Exploring the Molecular Basis of VHL – Tumor Heterogeneity in VHL ccRCC | Medical Symposium 2024

[00:00] Hello, hello. We are already running a little bit late so I invite everyone please to take your seats and we can start with the next session. So we will be in this afternoon first afternoon session we will be exploring

[00:20] bit more the molecular insights, get more molecular insights into VHL disease, particularly genetic, genomic data I assume, with a little bit of mouse modeling also as the selected abstract. And so that's it.

[00:40] It's a very exciting session that we have ahead of us. So I invite the first speaker, please, who is Tom Mitchell from the UK. And he will tell us about genetic profiles of kidney cysts and renal cell carcinoma. But oh, sorry. I'm reading the wrong title.

[01:00] He will talk about somatic mutations, but also of the influence of the tumor microenvironment, right, on VHL tumors. Please, the podium is yours. Okay, perfect. Thank you very much for inviting me to give this talk today to the organizers of the VHL Alliance. For those of you that don't know me, my name is Tom Mitchell. I'm an academic urologist based in the U.S.

[01:20] early cancer institute in Cambridge and it's a real honor to be here. We've heard such a diverse and influential talk so far so it's lovely to be part of that. I thought it would be useful again to frame the narrative and the and the inspiration really

[01:40] behind much of our work and I use this through the example shown here which is a 30 year old male patient who's already had a prior left radical nephrectomy for clear cell RCCs and in his remaining kidney as you'll see quite clearly on the CT scan

[02:00] multiple cysts and tumors present. So clearly his life's going to be beset by multiple interventions over time and we really need therefore to find ways to try and prevent the disease from forming in the first place to be able to control it better but also to be able to predict.

[02:20] Perfect, but also be able to predict how these tumors are going to behave in the future. And so as a surgeon, we often watch these patients who are under surveillance, and we tend to watch until a 3-centimeter-sized threshold is reached, and that's because

[02:40] based on lots of prior epidemiological data. Tumors three centimeters or less seldom tend to be metastatic. The risk is about 1%, but over time, that risk goes up quickly. And it allows us to stratify and for those tumors that are likely to become problematic.

[03:00] in the future they're going to be uncovered. And the problem with this is that growth of these tumors is highly variable. And I quote here the disarm registry, which although is not just for clear sol-RCCs, it's for a myriad of kidney tumors, a third of the

[03:20] these tumors showed no growth or shrinkage over time, whereas approximately one sixth of those had really quite uncomfortable growth or quick growth at about half a centimeter a year. And I've been interested in looking at, specifically working out what the factors are in these tumors that

[03:40] may constrain or facilitate growth. And earlier on this morning, we saw some lovely videos of the surgical resection of the tumor. And as surgeons, we get a lot of benefit from being able to handle the tumor, but also to give us a certain amount of confidence that we can detect.

[04:00] sex underneath the tumor and not be cutting through areas of tumor that's penetrated deeper into the tissue. And the reason why we can do that is there's a pseudo-capsule around these tumors and we feel this may be one part of the reason why growth may be constrained.

[04:20] look at the tumors under the microscope with a nice section through this tumor normal interface, what you see is a beautiful picture of your normal renal tubules at one side and then frequently see this dense infiltration predominantly of lymphocytes through to the fibromuscular stroke.

[04:40] coma before you come in and see typical, in this case, low-grade, clear-cell RCC cells themselves. And so we sought to understand the processes going on at this borderline of the tumors in much more detail by employing a multi-omic, multi-regional approach by taking

[05:00] from different regions, but also including the tumor normal interface, normal kidney, fat, and so on. And each of these tumor biopsies was then subjected to single-cell RNA sequencing, TCR sequencing, but also spatial RNA and whole exome sequencing. And if you look at the results of all this, you get

[05:20] a broad picture of all of the different cell types you might see in these important anatomical locations that can be finally annotated in terms of what they do. And it's when you start to compare where these cell types lie and how they're interacting with each other, you start to understand.

[05:40] understand the biology better. But of course, the first challenge was trying to work out exactly what the different cells were doing, and the most or the biggest challenge behind this was actually being able to categorize the clear cell RCC cells themselves, because if you've ever performed single-cell RNA sequencing,

[06:00] you see huge amounts of batch effect that depend on exactly how these cells have been treated, exactly what the microenvironment is like from patient to patient. We managed to derive some pretty robust metaprograms to allow us to categorize these cells.

[06:20] I'm going to concentrate on the two highlighted here. So this is the similarity to the proximal tubule. Now the proximal tubule is a cell of origin that we've talked about earlier today for CCR, CCN, therefore a similar profile. You'd imagine it confers a better prognosis, whereas the pro-invasive, pro-metastatic

[06:40] EMT profile, you'd expect to have a worse prognosis, which again, we're able to recapitulate looking at large-scale cohorts. If we then try and look and see where these different cell types are preferentially located, do you remember we had

[07:00] these bulk biopsies from different parts of the tumor, and you can calculate the observed to expected ratio. And we see by far and away the most enriched cell type at the interface with the proinvasive, progrowth EMT phenotype, whereas the proximal tubular-like type was very much depleted.

[07:20] that interface. And as I said at the start, we did some spatial transcriptomics from all of these patients, and this is an example slide of that here. On the bottom left, we've got normal kidney, the top right normal tumor core, and in the middle you've got the interface and quite clear.

[07:40] this EMT phenotypes are very much enriched at the interface, whereas for the proximal tubular-like cells, they're enriched further towards the core. And of course, the next question is what possibly could be driving this aggressive phenotype? And our first thoughts were that they're...

[08:00] might have been some sort of aggressive mutation present in these clones there. But we performed laser capture microdissection and whole exome sequencing, and these cells are the genomes of these cells were no different from elsewhere in the tumor. And so we looked to all of the cells of the microenvironment to see what difference

[08:20] also existed at the interface. And the most heterogenous, or the next most heterogenous cell type was myeloid cells, and in particular, tumor-associated macrophages. And we saw those one specific type of Tams that were very much enriched at the tumor.

[08:40] tumor normal interface and these were defined by AREG and IL1 beta expression as shown in the dot plots here. And we wondered therefore whether the ligands expressed by all of these macrophages might have the ability to cause this pro-metastatic EMT type.

[09:00] And so there's plenty of computational methods that one might use to look at this. This is called Neishnet, where essentially you can plug in all the ligands expressed by one cell type and look at downstream gene pathways in another cell type to see whether one could influence the other, and by far and away the most.

[09:20] influential in this model was I was IL-1 beta in promoting EMT in these cells. And so we next wanted to look specifically again at our spatial data to see whether these cell types are co-located. This is a different section from normal.

[09:40] kidney through to tumor core. Again, we can map PT cells, the proinvasive EMT cells, and also in this case, the loop of Henle cells. And we see these IL1 beta-expressing macrophage very much co-localized with the EMT tumor cells. In fact, in all of our sections, we had some sections of tumor

[10:00] core. We also saw pockets of EMG-high tumor cells, again, co-localizing with these IL-1 beta cells. And so we next wondered if there's any clinical evidence that one might use IL-1 beta to prevent tumor growth, and this is

[10:20] which is where we look back to the Kantos trials. So for those of you who haven't heard of this trial, the aim was to use IL-1 beta inhibition to prevent cardiovascular events, so heart attacks and strokes in an at-risk population. But there was this unexpected side effect where

[10:40] The new incidence of lung cancer was very much reduced at treatment. And these are thought to represent patients having very small tumors at the start of the trial which weren't apparent on imaging that were then prevented from presenting later on after the trial.

[11:00] And after this has been much more mechanistic detail on the role of IL-1 beta macrophages, first in pancreatic cancer where they've been found at the very early stages of tumor formation, where if you can suppress those you stop tumor growth.

[11:20] secondary in lung cancer in non-smokers where it's pollution that causes inflammation via the isle 1-peter axis which causes these pre-cancerous lesions to grow. And that's where I want to introduce this finding showed in this paper that pollution doesn't cause

[11:40] the mutations that drive lung cancers in this case. And if we look at the epidemiology of clear-cell RCC, so this isn't necessarily VHL disease, but in the sporadic form, all of the major risk factors being germline, genetics, obesity, or smoking, none of those are mutagenic.

[12:00] So, a single cell on its way to causing a cancer will have an equal risk of gaining the next stepwise mutation on the path to causing cancer. And it's likely that innate immunity, like I've shown before, is

[12:20] is playing a central role in this. And the risk factors, the only way these risk factors can act is by increasing the fitness of the somatic event that's present in one cell, and to put it very simplistically in

[12:40] In the VHL kidney, for instance, after your first hit, say, 3P loss and you've already got the VHL mutation, there will be a baseline growth rate, but if you increase any other patient-specific risk factors, then that growth rate will be elevated. You'll have a higher population of cells. You're more likely to go on the same level.

[13:00] to have the subsequent hit. Based on this, we've become very interested in tumor growth over time, and we're putting together a large cohort of patients who we've got multi-regional fresh frozen tissue, but also longitudinal scans.

[13:20] so we can start to tie together this interplay between what mutations are present and what risk factors are present to try and work out exactly how these factors increase growth. And I've just got a small amount of pilot data from this study where we see

[13:40] that those tumors with a clonal VHL event, so 3P loss and VHL mutation, they grow the slowest. That's by 1 to 2 millimeters per year, so typical growth rates of a small renal mass, whereas if you add in second hits, say a second hit in PBLM1, those tumors

[14:00] growth significantly faster, and then if you add on things like BAP1 mutations, then we see really very, very high growth rates. If we look at the data another way, by computing tumor size with growth rates, you see that actually these renal masses on the left, in this case, are not as high.

[14:20] case it's a 4 centimeter threshold that we might call small renal masses or indolent. Actually some of these are growing at exceptionally fast rates. And so we need to be very careful with these patients actually that we do not miss the therapeutic window. If we leave it too late then metastatic.

[14:40] spread could occur prior to treatment. And so in summary, we really need to move from this reactionary approach to VHLRCC to a more proactive one and it's fantastic to see all these new agents coming through. It means that hopefully we'll be able to do that more routinely. The tumor microenvironment of our

[15:00] really plays a critical role in cancer risk and is a promising target for prevention. And then finally, if we're looking at growth, somatic mutations are the strongest predictor of how these tumors are going to grow. But there's a really significant effect from patient-specific factors and these most likely are acting.

[15:20] through the tumor microenvironment. I'd like to finish by just thanking everyone in my lab and the chasm, but particularly my clinical colleagues who allow me to spend my time running a lab and research, but also actually all of our patient donors who, without their help.

[15:40] they wouldn't be, we wouldn't be doing this work and helping them. So thank you very much. Thank you very much. Are there questions?

[16:00] Thank you very much for the presentation and I think that your focus also to patient characteristics is probably the future also from a clinical point of view. I was wondering whether you can go a little bit more in detail in terms of how long

[16:20] should be the period of those factors in affecting those modifications. I mean never in our literature is able to, we are not able to understand for example, you are a smoker for one year or 20 years, you are a

[16:40] You have hypertension, diabetes, or whatever. What do you think could be the way to approach the timing of the clinical risk factor on that effect that you showed us so brilliantly? Thank you. Yeah, that's a really nice question.

[17:00] And it's hard to answer, but certainly it's going to be your lifetime exposure. And in fact, if you wanted to do, if you wanted to look at, say, the effect of obesity on the changes in the tumor microenvironment,

[17:20] a better way to look at the lifetime exposure in obesity would be looking at the germline genes that predispose to obesity and therefore you're going to more accurately pick up one's lifetime exposure through genetic risk rather than the point estimate that we get in

[17:40] clinic. But if you look back for the traceryx timing work that we've seen already today, you know, these time periods are long. Typically in sporadic cancers it might be 50 years. In a VHL setting, it'll be much

[18:00] less, but still we're talking five to 10 years at least, I would think. Okay, thanks Tom, that was a really, really great talk and I was wondering what your thoughts are about the practicalities of thinking about this kind of a molecularly guided prevention to interfere with the.

[18:20] with a microenvironmental thing. So I'll embed, I would be a candidate. But how do you think we should develop biomarkers and things like that to target that to the right population? And do you think that would build on some kind of a genomic analysis of the tumors or predictive markers? And just thinking what the kind of practical implications of this is.

[18:40] kind of a model. Yeah nice. I hadn't thought about stratification. I felt this is more likely to be a central pathway that's involved in all all CCRCCs so you know I haven't yet looked at the at the

[19:00] the categorization. But if one were to look at the large trial data which infers, say, your CD8-high tumors and more susceptible to checkpoint inhibition, if you add in high macrophage.

[19:20] they're less sensitive. So perhaps in the therapeutic rather than the preventative pathway it'll be those tumors with high numbers of macrophages that one could potentially choose. I kind of have a crazy question.

[19:40] So I was wondering if anything is known about the mutations found in macrophages, for instance. So, you know, going along the lines of what we're learning based on what we heard this morning, but also in other cancers and in other normal tissues where we have, you know, increasing knowledge about the mutation.

[20:00] mutations being present and especially in a condition such as this one where you have an inflammatory macrophage component which presumably in some cases is there chronically. So would you expect that they are, do we know anything about mutations? So in fact I have looked into this.

[20:20] And it's an interesting question in that these macrophage populations would have been there for a long time because the tumor's been there for a long time and therefore you'd expect selective pressures to be there in the same way that we see selective pressures in the primary tumor. In fact, what we see is the normal chip drive is the clonal hematopoietic.

[20:40] amatopoiesis drivers that are more present and there's an interplay between clonal amatopoiesis and cancer risk. Having said that, I'm not sure whether that would have been occurring early enough in the early timeframes of the year.

[21:00] cancer development and certainly if you have a sporadic RCC and you're in your 7th or 8th decade, the likelihood is that you do have some evidence of clonal rheumatopoiesis.

[21:20] perfectly in time. Thank you very much. And sorry that I forgot to introduce you properly. So our next speaker is Alyn Roe and she is from the Urological Research Institute here in Milan in

[21:40] institution actually and she's a PhD molecular biologist and she will tell us a little bit more about the data that we are accumulating on mapping the transcriptome and genetic mutations of

[22:00] different lesions in VHL patients in the kidney, of course, still focusing on the kidney.

[22:20] to present our study of the infra and intertumor heterogeneity in clear cell renal cell carcinomas in front of the endow disease. So indeed one major obstacle for the clinical management for clear cell renal cell carcinoma.

[22:40] or CCLCC in VHL disease is a heterogeneity because it occurs at different level. It occurs for the prognosis among patients and also within the same family. For the natural history of each single tumor that can be seen during the surveillance with imaging.

[23:00] and also for the histology after surgical removal. So even if you are all CC and CC, there are different grades and stages, but also for the response to systemic therapies. And because systemic therapies are entering the clinics for controlling the tumor growth, it's become more and more important.

[23:20] more and more important to understand the interpatient, intertumor, and intra-tumor heterogeneity. So for example, for Belzotifan, which is a HIF2 alpha inhibitor approved by FDA for VHLCC, LCC, which are,

[23:40] not requiring immediate surgery. There was a study showing that on the 61 VHL patients where the analyze was done on the measure of the average of the MRI estimated tumor volume.

[24:00] Half of the patient had objective response. So we wondered why 50 patients of this patient are not responding as well as the other ones. And in responding patients are ultimately responding the same way.

[24:20] So, our hypothesis is that because we know that in VHL disease, the CLLC are clonally dependent. We assume that there are different somatic mutations that can induce differences at the level of 2.

[24:40] tumor microenvironment, cell signaling, metabolism, and tumor histology, and that can influence the response to systemic agent. So our objective is to establish new tool to overtake this heterogeneity and for guiding clinical management in VHLCC.

[25:00] LCC patients. So we have done a study. We are doing ongoing study funded by the European Union and we are enrolling three VHL patients with at least three CCLC patients.

[25:20] We are taking four biopsies per tumor, and so we're analyzing 36 geographically predetermine areas. So it consists in first doing multi-paramedic presoldry MRI.

[25:40] And after the surgery, we take out the different tumors and we orientate the tumor in the way that we can correlate the data with the tumor and the imaging. And then we separate the tumor in.

[26:00] for quadrants, and in each quadrant we take biopsies. So one biopsy is fixed for histology and immunosucleochemistry. One biopsy is now frozen and we simultaneously isolate DNA and RNA for whole genome sequencing and RNA sequencing.

[26:20] And one biopsy is used fresh for single cell RNA sequencing. And our ultimate goal will be to do patient D-wave organelles on fresh biopsies. So I will show you the study that we have done on the first patient.

[26:40] with 29 years old female affected of VHL disease. So here is the axis MRI for LaoVu, and you see here the different tumors, the three tumors that are here with the VHL.

[27:00] with the imaging, and we also analyzed another exophatic tumor that was too small to analyze with the MRI. So first, the image was oriented here with the same criteria that have been used to do the sample.

[27:20] sampling and after we have divided in quadrants using the same color codes that have been used with the pathologist to take the different biopsies. And in this way it has been done with the images intelligently.

[27:40] analysis and also inter lesion analysis for each tumor. So he is the images for MRI for different features. So you have here the three tumors analyzed for

[28:00] C2, which allowed us to outline the different tumors. T1 and Zor, which is the time when accumulates the liquid of contrast, so it's corresponds to the vascular, displays the vascularization. And ADC for appropriate conditions.

[28:20] diffusion coefficient. So it's a calculation of the possibility of the molecules of water to fully move. So the more ADC is high and the lower is the cellularity and the higher is the fluid.

[28:40] So and you see here always a color code of the different quadrants. So first for the infralation analysis, we measure so this difference criteria. Here in this diagram you have the number of voxels for each.

[29:00] intensity here on the X axis, and you see that there was heterogeneity for the T1 angiole for the vascularization for the tumor L1 and L1, but not for L3. And there was heterogeneity for the ADC, so the same.

[29:20] cellularity or density for L1, but not for L1 and L3, so the other two. For the interlision analysis, we saw that there was a lower T1 and L2 for an higher ADC for L1, in comparison to L1.

[29:40] AR1 and AR3, meaning that there was lower vascularization and lower severity for this lesion. So the imaging characteristics of these religions is this patient with VHF syndrome suggests the presence of both

[30:00] intra and interlation heterogeneity. So when we did the histological analysis of the different tumors, so these three which have been analyzed for MRI and the exophilic one, so in the ones analyzed for MRI and for neurosclerosis.

[30:20] were from 1 to 1.7 centimeters. The fatigue one was 0.5 centimeters. So we took four biopsies for the three first and we could only take one of the exophilic ones and the pathological analysis reveals that the

[30:40] were all PTU and A and when we check all the different areas, quadrants, they were all CC and CC and they were all of grade 2. So there was a homogeneity at the level of the histology, the grade and the stage in the different tumor

[31:00] regions and when we checked also not only with surgical pieces but also for the biopsies in order to validate the technique we saw that there was a good complement of good correspondence between the biopsy and surgical

[31:20] pieces that the percentage of tumor cells were really similar between the biopsy and the surgical pieces. You see here, here, and here. And also that we also found that there was a low number of tumor cells in the wall.

[31:40] We have L1 tomorrow that were corresponding to what was seen with the MRR analysis. So we also found that with the biopsies, all the analysis with the biopsies were that all were CCL, CC, and grade 2, so meaning that it was

[32:00] also a good correlation of the histology, the grade, and the stage between the biopsies and the surgical pieces. So then we tried to do single cell RNA sequencing analysis on a peritumal piece and also 12 biopsies and that was carried out.

[32:20] With these 12 biopsies, we analyzed only five with RNA sequencing that we could go on with RNA sequencing. We obtained 9,452 cells after the filter qualities.

[32:40] We obtained that we had the main of the cells which were coming from the piece, so the peritomol piece and one from the biopsies. Even if we got one cluster with a CAA cluster.

[33:00] A9, expressing cells, and these clusters was corresponding to the cells with a loss of 3P. So in the integration analysis, most of the cells came from the peritumor piece, and so they were technical issues to establish single-cell segues from biopsies.

[33:20] So we are planning now to do special transhiptomics or to do single nucleos coming from the FFP. So when we looked at the whole gene of sequencing, so it's what we saw.

[33:40] So this morning we couldn't see any mutations in the canonical CCL drivers in any of the regions analyzed in the different tumors. And we only see ket1 mutation in mutation ket1 and CDC.

[34:00] And when we checked for the CNV, we saw that this one are representative of all the different quadrants for each tumor, which is the same. So there was three pillows in all the different biopsies.

[34:20] different tumors and except for the 5 cube gain in one of the tumors, we didn't see any additional gain or loss of chromosome and when we checked for the point of break of the of the chromosome 3, we

[34:40] We saw that they were all different in the different tumors, meaning that all these tumors are clonally independent. And when we looked at the different mutation signature, we saw the different cosmic signature corresponding to the SSB quaranta

[35:00] 40B and 48, so the C, LCC-specific, and also the SCC-S5 and 1 corresponding to the clock line mutation mechanism, and also SBS4, which corresponds to tobacco smoking even if this patient is not in the hospital.

[35:20] is not smoking. And when we check how this distribution of this different signature, we saw that this signature activity is conserved across tumors, we didn't see any differences. So in conclusion,

[35:40] or protocol

[36:00] There is no intra-tumor heterogeneity at the level of histology and at the level of genomics there are no mutation in the canonical CCL drivers and all the tumors have 3p loss with different breakpoints, meaning that they are clonally independent and there is also conservation of the mutation now seen.

[36:20] signatures equals tumors. So I would like to thank Professor Andrea Salonia who gave me the opportunity to work on the VHA disease. Humberto Capitano who is a PI of the study and Dr. Alessandro Laker with the head of the University.

[36:40] of the VHA program, Francesca Cora with the student, PhD student for kids with me and all our collaborators working on this study and also the patients. And you for your attention. Thank you. APPLAUSE

[37:00] Are there any questions? I may start with one even though I am involved in this project, so I should know, but I don't. So do we actually look also at in-del mutations?

[37:20] When we do sequencing? Yes. We have checked on the ulcerative mutation and the distribution was really similar in all the different biopsies. So what do you think are the driving mutations?

[37:40] besides VHL loss, obviously? I really don't know. So there can be different explanations for this. One can be that, for example, with the re-win model, you don't have the 3P loss and in this case we have the 3P loss. So maybe this can be the

[38:00] driver, maybe the haploinsufficiency of the drivers may be sufficient to be the beginning of the tumor. Otherwise, it's not that there are other drivers that we are now looking at and that we are missing because we are not looking at the good content.

[38:20] date. Thank you. Are there any other questions in the audience? If not, thank you very much.

[38:40] Can everyone hear me? No? Okay, let's try again. Great. So we continue with the third presentation of the session. It's a real pleasure to introduce Samra Duralic to present to us. She is a clinician scientist, so she works with us.

[39:00] clinically at the Royal Marsden Hospital and scientifically at the Francis Crick Institute. I think anyone who studied clear-cell renal cell carcinoma will certainly know of her work and her name. She's famous for having driven a lot of the Tracer X studies and studying T cells in human tumor. So it's a real pleasure to have you here somewhere and we look forward to hearing about clonal

[39:20] evolution in VHL disease.

[39:40] context, but arguably, clear cell renal cell carcinoma is the best model to understand clonal evolution more broadly. And that's really because of this very first event, which Tom Mitchell and I collaborated on a number of years ago now to characterize as something that happens many decades.

[40:00] before malignant transformation. So this is loss of the chromosome 3P, and not ubiquitously, but it frequently happens through a process of chromatripsis, which is a very clustered genomic rearrangement. And as Tom described, this happens, this results in a translocation between 3.

[40:20] and 5Q. It's an unbalanced translocation, so there's a 5Q gain and loss of 3P. So I'm mentioning this now because further down the line, I will use the presence of concurrent 3P and 5Q as an indication of chromatrypsis. But what's really important, of course, is the genomic content of the short arm of chromosome.

[40:40] 3, which contains these four tumor suppressogens that we keep hearing about, VHL, PBRM1, CetL2, and BAP1. And of course, this narrows the path of clonal evolution now because the loss of the remaining allele is the easiest way for these tumors to now proceed. And of course, as you all know- oh, I'm going in the wrong direction- as you all

[41:00] Although the next event that happens is usually in activation of the remaining allele of EHL, and in an inspiratic setting, almost 90% of patients will have had a mutation or methylation of EHL. That will give us the state of pseudohypoxia. And although the tumors have now achieved this very important cancer hallmark, in order

[41:20] to progress further, they usually will inactivate one of the remaining tumor suppressor genes. So certainly in very mature and certainly in metastatic sporadic kidney cancers, we see some combination of these four genes. So we've got these very stereotyped, very narrow genetics, and yet all of us who sit in

[41:40] in the clinic and see these patients know that there's huge clinical diversity in terms of the behavior. So how do we reconcile these two things? And that's really, that was the question that we set to answer many years ago in the context of the tracer-extrenal program. And the question is in this, the answer is really in this room.

[42:00] remarkable dependency between these genes, and highly conserved patterns of their ordering mutual exclusivity or co-occurrence, which gives us these three evolutionary modes, one which is very linear, where we only see VHL in three PLOS, another which is more Darwinian. What I mean by this is the evolution

[42:20] is very slow, but over time it picks up, there is lots of subclonal diversification. This is driven through PBRM1 and then subsequently there is a number of downstream trajectories. And then lastly we have these born to be bad tumors. So this is supercharged evolution. The tumors acquire many different driver events early on.

[42:40] levels of aneuploidy through generation of chromosomal instability. And these are the patients that we really worry about in clinic day present with showers of metastases. So now that we've established this, the sort of more philosophical question if you like, if you have only a narrow set of the

[43:00] these evolutionary solutions for tumors that are starting out, to what point is this predetermined in every patient? And this is really a broader question in the context, sorry, I seem to be going in the wrong direction all the time, it's my age, of an evolutionary

[43:20] biology and we are quite inspired by this Goldian view in evolutionary biology which is the concept of replaying the tape of life and of course this experiment can't be done but if you did replay the history of the evolution of species and you let it play out again if it was entirely determination

[43:40] with these evolutionary replicates, you would always get the same outcome. So you would always get homo sapiens, for example. But if there are contingencies along the way that prevent the same outcome from occurring, you would be arriving at a set of different solutions, but they would be different each time. So inspired by this,

[44:00] We turned to VHL related cCRCC as a concept of repeated evolution in a lifetime And above here, I'm just repeating this of canonical trajectory. We see in the sporadic context and of course what happens in The context of VHL disease is that every proximal tubule cell is at risk because once they start

[44:20] picking up that 3-P loss, which we know begins to occur around adolescents. Every single one of these cells might receive a second hit, and if it escapes, drift, if it manages to fix, then you're going to get clonal evolution. And this very simply put is the reason why patients have got such a high burden of both

[44:40] tumors, but also cysts and mixed tumors. And so from the point of view of understanding these repeat clonal evolution experiments, this is really an intrinsically controlled test of cancer evolution because we fixed the host genetics and we fixed the tumor microenvironment to a degree if we assume

[45:00] all these tumors are sharing a niche rather than having individual niches. There is a related question that can also be answered in the context of VHL disease, which is how does the cell of origin in the tissue context constrain clonal evolution if you fix that very first event. So in this context,

[45:20] We can look at pheochronocytomas, pancreatic neuroendocrine tumors, and manioblastomas, and say if you have the same host and if you fix the founder event, which is the VHL gene, do you then find that in the same patient you are seeing the same trajectories? And we've been extremely fortunate and really

[45:40] deeply honored to work with Marston Linehead and NIH and his team on this cohort of VHL families. So over 130 patients, almost 1,400 samples from these patients, accumulated over 60 years of surveillance. And this work has been beautifully led by clinician

[46:00] scientist in our group, Scott Sheppard, who is in the audience. So Scott will be able to answer many of your questions later on. So this is just a snapshot of the clinical pathway of these patients. And of course, this is the problem that we're all trying to address. Each one of these lines in the swimmers' plot is a patient, and each one of these

[46:20] little symbols is an intervention. And just to give you an example of one of these patients recruited at NIH, you know, their intervention started at the age of 27 and at the last follow-up at 60 they were having a radical nephrectomy. And of course this audience knows entirely how the intervention is.

[46:40] interventions are determined, the monitoring as far as CCICC is concerned is radiological surveillance and when any one of the tumors reaches the magic three centimeter surgical intervention is performed. But as you know, the surgeons will attempt to develop

[47:00] as many tumors as possible at this time. So this is how we end up, or how Dr. Linehan's theme ends up procuring many different tumors and then relating them back to this radiological image. Now I included this slide actually in response to the debate that we've been having throughout the day, which is are these VHL-related CCRCCs similar?

[47:20] discoloration, discoloration, and the answer in this cohort is yes. So here's 3P loss in the top line, PBRM1, CD2, BAP1. And you can see that these early stage CCRCCs in these patients are evolving through selection of those canonical tumor types.

[47:40] suppressor genes. Now, you can also see that many of them do not have a secondary driver, but this is in keeping with the small size of these tumors. We know from our previous work in the sporadic setting that the number of driver mutations is a function of tumor size, and most of these tumors are 1 or 2 centimeters in size. So I think that explains the relative

[48:00] lack of secondary drivers. So I'm going to talk to you just briefly through the five very simple questions that we've asked within this cohort. So the first first was knowing that those evolutionary trajectories are conserved and that post loss of VHL, tumors can go either via data.

[48:20] that's PBRM1 or BAP1 trajectory, do we see a convergence in patients? So what that would look like is that in a particular patient, they would be kind of addicted to one pathway as opposed to another. And the answer is no. In fact, there is no systematic repeat selection of those secondary driver-sematic mutations, and there is

[48:40] no evidence of intrapatient convergence. There was one exception, however, and this surprised us on two level. One was that the exception occurred in patients whose tumors only ever followed this pathway. So across dozens of tumors, they always followed this pathway. The second surprise was that this pathway was selection of a somatic mutation.

[49:00] in VHL, which obviously isn't a canonical pathway. We're expecting to see 3P as the second hit rather than another VHL mutation on top of the germline mutation. However, we were able to rationalize why this was happening because we saw this pathway enriched in patients with germline co-deletion.

[49:20] So as you know, a proportion of patients with VHL disease have whole deletion of VHL, and some of them will have a deletion of the neighboring gene, which in this case is BRIK-1. And there is preclinical data that shows that BRIK-1 completely, BRIK-1 deficient cells can't form tumors, suggesting that if you deleted both copies of BRIK-1, you could not proceed with it.

[49:40] tumor formation. And essentially that's what seems to be happening here. So this is the canonical trajectory. You lose 3P in this background of VHL and Bric I code deletion. If you lose 3P, you lose two copies of Bric I, and this is a nonviable cell, essentially. So the only way for tumorigenesis to proceed here is to inactivate.

[50:00] the second copy of EHL through somatic mutation or methylation. And when we went back to the entire cohort, we did see some of these gaps in other places. So this is in patients who don't have a bric codeylation, who also don't seem to have 3P loss. And when we look at these patients, some of them have no second copy of EHL.

[50:20] hit at all. Some of them also have second hit through VHL mutation. We don't know what the exact constraint is here that's negatively selecting against 3P loss. But we also see these as evolutionary dead ends, essentially because they haven't lost 3P. They're not inactivating any of

[50:40] those other tumor suppressor genes, which are so important for driving cancer evolution forward. They have low levels of chromosomal instability. They have small tumor diameter. So we really see this profound relationship between the clonal evolutionary pattern and the phenotype. And so speaking of phenotypes, this also I decided to include reflecting on what we've been discussing this morning.

[51:00] lovely data from Francesca on cystic lesions. And of course, we're all wondering whether these phenotypes are continuous. And those of you who know this disease much better than me tell me that some of these cysts always remain cysts and others become mixed and become solid over time. So we were able,

[51:20] annotation to separate all these different lesions out, and we can see that renal cysts have a much less evolved genetic profile. As you can see here, we are not really seeing mutations of those secondary drivers in any of the renal cysts. But when we look at the copy number profile, so these are losses in blue, gains in red across the genome.

[51:40] I noted at the beginning that when we see 3P and 5Q, that's the sort of canonical pathway that we also see in the context of sporadic disease. In cysts, we see that the frequency of 3P, 5Q is much lower. And in fact, there are other copy number events that are mutually

[52:00] is exclusive with that route, which happened on chromosomes 17 and 19. So, in cysts, there seems to be a divergence of the ones that follow the canonical trajectory that eventually become mixed lesions, that eventually become solid tumors, and those that don't follow the canonical trajectory where 19

[52:20] and 17 loss are in fact enriched. And perhaps these are likely the cysts that remain as cysts over time. Now the other phenotype that's very important when it comes to management of EHL patients is the risk of metastasis. And of course we've looked at this repeatedly today, which is the

[52:40] merges of metastatic competence really is something that comes along with increasing diameter, hence the three-centimeter rule. And we had a hypothesis because of our work in the context of sporadic CCRCC where we know that metastasis-competent clones from the primary tumor site have got a very specific genetic profile. They have high

[53:00] levels of aneuploidy and they specifically have fixed events such as chromosome 9P loss and chromosome 14Q loss. And it was very pleasing in the work led by Giovanni Ginese, who's also in the audience, that we see the same convergence under 9P loss.

[53:20] in genetic models. So the question is, is aneuploidy the bottleneck for emergence of metastases? And is the reason why we don't see them below 3 centimeters? Because that's the population size at which you begin to fix aneuploidy. So first of all, the level of aneuploidy is related to increasing tumor diameter.

[53:40] But when we look, this is in the VHL cohort now, specifically for loss of 9p, we do see it even in very, very small tumors that are just a few millimeters in size. And this tells us that the genetic event alone is not sufficient to do this. So why are these tumors not metastasizing when they've become

[54:00] metastatically competent relative to this genetic profile. Well, more recently, we've been looking at the phenotypic features of these metastasizing clones, and they do have features of heightened proliferation, also immune evasion through suppression of C-Gast-Think pathway, and in the large sporadic tumors, we find them in the center of the tumor.

[54:20] in large areas of necrosis, and this isn't something that's typical of that early stage RCC tumor that we see in VHL disease. And perhaps to illustrate this with a patient case, as was noted by Francesca and others this morning, most of the lesions that we see are

[54:40] completely clonally independent. However, with some patients where Dr. Linehan's theme have acquired lesions across time and across space, we've found that they are clonally related. And essentially they present these persistent clones. So these are recurring clones. So in this particular patient you can see this is three centimeters, partial nephrectomy, and

[55:00] There was a tumor nearby, later on less than 3 centimeters resected. Then a contralateral tumor resected and metastases in spite of this tumor being much smaller than 3 centimeters. And when we look at the molecular profile, actually what this really represents is a selection of this 9-pelos clonus.

[55:20] the primary tumor site here, which then recurs, it's by now caused this clonal sweep, it's entirely competent in terms of metastases, and it seeds the contralateral kidney and then it seeds metastases. So when we go back at the entire cohort and we look at the presence of these persistent clones,

[55:40] We see that they drive faster tumor growth and they correlate with metastasis risk. So we heard from Alessandro this morning about the patients that do present with metastasis sub 3 centimeters and this might be a representative of these recurring clones.

[56:00] So I think I've actually run out of time, haven't I? So let me just go to the end and thank everyone who's been involved in this. So of course Dr. Linehan and his team, everyone in our lab and as I said Scott Shephard is the

[56:20] person who's lead on this and the rest of our team, but most of all patients and families who facilitate this work. Thank you.

[56:40] questions I think. Yeah. Thanks a lot. I mean I think this this this data may be deserved another Congress which is devoted only to that so this is really a dream. I'm thrilled by the the concept of German constraint and impact on

[57:00] on phenotype, which is clearly a key concept in those patients. But on the other hand, as a clinician, how do we reconcile this idea with the heterogeneity that we observe within the same family of patients, which, by definition, is not a good idea.

[57:20] definition shares the same general annotation. What is missing between those two ideas? Sure. That's a great question and it's a very broad question, so I'll try and pick off a couple of points. So first is we obviously don't see within-patient conversion.

[57:40] So that's telling us that the way that the tumors evolve, if we take, if we accept the clinical phenotype is that they get CCRCC, the way that they evolve at least in these early stages is fairly stochastic. So whatever mutation they come across, if it's a fitness-affording mutation, that's what they select. The data I just showed you is that the tumor is not a tumor. It's a tumor.

[58:00] didn't manage to show at the end is looking at the tumor microenvironment and actually we see that the tumor microenvironment to our surprise is patient-determined as opposed to individual tumor-determined and that tells us two things one is that there are patient-specific influences on the tumor microenvironment

[58:20] that are probably germline determined. And second, that there is no such thing as a sort of permissive tumor microenvironment for a certain type of trajectory versus another. So these tumor microenvironments can support either PPRM1 or VAP1 trajectory. And at this early stage,

[58:40] these tumors are not shaping their microenvironment, which isn't the case much later on. They do. Now when it comes to the germline constraint and the differential across families, I think the fact that there is no intrapation convergence probably tells you

[59:00] about just how much stochasticity there is in the whole system. You know, once we didn't see intrapation convergence, we weren't expecting to see family convergences either. And then lastly, I would invite you to have a look at a paper by a colleague of ours, Greg Findlay at the CRIC, whose performed saturation genome editing of the VHL genes.

[59:20] and has provided us with functional scores to try and understand how the details of the exact VHL mutation might influence clonal evolution. And we see that the likelihood of cooperating with a particular mutation and the likelihood of tumor growth is also encoded in the type

[59:40] of the NHL mutation. So the subtyping is much more nuanced than what we've appreciated until now, which has mostly been like, oh, it's tissue-specific. You get these types of tumors or those types of tumors. So there's a lot more to learn, I think, from just studying the mutations themselves. Thanks. It's time for one more question, Zachary. Okay, thank you, Sumitra. That was really, really impressive.

[01:00:00] I have a question about the original three trajectories that you described based on already non-syntromic cases. So do you see those three classes come up in these early regions and in particular the born to be bad thing, so do you actually see very small teams that have the

[01:00:20] on to be a bad genotype. Yeah sure so we certainly see BAP1 driven tumors. In fact there was one patient who had four PBRM1 and four BAP1 driven tumors. That's when we became convinced that there wasn't going to be an exclusive selection. What we haven't seen so far are the multiple clonal drivers and that makes us think that

[01:00:40] perhaps the order of events for that particular trajectory in the sporadic setting bypasses VHL initially so that you get 3P loss, you then get PBR1 and BAP1 in the same cell, which has a fitness disadvantage, but once they pick up VHL as the final event that really completes the clonal's

[01:01:00] sweep and that's why they've got everything they need. Now because here the constraint is putting VHL first, it might be the reason why we're not seeing them. Right, so the born to be bad don't start off as the indolent VHL alone, but they start from different... So born to be bad is a broader category, BH1 is also born to be...

[01:01:20] bad sorry to be, it sounds like I'm going to spring into a song, but, you know, BAP1 also has that, you know, BAP1 mutant clones will outcompete everything else through fixation of aneuploidy, and VHL wild-type tumors also actually have that early fixation of chromosomal instability.

[01:01:40] But when you get multiple tumor drivers that are normally mutually exclusive, such as PBRM1 and BAP1 on the trunk, that particular trajectory we have not seen in VHL disease so far. And we think that the rationale is the order of events. Thank you. You bet. Yeah.

[01:02:00] So we move on to the final presentation, Francesca Colmore, who is someone who I didn't have to look up her CV because I see her about 300 days of the year in the laboratory. She's a PhD student in my group. Francesca was actually born to be bad I think, but I forced her to work on a Darwinian evolution tumor progression where she's been studying

[01:02:20] role of VHL, polybromo, and KTM5C in trying to understand the functional consequences. I think actually it links a lot of the things we've been talking about in terms of aneuploidy and DNA mutations. So, Francesca, it's all yours. Thank you, Jan, for the really kind introduction. I would like to thank also the VHL Alliance organizers to

[01:02:40] give me the chance to talk about my results in my homeland and to be surrounded by this many experts in the field. Today I'm going to talk about three tumor suppressors that you might know, CCRCC, VHL, KD5C and PBR1, but what you might not know is

[01:03:00] is that their combined loss can cause CCRCC in mice. As you might know, clear-cell, renal-cell carcinoma is the most common kidney cancer. It is the sixth most common cancer in female and the 10th most common cancer in female. It occurs twice as frequently in men.

[01:03:20] than in women. It can have a sporadic origin, but it can also be a consequence of the familiar VHL disease. In either case, the triggering event of the tumor process is the bialylic inactivation of VHL. CCC exhibits are really wide.

[01:03:40] spactogal mutations also in genes involved in the epigenetic regulation. Among these, KDN5C is the one with the most striking mutation rate difference between men and women. As you can see, it's my higher, much

[01:04:00] higher mutatin in men than in women. And this might be explained by the location of KDN5C in the genome and its tumor suppressor role in CCRCC. KDN5C is indeed locating on chromosome X, but he escaped the X-in activation, meaning that female cells have two copies.

[01:04:20] two active copies of OKD5C, while male cells only have one copy of OKD5C. But it's intriguing to notice that on the Y chromosome, male cells also present OKD5D, which is our homologue of OKD5C. The two proteins shared, indeed, OKD5C.

[01:04:40] percent of amino acid identity beside the biochemical function of demetylazin, the lysin 4 of istone 3. As many genes located on chromosome X, kinin 5C is also really important for self-functioning, leading to the idea that male cells might compensate the only copy of kinin 5C.

[01:05:00] of Canadian 5C with the pressure of Canadian 5D that works as a sort of second surrogate allele of Canadian 5C. In the context of CCRCC, then, male cells are more, are less protected from the loss of function of Canadian 5C and they are left with Canadian 5D.

[01:05:20] compensating for the loss of K5C. In this context, it's intriguing to mention that 40% of male CCRCC demonstrated the loss of Y chromosome, leading to the loss of K5D2. When we look at frequency of K5C mutation, we notice that it's often mutated with polybromo

[01:05:40] and VHL. And the single sampling of the single biopsies of primary CCRCC didn't identify many clones carrying KN5C mutation, but sampling multiple regions of primary

[01:06:00] tumor, identify intratumoral clones with kenufib C mutated alone with VHL loss or mutated in combination with VHL loss and PBL1. And this multi-region sampling was also performing metastasis and you could also see that these mutational

[01:06:20] status of cadmium 5C is maintained also in metastasis. We then aim to mimic the mutational status of cadmium 5C in cohorts of female and male transgenic mice carrying Kri-Lokspi system. It resulted with the loss of cadmium 5C.

[01:06:40] 5C alone or in combination with VHL loss and PBR1. And this loss of course is tamoxifen-inducible and tubal-specific. With this system we aim to understand whether the double loss of cation

[01:07:00] to reverse the sex bias frequency we see in CCRCC, in human CCRCC. After tamoxifen-induced induction, via ultrasound, we follow the cystic progression in these animals. And it's really intriguing to notice how the

[01:07:20] triple mutant female animals were able to develop an iron number of cysts with an earlier onset not only compared to the other genotypes within the same female cohorts but also compared to the male. We then decided to investigate and characterize histologically potential tumors in this area.

[01:07:40] animals and as you can see here and this is even more surprising is that only triple mutant animals were able to develop CCRCC with females showing an iron number of tumors per mice. We then characterize histologic

[01:08:00] Similarly, the tumors we could see in these triple mutants and we could identify two tumor progression. The non-cystic tumor progression where you have the tubules transforming into a neoplasm and then to a bigger, a small and bigger tumor. And then a cystic progression where the typical cyst

[01:08:20] will then transform to an atypical cyst. And I would like to point your attention on this higher magnification of this event I have just described. So here we see the healthy tissue of a kidney, and here we have the cyst border. Here the cyst border.

[01:08:40] cysts start as a typical cyst lined by a monolayered epithelium, then transform in a typical cyst lined by a multilayered epithelium, also more disorganized. And then this epithelium transforms into an intrasystic cyst.

[01:09:00] where the tumor cells are not yet invasive, they are still contained within the cyst. And then somehow they acquire the invasiveness and they are able to invade the cystic capsule and the surrounding kidney tissue. And I would like to point out that this is actually the first time we are seeing this

[01:09:20] phenotype in our CCRCC mouse model. We then wanted to confirm that the tumor we're seeing were CCRCC as the QI expression was tubule specific, so we stained for proximal tubules, marker, but also for hypoxone and proliferation. Once we observed that

[01:09:40] only the triple mutation of these three genes were the only causing the CCRCC in mice. We wonder how these three genes mutation coparatly because CCRCC and we aim to study the functional consequences of the loss of these three genes together. We then took advantage

[01:10:00] of renal epithelial cells extracted by the same mice we used for the in vivo studies. In this case, the Kree combination was induced by adenovirus infection. We then characterized the DNA replication in these cells because from the literature

[01:10:20] we knew that the individual loss of the trip of these genes already caused an increase in replication stress. So we wonder whether the trip mutation could exacerbate this phenotype in our cells. We quantify the replication stress by measuring the speed of the forking

[01:10:40] of the replication fork, seeking for replication fork stalling events. So slowing down of the fork. As you can see, upon single or double mutation of VHL and K5C, as well as upon the double mutation of VHL and PBN1, cells already experience an increase.

[01:11:00] in replication stress. What is interesting to notice is that upon the triple mutation, females are able to rescue the phenotype, but not the male cells. We then decide to continue this investigation by observing the DNA damage as we

[01:11:20] that unsolved replication cells can cause an accumulation of DNA damage. So we stained these cells for gamma H2X and we quantify the overall number of fosyper cells and here again you can see that upon VHL loss alone or cadine 5C individual loss but also in their combination we have an increase

[01:11:40] accumulation of DNA damage, but as well as both for female and male mice. But these phenotypes seem to be maintained for both the sex. When we look at the percentage of cells with an increasing number of foci from zero to more than 10, it is interesting to

[01:12:00] notice that female cells experience an increased DNA accumulation across the genotype, while in male, the DNA damage accumulation becomes evident only upon triple mutation. And we reason that if there is an incongruent effect,

[01:12:20] increase or the replication stress is not solved as well as the DNA damage is not solved, then the cells will experience genomic instability that can manifest with lagging chromosome, unattached chromosome, chromatin bridges, and macronucleus. So we decided to quantify these events in our cells.

[01:12:40] Here you can see that for the anaphases, across the genotypes, upon a single individual or double or triple mutation, cells both female and male experience an increase in aberrant anaphases, as well as micronucleation. Here actually, you

[01:13:00] can notice that the knockout male cells are the only one who's not showing any increase in micro nucleation, hinting at possible potential compensation mechanism played by the steel stand in kinase 5D. Beside the role that these three genes have in replicatory drugs,

[01:13:20] medication stress and genomic stability. We also considered the possibility that this, the role of these genes in replication in gene expression regulation. And we wonder whether the triple mutation can cause or can induce gene signature that are specific in this context.

[01:13:40] So we perform an in-sick analysis and I must say that we got the analysis back only two weeks ago. So the analysis is not finalized, but I would like to show what we have done so far. So we perform the Differentially Expressed Gene Analysis that allow us to identify uniquely their existence.

[01:14:00] genes within each genotype, but also to compare overlapping deregulated genes into different genotypes within the same sex so we can perform intra-sex studies, but also intersex studies. Here I for example show, I can show the

[01:14:20] overlapping upregulated or unregulated genes in triple mutant female and male cells. And as you can see, the overlapping is not really high. But if we look at the genes that are uniquely, for example, upregulated in the female cells and we perform gene-set enrichment analysis, we can see a

[01:14:40] already that there is an enrichment in gene set concerning the extracellular matrix deposition. But for now I would like to stop here to remind what this project brought to life so far. We could establish a new CCRCC mouse model in which we induced VHL loss along with P-1 and ketonein-freezones.

[01:15:00] And thanks to our in vitro system, we were able to study the functional consequences of the individual or combined loss of these three genes. We can notice that the triple mutant cells as well as the other mutant cells experienced an increased replication stress, DNA damage, and in general,

[01:15:20] and increased genomic instability, and probably this phenotype, along with the gene signature that are specifically induced by this stripal mutation, that they cause CC-SCC only in these animals. But for now, I would like to thank Ian for giving me the chance to work on this project. I would like to thank you.

[01:15:40] my colleagues that are super supportive and of course you guys for listening to my talk today and I want also to remind that if you are interested in the overlapping and an overlapping role of KTM-5 C and D you should definitely stay tuned for Ian's talk tomorrow. If you are more interesting in the molecular mechanism underlying CCI-CC

[01:16:00] you should stop by the posters of these beautiful people. Thank you so much and I'm glad if you have any questions. Thank you very much. I was actually told we saved one minute on the presentation so if there's one pressing talk.

[01:16:20] One pressing question we should ask it. Thank you. Finally, the third time. That's just one simple question. How did your model, in terms of penetrance as well as the latency of this model compared with the VHL PFC-RB model you established previously? So you asked, you asked in the penetrance of these

[01:16:40] animals. Yeah, penitrins also latency. How long is it going to take for them to turn? You could see it already I had it in the the sites. We noticed tumors after two years. The mice were two years old and the penitrans was 40% for both sexes. Okay, do you guys have a syngenic model for those?

[01:17:00] triple model, ketene 5C, VHL, and PD-R1. Sorry, again, sorry? Do you guys have a syngenic cell line generated from the tumors? I must say that we couldn't dissect so many tumors from this mice because the tumors were quite small, and we actually couldn't really detect them via ultrasound.

[01:17:20] We had to cut the kidneys open and then look at the, via histology, we could notice the tumors. And since they were quite small, we couldn't actually manage to dissect the tumors and get some cell lines from them. Okay, thanks. Okay, thank you everyone. Thanks to all the speakers for a wonderful session. And I think now we're

[01:17:40] We will go to