretty high affinity, just very similar to the satanic to LCK we showed Terry complex formation, a typical bell shape for a product versus the satanic there. So very stable high stability, sorry messing with this slide here. And ultimately we demonstrated a very potent LCK degradation. So this is, you know, Picomoro level degradation of LCK and we can also monitor PLCK with LCK is auto phosphorylates. So we use PLCK as a as a biomarker and I'll come back to that in the next few slides. So we have a highly potent and LCK degrader. The next question of course is really how selective it is. So this is a kinome for the Satanib and our PG product and you can see really very similar. So they both hit a number of kinases. But when you look into unbiased the proteomics data, then you can see really that mostly it's LCK degraded and with a couple of those direct analogs members of subfamily, subfamily up to 10,000 proteins or so identifying in this experiment. So we have a highly potent and selective degrader. So what does it do in vitro? So this is how dozen of patients arrived TLL cells and you can see in red our PG product, sorry the in blue our PG product in respect to the satiny shows considerable higher anti proliferative activity than than the satinib. And importantly, when we look into non cancerous cells like CD normal cells like CD 34 positive or PD and C cells, it shows no difference in terms of such toxicity and suggesting really still a much wider therapeutic index for our PG project. And we next tested the compound in vivo of course and you can see really in terms of leukemia burden. So this is the two models for two different patient derived models in mouse and you can see that our product in blue here considerably more reduces the leukemia burden compared to the satinibin cream and we are clean red and ultimately resulting in overall survival increased for our PG product in respect to the satinib. So this was a great, as I say really so early on coming back to this compound. It, it shows very good, excellent in in vitro potency. It did show in vivo efficacy as well, but only after IP administration. So where we were interested in so how we can improve its pharm kinetic properties for oral delivery. And again our optimization strategy was reducing the number of rotatable bones and policy this area and hydrogen bond donor specifically. So we designed a bunch of products around this scaffold focusing on the linkers and with the with the again in mind of reducing number of related to bones and pots of this area. I should say really that the difficulty was that mouse liver microzones and hepatosite housewives, well, they did not predict in vivo PK. So we had to test compounds in vivo and again we used our operated PK study, PK study generally testing compounds at 30 megs per cake or early and then measuring its four hours levels of the compound in the in the in the blood. And out of all the analogs really the best performer was the one that we identify early on. In terms of the linkers again it's it's the peridineal methylpeterazine linker and you can see the compound retained very good LCK degradation and anti proliferative potency and and shows very good oral bioability in mouse at least that's 58%. We demonstrated then that administration of this compound and it works in Viva as well. So it it shows near complete degradation of LCK in in a COP K1 cells administered to in in mice. So for for those chemists out there, so if you think you've seen before this linker, yes, of course you did. And this linker really is present in a number of venous clinical candidates and some others as well or some some similar kind of analogs of of what we call PMP short for repeteridineal, Matthew petridineal, Matthew comparison. And so we're it's emerging really. So this linker is now showing really still good properties in some other of our projects. And, and we started seeing potential, not just this one, but some others really as a emerging as a potentially privileged linkers. And, and these are the, some links that we try first now when we're exploring a novel project. So that brings me to the end of the story for today. Just to highlight that the vast majority of the greatest in a clinic are cerebral recruiting and they all contain glutaramide and glutaramide can often be chemically unstable. We found that venugutamide and female dihydroeosil, I didn't talk much about that really, but as well that they show considerably improved chemical stability and some other properties as well that ultimately result in them performing very well for in context over optimization for orderability. And really quite like now very often use this what we call now a PMPG building building block in, in context of exploring protectability for our novel projects. So with that, just to thank really this my colleagues, St. Jude, the chemistry of wonderful collaborators, ones for external collaborators really from MEGA. We couldn't do much of this without really support from from Mega and and Liz Kane and and Christine's really of data. So really thankful for that Biotechnet or talk risk synthesize the compounds and Neosphere really another phenomenal collaborator that produces proteomics data for us as Christine mentioned earlier. So I moved to now to ICR London where I will continue working on with my group on on the targeted protein degradation. We are interested in again continuously in product molecular glows expanding into novelty through I cases and anyone interested in a potential collaboration just give me a shout, would love to hear from you. With that, thank you in still being in the audience and again Amy and colleagues for the invitation. Great. Thanks so much. Lauren, I your picture is actually frozen on my screen, but we can hear you fine and you look fine. So no problem there. I'm we're happy to take some questions now. Zoran is still here, Chris, Kristen reaching is here and I believe Laura is still here as well. So any questions that anyone has in the audience, feel free to drop them in the chat or raise your hand and ask them verbally. I can start things out. So you know one for Zoran. So, you know, you talked about all the work you've done really optimizing these compounds for oral, oral availability. What are the remaining challenges you see there and and next steps for, you know, accomplishing this this task given these protects are, you know, beyond rule of five, not classical compounds. What are your thoughts? There, yes, it's, it's, it's fascinating really only a few years ago. So three or four years ago there have been this major conferences and there will be a whole session on are we going to be able to optimize projects for oral bioavailability. And and now that's rather common. It's not surprising we learn how to do that. It's incredible really the progress in that space remaining challenges of course, really there will be somebody mentioned early on selectivities really. So you know with the proteomics we can pick up, I mentioned early on ZPB 21, you know if this is your clinical candidate and you pick up really that normal new substrate there. So what do you do about that and how you assess the potential risks for clinical safety and so on Really. So that's what. So we will be hearing hopefully over the next year or two a lot from from those first wave of products and safety data coming back. So that's going to be a really exciting time and, and learning really still hopefully important learning ready for, for, for, for the whole field. And there there's huge amount of effort going on for E3 ligand over E3 ligases. So can we engage with the beyond? We had some discussion about, you know, cerebone and via child really. So what are other E3 ligases that can offer to us beyond what these two can offer like for instance, tissue specificity or or tumor specificity and things like that? So again, that's a huge area, lots of interest around that. And I expect over the next couple of years really still to, to hear more about that. Thanks. Yeah, I I think on the topic of, you know, E3 lagases beyond Sarah Blonde and VHL, you know what, what are you thinking? Maybe Kristen or Flora has comments, you know, as are you are you seeing some, you know, positive exploration in that space or where do you what do you think really, you know, how do we address that and really better understand how these how more E threes can be options for designing degraders? Yes, so it's it's it's it's huge conundrum really. So there are around 600 or so, some say more really 800 or so E3 leg is in human genome and we're exploring now in the planet at least really still mostly one of them. And, and the question is, OK, so how you pack which another one or two or or bunch really. So you start considering exploring for a for targeted protein degradation. So I've seen different approaches to this question. Our approach is that we first look for a, we would like to select ether ligases that are cancer relevant. We are focusing obviously on cancer. So we would like to identify those that are potential drivers, cancer drivers and and or at least a differentially expressed in tumor. So that's the first thing really first line of a kind of a selection. And the next very important as well is ligand ability. E3 ligases are complex molecules, very little known about biochemistry and biophysics really behind. So it's not going to be trivial really still kind of enabling any of one of them. So, so we would like to identify those potentially low hanging fruit in terms of legal nobility. So that's how we are approaching another approach really that people are also considering and and there's someone for examples already out there where you stabilize existing interactions between the physiological real EPA ligase with your targets of interest. That's another interesting approach that I think we'll see more examples in the future. Kristen or Fluor, any other comments? I think the, the only thing I would add, I mean, I agree with Zoran completely and I think there's huge potential to expand to additional E3 ligases. We've seen many successful examples of this already in the literature. But it's interesting that there, there aren't like clear trends like we've seen, you know, with Sarah Bond or even BHL emerge. So it's it's interesting to continue to watch the space to see, you know, if there are going to be even a handful of others that take hold for these types of approaches. Any other comments Flora? I think you guys covered it. I, I completely agree. And, you know, it's such a rapidly developing space that I think we're really excited to to see what comes out over the next few months. Thanks. Yeah. I think one question maybe Flora you would be able to comment on is, you know, as we kind of are, our AI and computational modeling tools continue to evolve and develop. You know, where do you kind of where do you see that role-playing in the future? And, you know, have you seen places where we already have a lot of promise coming from these, you know, these broad computational tools and AI type models that could facilitate some of this sort of, you know, next level of research? Yeah. I mean, I think this is a really exciting area for sure. If you have similar, for example, structures out there in the PDB or in a database, I think the structural modeling, things like alpha fold are really powerful in terms of docking and getting you structures that look fairly close to what we kind of see in the cryo M and crystal structure models in many cases. And similarly for these predictive models, I think where the gap is, is because it's a relatively new area where we're trying to do something new. Perhaps a protein hasn't been crystallized in a complex before, perhaps a new ligase or perhaps learning the rule of one of these new flavors of glues or one of these glues that targets a different E3 or a different component of the UPS. You know, to train machine learning models, to train docking models, you need high quality kind of large scale data that has positive and negative results. And so I think that's where all these resources that are being generated. I highlighted a few of them, but there are actually many out there on Bio Archive. Many sort of emerging in the literature are going to really move the field forward is by providing that uniform, large scale, unbiased data we can use to train these models. So I think it's going to be a really exciting area of growth for the next few years. Great. Thanks. This one's for Kristen. There's one question in the chat regarding looking at your different cell lines. Have you explored any options for like Multiplex detection of multiple targets within US the same cell line? Yeah, that's a great question. It's something we've thought a lot about. Can you hear me OK? It sounds like there's some feedback. Yeah, there's some feedback, but I I can hear you. OK. Yeah. And, and so there might be some ways we could consider that potentially coming down, you know, the pipeline here at Promega where we've actually got some new technologies that might enable that a little bit more easily and still retain kind of the same, you know, quantitation of degradation kinetics and all of those things. But yeah, it's, it's kind of limited to the availability of of tools. I mean, certainly you could tag a different protein with another, you know, you could tag with a Firefly potentially. We do have you know tools to assess dual manaloc or or high bit readouts in combination with a Firefly reporter that would be strictly endpoint. There may be other ways you could combine different reporters as well. We haven't explored a whole whole lot of that, but there is as I mentioned some potential new technology that might, might enable that type of multiplexing. Great. Thanks. So I just had one question, you know, what are your maybe your respective thoughts on combination therapies that use degraders alongside of more traditional small molecules or biologics? Is this an area that you think is going to open up a lot of new potential is kind of combining these together or you know, using them maybe when you reach a a point of resistance with the traditional small molecule or how do you see Protex kind of playing into the that space or, or glues or any types of degraders? I can start, I guess So what we realized, but when I say we really so the, you know, cancer drug discovery community and and clinicians of the past decade or so is that that it's ultimately the cancer is not going to be treated by silver bullets. There's no silver bullet cancer. And really we are realizing in again in over the past decade or so really still the tremendous challenges in terms of cancer heterogenicity and, and the the mutations of course, really the resistance development and so on. So it's, it's more than clear that it's going to be combination, different combinations that would targeting presumably different mechanisms that would would drive cancer. And it's it's an area that is really being active for inhibitors in recent years. People are now thinking about beyond two or three combinations and and pretty similar like in viral I can see really in in for tuberculosis or or HIV real is to where multiple inhibitors acting on the different mechanisms of action would ultimately show considerably better efficacy and and address the resistance. So, so that's going to be the way ready to go and and understanding the mechanism of resistance. So this is this is another part of really still coming back data coming back from clinical studies where this resistance is going to develop in what form really. And then considering that following that up with a rational design combinations and by all means, it's going to be really important just as much if not more really for a, for a. Great interest. Thanks, Aaron. Well, I'm not seeing any other questions coming in from the audience. So I think I'm the one that's got the feedback. So I looks like we have addressed everyone's questions. I want to thank you all for your time today. Really great presentations and discussion. I learned a lot. Hope everyone did too. Just highlighting that we do have two other days of our virtual event. So our first day was yesterday that is available on demand. And tomorrow we have a transitioning to a really interesting topic, topic of unlocking the potential of RNA as a drug target and some of the really innovative work that's happening in that field. So be sure to join us if you're interested in that. And thank you all for your time, time and attention. Just highlighting if there's any questions you have following this presentation, you can contact the presenters or also feel free to reach out to our Tech Serve team who is always happy to help. So thanks again and enjoy the rest of your day. Thank you. Bye. Bye. Thanks so much. Thank you. Yep. Bye, bye. Thank you. _1733989949557

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 2: Next-Generation Therapeutics: The Power of Induced Proximity and Targeted Protein Degradation

Efforts to develop treatments for previously "undruggable" targets have launched a new phase in drug discovery, highlighted by the rise of next-generation therapeutics. Induced proximity, a strategy manipulating molecular interactions for desired outcomes, and targeted protein degradation, leveraging the cell's machinery to degrade pathogenic proteins, are forging new treatment paradigms and redefining our ability to modulate previously intractable molecular targets. 

In this session we will hear from leading experts as they present the latest research in this fast-moving field and learn how these cutting-edge strategies are transforming how we are tackling diseases once considered beyond reach. Join the conversation and engage directly with the experts in a live round-table discussion following the presentations.

_1733989949777