Dr. Khashayar Vakili - Pediatric Hepatocellular Carcinoma: From Bedside to Bench

For the opportunity to speak uh here this morning. So I'm gonna talk about uh uh pediatric hepatocellular carcinoma. I have no financial disclosures. I'm gonna talk, uh, a little bit about surgical outcomes for HCC and then I'm gonna talk about some of the signaling pathways that we've been looking at and interested in, in the, uh, in the lab. And then I'm gonna uh talk about some of the strategies that we've been um trying to uh utilize in order to study uh some of these tumor cells uh in vitro as well as, uh, in vivo. Uh, hepatocellular carcinoma is an epithelial type of a liver malignancy. Uh, the, the most common, uh, uh, liver malignancy of epithelial origin is actually hepatolastoma, and, uh, you see that in kids under 5 years of age. Uh, they're, it's very responsive to chemotherapy and, uh, 5 year survival could be 85% or over. Uh, this is in, in stark contrast to HCC which is really not responsive to, uh, uh, chemotherapy in general, and It's really, uh, broken down into several subcategories. Um, this was a consensus, uh, conference by uh COG a few years back that sort of uh categorized HCCs in these, uh, categories. One is HCC arising in the background of cirrhosis, uh, which is actually the predominant form in the adults. Uh, I'll talk about that a little bit later. And then the other major one is the fiberla Mueller HCC that is generally seen in adolescents and young adults. And the HECNOS are not otherwise specified, it's just really a um Large sort of uh wide uh range of HCC's for, uh, before there was called the transitional uh type HCC and there was some hepatoblastoma component to it. But this has, this has really not been teased out, uh, significantly, just partly because of due to the small numbers of this type of tumor. So, um, generally, the mainstay of treatment for HCC, uh, given, uh, the fact that it's not chemo responsive, it's really surgical resection, uh, and, uh, And that probably gives the only chance of potential cure. However, when you look at um uh some of the recent studies that have been published, uh, which report their results with a resection for HCC, uh, and, uh, or transplant, you can see that the um range of survival for resection ranges anywhere from, uh, uh. In the teen range all the way up to uh 80, 80% or so. Um, but it's a very wide, wide range. Um. At most, uh, the, uh, you know, there have been some reports of liver transplantation that have had good long-term survival, uh, as opposed to resection, and I'll talk a little bit about the differences between, um, resection and transplant and what, um, benefits transplant may have over time. But overall, you can see that, uh, uh, the outcomes for liver resection for HCC is just generally extremely, extremely poor. Uh, and that's been reported across the board. Uh, here at our center, uh, Gabriel, um, who is, uh, Hong Bae Kim's fellow, actually has been working on looking at the, uh, outcomes for liver resection, um, for HCC, and here you can see this is the overall survival. And over a 5-year period, um, over, uh, over 5 years, uh, overall survival is around, um, 50 to 60%. However, you can see that the disease-free survival, um, drops down to zero after, um, just over a year and a half, meaning that basically all these patients have tumor recurrence and now they're just living with their tumor. And so this just illustrates what um you don't see in the other papers that, that um report patient survival, and that only report patient survival and not disease-free survival sometimes. So you might see patient survival at 5 years, it might be 40%, but they don't report the disease-free survival. So a large proportion of those patients might be actually living with, with tumor. Um, so that really, um, Overestimates the survival for HCC following recession. So there have been a number of um uh studies that I showed before. One of, one interesting study was uh this study which was published in 2015. Um, it really, it was asking the question of whether hepatic resections um can be justified. And this was I really looked at the adult population. It was a retrospective studies. It looked at 5 US centers and they compared 884 liver resections um versus 881 liver transplants and It was, um, it compared the cirrhotic resections to liver transplants, then you can see that there's a, a significant difference, um, obviously, um, for patients who undergo liver transplants versus cirrhotic resections and Part of this is obviously because the, the patients who undergo cirrhotic resections already have cirrhosis and they have liver disease, so the outcome in general is just not gonna be as good. However, when you compare non-cirrhotic resections to liver transplants, um, you see that still there's a significant difference between uh survival following liver transplantation versus, um, uh resection and, and non-cirrhotics. In, uh, children, uh, there was, uh, a, um, this is a probably one of the better studies that, that I found, uh, looking at transplantation for HCC. This is from a, um, from Europe. Um, that is actually did break down the, uh, underlying diseases and disorders, um, when they were transplanted for HCC. A lot of the other, um, studies really don't do this. So you don't really have the subcategorization of of the types of HCC or underlying, um, disorders in order to really draw some, uh, conclusions. But here They had, um, uh, they basically broke it, you know, they compared children and adults, and, uh, in, in children, they broke, you know, they compared the inherited and non-inherited, uh, diseases, underlying diseases. And inherited diseases, the majority of HCC's were in thyrosinemia, which is sort of a well-known, um, uh, predisposing factor to developing HCC's. In the non-inherited forms, uh, adults obviously have a, uh, higher, uh, prevalence of hepatitis B and C viruses, which, uh, which lead to HCC, uh, as opposed to the children which do not have. But they The majority of children here, they have unspecified and other, so it's a little unclear exactly what that is, but clearly you can see that there's a big difference between HCCs that arise in the setting of inherited liver disease versus not for. For these children. So, you know, given, given all that, really, the thought is, you know, I, I clearly, it's, uh, we've shown that at least others have shown that the resection is really has very poor um outcome in general. Um, we've seen that in our, in our center. Uh, so the question becomes, well, what's the benefit of transplant? The benefit of transplant is you're taking out the entire liver. Um, could it be that Uh, even though we don't see any, uh, lesions on imaging studies that there are small undetectable synchronous lesions that we just don't know about and they're gonna be there and they're gonna come back. So is it that that actually, um, is what, um, gives the advantage in doing a liver resection and you actually truly do get a curative, um, curative resection. So that's, that's really the uh one of the questions. And so what the question is, well, which tumors should we really considered for transplantation, even if they're surgically resectable. Uh, you know, obviously, as you know, um, Uh, some of the factors that have been reported to be associated with tumor recurrence, um, are vascular invasion, tumor grade, size, uh, multifocality, um, and especially if there's incomplete resection or positive margins, and this, this goes along obviously this, um, pertains to nearly, um, a lot of other cancers as well. It's really hard to tease this out, um. In, in the, in the pediatric population. Um, and I'll, I'll tell you, uh, why, uh, why that is. If we look at our transplantation here, um, our experience here, we've done 6 liver transplants, uh, for HCC. There were 2 of them were in the background of cirrhosis and, uh, And then 4 of them were without cirrhosis. One had fibrillamyar HCC, uh, 2 were HCCNOS and 1 multifocal well differentiated ACC. Um, one of them has recurrence, one has died of metastatic disease, and, uh, one of the patients who, uh, had ACC in the background of cirrhosis, um, had recurrence and died. So overall, there are, we have several um patients that are actually um You know, a couple of years out after having the uh transplant for HCC or at least maybe within a year or so, 1 year and a half. So we have to wait until we get more long term results. Um, these were really transplanted um within the past uh couple of years. Uh, we've done two multivisual transplants for fibrillomar HCC. One is doing well at about 5 years or so, and the other one, was actually multivisceral transplant, um, after he had recurred, uh, following resection, and he does have recurrent, uh, metastatic disease at this point in time. So, um, that being said, you know, we, uh, we've had some experiences with some patients that have done well, um, despite the fact that they have vascular, um, tumor invasion. Um, so, uh, this is an example of one of those patients who underwent a multivisceral transplant, uh, for HCC, and you can see the tumor growing down into the mesenteric, uh, venous system. Uh, he's several years out and he has done Uh, well, in, in general, he's had, uh, transplant issues with uh PTLD and immunosuppression issues, but overall he's done well and he, he hasn't recurred. We have uh transplanted a 17 year old with multifocal HCC uh that turned out to be well-differentiated HCC and she's several years out and she's doing well. And uh here's a child that had a uh left lobe tumor growing into the right atrium, and we ended up doing a resection on this child. Um, and the child recurred and died uh several months after. So I'm just showing you these cases just to illustrate that we sort of see a wide range of um patients, a wide range of pathologies, and some of them, um, despite the fact that they have vascular invasion and multifocal disease, have done well, and some of them, um, don't. And so it's very difficult to sort of tease out, um, who's gonna do well with transplant or resection or what, what factors are really involved here. Because the numbers are so small in the, in the pediatric world, as far as HCC goes, um, it really does become sort of an anecdotal type of an experience. So, in, in summary, we have poor long-term outcomes. Um, some of the data that I showed you does raise the question of possibly favoring transplant in certain cases, even if, if, if some tumors are surgically, um, uh, surgically resectable. And you obviously have to, everyone will balance this out against the long term survival with uh following transplants and the issues that arise from not only the, the surgery but also from immunosuppression uh long term. But even, even if you take that into consideration, um, you can see that that is just much better than what we have so far with these kids undergoing uh resection uh for HCC. And basically, you know, one of the uh important factors here is the fact that uh uh HCC is really not um chemo responsive and we don't really have good effective uh chemotherapy. So we really need to understand the, the biology of uh of these tumors and that's That's sort of been um something that people have been working on, especially in the, in the adult population. HCC is very um prevalent in the world. And so there's a lot of extensive research and data out there in the, in the, uh In the adult world, there's just not as much in the pediatric world. The numbers are just uh not, not as many. In adults, hepatitis B and C are, are, are predominant factors. Development of HCC, um, but you don't see that as, as frequently in the pediatric, uh, population. So, um, when we look at the, um, various, uh, potential, uh, mutations in the adult population, we see that, um, P53 and TERTS and the beta-catenin mutations are sort of the top genes uh in In the, in this kind of a study, we just don't have significant numbers in the pediatric population to do this kind of a broad, um, broad type of a study. Um, at this point in time. Uh, when we, so this really prompted, prompted me to start, um, developing a tumor bank in order to be able to study these tumors individually. So You know, the, my, my thought was, can we actually take each of these tumors, um, and study them, um, comprehensively, um, possibly in order to understand how they behave, why they behave the way they do. Cause as I showed you clinically, uh, despite the fact that, um, we saw certain clinical characteristics in the tumors, uh, whether it was vascular invasion or multifocality, it was still hard to sometimes predict their, um, clinical behavior. So I, I thought maybe, maybe we can try to understand each of, each of these tumors uh one by one if at all possible. So I began to develop a tumor bank, um, starting in 2012, 20013, and so far, um, been able to um collect 15 hepatolastomas, 7 fibrolamellars, and six other types of HCC so far uh from a uh variety of resections and, and transplants. And uh my thought has been to, to try to see if we can actually take these tumor tissues, perform, um, various molecular analysis on them, and, uh, and then also try to see if we can grow them in vitro and in vivo and to see if we can study them real time if, if, if at all possible. So I'm gonna talk about some of the uh efforts that, that have been uh undertaken in the lab so far uh in regards to um some of these uh these categories. First, I'm, I'm gonna talk a little bit about some of the uh molecular analysis that we've been doing uh in the lab. When, um, You look at the transcriptome of the various um uh types of tumors. This is, um, what, what we did here is that we actually took, uh, these are published. Data, uh, RNA sequencing data, and we did some bioinformatics analysis and this is sort of to be able to group the, the various types of, um, tumors and see how related or unrelated they are from each other. And here you can see the hepatolastoma is really um grouped together in this quadrant. Uh, the dark purple is a hepatocellular carcinoma. It's an adult hepatocellular carcinoma. And uh this light, this blue here, this is a fibrolamolar um uh ACC. Uh, so this is a way to, to sort of gauge as to in regards to the transcriptomic or gene expression programming, how closely related uh these tumors may or may not be. And we took one of our own, uh, samples, this, uh, ACC in this screen. This was in a 1.5 year old with an HCC of NOS and we just wanted to see where it fits um within uh within these tumors and it really, it, it, it's very close to um within the HCC component. So this really uh raises the question of now we, you see that the these tumors do tend to obviously from a transcriptomic program do bunch together uh although we see more variability in the ACC population and for example, the hepatolastoma population. And the fibrolamella are very tight, um, um, close, close to each other. They bunch together. But the question that we asked was, uh, despite different driver mutations, genetic events, or transcriptome, are there critical common pathways necessary for HCC development? Um, so as I showed you, there are a number of different mutations that can lead to HCC, but the question is, is there one critical pathway that might actually be able to Um, is actually necessary and critical, uh, for this. So one of the pathways that we looked at, um, and we, we, we're, we're interested in is the HIPAA pathway. The HIPAA pathway is a tumor suppressor pathway. It's, uh, it was initially, uh, discovered in Drosophila and it's involved in organ size regulation. And the HIPAA pathway, because it's a tumor suppressor pathway, when it's in the, on the on position, uh, it's, uh, it, it really, it's, it's suppressing the, uh, proliferation of the cell. Uh, when it's in the off position, that means it's suppressive action is taken off. Uh, YAP, which is a, uh, uh, co-transcriptional regulator, binds to teeth and results in proliferation. And uh uh epithelial to mesenchymal transition, um, as well as it, it gives sort of a stem nest to the cells. So when they look at these cells that have a lot of, yeah, uh, overexpression, they tend to have a lot of uh stem cell type markers. So, um, another study that was done was that they overexpressed, uh, yap in the liver, and after yap overexpression, the liver actually become actually hypertrophies and gets bigger. And after, if you let the yap, uh, overexpression go on, after a few weeks, actually there is development of HCC. And if you turn the yap off, it goes, it reverts back to, to its original size. Which, so based on this data, this is sort of became very, we became very interested and asked the question of is this pathway um inactivated in pediatric HCC? So is the tumor suppressor function of this pathway somehow interfered with in, in, in these tumors? And this is a work that was done by uh Michael Auaglia um when he was in the lab. So we looked at 77 patients with HCC. Uh, 4 of them have fibromylar type HCC. They range from 1.5 years of age to 15 years of age. And uh we looked at the nuclear yap um in these tumors and there was increased uh the yap going into the nucleus is, that's where it does its function. It binds to the DNA and, and that's how we can infer that the yap has increased activity um in these tumors. You may not be able to see it. This is a normal liver, non-neoplastic liver. Uh, but you have a, a lot more, uh, gap nuclear localization as well as some cytoplasmic gap, uh, in the tumors, uh, as as opposed to the, um, normal liver. When we looked at the transcript levels um of YA, um, we saw that overall there was a trend to increased um YAP expression. Uh, however, it should be noted that a lot of the regulatory mechanism of YAP is a post-translational modification. So even though, uh, it's transcript levels or gene expression may not be elevated, but it's, um, because of this post transcriptional uh regulation, uh, it can still be have significant effect on those cells. And uh we looked at the downstream target genes um and we did find the difference between the uh fibrolamyelar versus the non-fibrlayllar histology. It seemed like the non-fibrilloma histology had more Yap downstream gene effects than the fibrolaylar um histology in general. Some of the typical genes that we looked at are CTGF and CER 61, which were um well-known Yap Yap target genes. Uh, and then we did a, we did a HIPA array. We looked at a lot of the, um, HIPA pathway components, uh, using an array. One of the interesting things that we found was that there was a decrease in this, uh, tight junction protein too, and we basically saw this, uh, in 6 out of 7 samples which anytime you see this, this kind of effect, you have to, you know, ask the question of well what is this doing and what's the, uh, importance of this tight junction protein. There's not a lot known about this, this protein per se, but some of the reports do, um, suggest that, uh, it's an important, um, cell surface, um, or tight junction protein, especially in the, uh, uh, in regulation of proliferation and cell size regulation. There was one report of a pediatric HCC in a patient who actually had a tight junction um deficiency. So whether the deficiency really contributes to, to the proliferation of these cells or um or vice versa, it's, it's, it's unclear at this point. So in summary, um, our HIPPOYAP, uh, studies have shown that there's significant Yap nuclear localization in HCCs. We see a significant decrease in the tight junction protein too. I think I'd be interested in, in looking at this, uh, further down the road. Uh, and there we did see a difference in the, um, non-famulumar HTCs in regards to Yap target gene expression. Um. And uh, the HIPAA pathway uh has been looked at in many other cancers as well and it does seem to play a potential role in a number of other cancers. So, um, people have become more and more interested in, in, in looking at this as, as a potential therapeutic, um, uh, target, um, this pathway. Uh, there are a number of, uh, compounds that have been suggested to interfere with the apps effects, um. And uh I hope that we'll be able to actually test some of these compounds um down the road uh in some of our um in vitro and in vivo models. So, uh, you know how I mentioned the, the difference between fibromyellar and non-fibromyellar HCCs, um, Uh, the, one of the, one of the key, uh, histologic, um, findings of fibrillomylar HCC is really these dense collagen bundles, uh, that you see on histology. And so this is really a histologic diagnosis, um, that you see with the fibrillomylar, um. Tumors that this, the other types of HTCs do not have these, these um collagen bundles. But an important thing about Fulla Mueller HTCs is that uh is based on this finding, uh, which was published in 2014, um. This is a um the work from uh uh the Rockefeller Center and then Sloan Kettering which they found um the presence of this DNAJB1 protein kinase A chimeric transcript in in in in basically all the Farberella Mueller samples that they looked at. So what happens is that in chromosome 19, there's a deletion of 400 kilobase pairs. And uh there's a fusion of the DNAJB1 uh exon 1 or 2 to the protein kinase A. Uh, protein kinase A essentially just uses 11 exon, uh, one or two exons, and then you, and then, um, the patients end up with this fusion, uh, transcript. And fusion protein, um. When they looked at the, the rest of the genome, there were no other events. And, uh, here's a sample of the western blood against protein kinase A. You can see that in the tumors there are, there are two bands, uh, corresponding one to the wildlife protein kinase A, the other one to the, to the chimeric protein kinase A. So this was an important and interesting uh finding, uh, when we looked at our own um Uh, samples in our, in our, in patients with fibrillomylar HCC. Indeed we see as we saw that the majority of them with the histologic diagnosis of fibrillomylar HCC did have the, um, did have the fusion transcript as well. Uh, however, not all of them did, did, did so, and, uh, there's, there was, there have been another, there was another study that actually found that 80% of the fibroMR HCC that they looked at actually had the fusion transcript. So not every fibrolam HCC is going to have it, but a large majority seem to have it. So then, um, the question is, well, well, what is the, uh, what is this fusion, uh, protein does? I mean, how, how does it lead to cancer? Is it that the fusion protein actually has a completely new function? Protein kinase A does a lot, a lot of different things in the cell, but is it that suddenly now it has a new confirmation and, and actually functions in a new, new way? Uh, the loss of DNAAB1 allele is probably unlikely that would contribute, um, to, uh, carcinogenesis. Um. And then the other question that we have to ask is because of the protein kinase A and it's, it's effects on the cell, um, is it that, is there an increased activity, um, in the protein kinase A? Uh, protein kinasa's, uh, uh, just to give you a background, it's a cyclic AMP dependent protein kinase. Um, it's, it's involved in, in, in the signaling pathway that come down from the G-coupled, uh, protein receptors. Uh, it's function does tend to be cell type dependent, but it's involved in a lot of sort of, uh, metabolic and homeostatic, um, functions within the cell. So it's really hard to, you know, you can't turn off protein kina A in the cells, and really the cells do really need it, um, for basic function. But the question is, what really, then, then, then what's happening in some of these um uh uh tumors or what, why is this causing, um, uh, ACC to develop. Uh, one thing that, uh, we looked at was the CREB activation. This is a, a downstream factor of the protein kinaseta, uh, CRE, uh, a transcription factor. And this was the sort of the most, um, the first thing that, that, that, that we looked at, uh, it made the most sense to look at this, um, which Andrew Clarkey, uh, the, the postdoc in the lab has sort of been working on, on, on the, on this project and there's really no difference in, uh, in CRE, um, or, or phospho-CRE, which is the, uh, active form of CRE, um, in these, in these tumors. Then, um, we asked the question of, uh, well, what are the pathways might be of interest. And when we looked at the protein kinase pathway and we looked at our transcriptome data that I'm not showing here in regards to the fibroomy HTCs as well as some of the other published work, we, uh, became interested in P53 pathway, um, and we thought it would be worthwhile to look at this pathway in these tumors. So, um, we asked the question of the, what is the state of the P53 pathway in these fibrolumolar HCCs. Uh, just to, uh, refresh your memory regarding the, uh, function of P53. So P53 is a tumor suppressor pathway. As you know, um, you've heard of, you know, and I showed you in, in adult HTCs, uh, the most common, uh, mutation, driver mutation was a mutation in P5 P53 itself. And so P53 mutation results in cancers in various types of tissues. As you know, But in, in the fibrillomylars, um, P53 is normal, so it's a wild type, um P53 and it should function normally. And what its normal function is, and here's the P53, and what its normal function really is, is is to respond to any um genomic DNA instability or if there's any, any um Uh, radiation or uh ionizing uh injury to the DNA or any breaks in the DNA, uh, this DNA, what's called the DNA damage response pathway gets triggered because the cell has to make a decision of, OK, we, we have a problem, we have an Stability in our gene, uh, we need to repair it. So it either tries to repair, repair the damage, or if the damage is too much, the cell has to make a decision of actually dying or killing itself because it's just not a normal functioning cell. So this decision making can lead to apoptosis when the P53 pathway is activated such that it eventually um uh triggers some mitochondrial pathways that lead eventually to apoptosis and cell death. So this is an important um regulatory mechanism for the cell to be able to decide its fate. So when we looked at the, looked at P53, and this is a wild type P53 in these HCCs. Uh, we saw that there was, uh, an increase overall expression of P53 and 5 lomolar HCCs. We saw, uh, when we looked at the immunohistochemistry, here you can see, uh, this is a normal liver and this is a fabulomylar liver. You can see that, uh, there really is not much P53 expression, uh, within the cells at all. However, in, in the tumor cells, there's, there's a lot of, uh, cytoplasmic, um, P53 expression. When we looked at the, uh, phosphorylated um serum 15 or P53, which is an important, uh, um. Activator of P53. You can see that it, it shows a very interesting distinctive pattern of staining of the cytoplasm. And potentially of the mitochondria. And one was, this was um uh the immunohistochemistry was uh scored actually by Juan uh Petra who's uh one of the fellows in the pathology department um uh using this IRS score and then uh you see that there's an increased cytoplasmic and mitochondrial um staining within, uh, of the, of the, uh, phosphorylated P53 in these uh in these tumors. So, uh, then we ask the question, well, if the P5C3 here is potentially activated, is it activated because of an upstream factor, um, or regulator ATM, uh, which really, uh, is one of the early, um, events that recognizes any injury to the, um, to the genome. And uh indeed we found that there was ATM protein in the, in the fibrillomylar HCC. Um, here you can see the comparison between normal and tumor. And uh when we looked at the phosphorylated ATM and the fibrolaylar HCC there was also an increase in the um In the nuclear expression of the phosphorylated ATM. When we looked at the downstream target genes, uh, so P21 is a classic downstream target gene of P53, uh, and, um, but we didn't see any, any difference, um, uh, in, uh, in increased expression of, uh, P21. So Our results have shown that potentially the cellular stress signals are coming in into these tumor cells that somehow the cells are potentially recognizing this, that the ATM is getting activated and P53 is potentially getting activated. Um, however, it's, it, it, it's probably not doing its job of, of regulating these cells. Clearly the cells are not undergoing apoptosis because the cancer cells continue to survive. And um so that's, that we know at least it's not happening. Um. So we asked the question of um uh well, what could be preventing that from, from happening. Uh, one thing that uh is, has been studied fairly extensively is an interaction of MDMX and MDM2 which are inhibitors of P53. People have developed significant interest in In this, uh, sort of reactivation of normal P53 function for treatment of cancer in general. Uh, the reason being that, uh, it's, there's been a lot of work and some evidence showing that even in tumors that do not have P53 mutation. That the P53 pathways used by the cancer cells for their own advantage or that somehow the cancer cells are actually turning off the normal action of P53, uh, so that it doesn't result in cell death, senescence or apoptosis. So, there's this sort of movement towards, um, figuring out how can we sort of reactivate the effects of P53, uh, the normal effects of the P53. So, um, for our work, we decided to look at, uh, MDMX which is uh an inhibitor of P53. Uh, when we looked at the transcript levels of MDMX, we saw that in the majority, there's increased, uh, MDMX, uh, transcript levels. Uh, when we looked at the nuclear, uh, staining, uh, we saw that there was a significant increase in, uh, MDMX, uh, nuclear staining in the fibromlar HCCs. So, our summary of our P53 findings are that there, the, it seems like the DNA damage response is activated. Um, however, we don't see clear downstream effects of P53. Um, we do see an increase in the uh P53 inhibitor, MDMX based on its nuclear localization as well as its mRNA expression. And so currently there are uh several MDMX and MDA-2 inhibitors in clinical trials for various um cancers, cancers, uh, some of them are hematologic malignancies, but none of them are in liver tumors. But this is another um another uh target that I think, um, hopefully we can test out um down the road against some of these liver tumors. So, you know, I've talked to you a little bit about some of the molecular uh work uh that we've been doing. Um, now I'm gonna shift gears and talk about some of the uh culture work that we've been doing because we've gotten to a point where we can keep looking. We've been looking at these, uh, at, at the tissues from, from, uh, from patient samples and we've sort of found some potential interesting thing, the targets that we may, may, may want to test out, but we We really need to test this out in, in, in against human uh tumor cells. Uh, you know, a lot of this work has been done in, in cell lines and, you know, people have shown that some of this works in cell lines, but not until you can actually show it in, uh, in patient derived cells will it actually be of, of, of benefit or we can move it forward. So, um, Stephanie Kim, uh, has undertaken, uh, the majority of the work of, of developing, uh, cell lines from, from the patients, um. That undergo resection or transplant. Um, there's a picture of a hepatolastoma cell line, uh, that she's been working on. There's a picture of it, um, Fabula Mueller cell line. It's growing these cells is not easy. It's, it's a, it's a challenge. Um, it's, it's not an easy thing to do. Um, the, in general, uh, hepatocytes and liver cancer cells just do not like to grow. It's just not, it's not a normal environment for them, obviously. So it's not easy to grow these in, uh, in these types of static cultures in general. Uh, what, uh, what we've seen is that initially, you know, cells do grow, these, these do look like, um, uh, hepatocytic type, uh, tumor cells. We do see another population within, within, uh, some of our cultures, um, of these types of cells. Uh, these correspond mainly to, um, probably stele cells which are supportive cells within the liver, um. Over time as they grow, as the culture grows, they do become confluent like this. And uh the question is, is, you know, some of this is gonna be some of those stell cells that actually differentiate into myofibroblasts and when they get activated, they, they differentiate into myofibroblasts. But where are the tumor cells? It's really hard to know where the tumor cells are or are there any tumor cells. Um, when, uh, Stephanie looked, took some of the, the culture and actually looked for the fusion transcript, cause we know in our patient, the patient does have this fusion protein. And when we looked at the fusion protein within the culture, it seemed like there is some protein, uh, fusion protein there. So that was a little bit pro little promising. That maybe amidst all this, despite the fact that some of this may be myofibroblasts, there's still some tumor cells that are living there. And maybe the tumor cells do need some of this, uh, the cells sort of my my fibroblasts to actually um survive and grow. But we need to come up with ways to really find some of these tumor cells better. And, uh, uh, Ina, um, who's, uh, who's at, uh, who's been working on trying to find um uh uh to, to find a, uh, a way to actually um identify some of these cells. Um, she's in the process of, um, coming up with ways to be able to actually uh monitor this maybe with insight to hybridization. Against the fusion, um, fusion gene, uh, so that we can actually monitor the growth of these, uh, tumor cells as they get passaged, um, over time. So, um, the other thing that we've been doing with the tumor tissue besides the static cultures, we've been, um, implanting them in, in mice. These are, uh, so we take the tumor tissue from, from the patients and we put them in liver of, um, immunocompromised mice. So these mice do not have any immune system, so they will not, uh, the immune system doesn't react to the human, um, liver tissue. And uh this is a picture of a patient that had a hepatolastoma um implanted into the liver. Then after 11 months, it, uh, we noticed that there was a tumor there. That's a long time. Uh, it takes, it takes a while for some of these tumors to grow. Um, but this type of a, um, patient derived xenograft model is something that People, uh, are using to study various types of, uh, cancers and tumors. Um, it's a good way to test potentially, potential therapeutics. Uh, it's, it's always the downside of it is all, uh, one is, you know, obviously the time that it takes for these tumors to grow. Um, but, but it, it's a way that we can actually propagate so we take the tumor from this mouse and put it in other mice to be able to propagate the tumor over time and then to be able to study it. Anytime you propagate any tumor in a PDX model or cell culture, there's always the, uh, the question, the criticism of obviously those cells are changing their uh potentially their genetic makeup, um, and they become different types of cells as time goes on. Um, so, it becomes critical to be able to actually study this as close to your, um, initial tumor as possible, um, in order to actually draw some conclusions, um, regarding it if it's possible. So far since, uh, just a little over a year ago, uh, we have implanted, um, a number of mice. Um, we have 7, we tried this with 7 patients for hepatoblastoma 3 HCC, uh, fibrillomylar HCC and 2 with HTCNOS, um, from 12 patients that we have. So far 23 mice that are alive and that we've had so far 1 tumor grafting from hepatoblastoma. The fibrillamylar HCCs are still uh not that far out so we'll see, we'll see how this is gonna, how this is gonna pan out. Over the next year or so. So, as I show, as I've shown you, you know, classic in vitro culture conditions are generally not ideal for um HCC. Um, PDX models are slow growing and generally have low engraftment rates. Um. Uh, So one thing that we've also focused our attention on is, is to um look at dynamic culture systems. Um, one thing that has been um shown is that uh some of the, uh, there's evidence that cells generally like to feel a certain level of sheer stress on them. And so having a static culture where you just have the cells in a, in a dish um is not as physiologic. And so, I, uh, I, I designed this, um, type of a chamber, um, in order to be able to study potentially, um, liver, uh, biology and liver tumor biology in, in sort of a, in a dynamic, uh, system. So this is a, this is a chamber which the idea is that we can put a liver slice between membranes and we can have inflow and outflow of media which is connected to an oxygenator um with the idea of being able to Keep these cells in culture, um, for a prolonged period of time, but at the same time to be able to impart sheer stress on the cells. Um, you know, people use, you know, sometimes they grow cells in, in, in, in, uh, culture plates and they are shaking the cells. So just for the fact that the media is actually moving over the cells, it gives them some sense of sheer stress. You can imagine in the body, there's a lot of, because of the blood flow, um, there, there, there is sheer stress upon the tissues in general. So that becomes trying to get this as most physiologic as possible. Uh, so, um, we, with the help of the, uh, uh, Si uh pizza center here with that they have the 3D printers. They were, they printed out, uh, some of these um uh incubators uh for us. And, uh, Stephanie has been working on um testing this out. Uh, we're still continuing to work on the um sort of optimizing the design of these uh of these chambers. Uh, here you can see that uh there are some tissue slice, some liver slices within these chambers that, um, she actually, um, we, she did one run of these, um, over seven-day period as a trial run. Um, and here you can see what has happened, what happens to the liver cells. The conditions were probably not optimal by any means, but at least we're, we're able to get the system going preliminary for, for, for a period of time. This is a dappy staining look at the, the nuclei. Of the cells. Um, obviously, by day 7 or so, there's, there's a very significant number of drop off of cells. Um, however, there's still some potentially some cells that might be viable. Um, the idea being that if we use this Um, this type of an approach with the, with the liver tissue slice and if we can grow some of the tumors on them, the matrix and the extracellular matrix of these liver tissue slices might actually provide a much more um Uh, physiologic state and, and, and, and the scaffold for the tumors to actually grow on this. But in their static cultures, we're trying to grow some of these um tumors just simply in in Metrogel or other types of material that are really basically broken up, ground up um extracellular matrix. So, why not try to use an extracellular matrix as close to, to, um, Uh, to what is normal anatomically as possible. And, um, and so that's sort of the, the idea behind trying to lose, use some of these, uh, liver tissue slices, um, in order to grow, grow these, uh, grow these tumors. So, in uh, in conclusion, um, the targeting HIPA pathway and P53 uh reactivation may be potential therapeutic strategies, uh, in pediatric HCC, um. One take home message here should be that it's fairly clear now that the presence of this fusion gene um really is uh is a, is a predominant finding in the majority of fibrilllaylar HCC's, uh, and that fibrillomyellar HCC should probably start being, uh, categorized as such whether it has this fusion transcript or not. And I think, um, right now, uh, Ilana Church and, uh, the pathology department is working on, on, uh, getting this, um, Set up and running as sort of a clinical um A report that the pathology department will actually be looking at this in order to report the presence of this gene in, in uh fro MLR HCCs. And uh because the pediatric HCC is so rare, um, really, I think that individualized tumor analysis is sort of end of one analysis which, you know, it, it can cause a lot of controversy because obviously people like to look at very, very big data and um uh we, we don't have the luxury of that um right now. And so, uh, I think this type of analysis might be still beneficial. To really try to continue to look at each of these tumors, uh, look at, uh, the various pathways. I didn't share with you that data that we've also been trying to, um, look at some of the transcriptomic data, uh, as well, uh, in, in some of these uh tumors. Um, but I think this, this, this will, uh, hopefully be continue to be beneficial, um. As we move forward. And uh I think one of the big hurdles right now is to, to be able to actually test some of the therapeutics and on these HCC cells. And so, uh, developing better strategies to grow these cells, as I mentioned with some of the dynamic cultures, uh, and using precision, uh, cut liver slices might be uh a way to move forward in, in trying to grow some of these, uh, uh, tumors in vitro. Uh, I, well, I'd like to, um, thank the, uh, members of the lab. Uh, they do all the work, uh, the hard work in the lab, and, uh, I'm, I've been lucky to have a great group of, um, people in the lab, uh, currently and, and previously who've, who've done all the hard work and I appreciate all their efforts, um. Uh, and uh they're, they're curious, dedicated, uh, group of, uh, uh, folks that have in the lab. So, um, I'd like to thank them for all their, all their efforts. And, uh, you know, I'd like to thank uh Doctor Shamberger and the department for um funding the, the research that I presented you here. Uh, obviously, without that, none of this would be able to, to get done. Um, we, uh, the, the MGH group, um, has been very helpful, um, over the years with a lot of the, um, basic science work that we've been, we've been doing. Uh, here at BCH, uh, Matt Warman has been a great help in a lot of the genetic, uh, work that we've been doing on some of the, on the tumors. Uh, Fernando Camargo, uh, is a great collaborator. He's an expert in the HIPO-YA pathway that we've been working on a number of different projects with. And Antonio and uh Juan have been a great help uh from the pathology department. Um, And uh I also thank uh Doctor Jackson and Falza for all their work in the Lab and maintaining the lab and the keeping everything smooth there. And uh uh we've we've developed a, we've developed a collaboration with uh Sanji Bazu Devan at Children's at Texas Children's Hospital. We've had Uh, lab meetings every couple of months. Um, he's interested in, uh, hepatolastoma, um, more than HCC, but they're trying to grow these cells in culture as well as, um, develop PDX model. So we've been able to exchange a lot of ideas, um, and our, our Our lab members have been um uh in contact uh over the past year or so and that's been a great, great collaboration that we have the same, uh, a lot of the same goals, um. So thank you for your attention. I'll take any questions. it's a very I have a couple of good questions. In with the patients. And the, the patients with that had. I OK. The patients that had received transplant always ended up having much Higher survival rates, but isn't that an effect of that those patients were highly selected and the ones that had resection included the patients with with all stages and not necessarily metastatic disease. Uh, right. So for the adults, um, they're pretty much very highly selected group because uh I didn't get into this, but obviously for adults to get transplanted. Um, there's some criteria, um, beyond which, um, based on the tumor size, um, you just, you can't transplant them. So this is sort of the Millan criteria that's been used, uh, which you have to have tumors less than 5 centimeters or, or 3, or less than 3 centimeters. So for the adult population, it's a much more controlled, controlled group. But the pediatric population is probably not as controlled, I would say, um. Uh, and so it's hard to know in some of the pediatric data that I showed in general, um. They, not all the, not all the studies really define the sizes of the tumors or what, what they are, so it's hard to tease that out for the pediatric population. I think the adult population has a much more homogeneous population. But even if you look at the pediatric population and you assume that there's a, uh, some variability, definitely they they probably don't have any metastatic disease for sure. But in regards to the size of the tumor. Or multifocality of the tumor or presence of vascular invasion, that data is just not teased out, uh, and to be able to gain that, that data from it. Um, and the second question is, in the fibrolamellar, uh, tumor system, it didn't seem convincing that the B53 was the, the factor that led to the neoplasia, those. Yeah, it's not. So what, what is, what's the ongoing hypothesis on what, what effect the fusion protein has on neoplasia? Yeah, so it, it, it probably, it doesn't, it doesn't lead to the P53, um, uh, it's just the cells are actually probably using the P53 or, or preventing P53 from doing its normal function. So, what it's actually, how it's causing the carcinogenesis, that's not, that's not clear. And uh we don't know, we don't really know that. There are a number of groups that are working on figuring that out um but it's not, it's not quite clear exactly how, how this is happening. We're trying to, in our lab, what we're trying to do is um Uh, trying to actually create this fusion transcript in in vitro. So we're trying to working on actually getting this expressed in cell lines and to see how this actually functions and in hepatocytes, as well as we've been trying to work with our uh the gene manipulation core in order to create a mouse model for this. So they've been trying to inject the embryos with a CRISPR Cas9 system in order to delete. Uh, delete this segment and see if we can develop a mouse model that we can actually then study. Um, because without having that, it's really hard to know exactly what the early signals are and how, how the, um, uh, cancer actually starts developing. So that's, we really have to start seeing those early Um, events. The P53 event is probably a later event. It's probably the cells have already transformed into tumor cells, and this is what we're seeing is just the tumor cells trying to prevent themselves, uh, from, uh, P53, uh, activating apoptosis in, in, in them. So it's not necessarily causing the cancer. It's just the way it's basically tumor cells hijack this pathway and use it to their own benefit. And this has been shown in other, other cancers as well. Um, so, uh, so it's not necessarily the cause, um, but it's just something that, that we see that potentially if we reactivate the normal activity of P53, um, uh, it could be that it might slow the tumor growth down cause we can certainly see that its cells giving a signal that maybe P53 is activated. It, it wants to do something, but somehow the tumor, the cell programming is such that it prevents its normal, normal function. Additional questions for Doctor Filli. Doctor Jackson. Uh, first of all, Cash, thank you for uh just a tremendous talk. Uh, it occurs to me that there's probably no one else in pediatric surgery who understands hepatocellular carcinoma, uh, the way you do, and, uh, I, I, I have to ask a self-serving question. We have a, uh, large cadre of patients in the intestinal failure program that have frank cirrhosis that are fully rally fed. Uh, the, uh, uh, survivorship of those has been shown by Doctor Brenna Fullerton to be greater than 95%, 5 year, and we live in fear of these patients, uh, evolving uh hepatocellular carcinoma. And to that, I was wondering, are there any biomarkers that we can follow to see the likelihood of that occurring or give us early warning? And secondly, do the adult data show, uh, give us some numeric indication of what the incidence is of hepatocellular carcinoma if you're cirrhotic versus if you're not cirrhotic. So in the uh adult the idea has been that if in cirrhotics, adult cirrhotics, it usually, uh, non, let's say non-hepatitis BC cirrhotics cause I think that the mechanism is completely different because you have, let's say an alcoholic cirrhotic, uh, it, it probably takes, uh, you know, the teaching is always or the thought has always been it might take maybe 15 years or so, 15+ years to potentially be at the high risk of developing, uh, HCC. I think um uh if you have a viral infection, it's a completely different story cause you see in, in hepatitis B patients you see um HCC without cirrhosis, um, so that can come up. As far as markers go, um, obviously alpha fetoprotein is a marker that can be checked. If, if it's elevated, um, It tells you that there's something there, but, um, a lot of the HCCs do not necessarily produce AFP so it's even if it's negative, it, it won't, it won't tell you anything. Uh, probably surveillance with ultrasounds, periodic ultrasounds is probably the best way to do that. There are some people working on some, uh, looking at some markers, circulating markers for HCC, um, but nothing that's clinically that effective, um. Thank you. Any additional questions? Yes, in the population of the HCC pediatric transplant patients, um, any, uh, differences in terms of cellular or molecular baseline differences between those that fare better in terms of disease-free recurrence down the line versus those that actually die or recur very early. Uh, there are some. I mean, there's been a lot of studies, especially in the adult population, there's been a lot of, a lot of studies that have been done. The problem is in the adult population, you have a much more heterogeneous group of patients. One of the things that we've also, um, noticed, but it's hard to draw a conclusion from, for example, if you have increased, uh, lowerchemic expression, um, actually that there's worse outcomes with that. Uh, with increased the app activity, people have seen that there's increased, um, uh, worse outcome. If you see beta-catenin, uh, expression, um, along increased beta-catenin, uh, nuclear localization along with the app that has a better prognosis. So there are a lot of, a lot of these things that have been studied and have been, people have looked at, but a lot of times it means. It's in a heterogeneous group of patients. Sometimes it's in cell lines. Um, so it's really, it's really difficult to, to draw that kind of a global conclusion at this point. Would some kind of meta-analysis combining across different centers be helpful to actually pinpoint some baseline factors that might be associated with better. Outcomes, in the transplant cohort. Yeah, we have to, we have to look at the same tumors. In order to do that, we have to actually look at the same types of tumors. So not all HTCs are gonna be the same. And you know, we say HTC, but they're really not the same. Their driver mutations are different, um, their transcriptome is different. So we really have to subcategorize them and really study them, um, uh, based on, based on the driver mutation, we have to bunch them together as, as much as we possibly can. In order to draw that kind of, to do a meta-analysis basically. OK, thanks so much for sharing all your work with us this morning. Thank you. Very