Good afternoon, everyone. Welcome to the second part of the session on fibrosis. The first speaker, okay, we will have questions only after the fourth talk, so you can write your question and ask afterwards, okay? So the first speaker in this session is Natalia Nieto from the University of Illinois, and the title of her talk is Role of SPP1 High in Macrophages in NASH. Great, thank you, Nissim, for the introduction. It's great to be back at the liver meeting. Finally, we made it. All right, so the title of my talk today is Role of Osteopontin High in Macrophages in NASH. I have nothing to disclose. Osteopontin, encoded by the SPP1 gene, is a soluble factor and an extracellular matrix protein. It's present in most tissues and body fluids, opsonizes bacteria for phagocytosis, promoting host defense. We know that global osteopontin knockout mice are impaired for clearing intracellular pathogens, and more importantly, the concentration of osteopontin in human milk, umbilical cord, and infant's plasma is very high compared to that found in tissues or in adults. And that made us think that perhaps osteopontin could have some protective effects. Over the past few years, my lab has studied the role of osteopontin in chronic liver disease, and in this project, we asked the question of whether macrophage-derived osteopontin could play a role in NASH. To answer this, we performed analysis of RNA sequencing from human liver of 10 control and 206 NAFLD patients, and found correlation between osteopontin and the NASH activity score and the fibrosis score. We then performed clustering analysis of human macrophages obtained from single-cell RNA sick from five healthy and five cirrhotic patients. Cluster 4 was identified as with high expression of osteopontin, but also with TREM2, which is a marker of lipid-associated macrophages. Importantly, osteopontin-positive TREM2-positive macrophages were increased in cirrhotic, particularly in NAFLD compared to an injured liver. To make sure that osteopontin increases in macrophages in NASH, we performed immunofluorescence and fluorescence in cytohybridization for osteopontin and CD68, which is a marker of macrophages. And we observed a progressive increase in the number of osteopontin-positive macrophages with fibrosis stage in NASH. We then looked at mouse liver. Mice feather control diet have significant expression in cholangiocytes and a little bit in hepatocytes. However, if you feed them a NASH-inducing diet, there is increase in hepatocytes, but more importantly, in macrophages that start forming crown-like structures, also noticeable on fluorescence in cytohybridization for osteopontin and a macrophage marker. We then were granted access to RNA sequencing from macrophages isolated from patients that have NASH, osteoporosis score 0 to 3, from this clinical trial run by Gilead Sciences, and these are non-treated patients. Strikingly, we observed that macrophages that had the highest expression of osteopontin correlated with the lowest osteoporosis score. So then we asked the question of whether osteopontin-high macrophages increased to protect from NASH. To understand the role of macrophages, of osteopontin-high macrophages, we reanalyzed a single-cell RNA-seq of mouse liver macrophages from three control and three NASH samples. These analyses provide a higher-resolution view and identified several clusters that express osteopontin with the SPP1-high macrophages, not only as the highest expression of osteopontin, but also expressing three additional markers of lipid-associated macrophages, TREM2, GP, and MB, and CD9. Coincidentally, 98% of the osteopontin macrophages were from mice fed a NASH-inducing diet. We then performed differential expression analysis of this cluster of osteopontin-high macrophages compared to the other clusters. We observed that there was a significant reduction in genes involved in inflammation, such as S100A4, interleukin-1 beta, or in fibrosis, but upregulation of genes involved in lipid metabolism, such as LPL, CD36, and various fatty acid-binding proteins, and also in genes involved in extracellular matrix remodeling, such as MMP12 and MMP13. So this made us think that perhaps the upregulation of osteopontin in macrophages confers metabolic, but not inflammatory and fibrogenic properties in NASH. So this led to the hypothesis that osteopontin-high macrophages regulate lipid metabolism and protect from NASH. To test this possibility, we generated osteopontin-flux mice, and we obtained osteopontin-stop-flux mice. We bred them with lysosame CRE to target all myeloid cells and generated osteopontin knockout control littermates that I will abbreviate as wild-type throughout the presentation, and also osteopontin-knocking mice in myeloid cells. And we performed immunofluorescence and colocalization to validate the targeting strategy. In addition, in the knocking mice, we also isolated the macrophages from these mice and performed Western blood analysis. So we fed half of the mice with isocaloric diet, that's the control group, or with a high-fat fructose and cholesterol diet that mimics the Western diet to induce NASH. And then six months later, we evaluated liver injury. So both genders of wild-type mice developed NASH, as shown by severe osteothesis, inflammation, and hepatocyte ballooning degeneration. However, the knocking mice were almost fully protected. This was also reflected in a significant increase in what we call the NASH activity score, liver triglycerides and serum transaminases in the wild-type with NASH for both genders, what was prevented in the knocking mice. In addition, the wild-type mice developed chicken wire fibrosis, but this was absent in the knocking mice. Then we asked the question whether ablating osteopontin in myeloid cells could accelerate the development of NASH. So here we perform a time-course experiment from one to six months. I'm showing the first two time points. And basically, both genders of knockout mice developed signs of osteothesis and inflammation at one month, but then developed NASH at three months. And therefore, osteopontin knockout mice developed NASH faster. To validate our targeting strategy, we performed immunofluorescence for osteopontin in the red channel and F480 in the green channel. So the knocking mice with NASH showed massive induction of osteopontin in myeloid cells. And the knockouts did not have expression. This immunofluorescence, however, revealed something else, which is that while in the wild-type mice, you have an induction of osteopontin in neighboring cells, mostly hepatocytes, in the knocking, this is absent, but is reverted in the knockouts, and it's even higher, suggesting that overexpressing osteopontin in macrophages might be able to govern also production of osteopontin by hepatocytes. So to validate this, we isolated wild-type hepatocytes, plate them in culture, with condition medium from macrophages that were either from a wild-type with NASH or a knocking with NASH. And the middle panel shows that hepatocytes developed osteothesis and increased osteopontin under wild-type conditions, but this is prevented by the knocking macrophages. To understand the mechanism whereby the knocking mice with NASH were protected from osteothesis, we performed liver lipidomics analysis. Peak intensity revealed that 35 triglycerides were significantly changed in knocking compared to wild-type with NASH, of which 22 decreased and 13 increased. We then rearranged these triglycerides by the total number of carbon atoms and the total number of double bonds, and what we found is triglycerides containing long-chain saturated and monounsaturated fatty acids decreased in the knocking mice compared to the wild-type with NASH. And these fatty acids are preferred substrates for fatty acid oxidation, and therefore suggest that maybe they are efficiently removed by the mitochondria. To determine whether there were metabolic pathways that could account for the decrease in triglycerides in the knocking mice, we performed metabolomics and correlation analysis, and we found 88 metabolites that negatively correlated with the reduced triglycerides. We then performed enrichment analysis of these 88 metabolites using the small molecule database in Metaboanalyst and found that there were at least three mitochondrial pathways, that is, ammonia recycling, carnitine synthesis, and urea cycle, that were significantly upregulated in knocking mice with NASH compared to wild-type. Indeed, urea cycle was upregulated, as shown by a decrease in blood ammonia in the knocking, increased metabolites of the urea cycle, and increased production of urea. Now, if you look at the urea cycle, arginine can be converted to urea through the action of arginase, but it can also be recycled back to citrulline through the action of nitric oxide synthase. And when nitric oxide synthase is increased, it generates nitric oxide, and nitric oxide leads to three—to a post-translational modification called nitrosylation. So we performed three-nitrotyrosine immunostaining in the livers from our mice, and while wild-type mice have increased three-nitrotyrosine both in hepatocytes and non-parenchymal cells, the knocking mice were significantly protected, suggesting that another mechanism whereby the knocking mice could be protected from NASH is by decreasing nitric oxide pressure. Because arginase is the last enzyme in the urea cycle, we then evaluated the expression of both isoforms. Arginase 1, the cytosolic isoform, was slightly increased in the knocking mice. However, arginase 2, the mitochondrial version, is not only increased in the knocking mice, but is sustained throughout the NASH feeding regimen. We then validated this by immunofluorescence for arginase 2 in the livers of wild-type and knocking mice with NASH, and showed significant induction in the knocking mice. The induction occurred mostly in hepatocytes, as shown here in primary hepatocytes isolated from these mice. Now the literature suggests that global ablation of arginase in mice makes them develop spontaneous steatosis. This is work from Ariel Feldstein. However, range reduction of arginase 2 inhibits steatosis caused by a high-fat diet. We then got a hold of RNA sequencing from total liver from the same clinical trial I alluded to earlier. And here what was striking is that the highest the arginase 2 expression, the lowest is steatosis and the fibrosis scores, suggesting that perhaps arginase does play a role in steatosis, and as a consequence, it could also lower fibrosis. Because knocking mice were showing decreased steatosis, we then looked at fatty acid oxidation, and the first step really is the entry of fatty acid coase into the mitochondrial matrix to then start the beta-oxidation spiral. So we did metabolomics analysis in these mice, and the knocking mice showed an increase in all carnitines, particularly acyl carnitines, suggesting that the carnitine shuttle is effectively working. We then measured fatty acid oxidation in primary hepatocytes isolated from either wild type or knocking mice that were treated with palmitic acid as a representative fatty acid, and we did this by measuring oxygen consumption rate in the seahorse. So under baseline conditions, in knocking mice, there is an increase in basal respiration, maximal respiratory capacity, spare respiratory capacity, proton and proton leak, suggesting increased fatty acid oxidation, and this was further enhanced by treatment with palmitic acid. To make sure that there was a connection between arginase 2 and steatosis, we then targeted arginase 2 with an siRNA in hepatocytes and repeated the measurement of fatty acid oxidation. So in the presence of palmitic, so both hepatocytes are treated with palmitic acid, and what ablation of arginase 2 did was to dampen MRC, SRC, and proton leak. And if you look at the pictures enclosed below, particularly the knocking hepatocytes, so under control conditions, even if treated with palmitic acid, they can store very little amount of lipids, but if you then knock down arginase 2, lipids or oil redo goes back up, suggesting a connection between arginase and steatosis. To determine or identify whether there were candidates signaling molecules from osteopondin high macrophages to hepatocytes that could regulate arginase, and as a consequence, fatty acid oxidation, we perform RNA sequencing of macrophages isolated from wild type and knocking mice with controlled diet and NASH. And the upregulated genes were uploaded into the mouse secretome database. That led us to identify 16 genes, including osteopondin, that were upregulated in the knocking mice with NASH compared to the wild types. Then using Cytoscape, we constructed a regulatory network from the 16 genes to arginase 2, extracting the molecular interactions from IPA to visualize subcellular localization and determine node connectivity. From the 16 secreted factors, we then narrowed it down to six, out of which oncostatin, thrombospondin 1, and osteopondin had the highest connectivity in the network. And we focused on the first one, oncostatin, because it had the highest differential expression. We then evaluated oncostatin, the presence of oncostatin in the culture medium from wild type and knocking macrophages, so it's present there. And then we treated primary hepatocytes from wild type and knocking mice with oncostatin. That led to a significant induction of arginase 2. And then when we treated primary wild type hepatocytes with palmitic acid in the presence of oncostatin, we doubled the MRC, SRC, and also increased proton leak, suggesting that when you increase osteopondin in macrophages, they produce more oncostatin, among other factors, that target hepatocytes to increase arginase 2, and that increases fatty acid oxidation. And I'd like to conclude my presentation with this slide, because multiple people have asked me about the osteopondin high macrophages, if they are involved in other liver disease. And the answer that I can provide is yes. So we have tested it in the context of alcoholic liver disease with the liver-to-kidney model, or in the context of HCC. So they are significantly protected. They develop minimal amount of tumors compared to the wild type and the knocking mice. So in summary, osteopondin high macrophages in the liver secrete multiple factors. And we have identified that at least one of them, we have to test others, can target hepatocytes. On the one hand, there is an increase in deficiency of the urea cycle that lowers the nitric oxide pressure and makes the hepatocytes healthier. And therefore, they make less osteopondin, which could be contributing to protecting from steatosis and fibrosis. But on the other hand, on costatin, induces RGNase 2, which enhances fatty acid oxidation and decreases steatosis, protecting from NASH. Now the key question here is, does this have any therapeutic potential? And of course, we are generating osteopondin knocking mice that are conditional to target macrophages only with a CX3CR1. And they are right now heterozygous, so the experiment will be done. And we can also perform bone marrow transplants. Because we were approved for performing liver transplants in the lab, we did a pilot experiment. So what we did was to get a liver biopsy from a wild type mouse with NASH. It had been on diet only for two months, but it shows enough steatosis and inflammation. And then we transplanted that liver into a knocking mice under the assumption that there will be some repopulation of the liver with macrophages that would have high expression of osteopondin. And we put the mouse on diet for a week with a NASH diet. So they are still under a high-fat diet. And so what we see when we then sacrifice the mouse is that there is a complete resolution of steatosis and inflammation, suggesting that osteopondin-high macrophages are highly protective. And I would like to conclude my talk thanking all my lab for their great efforts, particularly Hui Han. He's the mastermind of the project. Xiaodong Ge for his great immunofluorescence, Sukanta Das for the alcohol model, Zulong Sun who performed the liver transplant, he's sitting right there, Sai Kumakula for the seahorse experiment, he's sitting next to him. And then previous members from the lab for their great support, and also the funding from a UIC pilot award, and also support from our core facilities. Thank you. Okay. Thank you, Natalia. I would like to introduce the second speaker in this session, Ennis Costellari from Mayo Clinic Rochester, and the title of her talk is Glycolysis in Hepatic Stellar Cells Promote Fibrogenic AV Release and Liver Fibrosis. Thank you. Hello, everybody. First, I would like to thank the organizers, our two SIGs, Liver Fibrosis SIG and Cell Biology SIG for this session. So I will talk about glycolysis in hepatic stellar cells and EVs. So I don't have to introduce a lot liver fibrosis in this audience, just a brief introduction. So fibrosis and cirrhosis are leading cause of death in the U.S. It is basically a scarring process, so it's meant to be a good process until it's not. If the injury continues, we have the hepatic stellate cells that start to proliferate and migrate, and we know that the main factor that induces the proliferation and migration of HSCs is PDGF. And then they also produce matrix, high amounts of matrix. The main factor is TGF-beta. Now all this proliferation, migration, and matrix deposition need bioenergy. So they have these HSEs, activated HSEs, they have increased bioenergetic needs. Now regarding this, activated HSEs, it has been shown by the group of Dr. Anamaya Deel in 2012 and recently, more recently, by the group of Dr. Mercedes Fernandez in Spain in 2020. They have shown that these activated HSEs have increased expression of glycolytic genes. Now what is glycolysis? It's a metabolic pathway that produces energy, fast energy, by converting glucose into pyruvate. So these cells that are proliferating and migrating, they need to have ATPs, they need to have energy very fast, very quick. And in this pathway, the first rate-limiting step enzyme is hexokinase. And I will be focused on HK2 because it's the hexokinase that is expressed mostly in HSEs. The question that we asked in our group is, does glycolysis affect the release of fibrogenic signals for liver fibrosis amplification? And included in these fibrogenic signals, we do have extracellular vesicles. And for those that don't know very well extracellular vesicles, I will give a brief introduction. So there are vesicles that are released presumably by any type of cell in the body. They are important for cell-to-cell communication because they are carrying proteins, DNA, mRNA, miRs, and all kinds of molecules in them. Depending on how you isolate them, how you purify them, you can classify them as large or small extracellular vesicles. And we will be focused on the small extracellular vesicles in this talk. Now EVs have been shown to be involved in liver disease pathogenesis. So we have shown previously that activated hepatic stellate cells release fibrogenic extracellular vesicles, or EVs. Injured cholangiocytes, injured hepatocytes, they all release pathogenic extracellular vesicles, and some of them are also pro-inflammatory, especially NASH. So keeping all this in mind, the aim of the present study that I'm showing here is to understand the role of glycolysis on the release of fibrogenic EVs and liver fibrosis. In 2015, we did a very quick experiment. We screened a bunch of growth factors to see which of them induced extracellular vesicle release from hepatic stellate cells. And we ended up with PDGF-BB, or PDGF, that increases this EV release. More recently this year, we performed, us two, like a lot of people in this meeting, we performed seahorse assay to test the glycolysis. So we did a glycolysis stress test. What we did is that we treated our cells with PDGF for one hour, and then we induced glycolysis by treating them with glucose. And as you can see here, there is an increase of the ECAR, or extracellular acidification rate. Now in magenta, you can see that the PDGF here increases more this ECAR than the vesicle, meaning that PDGF induces more glycolysis. This was further increased by oligomycin, which is an oxidation phosphorylation inhibitor. And then it was reduced by 2DG, which is a glycolysis inhibitor as a control. And for this slide, the last piece of data for this slide is that this PDGF-mediated EV release that we have here was abrogated when we inhibited glycolysis by downregulating HK2, so HK2 being the first rate-limiting step enzyme. As a summary for this slide, we can say that PDGF induces glycolysis, which is involved in EV release from hepatic stellate cells. Next question was, how does glycolysis promote EV release? What is the mechanism? For this, we performed a bulk RNA sequencing of hepatic stellate cells treated in the absence no glucose, in the absence of glucose, or in the presence of glucose here. And we ended up with many genes that were differentially regulated. We checked specifically the vesicle trafficking pathway. And we saw that the first gene that was upregulated from this pathway was RAP31. RAP31 is a molecule that is involved in Golgi vesicle trafficking, but not in EV release. So what we did is that we checked the effect of RAP31 on PDGF-mediated EV release. And as you can see here from this graph, the PDGF-mediated EV release here was abrogated when we knocked down RAP31 by sRNA. So from this slide, we can say that PDGF induces glycolysis, which induces the expression of RAP31, and which is involved in EV release. Next question is, are these glycolysis-dependent EVs fibrogenic? For this, we performed a proteomic analysis utilizing tandem mass tag, or TMT, spectrometry. And the conditions of the cells were no glucose, so glucose-free media, glucose or glucose plus 2DG, where 2DG, again, is the inhibitor of glycolysis. And we ended up with more than 1,300 proteins in the EVs, where 98 of them were differentially expressed throughout these three conditions. And in these proteins, we saw that there were fibrogenic proteins, such as TGF-beta-I, collagens, and other proteins. So we think that the glycolysis-dependent EVs are fibrogenic. And the last question that we asked for this talk is, what is the role of HSC-specific glycolysis on liver fibrosis? And for this, we generated a mouse where we deleted HK2, both of the alleles of HK2, selectively in HSCs. We purified, or we isolated, the hepatic stellate cells from control mice, or from mice with HK2 deletion in HSCs. And we checked HK2 mRNA levels, and we saw that there is a decrease. So we confirmed this decrease of HK2 expression. And then we injected these mice with CCL4 to see if there is any effect on fibrosis. Compared to the olive oil control mice, you can see that there is an increase of the serious thread staining in CCL4-treated control HK2-flux-flux mice. But this serious thread is abrogated when HK2 is deleted selectively in HSCs. And this was also confirmed by collagen-1-alpha-1 mRNA expression, where you see an increase of collagen-1 expression in mice treated with CCL4, control mice treated with CCL4, and a decrease of this collagen expression when we delete HK2 selectively in HSCs. So as a summary, we can say that PDGF induces glycolysis, which induces RAB31 expression, which is involved in EV release. These glycolysis-dependent EVs are fibrogenic, and glycolysis is also involved in liver fibrosis. And as key takeaways, there is a novel role for glycolysis in fibrogenic EV release. There is a new mechanism linking metabolism to the regulation of RAB proteins involved in EV release. And of course, all of this participates to our understanding of the pathobiology of liver fibrosis amplification. And if you want to know more about this story, there is a presentation from Dr. Shalil Kanal on Monday, November 7 at noon at the presidential plenary. And with this, I would like to acknowledge the people from my team, Dr. Shalil Kanal here, who has done most of the experiments. I would also like to thank Dr. Vijay Shah, my mentor, and the whole Shah Lab. Also, Dr. Gorss, LaRusso, Mali, Ibrahim, and Kang for their input, valuable input. The GI coffee group, where this is a fun group, but not only. We do science, too, because Dr. Wale Bamidele is the one who is helping us with the seahorse assay. And I would like to thank you, Dr. Michael Simons from Yale, for the HK2 Flux-Flux mice, the Mayo Clinic course, and, of course, all the sponsors, ASLD Foundation for the Pinnacle Award in 2019, Gilead for the Research Scholar Award, and, of course, CSIC and NIH. Thank you for your attention. And with this, I would like to introduce our next speaker, who is Dr. Vijay Shah from Mayo Clinic. He will talk about the sweet side of mechanosensing. Okay. Thank you, Ines. So today, the story I will tell you is about in the endothelial cell and fitting the theme about glycolysis, how glycolysis contributes to endothelial cell function in liver fibrosis. None of these disclosures are relevant. So portal hypertension is a major problem in patients with end-stage liver disease. Of course, it accounts for much of the morbidity and mortality that we see in patients with advanced liver disease. And in terms of the process that's occurring within the liver, there's an imbalance between the endothelial cell, shown here in purple, and the hepatic stellate cell, shown in the yellow. And the endothelial cell, we think, plays a major role in maintaining hepatic stellate cell quiescence, and in turn, angiocrine signaling or changes in homeostatic paracrine release from endothelial cells, in turn, facilitates stellate cell activation in processes of liver injury. So the hepatic sinusoid, with its close proximity of liver endothelial cells and pericyte-like cells, acts as what might be referred to as a microvascular niche involved in early liver fibrogenesis and portal hypertension. So this is a schematic of the hepatic sinusoid. You can see the location of the endothelial cells and the hepatic stellate cells, and the process by which early liver injury could increase liver stiffness due to cell extravasation and edema, and that the endothelial cell is ideally positioned to sense this mechanical change in force, both from the adluminal side as well as the abluminal side. And I won't speak today about mechanical forces within the vasculature, but you can imagine similar processes may be relevant with shear stress and other mechanical forces within the lumen. So the mechanisms by which endothelial cells sense mechanical strain and mechanotransduction occur through integrins and the focal adhesions within which they reside. And in turn, endothelial cells release molecules such as chemokines, and this is referred to as angiocrine signaling or an angiocrine program. And often this can facilitate liver fibrosis, portal hypertension, and inflammation. So the mechanisms by which this occur aren't fully understood, and I will show you some mechanisms that contribute to this process here. So we tested the hypothesis that stiffness induces focal adhesion changes in endothelial cells that may promote a specific angiocrine program, thereby linking mechanical force from edema, et cetera, to inflammation and increases in portal pressure. So the first question we asked was, what proteins might be recruited to focal adhesions in response to stiffness in liver endothelial cells? And then the next question we asked after that was, how do changes in focal adhesions promote an angiocrine program in response to this stiffness? And so a couple of the techniques I'll show you here, we did focal adhesion isolation from cells that were plated on gels. So this would be normal stiffness as well as stiffness that might approximate a cirrhotic liver. And then we also performed focal adhesion isolation by culturing the cells and then essentially removing the cells and allowing the focal adhesions and their associated proteins to maintain their adherence. And then those were subsequently collected and analyzed by Mass Spec. So this is the initial data set showing you proteomic analysis of isolated focal adhesion proteins. And surprisingly to us, what we found was there was a large number of glycolytic proteins that were enriched within these focal adhesions on stiff matrices. And so this is the pathway analysis showing enrichment of glycolytic proteins. And one of them in particular was phosphofructokinase, which similar to hexokinase 2 is also a key mediator within the pathway for glycolysis and is a rate-limiting molecule in that step. So we examined further about PFK and its role in focal adhesions. And we were able to show that, indeed, PFK co-localizes with vinculin, which is a marker of focal adhesions in endothelial cells that are on stiff membranes as opposed to those that are on a more homeostatic membrane. And furthermore, if you disrupt metabolism using pharmacologic compounds such as 3PO, while you're able to disrupt the actin polymerization and focal adhesion formations that you see, thus suggesting a role of metabolism in the actin polymerization process that occurs in response to stiffness. So how do changes in focal adhesions promote an angiocrine program in response to this stiffness? So we wanted to examine what genes are changed in response to stiffness. And then the effects of stiffness-induced glycolysis and how this was occurring. Today I won't speak in great detail about this because of the time, but this is now published and it's in Journal of Hepatology in a recent issue. You can see some of the more detailed analysis. So what was the angiocrine program that was activated in response to stiffness? This program was CXCL1-dominant neutrophil migration pathway, which was identified by RNA sequencing from endothelial cells, again, on a stiff matrix versus on a soft matrix. And again, a number of pathways that were activated, but one of them that piqued our interest was this chemokine pathway, and especially CXCL1. And I will comment that in many of the studies there's actually a group of chemokines that are activated in response to stimuli, whether it's mechanical force or actually other stimuli as well. And I'd refer you to a paper from Mengfei Liu, who's in the back of the room in Nature Communications looking at a number of chemokine pathways in endothelial cells that are upregulated in response to cytokine activation of endothelial cells. So here to link metabolism with this chemokine upregulation, again, we use the compound 3PO to show that the upregulation of CXCL1 could be attenuated by 3PO, which inhibits glycolysis, but on the other hand, molecules which inhibit oxidative phosphorylation don't have a similar effect. And furthermore, if you block actin polymerization, again, with pharmacologic compounds, again, you're able to attenuate the downstream transcriptional regulation of chemokines in endothelial cells. So to summarize this set of data, stiffness activates focal adhesions. It recruits glycolytic enzymes, such as phosphofructokinase, to the focal adhesion. And this stimulates actin polymerization, which leads to epigenetic changes that in turn activate CXCL1 chemokine and other sets of chemokines in that group. And this in turn stimulates inflammation and portal hypertension. So this work reveals new chemokine targets for chronic liver disease with a focus on glycolytic metabolism. So we wrapped up with some in vivo studies in mouse models to explore this further. So one of the models we developed some years ago by Doug Simonetto in the lab was an IVC ligation model in the mouse, which essentially replicates a Bud Kiari syndrome type process and stimulates mechanical forces in the liver characterized by stasis and stretch within the vessels. So using this model, we used a HK2 mouse that was deleted of HK2 within endothelial cells using VCAD CRE mice. And this is the same mouse that Ennis showed, except this is an endothelial cell deletion instead of stellate cell deletion. And what you can see here is that when you induce portal hypertension with this IVC ligation model, you upregulate CXCL chemokines as we showed in vitro. And when glycolytic metabolism is attenuated in these mice by knocking down HK2, that you attenuate CXCL1 production. And you see similar effects on portal pressure as well with increases in portal pressure in this model, the IVC ligation model, but attenuation of that process in the absence of HK2 in endothelial cells. So endothelial cell-specific knockdown of glycolysis attenuates IVC ligation-induced CXCL production and portal hypertension. And then finally, using a second model, well, this is still the IVC ligation model showing you that there's an attenuation of neutrophil infiltration in the model. And then finally, using a carbon tetrachloride model, and in this time, a pharmacologic compound to inhibit metabolism showing you attenuation of liver fibrosis. So this is a summary of the data sets showing you that under normal homeostatic conditions that there's no CXCL1 activation in endothelial cells. However, in response to mechanical forces such as stiffness, that there's a recruitment of glycolytic enzymes to the focal adhesions. And this, in turn, generates energy that's needed for actin polymerization, which in turn allows, and I didn't show you all of this data, but NF kappa B translocation into the nucleus and a number of epigenetic regulatory events that activate CXCL1 production from endothelial cells. So with that, I will conclude and thank especially Thomas Grutter, who's not here. He's back now in Switzerland, but he spearheaded much of this work. Thank you. Our next speaker is Dr. Nissim Hay from the University of Illinois in Chicago and he will talk about Hexokinase 2 is required for liver fibrosis through histone lactilation. Okay, thank you. So first slide, hexokinases catalyze the first step in glucose metabolism, first committed step in glucose metabolism. There are five known hexokinases that are expressed in mammalian cells. They have different level of expression in different tissues and a different affinity. Glucokinase which is exclusively expressed in the liver and pancreas is a high KM hexokinase. Hexokinase 1 and hexokinase 2 are low KM, high affinity hexokinases. Among all these hexokinases, hexokinase 2 has two tandem catalytic domain. Hexokinase 1 is the most ubiquitously expressed in mammalian tissues. Hexokinase 2 is expressed predominantly in adipose tissue and skeletal muscles. However, when normal cells convert to cancer cells, they induce quite dramatically hexokinase 2 level. So for example, in liver cancer, hepatocytes express only glucokinase. However, when they are converted to tumor, they shut down the expression of glucokinase and start expressing hexokinase 2. And if we delete hexokinase 2 in the liver, we markedly diminish hepatocartinogenesis. Same is true for hepatic stellate cells. These are highly proliferating cells and when they are induced, they also induce quite dramatically hexokinase 2. Together with induction of glycolysis. At the level of mRNA, hexokinase 2 is quite dramatically induced. Hexokinase 1 is not significantly induced. And the induction of hexokinase 2 is similar to fibrotic gene induction. If we delete hexokinase 2 in hepatic stellate cells, we decrease dramatically glycolysis and the expression of alpha SMA. We then crossed HK2 flux mice with LRAT3 mice in order to delete hexokinase 2 specifically in stellate cells. And this was shown before. You can see that the deletion of hexokinase 2 inhibits quite dramatically fibrosis and the expression of alpha SMA. This is also shown here by Western blot. So the question is, how does hexokinase 2 affect liver fibrosis? First I would like to introduce you to histone lactylation. The same as histone can be acetylated, it can also be lactylated. And this lactylation of histone opens the chromatin and allows transcription. Indeed, when we activated stellate cells, either by carbon tetrachloride or spontaneously in tissue culture, we get induction of histone lactylation using this antibody specific for histone lactylation in both cases. But no significant change in histone acetylation, maybe a little bit here. Now, when we inhibit LDH that produce lactate in the cells with oxamate, we dramatically reduce the secretion of lactate by the stellate cells and similarly histone lactylation. We also decrease the expression of fibrotic gene, alpha SMA, collagen 1A, and TYM1. However, when we add back lactate to the cells, we recover the expression of the genes, but not when we add acetate. When we delete hexokinase 2 in hepatic stellate cells, we also decrease substantially lactate secretion and histone lactylation. And this can be recovered by adding lactate to the cells. What is shown here is a heat map of histone lactylation at the promoters of genes. These are quiescent stellate cells and these are activated stellate cells. When we delete hexokinase 2, we markedly decrease histone lactylation at the promoters. But when we add back lactate to the knockout cells, we recover this lactylation. So, when we knock out hexokinase 2 in the cells, we can recover, we can decrease the expression of fibrotic genes and this can be recovered by lactate, not by acetate. We also studied the functionality of stellate cells. Apparently, stellate cells, when they are activated, they are able to invade matrigel. And this is the essay here showing that hexokinase 2 deletion prevented the migration. When we add lactate, it recovered this migration, but not when we added acetate. And this is showing here the histone lactylation at the promoters of alpha-SMA, collagen 1A, TMP1, and MMP19. Here at the alpha-SMA promoter, you can see marked increase in histone lactylation. This is quite dramatically decreased upon deletion of hexokinase 2. And when we added lactate, we recovered histone lactylation. Same is true for collagen 1, TMP1, and MMP19. Finally, systemic all-body deletion of hexokinase 2 does not cause adverse physiological consequences. When we delete hexokinase 2 whole body systemically in two-mouse model of fibrosis, carbon tetrachloride, and bile duct ligation, we decrease substantially fibrosis, suggesting that targeting hexokinase 2 is therapeutic for fibrosis. So in conclusion, the induction of gene expression during activation of HSC cells is mediated by histone lactylation. HK2 is induced in HSC cells during liver fibrosis. HK2 is required for liver fibrosis. And hexokinase 2 mediate gene expression of HSC cells during liver fibrosis by histone lactylation. And systemic inhibition of hexokinase 2 could be therapeutic for liver fibrosis. And thank you for your attention. Thank you. And now the panel discussion is open. So if you have any questions, please reach out to the microphones. I have a question for Dr. Castellari. Is there a rationale for why you used olive oil in your experiments? And would you see the same effect with a different type of fat? I think you can use corn oil as well. Olive oil is where we dilute usually in our lab, the CCL4. And that's why it's just a vehicle. But I think I've seen labs using other types of oils. So my question is mainly about the SPP1 story. So in the SPP1 story, I found an interesting finding that a microphage SPP1 upregulation, which is achieved by Norkin, can protect Nash-associated metabolic syndrome, mainly fatty liver. And also in the conclusion slide, I see that a microphage SPP1 upregulation is suggested to reduce liver fibrosis. So actually, you know, OPN is well-known as a fibrosis promoter. So how can we explain the discrepancy between the role of OPN in microphage and the general role of this OPN in promoting the fibrosis? Yeah, so your thinking is correct. And I think it's good that you're bringing up that question. So what we think is happening is that macrophages that are high in osteopontin increase the metabolic rate of hepatocytes. They are able to cope with fat a lot better. They actually do not develop steatosis. They develop minimal because you saw both the primary hepatocytes as well as the in vivo model that there is barely any steatosis. So by improving the healthy state or by making the hepatocytes remain healthy, hepatocytes make less osteopontin. And therefore, the macrophage is overriding the effect of the hepatocyte of fibrosis of hepatocyte-derived osteopontin. So you think it's just a different context, healthy context, and the different context that makes the role of OPN can be different in either inhibiting fibrosis or promoting fibrosis? Well, so it's different cellular sources. Obviously, the osteopontin that is being made by each one of them is still to be defined. Personally, I think that the osteopontin that is being made by hepatocytes, it's clipped in a way that kind of protects the RGD sequence, and therefore, it can bind integrins, whereas the one that is being made by the macrophage may be highly clipped, but also highly phosphorylated. So that's the type of protein. But what I think is happening here is that the effect of the macrophage is so profound in the hepatocyte that by way of maintaining a hepatocyte healthy, you prevent osteoptosis, and therefore, the cascade that will lead towards fibrosis. We have also constructed the regulatory network from the 16 genes that we identified in the RNA sequencing towards collagen, and we have a couple of candidates that we are going to be testing. So there is a possibility that the macrophage has a direct effect on improving the healthy state of the hepatocyte, but also that these releasing factors that are also improving the state of the cell itself, and therefore, they are making less collagen. So you are considering that the OSPP1 when expressed in macrophage, it can be, how to see, can be clipped into a different form in macrophage compared to hepatocyte or hepatic stellate cell? Correct. So it's cell type specific, and we are not only seeing this in the liver. We are seeing this in other organs as well. So the effects of osteopontin in any disease are cell specific and organ specific. So how much molecular weight you have found with this OPN in Western blood? So here's the thing for everyone who works in the osteopontin space, I mean, it's not a protein that you could really pinpoint very well in a Western blood because it's heavily clipped. So the extent of cleavage varies from context to context. So in the case of hepatocytes, we typically see like anywhere between two, three bands. In the case of the macrophage, we see more. So it's possible that it's highly clipped, and it's clipped in the RGD site, and therefore, it would not activate the stellate cells because they have an integrin binding site. Okay. Thank you. And another question. I apologize. I think we need to let a little bit of space for others. Yeah, we can talk later. Never mind. Yeah. Maybe you can continue later. Thank you. Thank you. All right. Dominic Helen, ME University. I had a question about lactylation. So what happens to H3K27 acetylation at the same sites that are lactylated? Right. There might be a competition between acetylation and lactylation. And actually, during fibrosis, HDAC are induced and deacetylate histones. And is K18 the only place where you see lactylation? I thought 23 and 27 as well. Yeah. You can see in other histones. The antibodies are not very good for other histones. But this one is good, so we show this. Yeah. Interesting. Thanks. Thank you for a great presentation. My question is, I guess for any of the last three speakers, you show us the importance of glycolysis in HEC cells activation or in mechanosensation on the endothelial cells. I understand that HK2 is the rate limiting step, but I'm just wondering if you have seen increase in the transport of glucose, GLUT1, GLUT2, or even relocalization inside the cell to the focal adhesion areas or the plasma membrane. Right. Actually glucose transporters are not rate limiting because glucose transporters in principle could be reversible unless glucose is phosphorylated. So the phosphorylation of glucose, in my opinion, is much more important than the glucose transporter, but GLUT1 is also induced. We haven't done the experiments ourselves, but Ningling Kang at University of Minnesota, I know he's working on that, looking at that. Robert Schwabe, Columbia University. Very nice presentation. So I have two questions. One, I wanted to follow up on the lactylation and what is the half-life of the lactylation and do you think stellate cells could also be just metabolically unfit so that it's not all the lactylation but some direct metabolic effects that you see in the HK2 knockout? Correct, but it appears that high level of glycolysis is required and as we showed the expression of genes is requiring glycolysis due to lactate. Now lactate was predicted to play a role in fibrosis before, but nobody knew what is the mechanism and here we show that it very well could be through histone lactylation. Now there are more into that because you can imagine if other cells in the liver secrete lactate, the expression of genes may not be dependent on exokinase 2 in the fibrotic cells because lactate can compensate for the loss of exokinase 2. Okay, so my second question is to Vijay Shah. So everything you had pinpointed towards CXCL1 and then you switched to the therapy. So I was missing two parts. Maybe these are tough questions, but that's what we're here for. One is, do you think the endothelial cell is the main source of CXCL1 or if not, you know, other cells would make it and then why don't you have portal hypertension? And my second question is, have you tried to actually go in with blocking, you know, with a let's say receptor agonist for CXCL2 or something like that and do you see decreased portal hypertension? Yeah, so those are both great questions. So the cell source of CXCL1 may be animal specific or species specific. Certainly in some of the mouse models, hepatocytes are making a lot of chemokines as well. And certainly in humans, we see a lot of CXCL1 production from endothelial cells when you look at single cell sequencing from human endothelial cells. The other point with CXCL1 is probably being regulated with groups of genes. So not just CXCL1, but several other chemokines that are commonly regulated. And the other point is immune cells are going to make a lot of CXCL1. So we've tried to focus on the early steps of the process before you have a lot of immune cell infiltration because certainly once that's happening, you'll have many targets of CXCL1. In terms of compounds, probably in humans, you know there's CCR2 inhibitors. There is also senacrivirac. So some of these are going to be inhibiting this pathway, although it's hard to say in humans which cell is the regulator. Hello, nice presentation. So this is Ramana from NIH. So my question is for Ms. Costolari regarding the EVs. So you said you focused mainly the small EVs for this project. Can you please tell us what's the range, size of these EVs, how small are, is there any range or they're all the same? So there is a range. We were focused on small EVs because this is where we found differences. We did not find differences for the larger EVs. And the range of the small EVs is about, let's say, 80 to 300 nanometers. For sure they include exosomes and microvesicles, but it's more difficult to distinguish based on the biogenesis than based on the size. Thank you. Okay, so thank you for attending. Thanks to all the speakers this morning and this afternoon. Thanks to the organizers, both SIGs, Cell Biology and Fibrosis SIG. Please become a member. They are great SIGs, the greatest SIGs of ASLD. And we'll see you in the next sessions. Bye.

fibrosis

osteopontin

macrophages

NASH

liver injury

lipid metabolism

glycolysis

hepatic stellate cells

extracellular vesicle

histone lactylation