So, good morning, everyone. We welcome you all in person and virtual attendees to this symposium, Co-Clinical Models to Interrogate Neoplasia and Proneoplasia. First, we would like to thank the Endocrine Society and Dr. Hydera Del Rivero, who's in the audience here, of the National Cancer Institute at the NIH for organizing the symposium. My name is Sunita Agarwal, Dr. Kate Lyons, and I will moderate the symposium. I'm a basic science investigator with funding from the Intramural Program of the NIDDK at the NIH in Bethesda, Maryland. I have a long-standing interest in investigating the genetics, epigenetics, and biology of tumors associated with MEN1. I will now hand over the podium to the co-chair of the symposium, Dr. Kate Lyons. So, good morning, everyone. So, yes, I'm Kate Lyons, so I'm a research fellow, so basic scientist working at the University of Oxford, and my research is centered around novel therapeutics for pancreatic neuroendocrine tumors predominantly, with a particular focus on MEN1 and epigenetics. And it's my great pleasure to introduce the first speaker in this session today. So, the title of the first talk is Notch One Signaling Pathway in Neuroendocrine Tumors from Pathway Analysis to Personalized Medicine by Dr. Herbert Chen. So, Dr. Chen is Chair of the Department of Surgery at the University of Alabama at Birmingham and the Surgeon-in-Chief of UAB Hospital and Health System. So, Dr. Chen's lab has been continuously funded for over 21 years, and his team studies the role of notch signaling in thyroid and neuroendocrine cancers, and Dr. Chen has also been recognized as a champion for diversity, equality, and inclusion. So, Dr. Chen, please start your presentation. Great. Thanks so much for having me today. I'm so glad to be here. So, I think we're going to have a focus on neuroendocrine tumors today. Since I'm the first speaker, I'll just briefly, for those of you who don't take care of these patients regularly, neuroendocrine cancers are hormones creating neoplasms that develop within multiple organs, including the lung, pancreas, GI tract, ovaries, and thymus. They're a diverse group, and often they're called by other names such as carcinoid, pancreatic islet cell, medullary thyroid cancer, pheochromocytoma, and although their incidence is relatively low, between 5 to 15 per 100,000, they have significant clinical manifestations, and they are the second most common tumor to metastasize in an isolated way to the liver. And so, as surgeons, we see a lot of patients with the disease, as surgery is the only curative treatment right now, and we see a lot of patients who present to us with metastatic disease for which sometimes there's a surgical option, but many times there is not. So the topic of my talk today is a pathway that I've studied for a few decades now, the Notch signaling pathway. And our interest began, my interest began in Notch actually as a postdoc at Hopkins working with Matt Ringel here, who's in the audience. And basically, there are four Notch receptors, Notch 1, 2, 3, 4, and they are very similar in structure, as shown in this picture here. And they bind various ligands, including delta-like, DLL, JAGID, as shown here. And they have diverse functions, both in development and many processes, but in particular, in cancer. And they've been well-studied in many cancers, and I'll shed some light on some of the work we've done in Notch signaling in neuroendocrine cancers. Many people are familiar with how Notch signaling initiates, but I'll just, for those who are not, as mentioned before, Notch is a full-length membrane-bound receptor. And when one of the ligands that I had mentioned before, delta, JAGID, when it binds the full-length Notch receptor, it initiates a series of three proteolytic cleavages, mediated by the three enzymes listed there. And what that results in is freeing of the Notch 1 intercellular domain, the NICD, or the active form of the molecule. It's released from the cell membrane, then translocates to the nucleus, where it binds its binding partner, CBF1. And together, they target various Notch response genes, such as the ones listed there, the most commonly known as HES and so forth, and initiate a series of different molecular events. So our interest in Notch, as I mentioned, goes back a long ways. And first of all, we were very interested to see what the role of Notch 1 was in neuroendocrine cancers, particularly medullary thyroid cancer to begin with. So this is a very old Western blot, which looks at the amount of active or cleaved Notch 1 NICD, shown in the top row. And then along with the measure by Western of chromogranin A, and ASCL1, which is also a neuroendocrine marker. And you can see in this Western blot, the, is there a pointer, just, oh, okay, thank you. Oh, use the screen. All right, very good. Oh, here it is. So TT is a medullary thyroid cancer cell line. So shown on this Western blot, you can see that all the medullary thyroid cancers here, MTC, and these are human samples that we obtained from the operating room, all universally lack the active or cleaved Notch 1, and confirming their neuroendocrine tumors by the high presence of chromogranin and ASCL1. And we did have a group of control tumors, papillary thyroid cancers, which actually do have levels of Notch 1, as shown here by Western, and are not neuroendocrine tumors. So this observation in medullary thyroid cancer, which we also observed in other neuroendocrine tumors, showing that we analyzed, I think, about 100 tumors, that we could not detect any active Notch 1 presence, led to us to make our first hypothesis, is that perhaps the lack of Notch in these cancers suggested that maybe that Notch 1 was a tumor suppressor. So then, of course, the next set of experiments done was to then ask the question, what would happen if we reintroduced the active molecule of Notch into neuroendocrine cancer cells? So we took the medullary thyroid cancer line, TT, and we stably transduced a doxycycline inducible Notch 1 construct, or an NICD construct. And shown here is the Western blot on the very top here of that cell, the one that's been transduced with the inducible Notch 1, NICD, and you can see, as you add doxycycline, you get a dose-dependent increase in the expression of Notch 1 within those cells. And then we have a vector control shown here on the right. And then in panel B was the effect that Notch 1 had on cellular proliferation as measured by an MTT assay and also a viable cell count. And shown on the graph on the left, you can see the control line, which is the top line, without any doxycycline present, the cells proliferate in their normal manner. But as you increase the dose of doxycycline, you get a dose-responsive suppression of proliferation to the point where when you treat the cells with the 0.8 micrograms per mil of doxycycline, the cells essentially stop growing. And shown on the right is the control with the vector only, showing that the doxycycline really has no effect on cells. So we did this in a variety of cell lines, and it was almost universal that whenever we induced Notch 1, the neuroendocrine cells stopped growing. So that led to our hypothesis that maybe Notch 1 is a tumor suppressor in neuroendocrine tumors. And the work that I showed you was in medullary thyroid cancer, but we subsequently did work in some of the carcinoid lines and also in pheochromocytoma lines and showed it was whenever we induced Notch, the cells stopped growing. But what's so interesting, so then we said, OK, Notch has to be a tumor suppressor. So let's start looking for some mutations which show the Notch is inactive. But we did a mutational analysis, and we found no Notch mutations in any of the tumors. So at that point, it should have given us some pause. But we thought, well, maybe there's another way it's inactivated, maybe it's not. But let's go with our initial hypothesis that Notch 1 is really a tumor suppressor and think of a way we could use that knowledge therapeutically. And so we began trying to think of ways, is there a way we could activate Notch in neuroendocrine tumors as a therapeutic target? So our next step, and this is when I was a faculty member at the University of Wisconsin, we then decided, well, let's see if we could screen for any small molecules which could activate Notch in neuroendocrine cells. So we used the resources at the Small Molecule Screening Facility at Wisconsin, where we analyzed a little over 7,000 compounds from the known bioactive library and NCI libraries 1, 2, and 3 of compounds which had been tested in humans, had a safety profile, but perhaps were used for other diseases or not. But then with the goal of finding a compound which was safe to use in humans that we could hopefully activate Notch with. So this is some of the results of our study, which we did a primary screen and a secondary screen. And this is a secondary screen of the 25 compounds with some Notch 1 activating activity. And how we did this assay is we created a cell line which had a luciferase reporter that was fused to the CBF1 site, which Notch binds. So if we add the compound which causes Notch to bind the CBF1 construct, it would cause luciferase production as shown here. And of these 25 compounds, and again, we screened over 7,000, the one that had the most amount of luciferase activity was VP or valproic acid. So having found this, we then took, of course we wanted to confirm that that's what indeed happened in cell lines. And so here we took the GI neuroendocrine tumor line, the bond cell line, treated them with various doses of valproic acid. And you can see on the western blot on the left that increasing doses of valproic acid caused an increasing expression of the full length and the active form of Notch 1 NICD. And it correspondingly inhibited ASCL1, which is what active Notch does. So confirming that it was indeed inducing Notch 1 signaling and hitting its downstream targets. And then we did a xenograft model where we actually put bond cells in nude mice and treated them with doses of valproic acid, which we would use to treat epilepsy with. And shown that on the graph above, tumor volume increases in the control situation, but essentially, but not quite, to stay suppressed in the animals treated with valproic acid. So we wanted to go to a clinical trial. So we indeed did a clinical trial of pilot phase 2 study with valproic acid for the treatment of low-grade neuroendocrine cancers. And the graph on the small trial, but the graph on the right shows a progression of free survival. But essentially, the valproic acid really had no clinical effect, which sort of stumped us because we had seen such a significant growth suppression in cell lines and in animals, which obviously we were disappointed with at the time. So after we did that trial and after we began to think, we really questioned what the role of Notch 1 is and really went back to the original question, why are there no Notch mutations? And if it was a tumor suppressor gene, why did every single cell line that we looked at retain the ability to induce the native Notch 1 signaling? And so we asked the question, there must be some important role for Notch 1 in these cancer cells that it wouldn't delete or stop expression of this active protein, which appeared to stop cellular proliferation. And so when we went back to this original study, you can see that we only did this original study for about six days, and we terminated the experiment. So we then did this experiment again, where we actually kept the experiment going, but then we withdrew the doxycycline from the cell lines. And what we found is the cells were actually still alive, that they started to grow normally after we withdrew the doxycycline. So really, what the cell was doing is using the Notch as just a switch to turn on and off proliferation. And they weren't killing the cells at all, but they were all still viable. So then we then asked the question, OK, if Notch then has some important role, let's try to get at that. And then we decided to use CRISPR-Cas to then knock out the Notch 1 gene in the neuroendocrine cell lines, and we did the bond cell line first. And what we did is we went through a number of clones. For those of you who have done this, knows it's a little bit of a process to find that clone with the gene missing. So obviously, we did a number of clones. We did the sequencing to confirm that the Notch gene had been made inactive. But we also confirmed, because we knew, as I showed before, that when we treated the cells with valproic acid, they could induce the native Notch 1. So shown on the left here is just a regular wild type bond cell, which we treat with valproic acid. And you can see the Notch 1. And you can see in our knockouts there, clone 2, 4, and 8, especially clone 2, there's no Notch present after treatment with valproic acid and sequencing, and did confirm that this gene was inactivated. So we created this Notch 1 knockout clone. And then maybe to our surprise, the cell looked perfectly fine. And so then we began, OK, let's see if there are any different behaviors that we observe in the cell. And so one thing we observed with bond cells when they grow in culture, and this is shown very heavily seeded, is by 14 days, although you can see the tissue culture plate is completely packed, the cells sort of pile up on each other and continue to grow. But in the Notch knockouts, they completely stop growing when they reach contact inhibition. And that is sort of shown here on the graph on the A point, where you can see that the wild type cells continue to proliferate, where they totally stop in the knockout shown in the solid black bar on the bottom. And then at that time, we decided to measure what the Notch expression was. And so Notch expression in the wild type, when they become over-confluent, was markedly enhanced, as shown by the green bar there. We also showed that in the knockout using a migration assay, where wild type bond cells are pretty, have the ability to migrate, the knockout cells did not. And they also had reduced expression of slug and snail. Interestingly, at the same time, we were studying resistance to treatment on a totally separate project trying to treat bond cells long-term with everolimus to see if we could establish a drug-resistant line. And this was one of our collaborators, and he just happened to look at the Notch expression and found that in his resistant cell line, which had been treated chronically with everolimus, Notch 1 was also being induced. And then treating cells with rapamycin the same way also showed that when you had rapamycin-resistant cells both established with a low dose of rapamycin and a high dose, that they markedly increased Notch as well. So when we then examined the wild type bond cells with the Notch 1 knockouts, when they treated with various compounds, such as an mTOR inhibitor, or actretide, or wild type bond cells, the IC50 for rapamycin is about 43 micromolar. The knockout cells were much more sensitive to any type of compounds that we treated with, such in this case there, IC50 was only 3 micromolar. And you can see normally, and I'm sure many of you have done this, when you treat bond cells with actretide in culture, they basically do nothing. But the knockout cells did have a reduced viability over time. And again, looking at wild type versus knockout cells and comparing their expression of a cancer stem-like features, such as expression of ALDHA1, CD133, CD44, SOX, and OCT4, the knockout cells had all significantly lower levels of these markers of cancer stem-like features. And that was also represented on the graph on the far right. So this got us back to our original question. We had, decades ago, thought that NOTCH1 was actually inhibiting the proliferation and trying to kill the cell, but maybe what, after these latest observations, maybe what NOTCH was doing is promoting cancer cell actually survival during stressed events, such as treatment with cytotoxic agents, potentially reducing the proliferation to protect the cell against these agents. And that NOTCH was promoting stemness and perhaps drug resistance, as shown by its increased expression in our drug-resistant cell lines. So then we actually asked the opposite question. Should we, in fact, inhibit NOTCH instead of activating it? This is the same pathway that I showed in the very beginning. But what we, as I mentioned, the activation of NOTCH occurs after cleavage. And so there are many inhibitors to NOTCH1 activation, such as gamma secretase inhibitors, which inhibit one of the cleavage steps in the process. And therefore, we decided, well, let's see what happens when we treat cells with gamma secretase inhibitors. So both we did in in vitro studies, shown above, where we treated BON and QGP cells, both with lenreotide and a gamma secretase inhibitor, nirogastat, and then looked at viability. And you can see the addition of a NOTCH1 inhibitor markedly reduced proliferation and viability in both these neuroendocrine cell lines. And we also did experiment in vivo using xenografts from both cell lines, BON and QGP, and basically confirmed when we treated the animals, tumors with, that were treated only with lenreotide or nirogastat, basically had the same proliferation. But when we treated with both together, the proliferation was much less. So in summary, I've captured a little bit of the work we've done over the years. But really, I hope that you think that NOTCH1 signaling is involved in the pathogenesis of GI neuroendocrine tumors, as well as other neuroendocrine tumors. But I think that therapy resistance may be associated with NOTCH1 activation and induction of stem-like cells, that NOTCH1 signaling promotes maintenance of a cancer stem-like cells from within, that down-regulating NOTCH1 with gamma secretase inhibitors results in decreased stem-like features and sensitization to somatostatin analogs and to mTOR inhibitors. And sort of the bottom lesson that our group learned from all of this work is that I guess that we probably initially should have dug a little deeper when we found our initial observations, which led us, perhaps, to pursue some research that didn't turn out exactly the way we thought it would be. And I think that with hopefulness, we're actually pursuing the opposite direction, which I think hopefully will, again, get us to the target of trying to find some type of new therapies for patients with metastatic neuroendocrine tumors. Just want to acknowledge all of our collaborators listed on the left, our past and present funding on the right, and as well as our neuroendocrine research group at UAB. Thanks so much for your attention. Yes. Thank you so much for a really interesting presentation. So we're now open for questions. So for those of you online, please go ahead and submit any questions you have, and we should get those coming through here. If you're in the room and have any questions, please come up to the microphone. Or whilst you're walking, I can ask a question. I was wondering whether you had any indication of how, say, like the genetic background of the tumors could be influencing the Notch signaling pathway. So for example, patients with MEN1, is Notch going to behave differently to say if you've got an ATRX, DAX, or mTOR mutation? Yeah, no, I think those are great questions. So we have not, so the results that I showed with our strategy to inhibit Notch, as opposed to activating it, are really new, and a lot of what I showed is unpublished. And so we've not gone back to then look at, to see any associations with the tumors themselves and the function of Notch, but that is something we're obviously very interested in. Please go ahead. Nice talk, Herb. So your earlier studies with the valproic acid treatment worked pretty well in vivo in the animal models. I'm kind of curious, given your hypothesis, that you're actually, you're probably doing two things. You might be inhibiting proliferation, so that's what you observed in the animal model, but ultimately you're up-regulating your stem cell-like features. Do you think that if you went longer in your animal model, and I don't know, I don't remember how long you went with the valproic acid treatment, if you went longer, you would actually see the development of resistant tumors that actually are more aggressive than the original non-treated tumors? No, I think that's a great question, Dawn. We have not done sort of longer animal experiments than we were thinking of doing it. It's hard to, I showed just our xenograft model where we put the tumor sub-Q, which we could potentially do it there. With our liver meds model, the liver meds get quite extensive, and so we often have to sacrifice those animals within a finite period of time, but we could definitely try to do a long-term treatment with valproic acid, and I think that would be a great study to do. Yeah, I'd love to see your secretase, gamma secretase inhibitor as well in some metastatic models to see whether that would actually be really beneficial in blocking that metastatic phenotype. No, I totally agree. I think that we, as you know, we had put that in for part of our support proposal, and so we'll see how that goes. Good luck. Thanks. I have a question. So, really nice work on NOTCH1, how it acts as a tumor suppressor, and the regulation of NOTCH1 in endocrine cells, so any thoughts or evidence as to what is responsible for this NOTCH regulation, and do you think that would be useful to understand what drives the resistance to drugs that over-regulate NOTCH1, I mean, up-regulate? Yeah, so I think in different tumors, NOTCH has been shown to be a factor that promotes proliferation of pancreas cancer a whole bunch, and it's also been shown to increase stem-like features in other cancer cells, so I suspect it's similar, but it's so interesting, it's totally different than what happens in the other tumors where NOTCH appears to induce stem cell features, but also induces the ability to be, have a more aggressive phenotype and proliferate more rapidly, but in our instance here, it almost does the opposite, and so the only way I can sort of think about this is that neuroendocrine tumors are sort of very different than other type of cancers, right? They can, their pace of growth can be very slow in the presence of metastatic disease. Most of the time when you have metastatic disease, the pace is pretty fast, right? When a patient develops metastatic disease, they tend to, but neuroendocrine tumors, as we know, patients can live a long, long time with metastases, so, and this is just me conjecturing up here, is that maybe what the cancer cell is doing in this instance is it's NOTCH is promoting survival, but it's also, I would think that if the cancer cell kills the individual, that's not good for the individual, they're not good for the cancer cell, right? So maybe it's some type of, I don't want to call it a synergy, but some type of way that it's allowing the host to survive so it survives, that's just. It's more a question of the mechanism coming from basic research background. Is it histone acetylation that is being affected, or histone methylation, or is it DNA methylation, or phosphorylation, or anything like that? So we know the way that valproic acid is inducing NOTCH, at least we studied it, it appears to be acting, we published a paper on it that I didn't really talk any about, but there's a specific area within the NOTCH1 promoter that valproic acid is acting at, that if you eliminate that area, so it seems to be a transcriptional event. Thank you, yes. Perfect, well, thank you very much again. I think we, in the interest of time, we'll move on, so yes. So the title of the second talk in this symposium is Novel Therapies, Targeting the RabL6A-RB1 Axis for Neuroendocrine Tumors by Dr. Don Kell. Okay, let me correct myself, and I'll just read my notes. So it is Dr. Don Kell, that's our second speaker. So it is Dr. Don Kell, that's our second speaker. So Dr. Kell is a professor of neuroscience and pharmacology and professor of pathology at the University of Iowa Carver College of Medicine. Her research is funded by several multi-PI grants from the NIH NCI. Dr. Kell's team is working with clinical colleagues to translate basic science research discoveries into new treatment options for patients with neuroendocrine tumors and malignant peripheral nerve sheath tumors. So Dr. Kell, please start your presentation. Thank you very much, and thank you to Heidi and the organizers for this invitation to talk to you today and tell you about our studies that identify new therapies for targeting and treating neuroendocrine tumors by targeting the RabL6A-RB1 or retinoblastoma tumor suppressor locus. I have no disclosures. So we all know that neuroendocrine tumors are challenging cancers. They are slow growing, they resist standard chemotherapies, and we know that we need new combination treatments for eradicating and combating advanced disease. And to do that, we really need to better understand the different genes and pathways that are driving the development of this disease. So for today's talk, I'm gonna focus mainly on pancreatic neuroendocrine tumors. There are a number of really good treatments for patients with PNETs or pancreatic NETs. And besides PRRT using radiolabeled somatostatin analogs, we have targeted kinase inhibitors that include those against AKT-MTOR, receptor tyrosine kinases like VEGFR that promote angiogenesis in the tumors. And in recent clinical trials, there have been studies of inhibitors of CDK4 and 6, cyclin-dependent kinases 4 and 6. Unfortunately, with these inhibitors, they are only modestly effective, and we know that patients will invariably or inevitably progress once the tumors become unresponsive. So the key is is that we need either new treatments or perhaps novel combinations of existing drugs to combat these tumors. This slide is a good overview of what I'm gonna talk about today, which is our approach. It's an integrated approach to discovering new drugs and or combinations of drugs that could be used to treat PNETs. And so we've taken the approach where we're combining information gained from genetic profiling of patient tumors, mechanistic studies performed in cell lines, as well as cell-based kinome and phosphoproteome data that tells us what are the protein changes that are occurring within PNETs. And we hope to combine that with ongoing proteomic profiling of patient PNETs in order to narrow down and refine our list of factors that we think are most associated with the pathogenesis of these tumors, and in that way refine our list of tumor-treating drugs in order to have a more reasonable number of drugs that we can subject to testing. So we've begun by first looking at patient tumors and conducting profiling. And this is a collaborative study with Jim Howell's group at the University of Iowa where they performed RNA sequencing on a large number of patient tumors, both the metastatic lesions in the lymph node and liver as well as the primary tumors. And we subjected that data to Ingenuity Pathway Analyses, or IPA, and identified a large number of key pathways and targets that are overactivated in the metastatic lesions. That's a metastatic gene signature. And some of those were expected, like PI3 kinase, mTOR, and HDAC activation, as well as some newer targets and pathways like cyclin-dependent kinase 2 and NF-kappa-B. We also subjected the data to something called Connectivity Map Algorithm, which is outstanding for predicting drugs that will be effective against your tumor cells based on their transcriptome. And we identified a fair number of FDA-approved drugs that target a number of these factors, like CDKs and HDAC. And then we confirmed and validated the efficacy of those drugs using the two pNet cell lines that Herb just talked about, BON1 and QGP1. So something else that came out of that RNA sequencing was a RabL6A gene expression signature, which indicated that it was upregulated or overactive in the metastatic lesions compared to the primary lesions. And RabL6A mRNA was upregulated, as well as its pathway, which is indicated by the circle around the gene. So anything in red is upregulated, and anything with a circle is indicating that the pathway is also activated. So we had RabL6A upregulation, as well as its downstream oncogenic effector pathways, like AKT, mTOR, and MYC. And we were interested in RabL6A because it is a new Rab-like oncogenic GTPase. There isn't a lot known about it, but it is a marker of poor survival in a number of different cancers, including breast, esophageal, lung, and additional cancers, as well as pancreatic ductal adenocarcinoma. And this is a study we did a while ago where we looked at resected tumors, and we found that tumors that lacked RabL6A connoted a much better survival for those patients compared to those patients whose tumors had any level of expression, low or high, of RabL6A. The other thing that we noticed by conducting immunohistochemical staining of tumors was that grade one and grade two pancreatic nets expressed very high levels of RabL6A, as did non-functional insulinomas. And we also saw, by doing knockdown studies of RabL6A, where we used two different shRNAs against the gene, called KD1 and KD2, that when we get rid of RabL6A, we significantly reduce the proliferation of the cells. And this is over a sustained period of time, over three weeks. Conversely, if you overexpress RabL6A, and we then put the cells back into mice by xenograft implantation, you can see that RabL6A overexpression promotes the growth of bond one xenograft tumors, compared to GFB control. In other studies, we found that RabL6A is important for promoting the angiogenesis of pancreatic nets. And we did this by using the well-established RAT insulin promoter, or RIPTAG2 model of insulinoma. So my student, Chandra Maharshan, used the RIPTAG2 model, and for those familiar with it, it's an outstanding model where you have hyperplasia in the islets, and then you have angiogenesis that's occurring within a defined period of time. And ultimately, you get malignant tumor formation. And these are insulin-secreting insulinomas. What we did was we crossed those animals with RabL6A knockout mice, and we wanted to test the idea that loss of RabL6A would slow down or reduce the formation of tumors. And in fact, that's what we did see. I'm not gonna go through all the data, but just summarize it, that RabL6A loss reduced islet cell tumor size and rate of formation. It slowed the development of angiogenic islets, and it also improved the survival of this insulinoma model. So that was really powerful in establishing the physiological or biological importance of RabL6A in PNET pathogenesis and the relevance of targeting it in tumors. Now, we don't actually have drugs that target RabL6A, so we're going to talk about how it functions so that we can try and target downstream regulators of RabL6A. So my postdoc, Salma, did a beautiful job in leading these studies at trying to determine how does RabL6A actually promote the progression of PNETs? And I'm gonna summarize a number of different studies that we've conducted where we found that RabL6A controls multiple clinically relevant pathways. So the first is that we found RabL6A down-regulates CDK inhibitors, and as such, activates cyclin-dependent kinases four and six. And we can tell that by looking at the phosphorylation of the retinoblastoma target. And when we have RabL6A, RB is phosphorylated, which means it's turned off. But when we get rid of RabL6A, we accumulate hypophosphorylated RB, which is growth suppressive. So that's a really important and very relevant drug-targetable pathway. We also identified AKT mTOR signaling as another target. And so when you look at AKT phosphorylation, it can be regulated by the kinases that put the phosphate on AKT, but also the phosphatases that remove it. And we found that when we knock down RabL6A, we lose phosphorylation on AKT. That's consistent with the cells stopping growing. But if we actually inhibit PP2A by using okadaic acid, we restore that phosphorylation of AKT and in the control cells even to higher levels. This shows us that RabL6A normally inhibits the PP2A tumor suppressor, and that that contributes to the activation of AKT and mTOR. That's pretty exciting because there are some new drugs that are being developed by a collaborator of ours at Michigan, Gotham NARLA. These are small molecule activators of PP2A, which instead of inhibiting kinases, what we're trying to do is activate phosphatases. And in so doing, you would think that you would then down-regulate AKT mTOR signaling. So we have done that in vitro and shown that that works. And I'm just gonna skip to the end part and show you that it also works in vivo against xenograft tumors. So compared to control tumors that grow quite rapidly, in the blue curve, you can see the activity of activating PP2A, and that's really quite robust and much better actually than inhibitors of AKT alone. So this was exciting, but we've also noticed that RabL6A has many targets. And in fact, gene sequencing studies would suggest that there are over 700 different genes that are dysregulated when you alter RabL6A. And we noticed that MYC was one of them. And so we looked at the MYC protein, and we found that if you get rid of RabL6A, you also lose the expression of the MYC oncoprotein. And that's pretty compelling. So MYC would be another target of RabL6A, which is clinically actionable. And we verified that by looking at normal islets versus knockout islets and showed that MYC transcription, and in other studies, the protein, are downregulated when you lose RabL6A. And this is relevant clinically because there are drugs that will affect MYC activity, and one of those is CPI-203. And you can see in the black line with the filled circles, these are controlled bond cells or QGP1 cells, and they respond pretty well to drugs that will inhibit MYC. But if you get rid of RabL6A in the two open symbols in those curves, you can see that the cells no longer respond very well, and they survive better in the presence of a MYC inhibitor. So RabL6A controls a lot of clinically relevant pathways. This is important for PNETs because patients are treated with mTOR inhibitors like Everolimus, as well as angiogenesis inhibitors, like those targeting VEGF receptor, and clinical trials are evaluating monotherapies against CDK4 and 6. Now, unfortunately, we have modest efficacy of those current drugs, of mTOR and VEGFR inhibitors, and the monotherapies against CDK4 and 6 have failed in the clinical trials. So we want to take what we've learned from our studies to hopefully use that knowledge and develop other targets, or perhaps other combination therapies that we could target these tumors. And that's really our hypothesis, is that we can take multiple other targets that are regulated by RabL6A, and if we inhibit several of them together, we will have synergistic antitumor activity. So our idea is the monotherapies to CDK4 and 6 are not working, but what if we combine the inhibitors of CDK4 and 6 with perhaps inhibitors of MYC, or inhibitors of MEK. I haven't shown you that RabL6A also activates MEK signaling, or activators of PP2A. And I'm really gonna focus on MYC and MEK for this talk. So two other lines of evidence also supported this idea to focus on these combination therapies, and that was data from our PNET kinome studies, and our PNET phosphoproteome studies. So we took cells that express RabL6A and are growing, versus cells that are arrested because we got rid of RabL6A. And we performed global kinome analyses with our colleagues at UNC, because we can't do that, they're great at doing it. And we identified a number of kinases shown by those red dots that are regulated by RabL6A. And excitingly, two of those 16 hits were cycle-independent kinases 6 and 16. And these are both regulators of RB1. We also did phosphoproteome analyses with our UNC colleagues in the same cells. And we subjected the data to IPA analysis, and we identified many of the same targets that we did before. So CDK4 and 6 was a major regulator that was pulled up by the phosphoproteome analyses, as well as some of these other targets that I've talked about, like MYC and MEK and PP2A. So we've used that data, this omics data, and the mechanistic data, and the kinome and phosphoproteome data to identify and refine our list of potential drugs for treating PNETs. And we were thinking, okay, palbociclib is topping this list. And we know that they're not working very well as monotherapies. They are not effective in treating PNETs. But what if we combine them? So we wanted to evaluate palbociclib combinations with inhibitors of other RABL6A regulated enzymes. So here you can see synergy plots, 2D and 3D, of low doses of palbociclib with the MYC inhibitor CPI-203. And anything in red indicates synergy. And you can see pretty good synergy between these drugs. We also combined palbociclib with a MEK inhibitor. And you likewise see terrific synergy in cultured cells. And this is in both BON1 and QGP1 cells. So that made a lot of sense, because there's accumulating evidence in the literature that this should work. The data have shown that if you inhibit MEK, which is an upstream activator of CDK4 and 6 kinases, as well as use direct antagonists of CDK4 and 6, you synergistically reactivate RB1, which leads to tumor cell stasis through senescence. And that in vivo will lead to an antitumor immune response, where you get increased activity and infiltration of T cells in the tumors. And we thought MYC inhibitors should also work, because it's downstream of MEK and upstream of cyclin D and CDK4. And that's what we evaluated here at the biological level. You can see that the individual drugs alone, shown with palbociclib or the CPI drug, lead to some cell cycle arrest. But the combination in red is synergistically more effective. And we also saw greater cell death with the combination. That correlated at the molecular level with significantly reduced expression of MYC, which is what we would expect to see with the combination of drugs, as well as reactivation of the RB tumor suppressor, which you see by hypophosphorylation compared to the individual drugs. So we took it in vivo, and we used bond I xenografts. We want to use a lot more models, but models have been limiting in the field. Here we have bond I xenografts, and the controls shown in the black circles grow very rapidly. But you can see that the individual drugs have a modest antitumor effect, but the combination is better. Still, it was only a two-fold extension in survival with the combination. So we were kind of disappointed by this combination of palbociclib and the MYC inhibitor. And we thought, what if we went back to this MYC inhibitor and palbociclib combination? And in fact, again, in vitro, we see good synergy in both cell lines with this combination. And we saw reactivation of RB, as indicated by the hypophosphorylated form. We also saw significantly increased cell cycle arrest with the combination, and much greater cell death with the combination. We also examined long-term survival of these PNET cells treated with the drugs by using clonogenic survival assays. And you can see there's a lot less survival with the combination. And that correlated with not only fewer colonies, but also smaller colonies. So again, we took this in vivo in our bond one xenografts. And you can see that there is a marked delay in the progression of these tumors. So instead of just going out to 30 days, like we did with the MYC inhibitor combination, now we're going out to 120 days, where these tumors, they aren't eradicated, but they are delayed. And I'll point out that these are immune-deficient NSG mice. So we don't have an intact immune system in these xenograft models. But we were pretty darn excited to see a six-fold extension in survival. So ongoing and future studies. We want to expand our analyses. We've been getting some nice data in patient-derived spheroids through combinations with Jim Howe and Po Hin Ear at our institution. We're also, we want to get in vivo and do some PDX model treatments. That's ongoing. And my research technician, Courtney, has developed some in vivo metastasis models, which we recently published on with luciferized QGP1 and bond one cells. And I don't have the data here, but actually we're seeing pretty good synergy, anti-metastatic activity of the drugs against bond one metastases, and moderate activity against QGP1 metastases. We also want to use this combination in immune-competent mice that develop de novo PNETs. And we're collaborating with Steve Labudi and Zhijing Wan at Rutgers, where we're going to be doing the treatments with the CDK4 and the MEK inhibitors. And in fact, it's having dramatic effects, so we're pretty excited by that. And I would say that we have some compelling data in another tumor model, sarcoma model, of MPNST, malignant peripheral nerve sheath tumors. My students showed that this combination, shown in purple, will actually lead to tumor regression in immune-competent mice, whereas the single drugs alone do not cause regression. And we're excited that in our immune-competent mouse model, of PNETs, we will see something similar. And in fact, in the MPNST model, we have evidence for activation of adaptive immunity in the tumors that regress. So we're excited to continue our studies. We will be moving on to doing, oh, it's not going there. We'll be moving on to doing analyses of mechanisms of resistance by doing both genetic and protein profiling of those tumors. And with that, I'd like to thank the people who have done the work, especially Salma, who has moved on to industry, but she did a terrific job in the lab, and our collaborators. We couldn't have done this without our NETS4 colleagues at Iowa. Thank you. Thank you. Thank you, Dr. Quill, for some really nice, informative talk that brings out a lot of questions, I'm sure, in all of us. So this presentation is now open for Q&A. Please use the microphone or submit your questions online. Thank you. First question from Dr. Chen at the microphone. Dawn, as always, fantastic work. Thanks. I have two questions. One is the, you've shown a lot of data on the therapeutic side, but going back to how RABL6 is induced, do you have any knowledge of why it's turned on or how it's turned on? Yeah, good question. And that's leading to my second question, because I know you identified it initially based on differential expression from metastatic versus non-metastatic tumors. And one of the clinical questions we always have is, for some of the pancreatic neuroendocrine, we wanna know which ones are gonna progress, which ones don't, because we observe and we just use information like size and stuff like that, which is not. So I'm just wondering if RABL6 would be something utilized on sort of the front end to determine which tumors need to come out. Obviously, the RABL1 positive ones would be, right? Great, great questions. I'll answer the second one first. So we are hopeful that maybe the gene signature, the activation of the RABL6A pathway at the transcript level could be a potentially good marker of tumors that might be likely to progress to metastasis. But we have yet to prove that. RABL6A is actually really highly expressed in normal islets, and so you can't use it as a typical biomarker at the protein level. So we've been disappointed by that, but I think at the gene expression level, if people are doing transcriptome analyses, that might be indicative of tumors that will progress. Now, as far as what regulates RABL6A, we haven't devoted that much effort to it. The funding agencies haven't given us money to do that, but we have found recently that we think it is repressed by P53, and so if P53 is absent, we're seeing higher levels of RABL6A. So that's a direction we're heading in, in both sarcoma and in neuroendocrine tumors. We have another question at the microphone. Could you please identify yourself? Absolutely. Hello, Dr. Quell. Ralph White III, University of Minnesota grad student. In regards to your work in looking at the IPA for your RNA-seq data, I know you spoke about all the upregulated genes. Were there any downregulated genes that contribute to any synergistic value to what you've studied thus far? Oh, that's a really good question, and since I was merely a collaborator on that study, which was led by Jim Howe's group, he would be more knowledgeable about that. There certainly were a number, a fair number of genes that were downregulated, and Aaron Scott is the first author of that paper, and I would encourage you to reach out to Aaron and to Jim and just say, hey, what are the different genes that are regulated, and if they can share that. It should be in the supplemental data that you can look at the list. It's so enormous, it's hard to wade through, but if you have particular genes of interest, I would think that you could probe that, and then you can follow up with Aaron and Jim. Awesome, and my follow-up question to the synergy studies, your choosing of dosage, did you just test the limits, or did you have a specific dose that you chose for each drug? Yes, it took a while to choose those doses, and so we've done a lot of dose response curves and assays. Salma, I think, was really exhausted of running so many, and it's also hard to document synergy unless you find the right doses, but that's why we did sort of the 2D analyses that I showed you. We're doing a wide range of doses. We then hone in on doses that we think give us the best synergy, and then we use those to do our molecular and biological validation. Got you, thank you so much. Thank you. You have another question at the microphone? Hi, Dr. Kell. My name's Omer Shareek. I'm a surgery resident at Mayo Clinic, but was a graduate student with Kate Lyons and Raj Thacker. A question about targeting RABL6A. Based on the preclinical models you're using, do you feel as though this may be most effective in patients with neck or grade three tumors as opposed to lower grade tumors? Kate and I have looked at BESS inhibitors, which, as you know, target the MYC pathway. I guess one of the comments that we receive is is this therapy more applicable to that subset of patients rather than all patients with PNETs? I think it's a great question, and we won't know until we get into the models that are better representative of the grade one and grade two tumors, because we're using BON1 and QGP1 PNET cells that are more grade three-like, non-functional-like tumor cells, and then we're working in that RIPTAG2 model, which is also much more aggressive. We are working with Steve Labudi's group, and they have a MEN1P10 double knockout, and that's more, they have like, I don't know, seven to 10% KI67 for their tumors, and the combination is working beautifully there, and we're really excited. Just a couple weeks of treatment is almost eradicating the tumors. So we think it can be useful early on, but we need to do better work in those lower-grade tumor models, and I think, you know, the criticism for the MYC inhibitors, I would suggest that a lot of studies that look at MYC protein in tumors are probably using concentrations of MYC antibody that detect really high levels and are not sensitive enough to detect upregulation of MYC in grade one and grade two tumors. So we're actually working on trying to better detect MYC upregulation, because certainly at the mRNA level, we are seeing evidence that the pathway is activated. So it's 10.15, and we may have to start the next one. If you have a very short question, let's do that. Thank you. I'll try and make it short. I just have a kind of clinical question. Have you looked at the systemic metabolic effects of the drugs in the animals and whether they get worsening hyperglycemia or hyperinsulinemia? That's a great question. So we are just beginning to get into that, and that's really thanks to my colleagues at Brodker's, Steve Labudi and Zhijing Wan. They're following the tumors by measuring insulin levels, and in fact, the drug treatments are really keeping that low. I think it's merely a feature of the fact that the tumors are going down in size, though. So how much we're affecting secretion of insulin, we don't know, but there's a colleague at Iowa who studies insulin secretion, and we're actually starting to work on that. So great question. Thank you. Thank you, Dr. Crow, and Dr. Kate Lyons will introduce the next speaker. Okay, so our third and final talk of this symposium is on the role of subcellular localization of focal adhesion kinase in thyroid cancer, and that's by Dr. Rebecca Schwepp, who hopefully is here. Yes, so she is a professor at the University of Colorado. She has successfully completed for the K12 ARI Challenge Grant and R01 funding through the NIH, as well as an American Cancer Society Research Scholars Grant and Pilot Awards. Her research is directed towards identifying novel molecular targets relevant to thyroid cancer with a specific focus on kinase signaling pathways. So please, we look forward to hearing your presentation if you'd like to start. All right, thank you so much for the kind introduction and to the organizers for inviting me to talk today about the role of subcellular localization of FAC in thyroid cancer. All right, so I have nothing to disclose, and then here's the QR code. And so, as many of you know, differentiated thyroid cancer is associated with a high prevalence of mutations in the MAP kinase pathway, and these include BRAF, RAS, and rearrangements in RETPTC. Overall, this leads to an activation of the MAP kinase pathway in about 80% of these tumors. And for a transformation of differentiated thyroid cancer to anaplastic thyroid cancer, additional genetic alterations are important, including loss of tumor suppressors. Let's see if we can see the pointer. Activation of the PI3 kinase pathway, as well as activation of different receptor tyrosine kinases. And while we've made a lot of progress identifying the genomic drivers of thyroid cancer transformation, there's still a major knowledge gap in the role of non-genomic drivers. And so, current treatments for thyroid cancer, standard of care includes surgery, and radio, followed by radioactive iodine and TSH suppression therapy. And for those patients that do not respond to standard therapy, approved targeted therapies for advanced patients include multi-tyrosine kinase inhibitors, serafinib and linvatinib, which also have anti-angiogenic activity. And then for anaplastic thyroid cancer, the combination of a BRAF and MEK inhibitor was recently approved. Although these therapies are very promising, these patients are rarely cured. So we still need to work on identifying additional therapeutic strategies. In terms of thyroid cancer survival, patients with stage one through three disease do relatively well, even those stage three patients that have invasion. It's really stage four where you have gross extra-thyroidal invasion and distant metastases that you see survival drop. And then finally, anaplastic thyroid cancer is really one of the most lethal human cancers with 80% lethality at about a year. And so for these more aggressive tumors, we would propose a rule for SARK and focal adhesion kinase signaling. And so what are FAK and SARK? So FAK or focal adhesion kinase is a non-receptor tyrosine kinase. It is autophosphorylated in response to growth factor receptor and integrin signaling. This serves to recruit SARK, which is also a non-receptor tyrosine kinase. And then SARK can further phosphorylate FAK. And then FAK serves as both a kinase and a scaffolding protein to recruit downstream pathways, including PI3 kinase, P130 Cas, and also the MAP kinase pathway, all to promote survival, proliferation, invasion, and metastasis. And so we and others have shown the inhibition of FAK and or SARK really inhibits these pro-tumorigenic functions in pre-clinical models. Let's see, so this just shows some of our previously published data that were led by my graduate student, Bertel Kessler. And using a FAK inhibitor to inhibit FAK kinase activity, here in orthotopic models, we actually see a pretty modest effect of our FAK inhibitor with maybe about a 20% reduction in final tumor volumes. And then we're also interested in inhibiting FAK expression, which would inhibit both SHRNA knockdown of FAK, would inhibit both the kinase and scaffolding functions of FAK. And in contrast, here you can see in this orthotopic model using the BCPAP cells, you know, these tumors really fail to establish. And if they do, they really peter off pretty quickly. And so at the end of the experiment, we really don't even have tumors to dissect. And then finally, using an inhibitor of Src, which would also inhibit FAK through its phosphorylation of FAK. This was using dasatinib. You can see a really nice inhibition of tumor growth, whereby after six days of treatment, these tumors are gone. But unfortunately, you know, these inhibiting FAK and Src have not been, have had limited success in the clinic. So it's important to understand, you know, why, it's important to understand the signaling mechanisms of FAK and Src, so we can identify potential mechanisms of resistance and then also to better target these kinases. And so for this, we developed a model of acquired resistance to the Src inhibitor dasatinib. And we previously published that these cells are more invasive in vitro. And in vivo, they actually form larger, significantly larger tumors. And so we went ahead and looked at what was going on with signaling. And we looked at, of course, our favorite protein, FAK. And as you would expect, we observed membrane staining of phosphorylated FAK here in our parental or control tumors. And interestingly, in our Src inhibitor resistant tumors, we noticed a decrease in membrane staining of this autophosphorylated FAK and actually an increase in nuclear staining. And so when we quantitated this, again, we saw this reduction in membrane staining of phosphofac and this increase, about a 20% enrichment in nuclear phosphofac staining. So we went to a panel, a small panel of patient tumor samples. These are differentiated thyroid cancer samples to determine clinical relevance. And as we would expect, we observed predominantly membrane staining in a subset of these samples. Then another subset had predominantly nuclear staining. And then another subset had both a mix of nuclear and membrane staining. And so this is a small panel. And overall, we observed phosphofac nuclear localization in 73% of these samples. And so we also noticed, and I don't have a good picture of this, in these nuclear samples, it exhibited more of a punctate staining. And so now Megan Kellett, who's a really talented MSTP student in my lab, went and took a closer look at this in our thyroid cancer cell lines. And for these images, these are the CUTC60 cell line, which is a BRAF mutant anaplastic thyroid cancer cell line that we developed in our lab. And so with this punctate subnuclear staining, this made us think of the nucleolus, which is a major substructure of the nucleus. And so here, she stained with our autophosphorylated FAC antibody in green. And as expected, we observed cytoplasmic and membrane staining, but also this predominant subnuclear staining here. And she co-stained with the nucleolar marker fibrillarin in red. And here on the right panel, you can see that, indeed, phosphofac in the nucleus is co-localizing with the fibrillarin in the nucleolar marker. And so she expanded this to a panel of four BRAF mutant thyroid cancer cell lines. And the co-localization is shown on the bottom. So I hope you can see that a phosphofac does co-localize with fibrillarin in this panel of cell lines. And it was also interesting that this phosphofac staining in the nucleolus was really present in nearly 100% of cells, so suggesting that it's probably not localizing or accumulated in the nucleolus in a cell cycle dependent manner. So it seems to have a constitutive function there. And so I started my talk by telling you that phac functions at the membrane in response to growth factor receptors and integrins. But there have been several studies that have shown it does localize to the nucleus. And in the nucleus, it has been shown to interact with P53 to promote survival. In addition, phac has been shown to interact with a number of different transcription factors to regulate immune-related genes. And together, it's thought to function to promote survival, immune evasion, and tumor progression. And so what about phac in the nucleolus? So there's only one previous study that we've been able to find that has shown potential role for phac in the nucleolus. And this was by David Schlipfer's group. And so in breast cancer, they had also observed the subnuclear staining in their breast cancer samples, their patient samples. And then they went on to show that phac associates with nucleolar proteins, including nucleophosmin, and then also nucleostemin, together in a kinase-dependent manner to promote survival of these breast cancer cells. But they didn't really hone in on exactly how nucleolar phac was promoting survival. And so what's the role of the nucleolus in cancer? So the nucleolus, as many of you know, is the primary site for our RNA synthesis and ribosome biogenesis, which are really necessary to fuel the growth and survival and metastasis of cancer cells. And so therapies, obviously, this is important. And so therapies targeting nucleolar functions are in clinical development. And so our findings that phac is in the nucleolus is really exciting clinically. And so our hypothesis is that nuclear phac drives thyroid cancer growth and survival through the regulation of nuclear proteins involved in ribosomal RNA synthesis. And so we looked at the sequence of phac. And in order to study nuclear phac, we were hoping to find a nuclear targeting sequence. And by both bioinformatic and manual analysis, we did not identify a nuclear targeting sequence. And so phac does have a nuclear localization sequence here in the firm domain. So we went ahead and mutated the NLS. And we called this our nuclear localization mutant, or NLM. And we re-expressed this nuclear localization mutant in cells where we had knocked down phac using CRISPR-Cas9. And so this really provided us a nice model to study the nuclear role of phac, whereas these cells are completely devoid of phac in the nucleus. And so this is just our knockout model. So here it shows we completely knock out phac expression in our CRISPR-Cas9 model, and that we also lose phosphorylation of phac at the Src-dependent and also MAP kinase-dependent sites. And looking at transformation, we showed that knockout of phac, really these cells fail to transform. And this is what we previously saw with shRNA knockdown of phac. And so characterizing our nuclear localization mutant, when we re-express our wild-type phac and our phac knockout cells, you can see cytoplasmic and membrane staining, along with nucleolar staining with our wild-type phac re-expression. And then when we re-express our nuclear localization mutant, we do lose nuclear staining of phosphophac. And then looking at phosphorylation, our wild-type and nuclear localization mutant are properly phosphorylated and at similar levels. When you look at the alpha tubulin, and then also expression is at similar levels. And then looking at downstream signaling of phosphoSrc with our phac knockout, so we do see a decrease in phosphoSrc levels. And then these are rescued by our wild-type and nuclear localization mutant. So it appears this nuclear localization mutant is signaling properly in the cytoplasm. And so then using our soft auger assays, again, our wild-type phac is able to rescue this loss of transformation with our phac knockout. And then we were really surprised to see that our nuclear localization mutant did not rescue at all. And the cytoplasmic functions of phac are still intact. So we went on to further characterize how phac is localizing or accumulating in the nucleus. And for these studies, we took a look at this 397 site as it's in close proximity to the nuclear localization sequence. And so we mutated the 397 autophosphorylation site, and then also made a kinase dead mutant. And so using our phac knockout models, we re-expressed our wild-type or our autophosphorylation mutant. And I hope you can see here that, as we saw previously, our wild-type phac is expressed in the nucleus and cytoplasm. These are lower resolution images. And that our autophosphorylation phac is excluded from the nucleus. And when we quantitated this, it was about a two-fold reduction in nuclear accumulation of the autophosphorylated phac when this site is mutated. And so then, looking at the functional role of when we mutate the 397 site and also the kinase dead, here, as expected, we lose autophosphorylation with these mutants. And we have expressed them at similar levels. And looking at transformation assays, we found that mutating the autophosphorylation site or expressing this kinase dead mutant of phac really phenocopies the nuclear localization mutant in transformation. So indicating an important role for autophosphorylation and kinase activity of phac for transformation, as well as nuclear localization. And so finally, we wanted to ask what the role would be by forcing phac into the nucleus. And for these studies, we fused an SV40 NLS to phac. And then we re-expressed this. And we're calling this our NLS construct. And using this, we mutated the autophosphorylation site alongside. And so this just shows Megan's confocal images. And we're here using a V5 tag, which phac is fused to V5. And so you can see in green that, as expected, we do see that phac is localized to the nucleus, as you would expect. But we really found these striking sub-nuclear staining patterns that did co-localize with fibrillarin in red in the upper panel. And then remarkably, when we mutate this autophosphorylation site, we really lost this punctate nucleolar staining of phac. So indicating an important role for this 397 site in phac accumulation in the nucleolus. And so now we're working on the functional studies for forcing phac into the nucleus and nucleolus. And so for the last part, I'll talk about our interactome studies. So to try to understand how phac is promoting survival in the nucleolus, we performed a BioID approach to evaluate the phosphophac interactome. And so for these studies, we fused phac to the promiscuous biotinylase, BRA. And then for this approach, you add biotin to the cells. And then proteins that come within 10 nanometers of this BRA phac complex are biotinylated. These biotinylated proteins are then isolated by streptavidin-coated magnetic beads. And then they're digested and identified by mass spectrometry. And so this has been really a breakthrough for the identification of kinase substrates due to the high off rates of kinases. And so this just shows the enrichment score for the GO terms that were upregulated in wild type versus our 397 phac mutant. And we were actually a little surprised to see an enrichment of nuclear pathways. And then based on our functional studies, we were excited to see the nucleolus there, even though it is further down on the pathway. But interestingly, it was ranked a little bit higher than focal adhesions, which is the canonical role of phac. And in addition, we performed this study in whole cell extracts. And given that the nucleolus represents only about 7% of the human proteome, I guess we weren't too surprised that the nucleolus was a little further down on the list. But we are repeating this BioID study where we can specifically localize and target phac to the nucleus versus the cytoplasm. So hoping for further enrichment there. So these are the interacting proteins that specifically interacted with the autophosphorylated phac. And all of these proteins are involved in our RNA or protein synthesis. One of these proteins, nucleophosmin, or NPM1, was actually shown to interact with phac in that breast cancer study that I showed you from David Schleifer's group. And then MibBP1 was actually shown to be a candidate phac substrate in an in vitro screen from Gelman's group. So strengthening in our findings here. And then Megan went on to do a string analysis of these proteins and found that the majority of these proteins do have proposed or direct interactions. And so we took a closer look at nucleophosmin or NPM1. So nucleophosmin is a nucleocytoplasmic chaperone protein that's involved in transporting proteins between the nucleolus and the cytoplasm in response to stress. The phosphorylation of NPM1 has been shown to regulate ribosomal biogenesis. And NPM1 itself is an endoribonuclease that cleaves and processes pre-rRNA. And NPM1 has been shown to be overexpressed in many different types of cancers. And so Megan used proximity ligation assays using antibodies to autophosphorylated phac and NPM1. And these images might be hard to see here, but she did see that NPM1 and wild type phac do interact as well as when we express our NLS phac construct. And the quantitation is shown here where when we re-express our wild type phac, we do see interactions with NPM1. This is enriched when we force phac into the nucleus. And then we lose these interactions when phac is excluded from the nucleus. And then since NPM1 has been shown to be a nucleocytoplasmic chaperone protein, one of our big questions is, how is phac getting into the nucleolus since it does not have a nucleolar targeting sequence? And so we asked if we knock down NPM1, which is shown here with three different SH RNAs targeting NPM1. And here's the scrambled control. We asked whether knockdown of NPM1 affects phac localization. And so here looking at autophosphorylated phac with the SH control, in green you do see phac nucleolar accumulation with the co-localization with fibrillarin. And then when we knock down NPM1, we do see a loss of phosphophac in the nucleolus. But you may also notice when we knock down NPM1, we do see a potential decrease in nucleolar size. So we are working on how to quantitate these findings. And so finally, we're trying to look at, of course, our big question is its function in the nucleolus. And looking at the regulation of rRNA, we first wanted to take a look at the regulation of pre-rRNA by phac. And so in the nucleolus, RNA polymerase 1, along with topoisomerase, which is one of our BioID hits, transcribe ribosomal RNA. And then NPM1 actually functions here to cleave some of these external transcribed spacer regions from the 47S rRNA. This is a very rapid process for the cleavage. And then finally, you have your ribosomal RNA units. And so we did qPCR looking at the ETS1 cleavage process, the processing. And with our phac knockouts, we do see a trend towards decreasing this processing of rRNA. As I said, this is a very rapid processing. So we are trying to develop, so we're developing different models, including dagrons to really deplete phac more rapidly so we can study this more accurately. And then we'll also do a flow cytometry approach, which has been shown to be even more accurate in assessing these processing steps. And just to show you some of our latest data, we're using some protax to try to deplete phac a little bit more rapidly. And Megan just showed that she can deplete phac using protax by about four hours in our thyroid cancer cells. So hopefully, that will give us more accurate information. And so in conclusion, we have shown that autophosphorylated phac, which is important for its kinase activity, interacts with key nucleolar proteins involved in rRNA synthesis. And that nuclear autophosphorylated phac drives growth and survival in thyroid cancer, which we've shown through our transformation assays. And so clinically, we're working to determine whether nucleolar phac may correlate with tumor aggressiveness. And so we're looking at tumor samples of different stages of thyroid cancer. And then also really understanding the mechanism of how phac is promoting thyroid tumorigenesis and its role in the nucleolus will really give us a new function to target. And so again, these studies were led by Megan Kellett, an MSTP student in my lab. And then I just wanted to acknowledge my entire team, which is incredible at the University of Colorado, our collaborators, and our funding. Thank you very much for your attention. OK. Thank you very much. So we have a few minutes for some questions. Looks like we have a few. Hi, Matt. Hi, Rebecca. That was great. Thank you. Matt Ringel from Ohio State. So a really nice date. I've got just a couple of quick questions for you. So have you had a chance to do the BioID or the PLA in a non-overexpressed system? Or is everything done in overexpression models? Yeah, I mean, so the BioID, you do need to express. The question is overexpression, whether you crispered in the BioID construct or whether that was a transfection overexpression or an infection overexpression. Right, so the BioID was not a fact knockout, since with the BioID, we're really just identifying proteins that will identify. I interact with the BRA. Yeah, I mean, we did try to express near endogenous levels, so we're not swamping the system and squelching and things like that. But that's a good point. We could definitely try it in our fact knockouts. Yeah, and then the PLA also, was that an overexpression or is that an endogenous PLA? As far as was that in our fact knockout? Yeah, when you showed the PLA for the fact interaction with it, when using the PLA to verify your BioID targets for protein interactions. Let's see, now I'm trying to remember that. That's why I'm putting it on the spot. I know, if I can go back and check my slides. Let's see. I think it might have been overexpressions. I think it was the V5 model, but I can't remember. Yeah, the, well, yes, yeah, we were overexpressing FAC, but we were not overexpressing NPM1. Right, right, right, so that's a fact, okay, great. Okay, so it would be interesting, I think, to see if the endogenous interaction, especially with endogenous FAC activation in the context of integrin, to see whether or not, just to see whether you can get a sense of the contribution in endogenous activation of FAC in terms of nuclear versus, but very nice studies and very nice data. Right, yeah, thank you. I completely agree. Looking at endogenous is very important, so. Hi, Rebecca, Dawn Quell from Iowa. Very nice talk. I was curious, since NPM interacts with P53, FAC interacts with P53, whether any of your phenotypes that you're observing are P53 dependent? Right, that's a great question, and you know, so this, I actually probably didn't mention this, so FAC has been shown to interact with wild type P53 to promote survival, and all of these cells that I'm using are actually mutant P53, and it's not known the role of FAC with mutant P53, so that's something that we would like to explore at some point, but yes, good point that NPM1 also interacts with P53, and so we're hoping to dig a little bit deeper, but there could be a connection there. Yeah, definitely, you know, some older studies from, I think, Pierre Colombo's group showing that NPM knockout is lethal because P53 becomes hyperactive, it's no longer restrained in the nucleolus, and so it would be really interesting to see what's happening there, as well as with a protein that I had studied for a long time, ARF, which also is nucleolar, interacts with NPM and P53, and so. Right. Very interesting data. Thank you, yes, we still have a lot to do. These are definitely new studies for our lab. Oh. I see you have one last question, quickly. More than a question, it's just a comment. I just wanted to thank all the speakers for the outstanding presentations, and as a clinician treating these tumors, I really can't wait to take all this to the clinic and develop clinical trials for these patients, so I really appreciate for outstanding presentations, and these are tumors in need, as you may all know, to develop other therapies for these neuroendocrine tumors, so I really appreciate, thank you. All right, well, thank you so much for organizing. Just want to thank all the speakers and for presenting the fascinating work, and we extend our congratulations to them for all the progress, and please remember that this discussion can continue about endocrine cancers in the Endocrine Cancer Special Interest Group and in the DocMatter community platform. Please also remember to visit the Basic Science Pavilion right around the corner in room B402. Thank you all. Oh, and also another, there's the rapid-fire posters at lunchtime today for the neuroendocrine tumor session, if so, shout out to maybe Swing by There as well for some more interesting science. Thank you very much. Thank you.

Co-Clinical Models

Neoplasia

Proneoplasia

Endocrine Society

Dr. Hydera Del Rivero

National Cancer Institute

NIH

Neuroendocrine Tumors

MEN1

Genetics

Epigenetics

Notch Signaling Pathway

RabL6A-RB1 Axis