Signal Transduction and Modeling in VHL Disease | Medical Symposium 2024

[00:00] Hello, everyone. Maybe we should try and get people in and sit down that we can get started with the session. We're running overtime.

[00:20] Thank you.

[00:40] child directly, I'm learning a lot because I think there's a lot of overlap between the disorders we study and these conditions. And so it's a true honor to be here and share the sharing with Dr. Srinivajan, who will start and open the session.

[01:00] I am Ram Stony Bhasan, medical oncologist at the National Cancer Institute. Why don't we get started with the session? The first talk is going to be present by Dr. Ching Zhang, a good friend and a colleague from the UT Southwestern, from UT Southwestern.

[01:20] southwestern is going to provide some new insights into VHL biology and Ching? Why don't you start with the forward guitar? Alright, thank you. Thank you for your invitation. It's really a great privilege to be here. So today, for this session, I'm going to start with introducing some new biology on vitro secondary access, especially in the setting of kidney cancer.

[01:40] So this is my disclosure. So at this moment, we're already pretty sure that VHL is the most important tumor suppressant in kidney cancer, obviously actually led by Tom as well as Samra a few years ago in this landmark Tresorax paper. But in terms of therapeutics, I won't quote Bill Canitz, my own mentor, saying

[02:00] that some of the mutations that arise later in the revolution exist because of the pre-existing mutation or trunk mutation preceding then. So if that's true, if you target those trunk mutations, potentially it can kill those tumors of selectivity. So also actually give an analogy in terms of this D'Aime trade

[02:20] by the 1866 German scientist, Ernest Heichel. So you think about this stowaining tree, this is bottom of the trunk and this is branch. If you talk about those branch events, there always can be resistance developing. On the other hand, if you talk about those trunk events, you will most likely eliminate tumor cell growth. This is also actually had

[02:40] to existing FDA-approved therapies, including ER, EJFR, B-RAP, B-C-O-AB, as well as AR signaling in different cancer settings. So those are successful examples lying in place. So today I would like to discuss a few potential strategies that we can target those trunk EJFRs.

[03:00] events in the case of the BHRA loss in kidney cancer. I want to talk about briefly four things we could potentially do. So one is going downstream such as targeting hip T-alpha TKI and then second is targets unique vulnerabilities in kidney cancer such as lipid droplets and certain enforcement I would like to talk in total.

[03:20] together as a whole because we can talk about sensitive anxiety as well as protax. And some of the colleagues here are also working on. So first, going downstream. So in terms of targeted therapy in kidney cancer, this is nice review by Tony Schor and Bill Kadin a couple years back in nature medicine and summarized very nicely.

[03:40] coming from upstream hip signaling, and then this will induce the downstream of the growth factors binding with respective receptors on the endothelial cells. And then previously the target therapy against TKI is the main of multiple tyrosine-canidase inhibitors that actually, major FPDJF are active.

[04:00] or FJFR or BHFR receptors. On the other hand, we also know that bazilifan is recently approved by FDA to treat either a virtual disease or advanced kidney cancer. So those are remaining to be the main approach beside immunotherapy and kidney cancer. But obviously, I quote Marshall Nyla, who originally cloned BHL.

[04:20] Now, this quote from 10 years ago is saying that on the other hand, we're dealing with drug resistance, either intrinsically or adaptive resistance. Also, there's actually lack of response from half of the patients, some maybe on the going treatments. So the question is, we are perfectly ignoring other perspective CCRs and CIs.

[04:40] that may need to be targeted therapeutically to achieve maximum therapeutic efficacy. The burning question in the field at that time when I started in Belsch-Leib was can we identify additional therapeutic target in RCC that is induced by visual loss? Consider visual isroelitis targeting HIF. The burning question is can we identify

[05:00] some other visual issues like substrates. And then the real million dollar questions, because we spend a million dollars on this, is how. So, cut a long story short, this is after more than 10 years' efforts, with global efforts from various different institutions across the globe, from China, Singapore, Basel, India, India, India, India, India, India, India, India, India,

[05:20] Boston, Vanderbilt, North Carolina, Texas, and some of the investigators actually in this audience. So we actually identified some of the novel significant access through these global efforts over decades. So in this case, I want to highlight some of the schematic we published previously after this long-term effort, led by two of the top

[05:40] in the post-op, former post-op in my lab. Right now they're actually running it on lab in China. So basically we identify two of the novel vitro substrate beside heftiophila. In this case, very little studies have been reported for those two proteins. Those are actually transcription factors. One is called SFN-BTN-1. As a polycholine company,

[06:00] protein. That's why it's zinc finger homobox protein too. In this case, one of the secondary acts through ZHT2 is activating NF-cobab, and the other SFN-BT1 is activating pHK sphingocanis 1, which is also later on collaborated by Ian Frues' lab

[06:20] on this kidney cancer research as well. So those two studies showing that besides canonical heptia alpha, there's also the third, the second and the third branch that is in parallel, is activating this mutual law of signaling. And then by crippling those signaling pathways, we could potentially

[06:40] in collaboration with heptiafas contributing to the kidney kind of tumor genesis. Most likely, importantly, I would like to mention those heptiafas-resistant CSRC cells, we actually, in active SFMP-T2R, Z-trix2 pathways, we can actually solve tumor growths, and it illuminates those primary tumor regions.

[07:00] growth as well as metastasis. So and then second part I would like to really quickly touch upon is targeting unique biology of the kidney cancer which we know that is clear sorena sircosanoma is majority of the kidney cancer illustrated by this histology where you can see this big lipid droplet formation.

[07:20] And in the field, we know that this lipid drop information is really important to provide bioenergetics as well as dealing with ear stress in kidney cancer development. So the underlying mechanism still remains somewhat unclear. So in this case, by performing genetic screening, by collaborating with very

[07:40] investigators, mainly from UT Southwestern. We actually, using the genetic screening, found that GMG-D-Seq, which is TOG-dependent dioxide genase, can form an actually complex with RBM39, which is transcription activator. By forming a transcription-activated complex, it's looking

[08:00] co-occupy those D-GAT1 signaling along with H3K4 trimestration marks. So D-GAT enzyme is a very important critical enzyme that controls the free fatty acid conversion to the triclosides. And then tricloside is obviously the component that actually induces

[08:20] lipid droplet formation. From this perspective, D-gal1 represents the most oncogenic events that is controlling lipid droplets in C-Cirrus acid development. So once we figure out the molecular mechanism, the most important thing is we can actually use D-gal1 inhibitors. In this case, we're using an isotopic mode of

[08:40] where it induced tumor growth and then in the control cells after five weeks tumor growth radially. But upon gut-winding inhibitor treatments you can see the slow of the tumor growth over time, suggesting this is one of the things we can potentially target in kidney cancer to eliminate lipid drop deformation as well as tumor genesis.

[09:00] The third and fourth one I would like to emphasize there is actually coming hand in hand because this is regarding with synthetic recitity as well as protax. So this is the review paper we published recently showing that the concept of synthetic recitity is patient in renal or carcinoma setting.

[09:20] In normal cell, you could think about we have VHO wild type, which is gene that we often talk about, and then gene A will be the one that will be the sensitivity, decidedly potential partners. In normal cell, if you're in activity gene A in this case because VHO is wild type, you don't cause the anxiety. On the other hand,

[09:40] Because most of the C-sarces cells, we know that it's vitro now. So in this case, if you inactivate gene A in this setting, so you can see that's actually the cause and desality. But the most important thing here is that because vitro wild type existing normal cells, by targeting gene A specifically, you only cause desality

[10:00] in the clear-sorrhizal sarcozonoma. So to cut the long story short, we actually identified one of the important kinases, so-called TPK1, as a sensitive-decided partner in kidney cancer. So TPK1, as we know, Pangchang's chance lab at UT Southwestern, it's a very important type I interferon response genes.

[10:20] When the organism sees the double-stranded DNA, they can sense by the ségat cyclinase, therefore regulates steam activation as well as typical phosphorylation on the downstream F3 pathways. Dimerization of the pathway will induce the nuclear localization as well as activation

[10:40] inhibition of type 1 interferon response to actually fight back the atopstranded DNA, foreign DNA invasion in the tissues. But in this case, in a tumor-intrinsic fashion, for example, we actually can deplete typical one using CRISPR-SGLRNA, so we can appreciate in UMRase-TIN-A.

[11:00] 2, which is VHL now, it can cause specific lethality. But if you restore the VHL in your MRC2 cells, you can see that those cells remain intact. And then this is also happening in another pair of cell lines we established. In this case, you can see specific targeting of those lethality in your MRC6.

[11:20] cell. The medical mechanism through collaboration with Dr. Kangong and PKU, Pei Jin University, we actually have found that VHLos can induce hyperactivation of phosphatid-pikin 1, therefore induce P62 activation and lead to tumor genesis. This actually corresponding with what's Bell cadence lab published

[11:40] showing that P6TQ, also called SQSTM1, is a pathogenic target of the 5Q copy number in kidney cancer. So obvious question is can we target it therapeutically? So through genetic model where we deplete typical in an inducible fashion in ostopic model, you can appreciate that those

[12:00] typical anti-depletion caused tumor regression, also showing quantitative down there. More importantly, when we look at spontaneous non-metastasis, depletion typically won't cause total abrogation of those non-metastasis spontaneously. So in this case, we add extra layer of the significant axis we're trying to target, which is

[12:20] phosphatipidyl 1 as well as typical antigens. Secondly, obviously, compound 1 is a long-specific typical inhibitor-targeted IKK-epsilon, which is not what we wanted. So we would like to develop a specific prototype that can be used to degrade typical 1. In this case, through collaboration with Lindsay Jans and the UNC and lower weddings, IUT Southwestern.

[12:40] We indeed develop such prototype molecules, which is used to degrade TPQI specifically in a dose-dependent manner. On the other hand, they wouldn't degrade the close family member, which is important for the immune signaling so-called IKK epsilon. So we also can actually calculate the DC50, which shows us how to calculate the DC50.

[13:00] showing about 100 nanomores, showing very potent. On the other hand, the most important thing for protock is actually specificity. In this case, we actually performed the global proteomics. In this case, you can see that TpK1 is one of the most abundant proteins that was being degraded upon the TpK1 proteomics.

[13:20] treatments. In my argument, there are also a couple of proteins that can be degraded as well, but those turns out to be typical one targets as we meditate experimentally. In this case, we actually try and typically protect in animal experiments. So far, the data looks promising. So anyway, and then last part I would like to touch upon is

[13:40] The potential problem targeting TPK1, because we know that TPK1 is important for the multiple cellular activities, including c-gastine, which is important to immediate cellular response to double-strand DNA. But the more important question is, can we identify the specific oncogenic events that is activating?

[14:00] typical in an olecranine fashion. In this case, we performed a candle-watt SRI screening basically using those 710 canyases, transphased CCRCC cells, and then did an individual typical one of phosphorylation, which indicate typical activity, followed by the quantification of specific signals, and then

[14:20] plus those potential candidate acts, and then eventually focus on individual candidate with individual validation in major candidate's aspects. The interesting part is that we identify a novel candidate, so-called DCLK2. So what is its candidate, DCLK2? DCLK2, it turns out, is a neuronal candidate, which is important.

[14:40] for the neuronal survival and the induced cross-confirmation. So this DCLQ2 kinase has a kinase domain at the base of various C terminal, but unfortunately there's no kinase substrate being identified previously. In this case, we published a molecular cell paper recently showing that various

[15:00] The explicit variant form of DCLK2 can specifically induce phosphatypical and phosphorylation, therefore induce p-sickening as well as c-c-c thermogenesis. So in summary, I would like to emphasize there's multiple sickening pathways which could be induced by VHLase, which we can target therapeutically. But more importantly,

[15:20] ongoing research besides those signaling we have identified. We also work on some of the mutual amsex regulation pathways, but more importantly, our lab has switched largely to using unbiased CRISPR screening to identify those potential novel targets. In this case, a study led by Jin Zhou, who has recently promoted a system-

[15:40] Professor, at ET Southwestern through collaboration with Celeste Sum and Jim Prokofievis and Serenus Maladis, we performed the general white CRISPR screening, where we infect those CCRCC cells with VHR now, subcunyest inject into the mice, wait for 6 to 8 weeks, isolates non-matastasis,

[16:00] primary tumors followed by deep sequencing. So we can appreciate about two different solids. We can see concordance of the top genes that is being actually enriched in the long metastasites. So one of the genes we identified eventually is called HLF, mainly actually involved in a hepatic leukemia factor in a leukemia setting.

[16:20] In this setting, we're using a stopping model of valedictus functionality, showing that CRISPR-SHRF knockouts can induce non-metastasis spontaneously, while not affecting primary tumor growth, but other hands, overexpression HRF can induce, can actually induce this non-metastasis regression.

[16:40] So another one of my times just ran out. In conclusion, I would like emphasis three signs. One is that we can target trunk amputation and visual loss. And second, we can use a synthetic docility. And third, CRISPR screening can help us identify some new therapeutic vulnerabilities using primary tumor growth or metastasis. So last, but last is.

[17:00] I want to thank all of the supports from collaborators of College of Long Way as well as my funding resources. I would like to take one or two questions if there's enough time. Thank you. Questions?

[17:20] Okay, it's unreal to the force of all this work. I'm wondering about the last bit that you showed with the regulator of metastases. How is this implicated in tumors? Is it, you know, transcriptionally upregulated? Can you tell us a bit?

[17:40] That's a great question. In summary, this actually, I didn't have time to show this. This HRF is actually silenced and then it's silenced contributing, promotes this lung metastasis specifically, but not liver metastasis. We also have the CRISPR screening showing the other set of genes that regulate liver metastasis and you know, independent.

[18:00] This gene silencing was controlled by the sweet-sneef complexes, for example, such as SMART4. So in this case, we use SMART4 inhibitors, we can stabilize HRLF, and then this stabilization

[18:20] upstream of the Swiss-Nif complexes. And downstream is actually regulating integrant pathways with the transcription activity of HLF. Yeah, I'm happy to show more data for you next time when you visit the UT Southwestern. Thank you. Yeah, thanks. Thank you, Ching. I mean that was a lot of data that you presented.

[18:40] Let me focus my question on the D-GAT one story. Can you tell us a little bit more about what inhibiting it actually does, other than, are there any metabolic implications to inhibiting D-GAT beyond just altering lipid-droplet accumulation, one?

[19:00] And two, it looks like you have an inhibitor that goes after it. Are there any plans to develop it further? Right. How strongly do you feel that this is a good pathway to go after as a target of kidney cancer? Ron, that's a great question. It always seems very insightful. I just wanted to quickly point out, D-GAT1 has a family member, so-called D-GAT2.

[19:20] So, the previous publication already showed that D-GAT1 is not an essential enzyme because you deplete D-GAT1, the mouth is totally fine. But D-GAT2, on the other hand, is an essential kinase. Even though D-GAT1 and D-GAT2 have similar names, they only share about 30% sequence homology. In this case,

[19:40] That's why it's very easy to identify a specific D-GAT1 in hipters, but not targeting D-GAT2. So in this case, we argue D-GAT1 in hipters on its own will not have toxicity in normal tissues, given the reason that D-GAT2 is an essential enzyme. In terms of your first question,

[20:00] regarding D-GAT1's function beside regulating lipid droplets? That's a great question. I'm not sure exactly how also signaling D-GAT1 mediates. The most reported or canonical function of D-GAT1 is to control the lipid drop formation as far as I know.

[20:20] But maybe there's a function not as aware of. But that's a great question. Thank you. Thank you. Can I add a quick one? Yeah, of course. Perhaps I missed it, but any of these TPK1, D-GAT1 inhibitors, do they impact on heath response at all? TPK1 actually is totally heath-endipated.

[20:40] So for example, if you deplete HIF2 alpha, you will not affect TPK1. Okay, and then TPK1 also not affecting HIF. So there's no, it's mutually exclusive. We're trying to combine TPK1 with HIF2 alpha inhibitors in those resistant cells, but most of the effect was actually mediated by TPK1 inhibitors.

[21:00] So, which means that typical one inhibitor, pass-inhibitor, or pass-inhibitor does not inhibit typical one inhibitor's activity. But typically itself, by either pharmacologic inhibition or depletion, will be sufficient to actually lead to the tumor growth regression. Yeah.

[21:20] That's a great question. Thank you. Thank you very much. We move to the next speaker who's going to be Ruhi Deere from the Baylor College and she's going to talk about linking VHL and CD2 in a common oncogenic pathway convergent at the metotic spindle.

[21:40] Thank you. Hi, everyone. I'd like to begin by thanking the VHLA for this opportunity to tell you a little bit about what's going on in the lab. We normally study chromatin-modifying enzymes outside of the nucleus in these new and normal functions that we're finding on the mitotic spindle.

[22:00] And today I'm going to give you just a small story where we think VHL and CetD2, which we know are moonlighting, at the mitotic spindle, are in fact converging on a common pathway. So we've seen this model several times over the past couple of days and essentially I just wanted to highlight that we have

[22:20] bileleic inactivation of VHL and loss of the 3P tumor suppressors, both in renal cell carcinoma and VHL disease. And the main question that we wanted to ask was what exactly is the loss of these three chromatin-modifying enzymes, the CEDD2P?

[22:40] BBRM1 and BAP1 really doing in addition to VHL loss. And to sort of give you a little history of what these enzymes do on chromatin, you have CEDD2, and I'm going to start from the right and move towards the left. We have CEDD2, which is a histone methyltransferase. It adds

[23:00] methyl marks on the histone tails. Here it's adding a methyl mark on H3-lysine 36. You have PBrM1, which is part of a larger complex called the P-Baf complex, which is involved in nucleosome remodeling. And then you have BAP1 right here, which is a D-Baf complex.

[23:20] T-ubiquitinase that removes the ubiquitin moiety from the H2A tail. But the one thing that I will mention here is that all of these enzymes, although I've just mentioned that they're chromatin-modifying enzymes, they do not share a common etiology on chromatin. And what do I mean by that? They're not in the same place at the

[23:40] same time. And why is this important? Because what we are observing is that there is this convergence of the 3P tumor suppressors at the mitotic spindle. What's known is that VHL was shown to localize the microtubules. We've shown CD2 as a writer. It lays the methylmoxonease.

[24:00] on microtubules at lysine 40. We have PBRM1, which reads that methylmark, and then you have BAP1, which is actually deubitinating gamma tubulin, which we won't get into today, but that's really interesting as well. And our goal in the lab was to sort of identify if there was a functional

[24:20] convergence with these enzymes all localizing at the mitotic spindle was actually happening for a specific reason. And you can imagine this is really important because if we were able to identify and explore different mitotic vulnerabilities, this would be a great opportunity to sort of take these cells.

[24:40] with the mitotic vulnerability and sort of just help them go towards mitotic catastrophe. And just to sort of step one, step back and tell you a little bit about ZD2 methylating microtubules, we were part of the group that made the original discovery of ZD2 being a methyltransferase.

[25:00] phrase on microtubules. We have an antibody that was raised to this methyl mark and what we see is that you see this beautiful localization or staining of methylation at the spindle poles here in the bottom row and at the mid-body and the mid-body is nothing but a structure with the spindle

[25:20] microtubules collapse during cytokinesis and that enables the separation of the two daughter cells. And if you lose CD2, you start seeing a lot of mitotic defects. Okay, you see a lot of lagging chromosomes, you see a lot of chromosome segregation defects, including chromatin bridges. And eventually you give rise to a lot of genomic instability.

[25:40] which is quantified as micronuclei here. But what's even more important and really exciting is that if you have monoleilic loss of Cd2, which is what you would see with loss of 3p, you see that Cd2 function is exclusively deficient for its cytoplasmic function. So you do not see methyl-oxygenase.

[26:00] circulation on microtubules, but it is perfectly capable of performing its chromatin function. So chromatin function remains unaltered. And so having told you that, I just told you that if you have mono or bilele class of ZD2, you should have aberrant mitosis and technically these cells should die. So the question then

[26:20] was how do these cells that have lost 3P, how do they survive with just one copy of ZD2? And to answer that question, we sort of move to our next favorite protein, which is aurorokinase A, and I'll tell you why it's our favorite protein. But what's known is that aurorin expression is high in multiple cancers, including

[26:40] clear cell renal cell carcinoma and in fact elevated aurora levels are associated with prophognosis and this is important because we had a previous publication where we actually showed that in VHL deficient cells you have elevated aurora expression and

[27:00] Here is just sort of one block where we had 786 O cells. These are clear cell, renal cell carcinoma cell lines where you see elevated aurora. And this elevated aurora can be rescued if you re-express VHL. Gosh, I'm trying to get this mouse to work. Where you re-express VHL and of course, because it's just a very long time, it's a very long time.

[27:20] its target for degradation. If you add MG132, which is a proteasome inhibitor, you should see rescue of levels, which is what we do. What we've also done more recently is using cells where we've got these cells actually from Dr. Genovese's lab at MD Anderson, and we also have a similar

[27:40] cell line from Dr. Ian Frew where you have 3P loss and we've tried to see what happens to aurora levels in these cells and what I've shown you here is we've actually measured the fluorescence intensity of aurora at the mitotic spindle. And if you see here we see a distinct increase in aurora levels both by immunofluorescence but also by western blots.

[28:00] in these cells that have 3P loss. And so the question that we wanted to ask was, could elevated aurora drive mitotic progression in these conditions where you have monolelic loss of ZD2? And why aurorokinase? So aurorokinase is a very important mitotic kinase. It is extremely critical for

[28:20] for centrosome biology as well as for bipolar spindle formation. It's elevated across multiple cancers and what I should mention here is that it's very tightly regulated. It's not just regulated temporally during mitosis but it's also spatially regulated. And what do I mean by spatial regulation? So you have different pools of

[28:40] within the cell. So targeting all of the aurora is going to be actually really bad for, in terms of toxicity. But what we're now realizing is that targeting certain portions or pools of aurora is going to be a far more precision medicine approach to regulating auroras.

[29:00] aurora levels in cells. And since it is in kinase, the first question we asked was whether aurora could phosphorylate ZD2. And to do this, we used in vitro kinase assays where we essentially took recombinant protein, both for aurora and ZD2, using P-32.

[29:20] to be

[29:40] in an in vitro kinase assay and we also see phosphorylation of ZD2. To determine the site that was phosphorylated by ZD2, we moved and did some mass spectrometry where we identified serine 2080 on ZD2 as a site that's phosphorylated. And we generated an antibody

[30:00] which we use for immunoblotting. And here what I've shown you is just an increased expression of GFP aurora in cells where we now see a nice increase in phosphorylation of ZD2. So the next question we wanted to ask is what is the functional significance of this phosphorylated ZD2? What is it really doing in the cell and why should we care?

[30:20] And so I just told you, CetD2 is a dual-function methyltransferase. It methylates chromatin and it methylates tubulin. And so we took these cells that we call here HECB9. These are HEC293T cells where CetD2 has been removed. And so we've rescued these cells with

[30:40] the full-length CD2 or S2080A mutant CD2. And we looked at what happens to the histone methylation. And it seems like there's absolutely no change in H2K36 trimethylation. Okay? We then moved on to see what was going on with tubulin methylation. And here in the lower panel, you can see that when you have the mutant

[31:00] you actually see less tubulin methylation. We then repeated this experiment in mouse embryonic fibroplasts, which had a tomoxifen-inducible system where we could get rid of the ZD2. And we've rescued these with either a truncated but fully functional ZD2 or a S20-ATA mutant.

[31:20] What we see is that in the wild-type expressing cells, you don't see much of a difference in terms of tubulin methylation, but you see a distinct decrease in tubulin methylation in the mutants or mutant expressing cells. So the next thing then was to see what's going on using microscopy at the mid-body.

[31:40] antibody and the methyl, using the methyl antibody that we described before. And what we see is that in the S20ATA, we actually see a reduction in methylation of the mid-body. Now this is just fresh data that we got last week and so we haven't quantified multiple experiments which we're doing currently, but we do see

[32:00] increase in methylated tubulin, we also start seeing an increase in chromatin bridges, which is a chromosome segregation error. And since we were seeing this, we also wanted to see what happens to genomic instability in these cells. And we took this population of cells expressing either the wild type or the mutant

[32:20] said D2, and we see an increase in micro nuclear formation. But since this was a population, we also decided to go ahead and select clones, and what I've shown you here are three clones from the wild type expressing cells and three clones from the S2080A mutant cell lines.

[32:40] Hopefully you can appreciate that there's a lot more micronuclei in these cells with the mutant rescue. Suggesting that this phosphorylation was in fact required for tubulin methylation and maintaining genomic stability. The other approach we've taken is using an erosivease.

[33:00] an aurora inhibitor to see if that would prevent CD2 methylation on microtubules. And here what we're seeing are human kidney cells where we have stained in green with the methyl antibody again. And what we're seeing here at very low doses of amylin, which is an aurora inhibitor, we start seeing a reduction

[33:20] in the methylation of microtubules. Now the reason we're using these extremely low doses of aurora is because at high doses of aurora, which is normally used traditionally in literature, we actually completely annihilate the normal mitosis. And so to prevent that artifact, we essentially use these low doses and you can see a nice reduction.

[33:40] in the methylation of microtubules all the way up to about 200 nanomoles of the inhibitor. So what I've shown you in summary is that aurora phosphatates sedate. I've shown you that its phosphorylation is required exclusively for microtubule methylation, but not so much for the methylation.

[34:00] much for its chromatin function. And this phosphorylation also seems to be required for maintenance of genomic stability. So what we've now gone on to do is ask if phosphorylation of CetD2 is required for its microtubule methylation activity, could we in fact inhibit aurora and cause cells to undergo mitotic catastrophe?

[34:20] And so to do this, we set up a colony farming assay like Ching just showed you where aurora inhibition was used again in extremely low doses with a CEDD2 inhibitor. So in the recent past, we've had the good fortune of having a CEDD2 inhibitor. And essentially, the second lane is with the

[34:40] sedative inhibitor alone. But then when we combine both the aurora inhibition with sedative inhibition, you start seeing a decrease in the ability to form colonies. This is true for A498 as well as the 786Os. These are VHL deficient cell lines, but they are proficient for sedative.

[35:00] We've also done this in a genetically modified system. This was a cell line that we obtained from Dr. Ian Froude's lab where he has A498s that are CEDD2 proficient and he's crisped out the CEDD2 and again, we start seeing this decrease in the ability to form a choline.

[35:20] disease. So this suggests that, you know, modulation of aurora is likely a good avenue to pursue. Now I wouldn't say that we would take these aurora inhibitors directly to the clinic because like I mentioned, they're extremely toxic and there would be absolutely no one who would be treating patients with an aurora inhibitor.

[35:40] is really crucial because, as I mentioned before, aurora is present in multiple pools in the cell, and it is absolutely critical that we understand what function of aurora is actually being modulated by some of these epigenetic modifying enzymes. And what I have not shown you here is at NSD1, which is also present on

[36:00] 5Q is also modulating aurora in a slightly different way. And so hopefully next time when I come back we can talk a little bit more about that. So I'll stop here and thank the team that's done all the work. Most of this work was driven by a postdoc manga Motrapu who has been driving

[36:20] This work, we have fantastic collaborators, both at Baylor as well as MD Anderson and University of Freiburg. And finally, thanking all of the funding sources for believing in this work and giving us an opportunity to try it. So thank you all and I'll be happy to answer questions.

[36:40] Thank you.

[37:00] No. No. Okay. So it's property driving from the current... Yes....or a candidate's activity? Yeah. Okay. So, CZ2 is lost in BHL diffusion tumors. So you are blocking aurora and then you are...

[37:20] saying that cells don't like it. So what are the targets of fororin, which are important for survival then? That is a good question. So we wanna look at the earlier events. We wanna look, before you've lost both copies of CetD2, because that was sort of the goal of trying to understand where these cells, how these cells are.

[37:40] actually surviving. But Aurora of course has multiple targets, right? So CEDD2 is not going to be the only target that Aurora has. And so I think with this study the goal was just to see what happens in 3P loss and that's the reason why we moved to that. But you're absolutely right. There will be many, many other targets that Aurora would be hitting. Thank you.

[38:00] No question. Maybe I can ask one that I suppose was in line with Rachel's question. Any role for primary cilia embazole body staining location of SCTD2? Just to make sure that I don't just jump because the moment I hear primary cilia I get very excited.

[38:20] Yes, there is a role for primary cilia, I mean, CEDD2 and the primary cilium, which we've not discussed. But as Rachel pointed out, apparently I'm wearing a shirt that has primary cilia all over it. So if anybody wants to talk more about cilia, I'm always welcome. Excellent, thank you so much. Thank you. Thank you.

[38:40] I invite our next speaker to the podium, Dr. Sakari Van Haranthe from the University of Finland. He'll be talking about transcriptional control of renal carcinogenesis. Dr. Van Haranthe. Okay, thanks so much. Thanks so much for the invitation. It's been a fantastic meeting.

[39:00] And I come from the University of Helsinki. And my lab is really a cancer biology lab. But we use renal cancer as a model for many of the reasons that have been highlighted throughout the meeting. Just like Samra said yesterday, that this is a great context to study tumor evolution. I think it's also a great context to try to understand basic.

[39:20] mechanisms of cancer biology. And so here we have a very, very reductionist view of cancer where there's a normal tissue, eventually it gets a mutant cell, and then something happens, it becomes a carcinoma and then metastases. And in the context of kidney cancer, I think it's been really well-

[39:40] it's a VHL null cell that initiates this process. And despite all these developments, there are a lot of open questions here. One of them is, for example, why is it that this leads to cancer in some tissues, but not in others? For example, colon cancers don't have these mutations. Why does it take so long for this process to happen?

[40:00] And then how can some of these tumors eventually become metastatic and others do not? And so it's in this context of VHL-Now oncogenic signaling that we try to understand how cancers develop and progress. And the way we kind of think about this is that a lot of this happens

[40:20] because of control of transcription. And one reason to think like that is that the tissues are defined by tissue-specific programs, defined by our DNA, and that leads to tissue-specific transcriptional programs. And it's in this context that the mutations happen, and I think this is one reason why this process is so very tissue-specific.

[40:40] Okay, so essentially we have a transcription factor network that then gets perturbed by a mutation. And again, in this context, we know that this is the HIF signal, this is a simplified version, I am aware of HIF I and all that. But in this context, it really is HIF II that then eventually somehow changes this program and then somehow this leads to metastatic.

[41:00] So I kind of entered this field from the end through work where we tried to understand how advanced renal cancers sometimes become metastatic and we tried to understand, we developed systems where we can study this and we identified the mechanisms where the VHL HIF2 pathway actually changes its output.

[41:20] through enhanced co-option or hijacking. And this is an example of that. So the CX-CF4 is a gene that promotes hematostatic fitness and we could identify these kinds of specific enhancements that is regulated by NF-KAP-B and in collaboration with A-HIF2-dependent enhancement then regulates metastasis and we've published these and there's some details.

[41:40] there. We've also studied the links between the tissue-specific transcriptional networks and the HIF2 signal and tried to understand whether their interaction would somehow explain, at least partially, why these processes are tissue-specific. Here we have identified, again, another enhanced level interaction where

[42:00] There's a paxate, which is renal in this factor, and a hip to bind at the same enhancer, which are upstream of second D1. And this is important for second D1 expression. And it happens so that there's actually a variant. So this 11, chromosome 11 locus has been linked to increased renal cancer risk overall, and there's a protective factor.

[42:20] that seems to disrupt this paxate binding site here. And we, again, experimentally tested this. And a prediction that arose from this work was that this risk is actually linked to CCRCC and not popular RCC. And so we collaborated with people, with Stevens Chanax Group at NCI, I think, and tested this.

[42:40] specifically and we can actually see quite nicely that the SNP, which affects paxate binding is associated with CCHCC and not piperine. Again, this is work published, but this is just these are examples of how we can use really this pathway to study mechanisms of tumor

[43:00] development and hopefully by doing this more and understanding better how these tumors develop, there might be also opportunities for intervention. Okay, so but now I'm going to talk about a slightly, again, different question related to the same thing and this is trying to address the question of why does it take so long for these mutant cells to

[43:20] to develop cancer. So and this came part of this work from this mouse model that was developed by our colleagues at the Cambridge at the time when I was there. Still, Athena Matagidou, who made a mouse model where they disrupted P1 and VHL, and they found that eventually these mice developed tumors, and this is what has been done.

[43:40] been shown by others as well, but there's a long delay in this process. And what we want you to simply ask, what happens during this time when there supposedly is silence essentially, phenotypically, but obviously something happens. You know the mutin cells are there and we know because VHL disrupts the transcriptional

[44:00] control mechanism so that we should be able to identify the cells right from the beginning and then phenotypic track those over time. And we did this by single-tonal RNA sequencing and also ataxic sequencing by just sampling these mice at different time points where we expected them to be kind of phenotypic more normal. And what we see here is just a U-map where we have the

[44:20] proximal, epithelial cells, S1, S2 and S3. These are normal ones. And what we could find is that all these different subsegments produced knockout cells. But through a careful inspection, we found that only the S1-derived cells knock out one cell that seems to progress.

[44:40] What we saw is we see this hip up-regulated population right from the beginning, which is expected, but nothing really much happened for a long time, but only several months later, most prominently 20 months later, we see some sort of progression. Okay, and we did all kinds of analyses.

[45:00] try to understand what this could mean in practice. And obviously this is transcriptional data so we can look at the signatures and genes and try to understand what that could mean in practice. There was two trajectories. The first one went down from this origin. We termed this population origin. And this was there right from the beginning. And then there's this other one.

[45:20] the trajectory that then ends up going to this endpoint with which is endpoint of this model. This is not necessarily very advanced, but this is still something that has started to move from the VHL-null cells that initially didn't do anything. So looking at the transcriptomics, what it seemed that this first translation from origin to translation is

[45:40] associated with very strong upregulation of transcriptional responses, and then there's further modulation of that when we move down towards the endpoint. The first transition is also very strongly associated with AP1 factor activation, and we could also look at correlation with clinical samples, to ensure that it's the endpoint that then starts to transcriptional.

[46:00] globally, a little bit resemble cancers. They are not cancers necessarily, but they're still progressing from there normal. Interestingly there were a few hits here that kind of caught our eye. KLF6 for example, we had through a completely different line of research identified as a factor that promotes renal cancer growth and actually there's an enhancer here.

[46:20] that is highlighted to think cancer is C, but it's also dependent on paxate and HIF2, and so this kind of fits with the idea that some of these processes are tissue-specifically encoded. And also, another one was that there's this, QIM1 is upregulated here at the endpoint, which is a marker of kidney cancer that can be also even used in the clinic.

[46:40] So this was quite encouraging. But also these signatures made us think of a field that we had not really talked about or thought about in the context of kidney cancer before, which was the field of renal injury, which is obviously well-developed and clinically important on its own right. But this is

[47:00] is an experimental thing that or approach has been used in the field quite a lot where people have clamped the renal artery and then released that to cause damage and then transcriptionally profile these cells for. And there's multiple papers that look for this and there's these waves of transcription responses that then initially

[47:20] change the injured cells and then allow them to come back, whereas if there's too much of damage, the injury programs fail. And this was interesting because when we look at the transcriptional, these are some of the examples. This is not our data, but our analysis is from this paper. So there's this really, these waves. It's massive upregulation of this, say, P1 factors and others, Kale-Ris-6.

[47:40] And then this second wave that comes later, which is, for example, marked by Sox-9. Okay, so this was interesting for this reason. When we go back to this trajectory and we look at some of those marker genes of this, they actually behave very similarly. So they get induced in this transition 1. But unlike in the injury models, they

[48:00] keep going up towards the endpoint. And then genes like Stoch 9 are only up regulated at the end of the endpoint. So this was clearly a parallel. It looked very similar, but it's also at the same time somehow different, but there's also the mutant cell. So this kind of makes sense. And we've looked at this in multiple different datasets.

[48:20] another dataset of renal injury. The top ones here, the cells are healthy, cells these are injured and this is just a signature that marks the first wave of injury and this is a signature that marks the failed type of response. But when you look at our endpoint signature it's kind of marking both the

[48:40] usually injured and the late or failed repair. So this again just goes to show that this is the same program, just sort of the same program, but at the same time also not the same, so it's somehow deterred. So we've now then tried to understand whether this is functionally relevant and tried to correlate further into clinical datasets. Also we've got some

[49:00] help from Iain Froust's team, who has a collection of different mouse models of renal cancer. And obviously these are things that are hard to study in humans because most parotidics are diagnosed late. And obviously the virtual syndrome is a context where we can see this early lesions, but this is actually from the mouse model. And again we see the same what we saw here.

[49:20] So we have Sox-9 expressing early lesions and they also have a skin 1 marker up. And then we have looked at human datasets and the oil phase. This is actually from a VHL patient, our Cambridge colleagues. So this is normal. There's hardly any Sox-9, but this is an early stage CCRCC and they have Sox-9 activated.

[49:40] clearly evidence from the mouse models and of all kinds of different mouse models and human datasets that this pathway is active in cancer. This is actually a metastasism that we had used earlier to understand the metastatic processes. They also express as ox-9. We can knock it out by CRISPR and we can see

[50:00] the phenotypes in mice and we have also other expression studies and things like that. So it seems that this pathway is functionally important for tumor formation. So I guess that's I think all the, oh one more point actually. So one thing that we noticed also through sort of control experiments

[50:20] because we have followed them eyes for a long time, we wanted to look at also what happens in the normal non-recombined renal epithelios. And actually what we found is that some of the same markers go up also over time in aged renal epithelium. And there's this data by Yang et al. from Science,

[50:40] This is obviously, again, hard to have different ages from humans, but this is public signal cell analysis data. And it seems that these young kidneys don't have as much activation virus than if you go to older individuals, you start seeing more cells that have these markers up. And I think this kind of links these signatures, the injury.

[51:00] and the tumorogenic processes and age in a way that I think suggests a model. And I think this is maybe not a, I can't say that this is a summary of what we have, but I think rather like a hypothesis that we should explore in the implications of this. Because we think about this in a cartoon, we have the normal epithelium and it sometimes gets injury.

[51:20] You could think of all kinds of risk factors and things like that that might cause renal injury. Most of the time, perhaps, it gets resolved and there's no major consequence. But it could be that when you have mutations, that affects the way this injury gets resolved. And perhaps that has something to do with how cancer develops, how perhaps the risk factors are associated with mutations, and how that could be done.

[51:40] would lead to cancer. So I think this is more like an idea that I think we should explore in this context going forward. Okay, so with that I think I'd like to just conclude that so from the earlier projects that I actually just kind of introduced as very, very, very important.

[52:00] shortly as examples, it's quite clear that the effects of the mutation, they do evolve over time. And it's clear that the first clones that have lost to the H.L., they don't the mutation doesn't have the same effect in those cells than it has in advanced cancers. And some clear examples, and we kind of tend to go quite

[52:20] deep and characterize specific examples as kind of conceptual examples quite carefully. And so we have evidence that it happens at the metastatic site. So basically some metastasis genes are clearly activated by the tumor-initiating mutations, but there's been epigenetic remodeling that.

[52:40] that allow that. And also we've linked this to tissue-specific programs which I think are quite interesting. As examples, now, PACS-8 is expressing other tissues as well, not just in the kidney. So this is not explaining everything, but there's clear and I think quite strong genetic evidence of the PACS-8, at least in this one locus of effects, the interaction with is inter-

[53:00] interacting with the HIF2 pathway and that I think contributes to the tissue-specific nature of this. How about these other dermatytes in the VH-relat syndrome? We don't know, but I think the same concept could quite well be true there. And then this second part I think links this kind of aging and tissue repair processes to the

[53:20] to the kind of mechanisms that then drive or allow the mutant cells to progress. And here I think there's quite a lot of open questions. We're trying to address some of those in this project, but I'm sure there's still plenty left. And I think this could be interesting, especially in the context of inherited mutations, and there's been a lot of discussion of how to perhaps prevent that.

[53:40] eventually, and to understand the mechanisms that allow some clones to develop, I think something that we should try to think about quite carefully. So I think this is all I wanted to say, just to thank the lab and the funding agencies and many, many collaborators that we have the privilege to work with. So thanks so much for your attention.

[54:00] stay some questions. We have time for a couple of questions. Great talk. I'm intrigued by the repair process of tubular injury and the

[54:20] SACS-9 rule, have you, because there is a switch between SACS-9 off and on in terms of tissue repair, have you actually examined this further in tumor progression? I think that's a great question. So we have not yet. So this is a very,

[54:40] slow model. So it kind of makes it a little bit hard to work with it. But I think this is exactly a good question. And we don't know, but I'm kind of guessing that somehow the socks off bit kind of might be inhibited by the mutations at least sometimes, and perhaps that could be. It doesn't even have to be a complete emission. It could be just the probability.

[55:00] of that reversal not happening gets less and that will eventually allow. And this is the kind of thinking that I'm kind of entertaining at least that it doesn't have to be all or nothing but rather a change in the probabilities of these MITN cells ending up somewhere or not. Thank you.

[55:20] Hi, this is very interesting and kind of insightful, but did you have a more kind of understanding of the injury response based on the genes that you presented? Like is it

[55:40] Is the injury response involved kind of ER response? Is it involved in DNA repair response? I mean what is, and the reason I'm asking these questions is can you tailor it, which is this is what you're going to do, to pharmacologic intonation?

[56:00] intervention?

[56:20] models in this and there's multiple papers that showed us very similar datasets and there's also kind of drug-induced damage that response is very similar but I don't think it's known what induces that and so again in the mouse model is something that's possible to study but it's quite challenging. We're trying to set up this kind of experiments now

[56:40] I cannot answer that yet. But there's a lot of things that are going on, so I think there could be definitely possibilities for that in the future. So one more thing. To what degree this gene expression profile of the injury that you see in the double knockout correlates with just a simple heft two?

[57:00] or a HIF1 or a hypoxia signature? Um, no. I don't think so. I'm not maybe sure. I understand exactly. Do you mean like in established CCRCC data, there's STPC correlation between HIF2 induction and injury?

[57:20] To what degree correlates with your double knockout transition signature? Right, yes, it does. Yeah, so it is present in the endpoint and it's, of course, mixed then with other injury signatures. So we see the hip signature already before we see the injury signature, but then at the endpoint those are mixed.

[57:40] And then we also see this failed repair and injury model at the end. So it is something that is not quite the physiological response. It is somehow different. And we're trying to understand why. So I just have two quick questions. One is about the, is in the context of familial.

[58:00] renal cell carcinoma entities. Would this be analogous in, for example, tubular sclerosis or other simples? Can we draw a larger context from that firstly? I think the data could well be in agreement with that idea. So because it is not in any way specified by the VHL mutations, but this is.

[58:20] seems to be like an epithelial injury response. We see, and E.N.s mouse models also have different genotypes, but they also show the same thing. So I definitely think this could be a possibility and something to keep in mind going forward. Yeah, absolutely.

[58:40] Have you measured a fibrotic upregulation of any flavor? And we have not looked at that yet, but I think that's a great question. Thank you. Thank you very much, Sakari. That was a great talk. Thanks. We need to move on. The next speaker is going to be John Nicola.

[59:00] Genovese from the MD Anderson in Houston, and he will tell us about cross-pieces functionalization of the cancer genome. How do I go forward with the slides?

[59:20] Is it just with the mouse?

[59:40] Okay, thank you. So first of all, I want to thank the organizer for inviting me. This has been really a great learning experience. I am still relatively new to the field. So just to give you a little bit of background, my laboratory is interesting.

[01:00:00] in understanding how solid tumors evolve, and we do so by using an in vivo cross-pieces functional genomic approach, mostly leveraging the genetically engineered mouse as a model system. So today I want to share with you some of the

[01:00:20] work that has been done by a very talented postdoc in my group, Luigiparale. So when he joined the lab, he basically asked a very basic question, which is specifically in the RCC renal cell carcinoma, which, as you know, is a relatively, usually relatively, indolent tumor.

[01:00:40] the emergence of clones with metastatic features. And although we have a very deep understanding of the genomic landscape of disease progression, and this is also thanks to the seminal work from some other TRACERx, Charles Wonton, and other people, we, at least at the time, were

[01:01:00] lacking a model system, an experimental model that would recapitulate evolution and the biology of advanced metastatic disease. So as you know, renal cancer are extremely heterogeneous from a clinical pathological standpoint and genetic standpoint.

[01:01:20] But I think one common theme is the fact that those tumors are driven by inactivation of several tumor suppressor genes, many of those on chromosome 3P. So those genes are often embryonically little or they really mess up the function of the organ when you condition.

[01:01:40] inactivate them in the kidney or in other organs. So making a mouse model of renal cancer has been challenging. So to address this question, we elected to use a CRISPR-Cas9-based somatic mosaic editing approach.

[01:02:00] Basically we crossed a tissue-specific query combination here under the PACS-8 promoter with a conditional Cas9 allele. So we basically activated the Cas9 during embryogenesis across the entire nephron and the reporter system which is activated by the RNA.

[01:02:20] by sequential exposure to a cre and a flip recombinase. So now we basically transduce this animal orthotopically by surgery or systemically with polycystronic adenosociated viral constructs carrying packages of gatarina against those genes or genomic regions.

[01:02:40] of interest if you want to do more complex chromosomal engineering, along with a latent flip recombinates that can be activated by creamed inversion of this cassette. So here basically the crear recombinates, activates the Cas9 in the nephron,

[01:03:00] removes the first stopper flanked by loxpicide in the tomatoe reporter and then only in the cells transduced by the virus activates the flip recombinase. Now the flip recombinase removes the second stopper from the tomatoe reporter which is flanked by the fret site. So it basically allows you to

[01:03:20] track the fate of the cells that you're editing. So last five years we tried a lot of genetic combination. We were able to model many tumor types. Here I'm reporting a model of a VHL driven clear cell renal cell carcinoma. You can see the all marks of the

[01:03:40] disease. The prominent neongeogenesis both at the microscopic level and at the microscopic level here by CD31 staining. The expression of the lineage marker, PACS8, and then if you look at the classical H and E slide, you will see the prototypical cytoplasmic clearing, both in low grade and in high grade.

[01:04:00] tumors. But I think the most important piece of information that we gained from this model was the fact that if you disrupt the cell cycle checkpoint through genetic engineering of a genomic region on the mouse chromosome 4 that is seen.

[01:04:20] centanic to 9p. So basically taking down cDKN2A and cDKN2B, now we're really capable of inducing tumors that are not only local invasive, but also metastatic. And the pattern of metastatic spread you can see is pretty similar to aggressive human disease with splintic metastasis.

[01:04:40] lung metastasis and we also had 10% of mice developing brain metastasis. Astrophotologically those tumors are poorly differentiated. We see a lot of sarcomatoid or rhabdoid differentiation. So this work was published last year and we basically were able to demo.

[01:05:00] that the progression to aggressive disease is associated with emergence of chromosomal stability and disruption of the interferon response to as an adaptive mechanism to chromosomal stability, to tolerate chromosomal stability. So then we

[01:05:20] I wanted to ask if there are specific vulnerabilities that are specific to highly aggressive disease. So we generated highly metastatic and low metastatic clones that we called M plus and N minus, and we perform a genome-wide CRISPR Cas9.

[01:05:40] screen. And to our surprise, running pathway analysis on the top candidate targets emerging from this from this screen, we found that the M plus highly metastatic tumors are highly reliant on oxidative phosphorylation and complex

[01:06:00] one, mitochondrial complex one. This was through both in VHL-driven and non-VHL-driven tumors, which for us was counterintuitive. So for a period of time, we were kind of stuck on those data. Then I have to say this was very fortuitous thanks to Jim Brogaro.

[01:06:20] as we got in touch with Ralph De Berard in his lab. So DIVIA in his lab basically did a very elegant experiment. So they basically they took patients with renal cancer and they infused those patients before surgery with heavy carbon-label glucose.

[01:06:40] we track the fate of glucose in the TCA cycle by mass spectrometry at primary and metastatic site. And what they found is very interesting and in line with our findings is that the incorporation of carbons, the right for

[01:07:00] from glucose in the TCA cycle is increased in metastatic sites and that metastatic lesions are more reliant on oxidative phosphorylation. Now if you look at transcriptomic data, you would also notice that, especially in advanced stage tumor, the presence of an ox-force signal is increased.

[01:07:20] of, a signature of i-mitochondrial DNA is associated with poor prognosis and is associated with the i-tumor grade. So I also wanted to mention that in my lab we actually validated these through special symphonies.

[01:07:40] special transcriptomic sequencing. And we looked at primary, matched primary metastatic site, and indeed we found overall at metastatic site a suppression of the if signaling signature and an activation of an oxfosignature. So these are all consistent with previous results.

[01:08:00] So, we joined forces with the Ralf Group and we designed some functional experiments. We used both pharmacological and genetic tools to suppress the function of complex I. We did that pharmacologically in M plus highly metastatic clones.

[01:08:20] And indeed in that case, we notice almost a complete suppression of metastatic progression and activation of complex I, hyperactivation of complex I in poorly metastatic clones. We did that by over-expressing east an adage de hydrogenase, which is extremely potent.

[01:08:40] And both in non-VHL-driven and VHL-driven tumors, we notice a profound suppression of metastasis when you block complex I. And this was associated with a decrease in cell proliferation.

[01:09:00] These work was published in August this year, and now we have a, we can add a new piece to the puzzle where we basically know that the progression from poorly metastatic or non-metastatic disease to highly metastatic aggressive disease is associated with a metabolic rewiring toward.

[01:09:20] oxidative phosphorylation. Still have time. So now I want to share with you a little bit of a different story. Just letting you know this is very preliminary, so please take it with a grain of salt. Here we basically wanted to study

[01:09:40] tumor predisposition syndrome using in use pluripotent stem cells. And you heard about this approach by other investigators yesterday. So the question here is what drives the tissue specificity of tumor predisposition syndrome susceptibilities?

[01:10:00] why some mutations like the one that we observe in the Liferome-Ani syndrome are more pleiotropic, affecting virtually any organ with different tumor types, mesenchymal versus epithelial, and why in other condition you see very specific tumor affecting very specific organs.

[01:10:20] So to study that, we used IPSCs. So we basically collected peripheral blood from patients with genetic syndrome. We collaborated with the GEO oncology clinic with peripheral blood.

[01:10:40] breast oncology, GI oncology, and also with the pediatric clinic now. So we established the IPSC and then we performed teratomyces. And the idea here is to use digital pathology, special genomics, special transcriptomic to look how the mutations are affecting the

[01:11:00] tissue commitment within ataratoma. So this is what you usually see in ataratoma. You see tissues derived from the three different germ layers. This is from an healthy control. To be honest, I had to

[01:11:20] roll up my sleeve to provide the blood for this. You can probably appreciate the typical southern Italian pigmentation here. But more importantly, at this point we were able to generate IPSCs from many type of patients.

[01:11:40] patients with cancer predisposition syndromes. And just to give you a little bit of preliminary data, again, take them with a grain of salt. Here I'm showing a teratoma generated from a BAP1 syndromic patient. We see a very strong bias towards

[01:12:00] neuroactidermal and neural crest differentiation. Here you have another set of teratomas from a VHL patient. This is a C393A mutation. Most of the teratomas were cystic in nature. However, in a couple of years,

[01:12:20] couple of them like this, M2 here, we saw somatic transformation towards a neuroendocrine phenotype. You can see the expression abstaining with chromogranin A here. And very interestingly, this patient had an history of a pancreatic neuroendocrine tumor. Now, it doesn't mean anything but for me it was

[01:12:40] quite interesting to see this phenotype. So the hypothesis that we're testing here is to basically look at how different type of second eats on VHL or other genes for other syndromes are affecting the lineage specificity, the lineage specification in.

[01:13:00] in those teratomas. And with these I want to leave some time for questions. I want to thank the people in my lab. They're a very energetic group. Collaborators at MD Anderson and across the institutions and the funding agency and I'm happy to take questions.

[01:13:20] super exciting Jennifer are there any questions that's really fabulous all of it in terms of the metabolic rewiring in the metastatic phenotype and I'm probably being over simplistic

[01:13:40] Given that you've got CDK into A as the driver there and its role in cell cycle progression, could this just be downstream of proliferative phenotype? And for example, in Scott Lowe's model, do you see the same metabolic search or is this really specific to BHL? That's a very good question.

[01:14:00] I don't think it's specific to VHL. What I think is that it's more specific to chromosomal instability. So when a tumor develops a degree of chromosomal instability, we see this higher dependency towards oxfos. It could be cell-to-cell.

[01:14:20] division dependent as well. I mean aneuploidy also drives. Potentially. So to be honest we haven't really looked at the mechanism of this switch. And a quick second question because with your teratomas, you said you're looking for investigating how the nature of the second hit.

[01:14:40] drives different phenotypes. So do you engineer it before you implant these or you put them in as hats and then you just wait for... We thought about engineering but we're just gonna be with the spontaneous model. So whatever comes out spontaneously and then we characterize with special genomics. Thank you.

[01:15:00] Thanks, Amr. Did you investigate what do tumors want out of oxfoss? Is it ATP? Is it an AD plus? Good point. I think it's just more food. They really need more, as you said, more ATP.

[01:15:20] more reductive potential. But to be honest, this is probably a question for Ralph and for myself. I'm not like that great of a metabolic person. Thank you. Hi, did Ralph look for...

[01:15:40] glutamine contribution? No, I don't think they, we didn't do it and I don't think they did it. I don't know if they're doing it. Because essentially the whole thing can be explained with more heif activity. Say again? The whole thing can be explained with more heif activity because heif would drive oxfos and

[01:16:00] and heave will drive also glutamine. Yeah, that's, I mean it's. Reactive corvoxidation. So it may all be compatible with the simple fact that when you go from primary to metastatic, your heave levels go high. That's very possible. On a more philosophical standpoint,

[01:16:20] noticed across cancer type, because we also work on pancreatic melanoma or other tumor, when tumor advances, they usually tend to lose some dependency from the driver initiating oncogen. And I also think that this is happening somehow in renal cancer. So they might be – I don't want to be blasphemous.

[01:16:40] ear, but they might be less dependent on the if signaling. In the humans they do accumulate mutations. We know this. So there is a possibility that they're becoming partially independent of the if, but you know in the setting of

[01:17:00] animal model that you are testing it with a specific inactivation of the pathways. I mean you don't really know if your metastatic disease is primary to more phenotypical independent. The only thing you know is more ox force. Yeah. No you're right.

[01:17:20] And the truth is that we didn't look at the mechanism. So I'm really waving my hands here. But yeah, this is a very interesting point to address, absolutely. Thanks. Just kind of an open question here about the regulation of tissue specificity in VHL disease.

[01:17:40] fascinating question. Absolutely. Do you think that my RNAs regulating HIF2 mRNA might be relevant to this whole story or is that something that we're not, I haven't heard anything, anyone talk about MRNAs in this particular meeting and I'm curious if that's even a topic that people are investigating. I mean yes totally

[01:18:00] possible. We just don't know and I will be open to any mechanism of regulation. BHEL loss wouldn't be relevant in tissues if if if if 2 isn't regulated. Yeah that's pretty, it's a fair point. Thank you.

[01:18:20] I invite our next speaker, Dr. Ian Froude, from the University of Freiburg. He's going to be talking about sex-specific differences in CCRs as he's separating the functions of KDM5C and KDM5D.

[01:18:40] afternoon everyone from my side. When I was trying to decide what to present to you today I decided something unusual for myself and that is to not talk about many mouse models that we're working on. We have several of them here as posters which we'd love to discuss with you including myself and the guys in the lab at their lunch break. But rather I thought what I'd do is follow on from the presentation that Francesca gave.

[01:19:00] you yesterday and tell you about an example of how we can use humans as a beautiful model organism to help us understand mouse cancer. And so there are two messages I'd like to give you here and for anyone who wants to check their emails for the next 13 minutes can listen now. We think sex-specific differences are something that shouldn't be ignored and we think that

[01:19:20] KTM5C and KTM5D are actually not the same. So to come back to what Francesca was telling you yesterday, obviously it's very well characterized that in sporadic cliosar, renal carcinoma, there's a very strong incidence bias. Men get cancers more often than women.

[01:19:40] And as I was preparing this presentation, I thought I'd just write a little line saying that it's the same in VHL patients with CCICC. And I started googling and looking and actually it turns out I couldn't find anything. And maybe the clinical colleagues can correct me here, but I think it's not really clear whether there are in fact differences in CCICC incidence in male and females in VHL patients.

[01:20:00] patients. And I wrote a quick email to Atina Gana, the colleague from Freiburg is actually here in the audience, and asked her this question. She said she didn't know, but she looked in the Freiburg registry of VHL patients, as shown in this plot here, and characterized only those VHL mutations that were defined as being loss of function in the recent Nature Genetics paper.

[01:20:20] screening for different functional variants. And what she saw was sent me back within about half an hour this nice graph that shows that at least for this subset of patients there is actually a difference. And I think this is a maybe to my mind an unanswered question and we'd love to talk to people in the field who maybe have similar data that we could think about.

[01:20:40] So, I think there is an intrinsic difference in the biology of tumors between males and females. And ignoring now for the moment any hormonal differences and physiological differences, as Francesca said, what's clear is that there are genetic differences. Most strikingly that KTM5C is mutated in male CCI, CCI, and in other cases, in other cases.

[01:21:00] disease. And as she said yesterday, that 5C in females, in women, there are two active copies of KTM 5C in males. There's one copy of KTM 5C in 1 of 5D. And it's thought in general that KTM 5D functions as a surrogate KTM 5C in men.

[01:21:20] the very high amino acid homology and the fact that both of them have this same biochemical capacity to change or demethylate the trimethylated mark on histone H3 lysine 4 to a monomethylated. And by acting directly at promoters and enhancers and changing the balance of mono and triads thought that the KTM5 is actually

[01:21:40] thereby regulate gene expression. Two other little important things. We know from many talks yesterday that KTM5C is not a truncal mutation but arises later. We see that, for example, in higher frequencies, subclonally, and in metastases. KTM5C also, along those lines, can be found together with

[01:22:00] many different CC-RCC mutations, so most prominently with polybromo and CT2, actually never with BAP1, but we also find out with P53 and MTOR and P10. So it seems to cooperate probably in different ways with different mutations. Another interesting thing is that when we look across all tumor types, KTM5C mutations, so the typical

[01:22:20] pattern that one might expect of a tumor suppressor. The mutations are scattered throughout the entire length of the protein. The black ones here are quite striking. They're truncating mutations that are predicted to cause complete loss of function. And what's interesting in kidney cancers is that these are more prevalent, these truncating mutations, in about two-thirds of the cases.

[01:22:40] quite a strong selection for really killing KTM-5 function. So that's what we decided to do in human CCRCC cell lines. We designed SGRNAs to inhibit or to mutate KTM-5C in exon-8, and then put these in a lentiviral, using lentiviral CRISPR,

[01:23:00] into populations of CCICC cells that we know through exome sequencing are actually KTM5C wild type. These are the four cells we've been working on here. You see KTM5C is nicely expressed and again when we generate these populations of cells in each of the different cell lines, we can nicely get rid of KTM5C protein expression.

[01:23:20] We deliberately work with whole populations of cells here because we want to avoid any potential clonal artifacts that can come from selecting out individual clones. What we see now in all of these pictures, all of these diagrams, you'll see the control cells in grey, the two knockout clones in blue. We basically see no difference.

[01:23:40] in proliferation in normal cell culture. There's a very small difference in the two female cells in terms of increasing colony formation. This is not true in the male cell line where KTM5D may be compensating. And I'll show you a little bit of data about that in a minute. We see differences in zenograft assays. So in 7, 8, 6, 0, there's no effect. In CACI 2, there's actually a block.

[01:24:00] or a slowing of tumor formation, there's an acceleration in A498 and 769P cells which don't grow as tumors normally also don't grow as tumors when we've knocked out KTM5C. So we see different kind of phenotypic effects, albeit fairly mild ones of KTM5C deletion or mutation in these cells. Transcriptionally we also see

[01:24:20] actually relatively mild effects. So there are very few genes whose expression is changed in the knockouts. And the interesting thing is that there's basically no overlap in the upregulated or downregulated genes between the different cell lines. And I think this is consistent with the fact that KTM5C is not really a primary transcriptional regulator but is probably recruited.

[01:24:40] to different sites, depending on the underlying state of the chromatin that's governed by many other factors. And consistent with that, when we localize KTM5C peaks in A498 or 7860 cells with cut and run, you see that there is a small overlap, but largely the binding patterns are quite distinct. So it's probably that KTM5C has cell types

[01:25:00] specific functions. We can confirm biochemically at least that KTM5C is doing what we think it should be doing. So when you do cut and run for all of the KTM5C binding sites, you can see that it is enriched in sites that have monomethalation, which is the mark that's actually kind of written by the D.

[01:25:20] methylase activity of KTM5C. Some of these are probably active promoters because they're high in trimethyl. And these sites are actually reduced, as we might expect, in the knockouts. What's kind of interesting is that when you just look at all of the sites across the whole genome of monomethalation, we see that globally these are actually reduced as well in the knockout.

[01:25:40] doubts, while KTM5C is not actually binding there. So either the cut and run antibody is not fully efficient or these are sites that don't kind of require ongoing KTM5 activity. Now you might expect that the differentially expressed genes should be enriched for binding of KTM5C, either at their promoters or enhancers, based on the common model.

[01:26:00] this diagram shows you that that's actually not the case surprisingly. So when we take in A498 all of the upregulated genes and look for promoters or enhancers or the downregulated genes, we don't see any enrichment for KTM5C binding, but we do see this sort of global effect on monomethalation. So we can't really say actually that KTM5C is directly regulating these

[01:26:20] these genes. Now of course you can't talk about KTM5C in males without talking about Y chromosome loss where KTM5D is located and the kidney in general kidney tumors actually show quite a high incidence of loss of Y chromosomes as shown in this cell paper. It's quite striking in papillae, CC.

[01:26:40] papillary renal carcinoma, but in CC-ICC it's about 40% of cases. What's not known is how this correlates with KTM5C mutation, and you can do that by simply looking at the expression of microemosome-encoded genes. They're basically also the same pattern. What you see in the KTM5C wild type is that there are many tumors

[01:27:00] that have very low expression of why chromosome genes compared to normal, we're only looking at male tumors now, and this number happens to correspond to more or less the 40% that has been shown by fish. What's very interesting is that almost all of the KTM5C mutant tumors have also lost KTM5D. So this is a state we decided

[01:27:20] to model, which I'll come to in a second. What I wanted to show you is that this Y chromosome loss is just a smaller side. We know that in the mouse model of VHLP53Rb deletion, we also see this gender bias, you know, the sex-specific bias, I should say, in mice, that males, male mice, get tumors earlier than female mice, they get more tumors. And what's interesting is that they also

[01:27:40] showed this pattern of about 35% in this case of loss of the Y chromosome apparently. So this is something that seems to be conserved across species and across cancer models. We therefore decided to now knock out KTM5D on top of KTM5C in the male cell lines and we chose 7, 8, 6, 0 which has kept the Y chromosome. It expresses

[01:28:00] KTM5D. Here we knocked out 5C, here 5D, and here the double knockout. And what you see again is basically no change in proliferation. Now we see actually this little increase in soft agar colony assay formation. But the reason I really want to show you this data is this surprising observation. When we put this into zenograph setting, as a

[01:28:20] showed you before, KTM5C has no effect. KTM5D surprisingly completely killed tumor formation. And this we could fully rescue by deleting KTM5C and 5D. So this clearly shows that 5C and 5D are not just siblings that do exactly the same thing, they're not compensating from one another. In fact, they're sort of antagonistic in that respect.

[01:28:40] So we see similar thing when you simply look at the tumors. So 7, 8, 6, 0 generally grow with this sort of mix of squamous and clear cell morphology. That's the same in 5C knockout. The 5D remnants of tumors look very more benign, and this appearance was rescued by the double knockout. There's a nonspecific trend to decreased metabolism.

[01:29:00] spontaneously from these subcutaneous tumors in 5C knockout. And again, this seems to be a little bit rescued by the double knockout. We see the same thing at the transcriptional level, that there are differing effects. So the genes that are upregulated in 5C or 5D are both distinct as well as overlapping. So they may have both overlapping and

[01:29:20] and distinct transcriptional functions. And one can see very nicely the dominance of KTM5C over 5D when you simply look at those genes that are 5C-dependent, shown here. They're obviously not affected by KTM5D, and if you knock out both again, you restore this. You keep this KTM5C path.

[01:29:40] pattern, but the opposite situation is when we have the 5D-specific genes, they're actually rescued by KTM5C co-deletion. And what's interesting is that many of these upregulated genes are putative tumor suppressors, so that may explain this effect that we're seeing. Again, when one tries to look at similarities and differences between 5C and 5D, we see there's a little bit of an overlap in binding.

[01:30:00] sites. There certainly seem to be binding to different sites, although I would caution here using two different antibodies, we don't know about their relative efficiencies in cut and run, so it's a little bit hard to discuss there. And without going through all of the data, we see exactly the same thing. The differentially expressed genes are not preferentially bound by 5C or 5D. So this speaks

[01:30:20] against this common model that is proposed in the literature. So this already brings me to the conclusion, and the take-home messages are that KTM5D is not simply the equivalent homologue of KTM5C under the X chromosome in females. We don't know biochemically how they're different. That's something we're really working on. But we do know

[01:30:40] that at least functionally when we mutate KTM5D, we have a KTM5C dependent change in tumor formation as well as gene expression alterations. Now this is a working model and I'm certainly drawing a very long bow when I try and claim this, but it's at least giving rise to these ideas that we hadn't really thought about before. And that is that in most

[01:31:00] cases at least in men of CCICC which are driven by all of the gene mutations we've been talking about, the X and Y chromosome are perfectly fine. So that's about 60% of cases, but in 40% of male cases an event happens that causes loss of the Y chromosome. And if we assume for a moment that as we hypothesize that KTM5D mutation is the same as KTM5D mutation,

[01:31:20] KTM-ers Y chromosome loss. We can't claim that and that's what we're testing. One might expect that this may actually act as a break or a proliferative block on the formation of tumors. We can speculate that one way of overcoming this is for the cells to select for mutation of KTM-5C, which would then allow tumor formation to continue.

[01:31:40] But what's also interesting is that there are many tumors where KTM5C is wild-type that have lost the Y chromosome, and something else has probably happened if this model is true that has let them get around this barrier. And we don't know what that is. It doesn't correlate with anything sort of commonly that we might think. Yeah, so that's something we're trying to understand, and I think it would be really wonderful to talk again.

[01:32:00] to all of the people who showed the beautiful work that they've done looking at evolution of CCICC to see if we can tease out a little bit more collaboratively maybe when 5C mutation occurs, when Y chromosome loss occurs. I think that would be a very interesting question. We'd love to discuss with you further. And the final take-home message is a more general one and kind of an appeal to everyone and that is that in

[01:32:20] In classical art, we would never confuse David and Venus as shown here. And I think in medicine in general, there's an obvious realization that men and women behave differently. But in the research context, in clear-cell renal cell cosmonomer, in VH-hole disease, I think we have to be very aware that there are very different physiology, there's very different genetics between males and females.

[01:32:40] It's really important to think about gender differences when we're looking at all of these things. And something that's actually I must say is not really reported on. There are many mouse studies, for example, where it's not even mentioned what sex the mice were. That's something that has to be improved upon. We have several examples. Basically, in all our mouse models, we always split the genders.

[01:33:00] As I said, we're very happy to discuss some of these interesting gender-specific or sex-specific differences in different mouse models that we have. And as a second aside, I'd love to talk to all of the kidney cyst people specifically at this poster, maybe if you have time in lunch, where we have another story that I actually would have liked to have presented today. So with 20

[01:33:20] seconds left I'll close it here. Thank you for indulging me and thank all of the wonderful people that I really have the privilege working with in the laboratory. Thank you very much.

[01:33:40] in 5D in terms of appliance effects on the H3K4 trim isolation, either on the protein level or on the chipsick level? Yeah, so in 8498, we very clearly see that at least at the chipsick level there's an effect. What's interesting is in 7860 cells, neither the single nor the double knockouts have a global effect.

[01:34:00] effect like we see in KTM in A498. We don't know any other cell lines. It may relate to possible compensation by KTM5A and 5B and different effects there. But we can't answer that question. It's a good question. Thanks.

[01:34:20] Is the PBRM1 dependent? Do you have a sense of whether that background is important in influencing this interplay with the KTM5C and ND? Or is it? We don't have any indication of that. The 786-0 model is polybromo wild type. It has P10 and P53 muting.

[01:34:40] presentations. A cell-line study of N equals 1 is very hard to conclude anything rigorously from that. We think we have some preliminary evidence from other studies we've been doing that knocking out polybromo may actually, as Ruiz showed and others have showed, may increase tendency for aneuploidy in chromosome develation.

[01:35:00] ability and select actually for Y chromosome loss. That's something we're working on now. This interplay between sort of these driver mutations then selecting for specific chromosome. And then the P53 mutant mouse model where you see worse outcomes in male animals, is that correlated with the third?

[01:35:20] where you see why chromosome losses in some way. No, we don't have the resolution to ask that. It's a good question. Exactly. Is why promoting or inhibiting tumor formation? We can't. No, I don't have that. It's a good question.

[01:35:40] Thank you. Wonderful presentation. Also, from all the molecular aspects. I will also give you a provocative point of view from a clinician that probably is less fancy regarding what you've shown, but the difficulty in treatment is completely different.

[01:36:00] different. So to operate on female it's much easier in terms of surgery, in terms of less toxic fat, less BMI, less comorbidities for a clinician to give an indication of an active treatment in a VHL patient. Female versus male is a completely different topic. It's very interesting.

[01:36:20] Thank you so much. Thank you. Super fun. So that concludes this session. I want to thank all the speakers for the wonderful lectures they gave and the audience for being.

[01:36:40] very engaging and stimulating a wonderful discussion.