Welcome to the Liver Meeting Debrief Session Two. Please welcome your moderators, Dr. Carla Brady and Dr. Debbie Shawcross. Good morning, everyone. It's great to see you all and the fantastic session that we've got planned this morning. So we've got six really exciting debriefs and we're going to begin with basic science. We've got Yaron Rotman, who is a senior investigator and chief of the liver and energy metabolism section in the liver disease branch of NIDDK, taking us through all of the best bits of basic science. I'm really looking forward to hearing what you have to say. Good morning, everyone. Here are my disclosures. Essentially, I have no personal disclosures, but I did marry well, twice. So when ASLD asked me to cover this and I had to pick, I didn't know how difficult it's going to be to pick only a few abstracts out of the very many fantastic works that I've seen. I sort of felt like a kid in a toy store or, for those who know me better, a middle-aged man in a coffee machine store. And I promise, this is going to be almost the last use of AI here. So to make this, just so you know what I'm going to do, I'm going to very, very briefly summarize some works. I'll tell you what's the question they asked. I'll tell you what is the model they used, because this is the basic science part. What are the very, very distilled key findings and why I picked it. And I'll always have the abstract number and the first author name on top. So if you find something interesting, please go and either talk to them, read the abstract, or send them an email. Okay, let's start. Portal hypertension. This is work presented yesterday by Dr. Ma from the Wachiri Lab at Yale. And they based it, so the basic for the basis for this study is that we know that in portal hypertension, the decreased blood flow through the portal system is accompanied by a compensatory increase in the hepatic artery flow. And they asked a very smart question, does that matter? Does that affect liver remodeling? The model they used is a rat with partial portal vein ligation. So they essentially can control the diameter of the portal vein and control the degree of portal hypertension. So what did they find? First, they found that the model worked, so they did see the compensatory increase in hepatic artery flow. And they found that there was upregulation of VCAM-1 on the endothelial cells in the hepatic artery. Remember, they did not do anything to the hepatic artery. This is just because of the flow. They then could see that this leads to recruitment of specifically osteopontin-secreting macrophages, which they could also demonstrate the presence of in human samples with portal vein thrombosis. And that these, we know that osteopontin is a major activator of stellate cells and leading to fibrosis. And finally, when they treated their animals with anti-VCAM-1, they could actually decrease the degree of fibrosis that is seen. So I thought that was interesting because we found that in portal hypertension, the increased flow in the hepatic artery is sufficient by itself to drive portal fibrosis through the VCAM-1 macrophage axis. And maybe VCAM-1 inhibition could be a therapeutic target in some of those people. So moving from pre-hepatic to post-hepatic portal hypertension, this is a study by Dr. Kato from the Takehara Group at Osaka. And this is looking into hepatic congestion and how does that lead to fibrosis. They used the mouse model and partial inferior vena cava ligation. And what they found in the model is that after the ligation, specifically in the zone 3 liver sinusoidal endothelial cells, they see upregulation of integrin alpha-V, leading to activation of the aptas pathway, secretion of CTGF, and that CTGF can activate the stellate cells and drive fibrosis. Then they very nicely were able to show that if they knock out CTGF specifically in the endothelial cells, they can improve the fibrosis, they can improve the portal pressure, and even prevent some of the cancers that develop in these animals. And they achieved the same results by treating them with an anti-integrin. And they did have a correlate in human samples from fontanel-associated liver disease. So I thought that this was a very nice way they've shown how the pressure itself, hydrostatic pressure, leads to changes in the endothelial cells that eventually drive a fibrotic process. And maybe we can use anti-integrins, for example, in bad Chiari syndrome or something like that to prevent damage or decrease damage. Let's move to cancer. This is a study shown yesterday by Dr. Lemaitre from the Dan-F. Sekaran group at Stanford. And so we treat some patients with hepatocellular carcinoma with transarterial chemoembolization. We know that in most cases this is not curative and the cancer recurs. And the question they ask is, how come those residual cells that then eventually become a recurrence are not hit by the immune system? What helps them escape the immune detection? So they started their model with generating doxorubicin-resistant hepatocellular cells and growing them in 3D co-culture organoids with macrophages. And they also used transplant of human cancers into mice. What they found in their organoids is that the resistant hepatocellular cancers change the phenotype of the macrophages to a more pro-tumor phenotype with expression of PD-L1 and a secretion of TGF-beta. And then they used their mouse models. So they took mice. They transplanted them with doxorubicin-resistant HCC. And they treated them, hitting both pathways with TGF-beta inhibitor and anti-PD-L1. And there was a remarkable reduction in the tumor burden in these animals. So this is a nice pathway to suggest how residual HCC after taste maintains immune evasion and potentially suggesting an adjuvant therapy in humans in the future. This is a work by Dr. Zhang from Seoul National University. And we will talk about DPP-4. Most of us, or the clinicians, know DPP-4 from the diabetes field. So DPP-4 is a peptidase that cleaves proteins and basically regulates protein levels by cleaving them. So something I didn't know until I read that DPP-4 also cleaves CXCL-10 and inactivates it. And CXCL-10 is a major chemokine that withdraws in both infection and tumor, anti-tumor activity. So then they asked if this is DPP-4 can impair CXCL-10, maybe hitting DPP-4 can help us in a cancer model. So they used mice. The mice were injected with HCC. And they were treated with either anti-PD-1 rigoraphanib or cytoglyptin, which is a DPP-4 inhibitor, or the combination of both. And very simply, you see a marked effect. So black is the control. Red is the combination therapy. And this is the tumor burden over time. And this is the tumor weight. So there's a remarkable effect as an add-on inhibition of DPP-4. And they also saw an increase, which you'd expect, the increase in effector immune cells into the tumor. So this suggests that DPP-4 inhibitors may enhance the response to immune checkpoint inhibitors in patients with HCC. And DPP-4 inhibitors are clinically available, are safe. I think they're going generic. So maybe this is applicable to humans. OK, alcohol-associated liver disease. This is a study that was brought to us by Dr. Nagesh from the Zonghi-Sabers group and from Beth Israel. And I completely forgot to mention, if you see this icon, this is work that was supported by the ASLD Foundation. So please donate. And we know that alcohol-associated hepatitis has several important features. There's steatosis. There's a lot of inflammation related to neutrophils. There's an activity of DNRRP3 inflammasome. This is a lot of work that's been shown by that lab over the years. And the question is, can we have one target to rule them all? Can we hit all of those with one thing? And the authors suggest that brutal tyrosine kinase may be the candidate. And to do that, they've used several models. They've used samples from human patients with alcohol-associated hepatitis. They used the mouse model, the classic NIAAA of chronic exposure with binge. And they used evobrutinib, which is a BTK inhibitor. So first, they could see that in human and mouse samples from alcoholic hepatitis, they see upregulation of BTK phosphorylation. And in vitro, they can see that this is directly affected by ethanol, that ethanol induces the phosphorylation, and that the phosphorylated BTK will stabilize the NLRP3 inflammasome. They then do the mouse experiment. They treat the mice with BTK inhibitor. And you can see that there's a marked reduction in damage, as suggested by ALT, in triglyceride accumulation, as well as they could show that there's a decrease in macrophage NLRP3 and a decrease in neutrophil metosis. So overall, it does seem that the BTK inhibitor can improve multiple separate features of alcohol-associated hepatitis in a mouse model. And again, these inhibitors are available clinically. They're currently used in cancer treatment. So maybe they will work in humans. I think it's at least worth trying. This is another study on alcohol-associated liver disease coming from a collaboration between Elena Kaffert-Penn and Waj Mahal and Richard Flavell at Yale. And they were asking about the role of lipids in alcohol-associated hepatitis. So what they did is they took blood samples from humans with alcohol-associated hepatitis, heavy drinkers who don't have hepatitis, and healthy controls. And they did lipidomics on those blood samples. And what you can see very broadly, without even looking at details, is the patterns are very different. There are several lipid species that are markedly upregulated, and specifically in alcohol-associated hepatitis. Then they did something very cool. They used a humanized mouse model. But the problem with most humanized mouse models is that they don't have a good, they're immune deficient. And this is actually an immune-competent humanized mouse model, because they transplanted the mice with both human hepatocytes, as well as with human cord blood. So you actually get the human immune system in them, at least partially. They exposed these mice to the NIAAA mouse model, and they actually see the same alteration of lipids that they found in the humans. So this makes it a useful model. Then they asked, are these lipids just a feature, or are they causative? And what they did, they used regular mice, not humanized mice. They exposed them to the alcohol model. And they also treated them with those specific lipids that they saw upregulated in the humans. And what they found is that they could recapture some, not all, but some of the features of worsening disease. They had higher ALT in some of those lipids. They had the similar feature, immunological features in the livers. So I picked this because I thought this is a very cool mouse model, and it seems like a useful model to study that disease. And also, because it's interesting, it suggests that the lipids that we see altered may actually have a causative role in promoting the disease progression. So this is a very cool mouse model. So let's talk about mouse study. So this is work from Dr. Daniel from the Harmit Marhis group at Mayo Clinic. And we know that if we look at the hepatocyte in mash, the hepatocytes are miserable with lipotoxic ER stress. There's an entire epigenetic regulation and change of hepatocellular processes probably to accommodate for that. And also that those miserable hepatocytes are releasing damage-associated molecular patterns through extracellular vesicles that affect other cell types and affect the rest of the liver. And a lot of this comes from Harmit's previous work. And how are all of these connected? So what they did is they used three models. They used extracellular vesicles from in vitro hepatocytes cultured with palmitate to generate lipotoxicity. They took livers from the classic mouse models of mash or mash fibrosis, and they also had patient sera. And in all of those samples, they could see specifically upregulation of a protein S100A11. Here I'm just showing you extracellular vesicles from the in vitro experiments and how it's upregulated by palmitic acid. Now comes the clever part. So they went into the genomic region upstream of near S100A11, and they found a putative enhancer site. And they could show that that enhancer site is specifically acetylated when exposed to lipotoxic stress. So they promptly named it the Lipotoxicity Influence Enhancer, or LIE. Then they had a nice, very nice mouse model based on CRISPR-Cas technology, where you very, very specifically block the ability to acetylate that enhancer. It's the Cas9 crab mouse model, for those who want to read about this. And when they block just that specific enhancer acetylation in those mesh models, they can actually see that there is a decrease, as expected, in S100A11 secretion, but also improved ALT and improved fibrosis. So this tells us several things. First, they actually have a very nice pathway how lipotoxic stress and through something I haven't shown you, through the ER stress, cleavage and activation of XBP1 that migrates to the nucleus, very specifically activates a very specific site, I said specific too many times, enhancer for S100A11. And they also showed that S100A11 may actually play a role in the pathogenesis of the disease. Okay, let's talk about toxic injury, or more specifically, acetaminophen. So this is coming from Dr. Kay from the Wenchies group at University of Pittsburgh. And we know that the liver, after acetaminophen injury, there's marked oxidative stress and injury. Now, proteins have different responses to oxidative stress, and one of the processes is S-sulfenylation, the attachment of a sulfonyl acid to proteins in the context of oxidative stress. And this is a transient thing. It gets attached and it gets removed. The major desulfonylation enzyme is Sulfi-rayoxin-1, SRXN1. What is the role of the process and what is the role of SRXN1 in acetaminophen injury? So the authors first look at samples from both humans and mice treated with acetaminophen, and they found upregulation of acetylphenylation, and they found, as in this immunohistochemistry, upregulation of SRXN1 in those livers. That then generated a SRXN1 knockout mouse, a PADSAT-specific knockout mouse, and what they could show is that if you knock out SRXN1 and expose the mice to acetaminophen, you actually have way worse injury. And the flip side is that if you overexpress SRXN1, you actually improve the injury. So what this is suggesting is that the sulfenylation of proteins in the context of acetaminophen-induced oxidative stress is damaging, that the upregulation of SRXN1 is a compensatory pathway to actually try and control that and contain that damage, and that maybe if we target SRXN1 some way, we may be able to help these patients. Another study looking at acetaminophen injury was brought to us by Dr. Jain from the Bushan Lab in University of Pittsburgh, and they were focusing on the EGF receptor on hepatocytes. So we know that the EGF signaling and EGF receptor are crucial for hepatocyte proliferation, and especially in models like partial hepatectomy where you need the liver to regenerate. Interestingly, in previous studies, the regeneration that happens after acetaminophen injury, you also need the hepatocytes to regenerate, seems to be actually inhibited by the EGF signaling. So they wanted to look at and explain this, and they used a hepatocyte-specific EGFR knockout mice, and they treated them with acetaminophen. And what they found is that their knockouts actually had remarkable improvement in resolution of injury. You can see that the wild types, this is the ALT in the wild types over time, and you can see that in the knockouts, the injury actually peaks at 12 or 24 hours and then decreases. And this is accompanied by remarkable recovery of glutathione stores, and we know that so basically they just recover better. Now what's the mechanism? They very nicely show that in the knockout animals, there's a decrease in phosphorylation of Junkinase and its translocation to the mitochondria. And that leads first to less mitochondrial damage, but also there's upregulation of mitophagy, so the damaged mitochondria can be handled in a safe manner. And that drives increased and faster regeneration. And of course then they treat the animals with EGFR inhibitors, and they can show that these animals have benefit both if you treat them very, very early, or even if you start treatment four hours after the exposure, which is a little bit more like the clinical feature that we tend to see. So overall I thought it's interesting because they show a pathway that EGFR, although it's useful in regeneration after a hepatectomy, actually contributes to the damage, and they show how by preventing the containment of mitochondrial injury. And maybe we can use EGFR inhibitors in the future in people who have acetaminophen overdose. Okay, let's go to the biliary system. This is very nice work coming from Dr. Ferreira from the Forbes Lab at the University of Edinburgh. So we heard a lot about senescence in the basic science presentations in this meeting, and it's been known that in biliary diseases you can see a senescent phenotype in biliary epithelial cells. And the authors ask a very interesting question, does that actually affect the hepatocytes as well? So first they looked at human samples, and they looked at human samples from biliary diseases. We have biliary atresia, we have PSC, we have PBC. In all of those they saw senescent biliary epithelial cells that expressed P21. But also in all of those they saw that hepatocytes upregulated and expressed VCAM1. We've already spoken about VCAM1 earlier. And you can see it here in purple around the hepatocytes in all of these three models. And they also found that there's clustering, that there's migration of neutrophils and clustering specifically around the senescent biliary epithelial cells, and specifically around the VCAM1 hepatocytes. They're here in yellow, the neutrophils. So then how does that work? They go in vitro, they generate senescent biliary epithelial cells, they co-culture them in a known contact system with primary hepatocytes, and they see that this is sufficient to induce the senescent phenotype is sufficient to induce VCAM1 expression in the hepatocytes, and also to induce and to upregulate programs that are associated with neutrophil recruitment. And they could then replicate it in a mouse model. So overall, I think they show a nice pathway where the senescent biliary epithelial cells signal to adjacent hepatocytes, adjacent hepatocytes upregulate the VCAM1, which then sort of calls the neutrophils to the region, and the neutrophils actually can target those senescent biliary epithelial cells. This could be beneficial or could actually be a harmful feedback loop. And maybe if we can target this biliary hepatocyte communication in some of the biliary diseases. From chronic to more acute, this is data by Dr. Kutsenko from Paul Monger's lab. So we know quite a lot about the role of beta-catenin, especially in the context of wind signaling and maintaining hepatocyte identity, and a lot of it actually comes from the Monger lab. But beta-catenin also has another role. Together with its cousin, gamma-catenin, they actually form the adherence junction in biliary epithelium, keeping those cells together. And the question is, how important are they there, and what's the importance of the adherence junction? So the authors used a double knockout mouse model, where they knock out specifically in biliary epithelial cells, the beta and gamma-catenin. You have to do this in an inducible system, because otherwise they would never develop. So this is inducible. The knockout is driven by exposure to tamoxifen, so you can actually control it. So you take these mice, you give them tamoxifen, you knock out beta and gamma-catenin, and then in four weeks, all of them die. And when you look in their livers, you see tremendous biliary infarcts. There's complete loss of integrity of the biliary system. This is dye injected into the biliary tree, and you can see how it leaks in the damaged animals. And you also could show it functionally, that there was a change in altered biliary flow in these animals. But all of these animals die. You want to study how to recover. So they could use the benefit of the tamoxifen-induced system to actually titrate the damage. So they used a milder version. They did not completely knock it out. And then some of the animals die, but some of the animals actually are able to recover. And when you look at those animals, what's very, very interesting is that you actually see changes in the hepatocytes. And some of the hepatocytes are starting to express biliary markers like A6, which is shown here in the figure. So although this is preliminary, this is, I think, very nicely shows the roles of the beta and gamma-catenin in maintaining bile duct integrity. It gives us a useful model for acute biliary injury that can be titrated. And it shows the importance of the transdifferentiation in hepatocytes and their plasticity in helping to recover from something that's actually starting at the biliary tree. Okay, viral hepatitis. So this is a study by Chloe Berm, a medical student from Charlie Rice's lab. And very simply, they wanted to identify host factors that regulate viral HBS antigen production and maybe could be used as targets in the future. They were specifically focusing on stages that are after the viral entry into the cell. And this was done by transfecting HUH7.5 hepatocyte cells with HBV pre-genomic RNA. So this is the, you transfected this step, you bypass all the entry sites, and you just go, you only study these levels. And then they used those transfected cells in a, essentially in an automated manner with a CRISPR-Cas library that individually knocks out every single gene in the genome. And you want to see what every single gene in the genome, what that would do to HBS antigen expression. So this is a very broad screening method. And looking at all of the genome, they found some targets that are increasing the secretions. And these are, when you knock them down, so these are antiviral, some targets that are decreasing secretions. These are, these would be pro-viral. Then they confirmed, they used a confirmatory assay of the 100 top hits. And they could confirm many of them. They could find, they found pro-viral, anti-viral, they even have very nicely NTCP, which is the entry. And you expect it not to be changed, because that's not the step that they use. And they could confirm this. And unfortunately, we don't have more data on the specific targets, but I thought it was just a nice start to something that may lead to new therapeutic targets. Okay, and we're going then to regeneration. So this is work that was presented by Professor Carr from the Institute of Liver and Biliary Sciences in New Delhi. So we know that when you regenerate in the classic model, the liver regenerates, a classic model is after partial hepatectomy, it's a multicellular event, the hepatocytes regenerate, but you need to build new blood cells, you need to have all the non-parenchymal cells. We actually don't know much about the liver lymphatics in that context. So the authors used a model of RET, undergo 70% partial hepatectomy, or a sham experiment as a control. And then they specifically looked and stained the lymphatic endothelial cells in these recovering animals, and you can see them in red. And what they found is that the kinetics of recovery really paralleled the kinetics of recovery of cholangiocytes. Here you can see them in green, of the biliary tree. And when they did single-cell proteomics on these recovery lymphatic endothelial cells, they could see that there's upregulation of Wnt7a. We spoke about Wnt signaling just recently. So Wnt is a secreted factor that's, so do they secrete anything? So then they take lymphatic endothelial cells, they grow them in culture, they use conditioned media, and they treat cholangiocyte organoids with that conditioned media, and they could show that that actually increases the proliferation of these organoids. And then they go back to the animal experiment. They block lymphangiogenesis after partial hepatectomy with anti-VEGF treatment, and they saw that this truly decreases the recovery of cholangiocytes in biliary tree. So overall I thought that, first, this is a good study to show us that we tend to neglect the lymphatics in the liver, but this is how important they are and how they recover after partial hepatectomy, but also that this is not just the lymphatics recovering, that again, we have this cross-cell type communication that's important for other things. For the final abstract of the morning from me, this is work coming from pharma, presented by Dr. Li from Regeneron, and this is dealing with humanized mice. So for those of you that don't work in humanized mice, these are essentially mice that are A, immune deficient, and B, there's something in them that kills the hepatocytes. People have used in the past TPA, and now I think most people are using the FAH knockout, which essentially is a model of tyrosinemia, and you can keep them alive until you change their diet, and then their hepatocytes die. And what people do is then you inject them with primary human hepatocytes that will then go home into the liver and engraft, so you get a mouse with human hepatocytes. That sounds fantastic, but these are very difficult models. Their engraftment can be slow, can be unpredictable, you will get a lot of variability, and you have some donors that are not very good donors, you see that they don't engraft well. And also, the engrafted hepatocytes tend to develop spontaneously lipid droplet accumulation, which suggests that they're not very happy. Now they had a very nice hypothesis, and they said, okay, you know, we have those human hepatocytes in the mouse, but the signals that they see are mouse signals, they're not human signals, so maybe we have an issue with mitogenic signals. And they show it very nicely here, where you take human hepatocytes, and if you treat them with human IL-6, you see that the pathway that, for example, this is STAT-3 phosphorylation, is affected, but if you treat them with mouse IL-6 or rat IL-6, you don't see that. So essentially, we put the hepatocyte there, but it doesn't see the signals it's used to. So can they improve on the process? So of course they can. What they did is they took this model, which is like triple or quadruple knockout, and they genetically modified them even further. So they made these mice express the human IL-6 instead of the mouse IL-6. And then hitting another mitogenic pathway, they treated these animals with C-met activating antibody, a human C-met activating antibody, which Regeneron has. And when you hit both of these targets together, not each one separately, this is what you see. You can see that when you hit both targets together, you have marked increase in human albumin in the serum. And if you look, the cells stained in brown here are stained for human albumin. And you can see that in this liver, there are few human hepatocytes, but here there are way more. Then they even go even further. Instead of hitting two pathways, they hit three, and they treated these animals with human endothelial growth factor. We just recently spoke about that. And you can see, really, this may be a typical best example, but still you can see marked increase in human albumin and engraftment. So I thought this is interesting because without repeating the word humanize too many times, this is actually a way to improve the model, make it more effective. And these are models that are very helpful in studying human disease in a mouse model. I'll end first by thanking all the authors for fantastic studies, not only the ones that I presented, but also other studies, and for sharing the slides with me. ASLD staff for help, ASLD Foundation for paying for a lot of this work. I also would like to apologize to the authors whose work I butchered and the authors whose work I couldn't show, to the audience who had to tolerate me. And I'll return to AI and say that if you could handle this until now, then you're welcome to enjoy your coffee. And it's quite remarkable what you can do with those software. So thank you, everyone. Thank you very much.

liver disease research

portal hypertension

hepatocellular carcinoma

fibrosis management

immune evasion

therapeutic targets