My name is Neil Henderson, and I'm a hepatologist based in Edinburgh, and it gives me great pleasure to give you a warm welcome to our liver fibrosis SIG on NASH fibrosis from Matrix to Medicine. I'd like to remind attendees to please feel free to ask questions in the chat area, as it would be great to build in some Q&A to our session. Liver fibrosis continues to be a very vibrant area of research and development in hepatology, and I'm delighted that we have an absolutely stellar lineup of speakers today. A really exciting range of topics will be covered in this session, spanning all the way from the cellular and molecular mechanisms in NASH fibrosis, to the cutting edge developments in translation of putative targets, development of biomarkers, and all of which are geared towards enabling the discovery of new therapies for NASH fibrosis. In terms of our speaker lineup, as I mentioned earlier, it's truly stellar, which is fantastic, Annamaye Deol will lead off with a talk on the molecular pathways driving fibrosis in NASH. This will be followed by Frank Tack, discussing the immune-mediated regulation of the fibrotic response in NASH. Buxiang Cai will then discuss lipid metabolism and hepatocyte death in the context of NASH fibrosis. And then Scott Friedman will give a talk about, is this the end of the beginning in NASH, and what's next for fibrosis in NASH? Then finally, Ekihiro Seki will discuss NASH fibrosis from biomarkers to treatment. So I hope you enjoy this session, and I'd like to thank all our speakers in advance for contributing such an exciting series of talks. Thank you. Good morning, this is Annamaye Deol, and I'm happy to talk to you today about molecular pathways for NASH fibrosis. I have no conflicts to disclose. Most people who take care of patients with NAFLD subdivide them into two different groups because the diseases have different prognoses. We know that people who have simple fatty liver rarely develop cirrhosis or liver cancer. In contrast, individuals with NASH are at increased risk for both conditions. It's unclear why some people with NAFLD get NASH and why only some people with NASH progress to more advanced stages of liver disease. And this has excited quite a bit of interest in the research community. My group has been pursuing the idea that the big difference here is whether or not hepatocytes are dying. That happens more in NASH than it does in NAFLD. And we think that might be important because we know that dying cells trigger wound healing responses in order to replace the cells that died. These wound healing responses are really very multifaceted. They're triggered by paracrine signals that are released from the dying cells. And the idea is to create a nurturing microenvironment that will promote the outgrowth of liver progenitors by recruiting immune cells to the liver, regulating matrix remodeling and controlling the size of myofibroblast populations, and stimulating vascular remodeling. Now, all of that is good as long as it's appropriately controlled. However, when it becomes excessive, all of these good things can turn into bad things. And this raises the possibility that fibrosis could actually be considered a biomarker for dysregulated repair and eventual organ failure. In fact, this idea was supported by natural history studies done in NAFLD, which showed that liver fibrosis severity is an independent predictor of liver-specific morbidity, mortality, need for liver transplant, and even all-cause mortality in NAFLD. Now, in the fibrosis scene, we're well aware that myofibroblasts are the major source of a fibrous matrix that accumulates and leads to cirrhosis, and that one of the major producers of these myofibroblasts is the hepatic stellate cell, a tissue-resident parasite. In order to try to understand better what signals might be enriching the liver with myofibroblasts, we took the approach of doing microarray analysis on a well-defined cohort of humans with non-alcoholic fatty liver disease. And we selected people who had very mild NAFLD as defined by stage F0 to 1 fibrosis, or individuals with more severe NAFLD, shown by F3 to F4 fibrosis. The heat map here clearly shows that these groups have very different gene expression profiles, and we were able to generate a signature that independently correlated with NAFLD severity. More interestingly, though, is when we drilled down into the gene expression signatures and tried to look for functional pathways, the top things that popped out were cancer, and the top ingenuity pathways were tissue remodeling, regeneration, wound healing, regenerative cells, and development. I will say that other groups have subsequently validated this with their own microarrays. And in fact, many people use our microarrays to justify the relevance of their animal models. So for example, Scott Friedman recently published a paper showing that ATG7 knockout mice, which spontaneously develop NASH and severe fibrosis and cancer, are very similar to the patients that we studied who have advanced fibrosis. So all of this led us to the following hypotheses. First, we said that pathways that regulate cell plasticity to accomplish organogenesis during development may reactivate to regenerate the liver during NASH. But dysregulation of these pathways promotes accumulation of cell types like myofibroblasts and progenitors that are necessary but not sufficient for effective regeneration and recovery. And instead, persistent growth of the repair-related cell populations actually impedes effective repair and promotes the development of cirrhosis and liver cancer. For the sake of time, I'm going to focus my discussion on two pathways we've studied the most, hedgehog and hippo. And let me cut to the chase quickly. What I'm going to tell you is that hedgehog and hippo yap interact to regulate liver repair. All liver resident cells have both pathways. The cells control pathway activity via autocrine and paracrine transplants that engage other morphogenic pathways like TGF-beta, NOTCH, WNT. And importantly, signaling sensitivity is modulated by inflammatory cytokines like TNF and IL-1-beta, as well as nutrients that are important in NAFLD like cholesterol, lipids, and glucose. The bottom line though is that any combination of things that result in activation of hedgehog or silencing hippo to activate yap lead to an oncofetal state and promote fibrosis and cancer. So now let me quickly show you some of the evidence that led to these conclusions. First, let me remind you about the hedgehog pathway. This is a pathway that regulates fate decisions in stem and progenitor cells. It's oncofetal, meaning it's active in development, downregulating at birth, reactivates in main cancers. We and others have shown that this pathway is activated in multiple liver diseases before cancer is developed. It promotes the outgrowth of myelofibroblasts and therefore is important for driving fibrosis progression in NASH. And it promotes the outgrowth of immature proliferative liver epithelial cells that give rise to the ductular reaction and also promote degeneration of hepatic adenomas and hepatocellular cancers. This is a schema of the hedgehog pathway for the sake of this discussion, just focus on the stuff on the right. Basically, the pathway is off until the receptor patched and engages hedgehog ligands. That interaction liberates another signaling molecule called smoothen from the inhibitory actions of patch leads to the activation of Glea family transcription factors and the transcription hedgehog target genes. A number of years ago, we were fortunate to be able to study patched Lax-C mice and these mice cells with active hedgehog activity are blue. We perfused the livers of patched Lax-C mice and much to our surprise found that the non-parenchymal cell fraction, including the stellate cell fraction, had blue cells. In order to make sure that this wasn't an artifact due to cell contamination, we quickly turned to a well-studied human myelofibroblastic stellate cell line, LX2 cells that came from Scott Friedman's lab. And we manipulated hedgehog pathway activity by treating the cells with either an inactive, smoothened inhibitor, tomavidine, or cyclopamine, which inhibits smoothen. And you can see that inhibiting the hedgehog signaling pathway in LX2 cells reduced Glea and prevented the cells from expressing a myelofibroblast marker, even though they were still viable. This led us to wonder if hedgehog signaling might regulate state transitions in stellate cells. In order to study this more definitively, we isolated stellate cells from the LX2 cells and stellate cells from mice and got highly pure populations. And when we took the stellate cells from smoothened Flux-Flux-Flux and treated them in culture with an irrelevant adenovirus as a control, we saw that compared to time zero, shown in gray, the cells that had been cultured for seven days had activated smoothened and Glea, indicating hedgehog signaling was turned on. They had become myelofibroblasts, high collagen and alpha-small expression. They had up-regulated a gene called SNAIL that regulates EMT and down-regulated some epithelial markers like adherin and desmoplakin. And of course, they were less coagulant PPRBM and LRAT were decreased. If we did the same experiment but took cells from the smooth Flux-Flux mice and treated them with an adenovirus-bearing CRE to delete smooth, we blocked the accumulation of smoothen. The hedgehog pathway did not activate as elevated as evidenced by reduced Glea. The cells reverted to quiescence, PPR gamma came up, LRAT came up, they became less myelofibroblastic and they regained their expression of epithelial markers. So we concluded that hedgehog is one of the things that regulates state transitions in stellate cells. So the question was, how does that happen? We performed microarray analysis of stellate cells that were freshly isolated or activated by culture and were surprised to learn that the most up-regulated pathways involved metabolism, leading us to speculate that transdifferentiation might be characterized by metabolic reprogramming. The first pathway we studied was glycolysis because there was clear evidence from our arrays that glycolytic gene activity was increased when myelofibroblasts transitioned from quiescence to a myelofibroblastic state. This seemed to be relatively specific because other metabolic pathways were suppressed like lipogenesis and gluconeogenesis. And the activation of glycolysis led the stellate cells to have a Warburg-like effect in which they actually generated and released lactate. To determine whether glycolysis was regulated by the hedgehog pathway, we treated the cells with various inhibitors all along the trajectory. And we were able to demonstrate that similar to inhibiting glycolysis directly by blocking HIF-1-alpha or using 2-deoxyglucose or LDHA inhibitors, we could get the same inhibition of glycolysis by neutralizing antibodies to hedgehog ligands or treating with pharmacologic or genetic approaches that inactivated smoothing. Now, this was an interesting finding to us because if hedgehog increased glycolysis, it made us ask, did it also have an effect on YAP and TAZ and glutaminolysis? Because in cancer cells, glycolysis actually stimulates both YAP and TAZ and promotes glutaminolysis. DK in our group published a paper in Gastro confirming that yes, indeed, hedgehog does activate YAP and increase glutaminolytic activity in stellate cells. He went on to show that this glutamate that's generated by glutamine breakdown gives rise to alpha-ketoglutarate that goes through the TCA cycle and fuels stellate cell transdifferentiation, proliferation, migration, and fibrogenesis. So the bottom line of all of this is that hedgehog seems to induce cancer-like metabolism in stellate cells. And that was reassuring because several years earlier, we had noticed that there was a co-localization of GLUT2-positive cells with the alpha-small-positive cells in cirrhotic livers. So we wanted to ask what activates hedgehog signaling in liver disease. And we and others have shown now that hedgehog ligands, sonic, Indian, and desert are produced by all types of resident liver cells, hepatocytes, ductal cells, epithelial cells, stellate cells, immune cells. In healthy livers, the level of synthesis is low and the ligands are released in exosomes or from hepatocytes and lipoproteins. In these structures, they're associated with other cargo that modulate hedgehog signaling and they're antagonized by locally abundant soluble hedgehog inhibitor called HIP. So hedgehog pathway is generally off in healthy livers. However, when liver cells become stressed, they dramatically upregulate their synthesis of hedgehog ligands. We now know that this is due to AP1 and TAS. And moreover, in addition to releasing them in exosomes and lipoproteins, as the cells die, the hedgehog ligands come out in microparticles and membrane blebs, and the associated signaling modulators are probably altered. And the net result is that you get an increase in local accumulation and activity of the hedgehog ligands. Let me show you now a couple of slides that we took from patients who were in the PIVENS trial, a study where patients with NASH were randomized to be a glutasone vitamin E or placebo. The top slide shows an immunostaining of the liver biopsy in one patient. Here, the green arrows are pointing to the clear cells that are ballooned hepatocytes, and the brown is cytokeratin 8 and 18. They're ubiquitinated and degraded, but retained in liver cells because of ER stress, and these are ballooned hepatocytes. The bottom slide shows the same slide, restain for sonic hedgehog ligand, and you'll notice that the balloon cells are packed full of hedgehog ligand. Interestingly, this ballooning and cell injury goes away with treatment, and along with that, the hedgehog ligands disappear. We thought this might be relevant because in the pretreatment livers, these ballooned hepatocytes are surrounded by hedgehog-responsive myofibroblast, shown in the brown cells with red nuclei, and when the hedgehog ligands go away with treatment, the hedgehog-responsive liver cells also disappear. So although that's correlative evidence, which suggests that hedgehog sigmarine might cause NASH or NASH fibrosis, really you need to turn to animal studies to prove that, and many other groups have done that now. This is a nice paper from Wang et al that was published in Cell Metabolism in 2016. They used various animal models of NASH and a combination of transgenic and pharmacologic approaches to prove that stressed hepatocytes activate TAS to produce Indian hedgehog, which then recruits inflammatory cells and promotes fibrosis in animal models with NASH, and that Indian hedgehog was necessary for this to occur because knocking out or inhibiting Indian hedgehog prevented the process. The same group has more recently published another paper where they have shown that targeting siRNA to TAS, to hepatocytes, has the same effect. Now, interestingly, although this talk focuses on fibrosis, I would be remiss if we didn't acknowledge that all of these studies have also shown that modulating hedgehog modulates the accumulation of immune cells in liver progenitors, and that may be because hedgehog actually regulates liver progenitor fate. A number of years ago, we showed that it increases the accumulation of immature ductal cells and that these ductal cells produce immune chemoattractants. We call this repair-related inflammation. It's now known that immune and immature ductal-type cells accumulate in NAPL. We call that the ductular reaction, and it parallels the severity of a hepatocyte injury in fibrosis to things which are associated with increased hedgehog, and this injury in fibrosis are also associated with increased cancer risk. It should be mentioned that YAP is known to cause hepatocytes to de-differentiate in proliferative ductal-appearing progenitors, and eventually this can promote cancer, and I showed you that hedgehog activates YAP in stellate cells, where we have also shown that it activates YAP in hepatocytes. To remind you, the HIPPO pathway is also an oncofetal pathway. It regulates fate decisions in stem and progenitor cells, but in this case, the pathway is inactive on development, and that enables pluripotency factors, YAP and TAS, to accumulate. The pathway is upregulated at birth, so that YAP and TAS are silenced, and in liver cancers, the pathway goes off, and YAP and TAS come up. During normal liver regeneration, this pathway is also regulated. The hippokinases are shut down, so that YAP and TAS come up, and the liver grows. If this is maintained, eventually the liver's developed cancer. I just showed you that YAP and TAS in epithelial cells promote the synthesis of hedgehog ligands and cause myofibroblastic accumulation and fibrosis in NASH, and now I'm going to show you some slides which demonstrate that YAP activity parallels the severity of the ductrular reaction in NASH. This is slides from representative humans with varying levels of NASH severity as you go from left to right, low on the left, high on the right, and what you can see is that as NASH gets worse, you get more accumulation of hedgehog ligands. The hedgehog ligands promote ductular cells, so you get more keratin-19-positive cells, and if you look carefully, these keratin-19-positive ductular cells are expressing YAP. In fact, there's a very nice correlation between the keratin-19 cell morphometry and the YAP morphometry. Fibrosis severity parallels YAP accumulation in humans with NASH, as shown on this slide. People with normal livers have little fibrosis, little YAP, and those with F2-F4 fibrosis have more YAP and more fibrosis. Of course, the same thing can be demonstrated in animal models. Again, as you go from left to right, less severe to more severe disease, you see an increase in YAP, progressive increase in keratin-19, increase in serious red staining and hydroxyproline, and there are very strong correlations between the increases in all of those factors, and YAP immunohistochemistry with R values that are between 0.8 and 0.9. So in summary, pathways that regulate cell plasticity to accomplish organogenesis during development do reactivate to regenerate the liver during NASH, but dysregulation of these pathways promotes accumulation of cell types that are necessary but not sufficient for effective regeneration and recovery. Importantly, myofibroblasts cause fibrosis and hepatocyte progenitors that interact with myofibroblasts to promote inflammation and cancer. Persistent growth of these repair-related cells therefore impedes effective regeneration and actually promotes bad outcomes. So why does the liver repair process become dysregulated in NASH? I think what we've seen is that diseased livers are enriched with permissive signals for developing livers. We talked about hedgehog being up, YAP being up, and they're enriched with these immature liver signatures because pathways that repress the immature state, like HIPPO, are suppressed. So is it possible that dysregulated regeneration is actually a failure to restore repression of the fetal state? And if so, why might that occur? Well, a couple of years ago, well, Noosa's lab actually published a very important paper demonstrating that an inflammatory cytokine, TNF-alpha, was sufficient to drive normal hepatocytes to acquire stem-like characteristics. If you took normal hepatocytes from a healthy animal, they couldn't grow as spheroids under anchorage-independent growth conditions. But simply adding TNF-alpha was sufficient to trigger the cells to become more fetal-like. They up-regulated YAP. They became less mature. They down-regulated their expression of albumin, aldolase, and other metabolic enzymes. And they actually also became more pro-inflammatory. Jenna Nguyen from our lab showed why this might have happened. She found that liver inflammation removes a break that normally constrains YAP test activation. She showed that TNF-alpha and IL-1 beta actually inhibit an RNA splicing factor called ESRP2. This is a very important splicing factor. It controls adult splicing in about 20% of hepatocyte genes. It's low in fetuses, high in adult hepatocytes. It does this by activating the most upstream hippokinase called NF2. And so when hippokinase signaling is up, because NF2 is active, YAP and TAS activity go down, and the hepatocytes remain mature and non-proliferative. However, in inflamed livers, or if you treat hepatocytes directly with TNF-alpha, IL-1 beta, ESRP2 is suppressed, NF2 becomes inactive, HIV signaling turns off, YAP and TAS become activated, and hepatocytes do differentiate and proliferate. We think this may have broad implications because NAFLD is the liver correlate of the metabolic syndrome. And we know that the metabolic syndrome is a chronic inflammatory state. It's characterized by dysregulated cellular and systemic energy homeostasis. And it's been known for years that TNF mediates insulin resistance. The metabolic syndrome leads to defective inter-organ communication by the liver, adipose tissue, muscle, pancreas, brain. And hedgehog ligands are actually produced in the liver and carried in lipoproteins to modulate metabolism in hedgehog-responsive parenchymal and stromal cells. And recall that morbidity and mortality in the metabolic syndrome are caused by progressive organ failure or cancer, things that we know occur when repair becomes defective. So this sort of raises the possibility of a chicken-egg type process. If you have the metabolic syndrome, you may have defective repair. If you have defective repair, you may develop the metabolic syndrome. And perhaps that contributes to all of these bad outcomes of metabolic syndrome associated diseases. I've given you a very hedgehog-centric world view. I would be remiss if I didn't acknowledge work from many other groups, including some very recent publications from people that I know in the audience, which have shown the importance of TGF beta, hyaluronan, CD44, osteopontin, NOTCH, signaling, FOXM1, either in stellate cells or hepatocytes. All of these things, when activated, can cause liver injury, inflammation, fibrosis, and cancer. But again, I can't prevent myself from reminding you that all of these are downstream targets of the hippo, yak, and hedgehog signaling pathways. So I'll stop here and be happy to take questions. Thank you. Good afternoon, everybody, dear colleagues, dear friends. My name is Frank Tucker. I'm the head of the Department of Gastroenterology and Hepatology at the Charité University Medicine in Berlin. It's a great pleasure to discuss with you at this sick liver fibrosis meeting the immune system regulation of fibrotic response in the context of NASH. These are my disclosures. We have received funding in the lab from some pharmaceutical companies to work on NASH and fibrosis. If you think about mechanisms of fibrosis progression in NASH, it's, of course, very important to consider that an apparent metabolism or lipotoxicity will drive an inflammatory response, which then would lead to the activation of hepatic stellate cells and subsequent liver fibrosis. However, we also know that inflammation has implications for metabolism, and some of the signals generated during metabolic processes are actually pro-fibrogenic themselves. One important cell type that is able to integrate all these different signals are macrophages, especially macrophages in the liver. We understand that macrophages are localized in the liver sinusoids and are able to receive signals not only from the liver or hepatocytes or fat-laden hepatocytes, but also from the circulation, particularly from the gut liver axis or from adipose tissue. It's important to understand that macrophages are not one population of cells. We distinguish different subpopulations of macrophages, for instance, the Kupfer cells that are localized at the sinusoidal endothelium to sense danger signals or tissue injury, and that could then provoke the infiltration of monocyte-ref macrophages from the bone marrow or the circulation that are recruited and very responsive to signals. For instance, if we analyze a mouse model of acute liver injury after acetaminophen poisoning, we can sort different populations of macrophages. And if you look at the monocyte-derived macrophages, compare them to Kupfer cells or other monocytes, then you understand that they have many, many pattern recognition receptors and polarization markers, making them very, very responsive to injury signals. The same holds true also in NASH models. For instance, here, the methionine choline-deficient model of NASH in mice after eight weeks, if we sort from the liver Kupfer cells or monocyte-derived macrophages, there are several genes differentially regulated. For instance, in Kupfer cells, markers that would be associated with innate immunity all would directly contribute to the metabolism of lipids, whereas in monocyte-derived macrophages, we have a lot of markers associated with fibrosis and also with angiogenesis. Of course, nowadays, to apprehend the heterogeneity of cell populations in an organ is not only done by FACS or sorting or RNA sequencing. We nowadays do single-cell RNA sequencing, meaning that we understand the transcriptomic regulation at a single-cell level. And that led to really new and fresh insights into macrophage heterogeneity in the liver. For instance, some examples of very recent work demonstrated then that in a lipid-loaden environment of the liver, the Kupfer cells cannot really self-renew. So, they die over time and are being replaced by inflammatory monocyte-derived macrophages that acquire a Kupfer cell-like phenotype, but are a little bit different and more inflammatory. On the other hand, we understand that these monocyte-derived macrophages have a phenotype that some researchers call lipid-associated because they express high levels of osteopontin, have a particular activation phenotype, and are also able to process lipids. Moreover, understanding macrophage heterogeneity at a deeper level also shows us that a lot of epigenetic reprogramming is taking place in NASH. For instance, the Kupfer cells undergo some epigenetic changes that decreases their survival time, whereas the macrophages or monocyte-derived macrophages acquire a scar-associated phenotype. And these are just some three recent papers in immunity dealing with liver macrophages in the context of NASH. So, you understand this is a very hot topic. On the other hand, there are many, many more single-cell RNA sequencing analysis in the context of NASH or HCC that provide really deep insight, particularly in immune cell heterogeneity in the liver. We have also contributed or had some projects going on in this area looking at single-cell RNA sequencing, particularly from liver macrophages or myeloid cells. And if we sort from mouse liver only the macrophage population, meaning we isolate CD45 positive cells that are negative for lymphocyte markers and negative for neutrophil markers, we still get a high heterogeneity of cell subpopulations from which some are monocyte-derived macrophages, here abbreviated MoMF, some are Kupfer cells, and some are dendritic cell markers. In the context of a Western diet model, meaning a NASH model after 16 weeks of feeding the mice a Western diet, we do see drastic changes in the liver. So, a particular enrichment in the monocyte-derived macrophage population. And we do understand that the monocyte-derived macrophages consist of several subpopulations. And in the comparison between normal diet and Western diet, it's very apparent that some markers are up and some are downregulated, particularly S100A8, A9, a marker of inflammation, is rather downregulated in the Western diet condition. These changes do not only happen in the liver, they also affect the bone marrow. And we understand the differential regulation of genes that is very much conserved between liver and bone marrow in the myeloid cell compartment. And we do see a NAFLD-associated myeloid phenotype in liver and bone marrow that is relatively stable and determines then the functional responses to injury. We're currently working on expanding our view by looking at different liver injury models at the same time, Western diet compared to simple fibrosis, compared to CDAA high-fat diet, which is a NASH model with really advanced fibrosis. And if we just look at those three dietary models, CHO diet, CDAA high-fat diet, and Western diet, we do see kind of changes in the macrophage compartment that are pretty much related to the severity of injuries, so very pronounced in CDAA high-fat diet model. And if we look at particular populations of monocyte-referred macrophages, we do see a pronounced phenotype in the most or more advanced NASH model. This is not only relevant for mouse models, also relevant for human disease. Here's a beautiful paper by Neil Henderson, who chairs this session on single-cell RNA sequencing from a cirrhotic human liver, where he could, with his team, identify several immune cell populations being regulated in cirrhotic versus healthy liver. And one particular aspect that I found very interesting is the identification of a scar-associated macrophage that was characterized by TRM2 expression, probably of monocytic origin, not necessarily profibrogenic. And I think more work is going on at the moment to understand this. What we are trying is also to understand the spatial distribution of the different macrophage subsets. So, what we established in our lab is a multi-parametric histological analysis with something between 15 and 20 markers for immunohistochemistry at the same time. And if we do analyze control mouse liver, liver fibrosis, and the NASH model, the CDA high-fat diet model, you see here the Kupfer cells, IBA1 positive, CLEK4F positive in orange, and you see monocyte-derived macrophages that are CLEK4F negative in red. A really strangely altered distribution of the monocyte-derived macrophages versus Kupfer cells in fibrosis, but particularly in NASH, where you really see a depletion of the Kupfer cells. And this can also be observed in human NASH. The beauty of our immunohistochemistry multiplex staining is that you can also apply it to human tissue. And in human NASH, you do see much less Kupfer cells and more monocyte-derived macrophages, particularly around the fibrotic areas. And this is what we believe is very crucial, that the recruited monocytes are actually more fibrogenic, are able to activate stellate cells. And this has also therapeutic implications because you could think about either repolarizing the macrophage compartment to a less fibrogenic phenotype or block the recruitment of monocytes into the liver. And this can be done, for instance, by Sinicreviroc, an orally available dual CCR2-CCR5 inhibitor. And I think we have really very convincing evidence from many, many rodent models of different types of liver disease that blocking the recruitment of monocytes into the liver has a very strong anti-fibrotic effect. And there are some indications that this can translate also in benefits for humans. These are data from a phase 2 clinical trial with almost 300 patients being biopsied. After one year of treatment with either placebo or Sinicreviroc, we see a significant proportion of patients improving fibrosis by more than once or by one stage or more in the Sinicreviroc-treated arm. After two years, the data are not as clear, but if you look at fibrosis improvement by two stages or more, there seems to be a trend that more patients benefit from CBC than from placebo. So how do these concepts fit in the current landscape of drugs and the development for Nafld and NASH? I think there are many drugs that try to revert the pathogenic liver metabolism by, for instance, targeting de novo lipogenesis or lipotoxicity or mimicking beneficial effects of bile acids, whereas other drugs, for instance, the CCR2-5 inhibitor Sinicreviroc, aim at disconnecting Nafld from inflammation and fibrosis, meaning that you still have lipotoxicity, ER stress, steatosis going on, but this would not lead to inflammatory responses and to the fibrogenic activation of stellate cells. However, I think if we want to have a sustainable beneficial effect for Nafld patients, we probably need to combine different pathways at the same time, and this would have implications for metabolism and inflammation and fibrosis at the same time. One key example that we studied in our lab is the PPAR agonism. There are specific agonists for PPR-alpha, delta, or gamma isoforms that are listed here in this experiment, and then there are PPAR agonists like linofibranol that target all three isoforms at the same time. And as you can appreciate here from our study, the pan-PPAR agonist was much more effective than any single agonist. And if we look at the macrophage compartment, we could appreciate that the combined activity of different PPAR isoforms was actually most helpful to reduce the activation of monocytopoietic macrophages in the liver. When we further dissected the different aspects of PPAR agonism in our liver disease model, it was very clear that the different isoforms acted on different cells. So the PPAR-alpha agonism was actually very important to reduce steatosis, whereas the delta agonism was more active in the macrophages, and the gamma agonism was very active in the stellate cells in reducing either steatosis, inflammation, and fibrosis. Possibly, you can also combine different types of drugs to have the most therapeutic benefit. And this is a mouse experiment that we presented at the digital ILC meeting from ESIL this year, in August, where we treated mice in an FLD model, the CDAA hybrid model that were subjected for 12 weeks, and we treated for six weeks in a therapeutic intention, either with an FGF21 analog or with a CCR25 inhibitor. And you can see here that FGF21 had a strong effect on steatosis and steatohepatitis, whereas the CCR25 inhibitor dramatically reduced liver fibrosis. And when we combined those two drugs, it actually dramatically improved the liver phenotype because it preserved the efficacy of FGF21 on steatosis, steatohepatitis, and combined it with the anti-fibrotic effect of anti-CCR25. And I think this is the direction that we should aim at when we try to understand what the immune response in the context of NASH-induced liver fibrosis means and how we can modulate that. And with this, I would like to thank the people in my lab and all the collaborators in Berlin and Aachen, our prior place, and the collaborators worldwide for very fruitful projects, very interesting on the interface of immunity, inflammation, and fibrosis, particularly in fatty liver disease. And I would like to summarize the immune system regulation of fibrotic response in the context of NASH. I think it's very important to see that inflammation and immune responses would drive liver fibrosis, but it's also important in the resolution of NASH. The fatty microenvironment of the liver has major implications, not only for the resident immune cells, such as the Kupfer cells, but also for the recruitment of inflammatory cells in their subsequent phenotype. Regarding the macrophages, we have different subsets in the liver that have differential effects on inflammation, injuries, phthalate cell activation, but also resolution of fibrosis. And this is true in mouse models as well as in human disease. I did not talk much about other immune cells, but they are also important and probably subject of ongoing studies in many labs, for instance, platelets, lymphocytes, dendritic cells. New therapeutic concepts attempt to revert metabolism or disconnect NAFLD from inflammation of fibrosis. And I think it's most efficacious if we combine those different concepts, either in one drugs or in combination therapies. The lines of research in the field is, I think, single cell RNA sequencing, spatial transcriptome imaging to really dissect the different function of immune cells in the context of the disease and translate that into biomarkers and therapy. With this, I would like to thank you and I'm open and happy to receive your questions. Hi, everyone. Good afternoon. I'm Bizhuang Cai. I'm an assistant professor from Icahn School of Medicine at Mount Sinai. Thank you, AASLD organizers, for this great opportunity. It's really my great honor to present our studies here. Today, I'm going to talk about hepatocyte cholesterol test pathway promotes non-alcoholic state of hepatitis. These studies were done by my previous colleague, Dr. Xiaobo Wang, from Columbia University Medical Center. Here's just a brief introduction of myself. I obtained my PhD from University of Nebraska Medical Center, and then I went for my post-doctoral training in Dr. Ira Tabas at Columbia University Medical Center. And recently, in this July, I joined Icahn School of Medicine at Mount Sinai as an assistant professor. My research interest is about the mechanisms of niche progression. I have nothing to disclose. Before I go to my talk, I would like to acknowledge Dr. Xiaobo Wang. He's an expert in the HIPAA pathway. He has shown that the HIPAA protein test is an important driver for niche progression. And he recently was able to link test to cholesterol metabolism in niche. And these studies have been published in Cell Metabolism. There are more and more evidence showing that cholesterol may play a role in niche progression. And as you all know, niche can be driven by both genetic factors, including gender, race, genetic polymorphisms, and environmental factors, including lifestyle and diet that can drive obesity and moderate gut microbiota. Cholesterol may function as a cofactor for both genetic factors and environmental factors. For instance, the impaired function of genetic proteins, such as PMP3, TM6, SF2, and FOB can induce the accumulation of cholesterol in liver by preventing the secretion of cholesterol. The liver has a critical role in producing and processing circulating lipoproteins. Healthy liver has the ability to maintain cholesterol homeostasis. However, cholesterol metabolism is dysregulated enough for a niche. There are evidence showing that in niche patients, hepatic cholesterol is induced. And in those mouse niche models, high cholesterol diets, including FPC diet, Western diet, combined with low-dose CCL4 and GAN diet can easily induce niche features, including steatosis, inflammation, and liver fibrosis. Cholesterol is thought to induce inflammatory and death pathways in hepatocytes and Kupfer cells. It can also activate hepatic acid cells by enhancing TLR4 signaling. And now we provide a novel mechanism by showing that cholesterol induces the stabilization of TAIS, which is an important driver for niche. Niche can be progressed from steatosis, and it can further progress to cirrhosis and hepatocellular carcinoma. The progression of niche requires multiple hits, including hepatocytes, oxidative stress, and hepatic inflammation, and particularly liver fibrosis. We have been thinking about transcription factors such as TAIS that may contribute to all these processes, since TAIS has been shown to induce lung fibrosis. We first found that TAIS protein levels were induced in both niche patients and most niche livers. To study the function of hepatocyte TAIS in niche, we silenced TAIS in hepatocyte by ligating SH-TAIS to a VAH1 factor. And then we fed those SH-TET-treated mice with FPC diet for 16 weeks to induce NASH. And we found TET silencing reduce the infiltrated inflammatory cells and reduce liver fibrosis as indicated by a decrease of serous restraining. It also reduced the number of tunnel-positive cells. So TET silencing suppress liver inflammation, liver fibrosis, and cell death in NASH livers. Mechanistically, we found hepatocyte TET can induce the Indian hedgehog IHH transcription and secretion, and the secreted IHH from hepatocyte will then act on hepatic-acidic cells to promote the expression of pro-fibrotic genes leading to NASH progression. So how is TET induced during NASH? Because our NASH diet contains a high level of cholesterol, we want to know if there is a correlation of liver cholesterol and liver TET. And we found in NASH patients, liver cholesterol is positively correlated with liver TET, meaning NASH patients who have higher level of liver cholesterol, they also have a higher liver TET. To study the role of cholesterol in NASH, we fed mice with diet that contains different concentration of cholesterol, and we found only the high cholesterol diet, which contains 1.2% cholesterol, is able to induce liver TET and liver fibrosis. Next, to study if cholesterol has a direct effect on TET, we treated the primary mouse hepatocytes and primary human hepatocytes with liposome-loaded cholesterol. And we found cholesterol loading is able to induce TET in both primary mouse hepatocytes and primary human hepatocytes. So how is cholesterol transported into the cells? We focus on a newly identified cholesterol transporter called Gram-D1 family protein. Gram-D1 family proteins are the cholesterol transporters that mediate cholesterol trafficking from the plasma membrane into the cell. And when we silenced Gram-D1 B and C transporters, we found liver-free cholesterol is reduced, and the liver TET protein levels will also decrease. So the intracellular cholesterol is critical for the increase of TET in NASH livers. And as expected, the decreased TET is associated with a decreased liver fibrosis in the Gram-D1 B and C silenced NASH livers. And this is our current working hypothesis on how hepatocyte cholesterol induce TET stabilization. The internalization of plasma membrane cholesterol through Gram-D1 caused an increase of CAMP and PK-dependent activation of IP3 receptor, leading to an increase of cytosol calcium, which can activate Rho-A. Rho-A will then inhibit the last dependent TET phosphorylation. Because phosphorylated TETs undergoes proteasomal degradation, activated Rho-A can prevent the degradation of TETs. So increased intracellular cholesterol will activate Rho-A and stabilize TETs, leading to NASH progression. Because the importance of intracellular cholesterol in the TET stabilization and NASH, we want to know if other regulators in cholesterol metabolism can affect TETs and NASH progression. We become interested in a novel cholesterol regulator called EHBP1. EHBP1 is an actin-binding protein. It couples endocytosis machinery to the actin cytoskeleton. And interestingly, there are several polymorphisms in this protein have been shown to associate with cholesterol. Although the mechanisms by which those SPMs affect cholesterol metabolism are unknown. The role of EHBP1 in NASH has not been described yet. And there is a single cell seq from human liver that shows that EHBP1 expression is higher in haplocyte than other cells. And in human cirrhotic livers, haplocyte EHBP1 expression is suppressed. Next, we want to know if EHBP1 protein levels are changed during NASH. So we check EHBP1 protein by three different kinds of NASH models. And we found EHBP1 is suppressed in all these three mouse NASH livers. So EHBP1 may regulate NASH. Next, we want to know if there is a link between EHBP1 and TASER. We science EHBP1 in primary mouse haplocyte. And we found SI-EHBP1 induce TASER protein levels. And because this project is still in a very early stage, we don't know the mechanism by which EHBP1 regulates TASER for now. Given that EHBP1 regulates the endocytic membrane trafficking, we propose that EHBP1 may regulate the levels of those cholesterol-related receptors, including LDRR, and those cholesterol transporters on the plasma membrane through regulating their endocytosis. So we think EHBP1-mediated endocytosis of those cholesterol-related proteins may regulate the intracellular pool of cholesterol, which may affect the stabilization of TASER and eventually NASH. With that, I would like to thank Dr. Eric Tabas, who is my mentor for my postdoc training for his support, and all his lab members, particularly Dr. Xiaobo Wang, who developed the FPC-NASH diet and performed the very beautiful cholesterol test study. And I also want to thank all the collaborators for their reagents and human samples and discussions on those projects. And finally, since the funding support from Xiaobo Zi Liver Foundation and my Canadian Archaeology Award, thank you so much. Thank you. Good afternoon. This is Scott Friedman, and I'm delighted to present this lecture as part of the Fibrosis Special Interest Group Symposium. My topic is The End of the Beginning, What is Next for Fibrosis in NASH? These are my disclosures. They're available also on the program website. These are the two major areas I'm going to discuss in the next few minutes. First to emphasize what we've accomplished, which I refer to as the end of the beginning. By that, I mean NASH is a growing unmet need. We recognize that. And more importantly, we now recognize that of all the features of NASH, it is fibrosis that most directly drives outcomes. So I'm going to underscore three concepts that are listed here. Stellate cell activation remains important. It's clearance and the senescence of stellate cells are an emerging focus. And also the stellate cells are heterogeneous and support normal liver function. In the second half, I'm going to point towards the future and underscore some of the unmet needs that you see listed here that I'll walk through in some detail. As we all know, stellate cells have been recognized for over three decades as the primary fibrogenic cell and liver. As you know, they are a subendothelial cell type found in normal liver interposed between hepatocytes and endothelial cells and characterized by the presence of retinoid containing lipid droplets around the nucleus. We and others have described for many years the response of stellate cells to injury, which we have called activation associated with fibrosis and disordered or dysregulation of cells surrounding the subendothelial space. This simplistic model has nonetheless proven quite robust in allowing us to understand both how the cell responds to injury as well as how the cells are cleared when fibrosis regresses. And I'll be saying more about that in a few minutes. At the same time, we're also beginning to apply more elegant methods of characterization of stellate cells morphologically. This is a photo or an image acquired by Dr. Sammy Wong in my group, who is using the glass liver method with Desmond and F480 staining to define both the appearance of stellate cells shown in purple, as well as their interaction with macrophages shown in green. And what you can see is this beautiful filigreed cell body pattern in a normal mouse liver, typical of the images of stellate cells from many decades ago. For example, this study years ago by Professor Kenjiro Wake using gold impregnation further identifies the perisinusoidal nature of stellate cells and their elegant foot processes that wrap around the sinusoid. I also want to acknowledge the memory and contributions of our friend Albert Geertz who died several years ago. Albert was instrumental in helping define key features of stellate cell heterogeneity as well as its morphology. And his work is long remembered and contributes to our understanding today. I first want to talk about a very interesting method we've been using in the lab to deplete stellate cells selectively. It's called the JEDI mouse, stands for Just EGF Death Inducing Mouse. And this is a mouse that has been engineered by my colleague, Dr. Brian Brown at Mount Sinai to express specialized T cells that will identify and target cells expressing green fluorescent protein for destruction. What you see here is the mature JEDI T cells interact with GFP, which is expressed on the cell surface by MHC class one that ultimately leads to T cell mediated killing. In our studies, we have used beta PDGF receptor driving GFP, as many of you know, this will express in stellate cells as well as other GFP, I should say beta PDGF receptor cells in other tissues. And so we use expression of GFP transgenically to render these cells susceptible to killing by these engineered JEDI T cells. This is a study demonstrating the impact of JEDI T cell administration on Desmond staining in a normal mouse liver. On the left hand side, you see Desmond staining in this elegant pattern throughout the sinusoid. At the same time, animals injected with JEDI T cells have complete loss of Desmond. And this is exactly the same methodology and image capture as on the left hand side, but really reflects almost complete loss of Desmond expressing cells. If we take a low power magnification to depict the diffuse nature of this loss of JEDI, of beta PDGF receptor expressing cells, you can see on the left is the PBS injected cells. I should say liver. And on the right is an animal injected with JEDI T cells. You have almost complete loss of Desmond expression, except for areas around the capsule and here in the center of the tissue. And this is probably an artifact. Again, this was work done by Sammy Wong. Sammy has gone on to more elegantly depict the loss of Desmond expressing cells using the JEDI method. This is 10 days post-injection. Here's on the left hand side, Desmond expression in the sinusoid and depletion renders these Desmond expression completely lost. If we use glass method as described a minute or two ago, now we can see the beautiful perisinusoidal foot processes of Desmond expression by normal stellate cells and CD31 expression denotes these sinusoids. Following T cell administration, there's virtually complete loss of Desmond and dilation and dysregulation in the appearance of these endothelial cells. This is accompanied by almost complete abrogation of Desmond quantitatively based on PCR in the JEDI injected animals, as well as complete loss of expression of beta-PDGF receptor. And these animals have impaired liver regeneration when HSCs are depleted and we're exploring the methodology underlying this. Currently, this was work driven largely by Youngman Lee, as well as Sarah Lamon and Maria Ibanez. Youngman has now moved to a faculty position at Vanderbilt University where she continues these studies. So let me return to the concept of stellate cell activation and introduce an additional element to this. As we've described for many years, stellate cell activation occurs broadly speaking in two phases. One is an initiation, which renders the cells responsive to a whole host of growth factors and cytokines followed by perpetuation, which is characterized by a whole host of phenotypic features that collectively enhance inflammation and scar formation. Beautiful work by Robert Schwabe, Tatyana Kisileva and their group, and earlier work by Derek Mann and John Iredale collectively defined critical pathways by which stellate cell resolution occurs in vivo. The original studies suggested that cells can undergo apoptosis or programmed cell death when fibrosis is no longer required. But in addition, Tatyana and Robert's work underscored the importance of deactivation as a mechanism to reduce the population of activated stellate cells. And then most recently studies have explored the role of senescence of stellate cells as a driver of liver disease and outcomes, specifically through the secretion of a senescence associated secretory phenotype that may contribute to fibrosis, inflammation and cancer risk. Those initial studies largely done by Amaya Lujanbio while in Scott Lowe's lab led to this more recent paper that we collaborated with Dr. Lowe's lab and publishing in Nature earlier this year in which they developed CAR-T cells that specifically targeted a molecule that was identified as a senescence associated cell surface marker on stellate cells, and that's EUPAR or uroplasminogen kinase activated receptor. And so with the assistance of Michelle Sadelang's lab at Sloan Kettering, Scott Lowe's lab documented the accumulation of senescence hepatic stellate cells and its contribution to a pro-inflammatory, pro-fibrotic environment. What you see here in the lower left-hand panel is beta-gal staining denoting senescence cells that were completely cleared following administration of CAR-T cells to deplete only senescent cells that express EUPAR. This was associated with a marked diminution in fibrotic areas, of course, a decrease in beta-gal staining, and interestingly, evidence of functional improvement by way of increased serum albumin in those animals in which the CAR-T cells had depleted senescent stellate cells. So this really brings us to where we're going next, and I want to start by talking about cell-cell interactions and networks. And to do so, I want to emphasize the beautiful work published in Cell Metabolism by Xiang et al. late last fall, in which they use single-cell RNA sequencing to identify the contributions of different cell types to the NASH phenotype. In particular, they used an amylin diet described here, similar to many other NASH-induced diets in mice, and they underscored the accumulation of a specific subset of macrophages known as TREM2-positive macrophages, which have a widening array of functions throughout the body, not only in liver, but also in adipose, heart, as well as other tissues, and including cancer. They also defined features of stellate cell gene expression in different subsets of stellate cells, as well as their contribution to contractility. And the data is summarized in part in this figure. First, they characterized the range of secretome, or secreted, I should say, secreted molecules. They called the hepatic stellate cell secretome genes and compared Chow and NASH diets. What they demonstrated was that a large number of extracellular matrix components were secreted during this NASH diet or NASH model, as well as protoglycans and molecules that contribute to extracellular matrix remodeling. Perhaps more interestingly, they began to define a theoretical network of receptor-ligand interactions that involved a number of different mediators and cell types, not only stellate cells, but also macrophages, B cells, dendritic cells, T cells, and sinusoidal endothelium. I'm highlighting here with the yellow box with the red outline one particular chemokine that's been strongly implicated in NASH pathogenesis, CCR2. And in this case, the informatics analysis indicates that there should be interactions between CCR2 comprising both dendritic cells, as well as T cells, and further interacting with other chemokines. Sorry, this is a chemokine receptor, and this is interacting with other chemokine signals, including CCL11 and indirectly CCL2 derived from stellate cells. They also looked at the impact of elevated expression of these so-called stellic kind genes with serum ALT. And what they showed is there was a progressive increase in a number of these stellic kind or stellate derived genes, and the more expression, the higher correlation with serum ALT. And this reinforces a number of studies that demonstrate stellate cells as a driver of that inflammation. Next, I want to underscore over the slide the emerging data about fibrosis resolution in proteases, particularly emphasizing a class of molecules known as specialized pro-resolving mediators. These are a class or classes of different mediators that are lipids, including lipoxins, resolvents, protectants, and merasins. This is driven largely by the pioneering work of Charles Sirhan at Harvard. And there have been very few studies looking at the contribution of these SPMs to liver fibrosis. There was one study published in JCI last year that I refer you to in which they implicated merasins specifically in fibrosis resolution and inflammation resolution. Once merasins and other SPMs are thought to be active, this is potentially resolving inflammation, leading to the loss of activated myofibroblasts in the ways I described previously, and ultimately activating ECM degradation. I want to underscore that while the biology of ECM degradation was very hot in the liver field 15 or 20 years ago, there's been very few studies recently that have returned to this important area. And so a key area for future research is not only to identify the sources and types of pro-resolving mediators in liver fibrosis regression, but also the identity and cellular sources of the different proteases and their inhibitors known as TIMPs that collectively regulate the degradation of scar. Let me say a word about novel therapeutic targets and combination therapies using this diagram to illustrate the different classes of targets that are being developed in the commercial sector. Certainly fat is an important component of NASH and thus a number of targets are being interrogated as potential targets for reducing liver cell or apatocyte fat. In addition, targets are being attacked that improve insulin resistance and glucose homeostasis, others that improve oxidant stress in the liver, and yet others that are anti-inflammatory targets or can be targets to reduce inflammation. Then finally, targets that represent direct stellate cell mediators that may reduce fibrosis directly. Currently, there are only empiric efforts to combine different groups or classes of these molecules to develop combination therapies. But I think a compelling opportunity for the future is to use more high-throughput methods to try different combinations at different concentrations to determine which achieve synergy when combined. One could envision using some of the newer organoid approaches. There are a couple of different beautiful papers, including one by the Takebe Group in Cincinnati, describing a multicellular Nash organoid. Conceivably, these or related organoid models could be used in a more high-throughput way to combine different classes of molecules or different, as I said, concentrations and seek synergy based on transcriptomic profiling. I think the selection and the implementation of combination therapies need to become far more rationalized and not simply empiric based on what drugs are available to a specific company. Increasingly, the microbiome is recognized as an important driver of disease. This is a beautiful paper recently from the Rohit Lumba Group and a large number of great scientists at UCSD, that suggests there is a universal gut microbiome derived signature that predicts cirrhosis. Certainly, we have lots of ways to identify cirrhosis clinically. But the real issue here is, what are the features or what components of the signature reflect direct effects on stellate cells? There's a rapidly advancing field identifying different bacterial types and different bacterial molecules that could potentially traverse through the portal system and activate stellate cells or other cell types directly in the liver. It's worth remembering that stellate cells are a rich source of pro-inflammatory and immunosuppressive molecules. They also are responsive in part through the LPS receptor complex to signals derived from gut bacteria, most notably LPS, which signals through TLR4, as well as other molecules involved in the TLR4 signaling axis, including Bambi, which was described by Eki Hiroseki and Robert Schwabe some years ago. It's conceivable that there are other bacterial components besides LPS, in particular, DAMPs and PAMPs that derive from the gut microbiome in a very specific way that translate into enhanced signaling or fibrogenesis or activation on stellate cells in liver. We need to understand a lot more about the components of the gut microbiome that specifically signal through stellate cells and what molecules they are signaling through when they arrive via the portal blood. Finally, I think there's a compelling need to understand better the links between fibrosis and regeneration. It's really laid out by this conundrum, which is that hepatic regeneration is a feature of healthy liver, and when livers regenerate in a healthy manner without injury, there is no fibrosis. Yet, contrarily, severe hepatic fibrosis represses regeneration. These are mutually exclusive. The severe fibrotic liver can't regenerate and the healthy regenerating liver doesn't develop fibrosis. This begs the question, are there signals in normal regeneration that suppress fibrosis, and what are those signals? For example, could there be molecules like endogenous HGF or suppression of TGF-beta or other molecules that are candidates for being both pro-regenerative and fibrosuppressive at the same time? Conversely, are there signals in the fibrotic liver that suppress regeneration, such that when fibrosis is attenuated, for example, when we cure patients with hepatitis C and fibrosis is no longer driven by a pedocellular injury, does that de-repress an endogenous hepatocyte program that leads to regeneration, which is exactly what we see clinically when patients are treated with effective direct-acting antivirals in hepatitis C or suppressive antiviral therapy for hepatitis B? Critically, can these molecules or signals that are pro-regenerative and antifibrotic be exploited therapeutically? In other words, can we mimic what the liver does endogenously when we cure hepatitis C? Or it's possible that just a pure antifibrotic alone could unleash or de-repress the liver's inherent regenerative potential. I think this is an area that we really need to explore, and it could explain in part the unique and powerful regenerative capacity of the liver that exceeds that of any other organ. In summary, I would conclude that we are, as I said, at the end of the beginning in NASH fibrosis. We're now ready to leverage our knowledge in pursuit of new diagnostics and therapies. But advances in several areas remain the goals of current and future studies. I've underscored the emerging networks of cell-cell interactions that really emphasize that cells don't operate in isolation, they operate in conjunction with other cell types, and these network of cell-cell interactions may ultimately yield a phenotype, as well as unearthing therapeutic targets. We need to make a better effort to revisit the importance of proteases, their identity, and their cellular sources, as well as mediators of inflammation fibrosis resolution, particularly the lipid molecules. We need to develop rationalized semi-high throughput approaches to try different combinations in organoids, for example, to yield greater synergy before we test them in animal models or other higher organisms. We need to understand the specific mediators of the microbiome that are driving the fibrogenic and inflammatory response by stellate cells. Finally, we need to uncover those mediators that link fibrosis and regeneration such that we can enhance regeneration while suppressing fibrosis. In summary, I think the future is quite bright. We've built on a 30-year legacy of advancing the science of hepatic fibrosis, and we're really ready now to move to the next level in advancing the development of diagnostics and therapies. Thank you very much. Hello, everyone. I want to start my talk, NASH Fibrosis, From Biomarker to Treatment. My name is Ekihiro Seki, Professor of Medicine, Cedars-Sinai Medical Center. I have a research support from Nippon Zoki. This research support is not related to today's my talk. This is the overview of today's my talk. I want to start from a brief introduction of NASH fibrosis. Next, I will discuss about recent biomarkers for NASH fibrosis. Then, I want to talk about our recent research, hyaluronic acid, also known as hyaluronan or HA, which is a biomarker for liver fibrosis. Our recent study indicated that hyaluronan is associated with a molecular mechanism and also is a potential treatment target for liver fibrosis. Non-alcoholic fatty liver disease, NAFLD, is a spectrum disease, which ranges from simple steatosis, NAFLD, to non-alcoholic steatohepatitis, NASH, which is steatosis plus inflammation and fibrosis. In the United States, 20-30% of adults have NAFLD. 15-25% of NAFLD will develop to NASH. Some researchers suggest some NASH develop directly from normal liver, not from simple steatosis. 30-40% of NASH will have fibrosis, and 15-20% of fibrosis will develop to cirrhosis. Cirrhosis is a risk factor for HCC, but recent studies suggested without cirrhosis or any stage of NAFLD, even simple steatosis stage can increase the risk of HCC. Fibrosis is the most important determinant of survival of NAFLD because fibrosis can predict future cirrhosis. Why this is important? Because complications of cirrhosis, such as portal hypertension, liver failure, primary liver cancer, are all life-threatening complications. Also, the exact molecular link is still unknown. Fibrosis is highly associated with the risk of cardiovascular disease and atherosclerosis. Importantly, cardiovascular event is a leading cause of death in NAFLD patient. Interestingly, not only the risk for primary liver cancer, NAFLD also increase the risk of non-liver primary cancers, such as colon, stomach, pancreas, breast, and the prostate cancers. Because fibrosis is a prognostic factor for NAFLD, staging fibrosis is important to decide disease management. Also, it is important to assess whether treatment is effective for fibrosis. For this, biomarkers for fibrosis staging are critical. However, currently available biomarkers still have limitations. Therefore, it is important to assess whether treatment is effective for fibrosis staging. Therefore, histological diagnosis using liver biopsy is still a gold standard. Liver biopsy is invasive, hard to apply for all patients, but still useful for histological diagnosis. Also, for understanding specific gene and protein expression using biopsy samples. Imaging technology is developing year by year. Now, transient elastography and MR elastography become very useful approaches for staging fibrosis in NAFLD patients. We still hope to have more sensitive and specific biomarkers using blood samples. This is a diagram of NASH fibrosis pathogenesis. Excessive intake of dietary fat and fat from adipose tissues accumulated to path sites, which increase the lipogenesis and decrease beta oxidation and mitochondrial functions, which then induce hepatocytes, steatosis, and injury. Fatty and injured hepatocytes release oxidative stress, dumps, and extracellular vesicles that stimulates immune cells such as Kupfer cells or directly activate hepatic stress cells. Activated stress cells produce extracellular matrix inducing fibrosis. Kupfer cells also produce cytokines to stimulate hepatocytes, promoting hepatocyte injury and steatosis. Also, enhanced stress activation and fibrosis. Extrahepatic organs such as intestine and adipose tissues also contribute to NASH and fibrosis. Especially, intestine-derived bacterial products and bile acids are associated with NAFLD development. Fatty and injured hepatocytes release AST, LT, side keratin-18 fragments, extracellular vesicles, microRNA. These can be measured in the blood. Activated stress cells produce extracellular matrix, collagen fragments, and hyaluronan, which also can be measured in the blood. AST and LT are gold standard, but one third of NASH fibrosis patients are with normal ALT. Side keratin-18 fragments can be measured by M65 and M30. Studies suggested M65 is useful to distinguish NASH from simple steatosis, and M30 is more useful for fibrosis. However, these markers are for hepatocyte injury, not specific to fibrosis. MicroRNAs and exosomes in the blood can also be used for biomarkers. Studies suggested miR-122, 34A, and 192 are upregulated in NASH fibrosis. Of note, miR-122 and 34A reflect liver injury, and miR-192 is more specific to fibrosis because TGF-β is associated with miR-192. Importance of exosome is still unclear. Exosomes can pack microRNA, so microRNA in exosome should be more stable. ECM fragments should be more specific to fibrosis. These include procollagen type 3 amino terminal propeptide, pro-C3, and hyaluronic acid. Also, there are scoring system such as NAFLD fibrosis score, APRI, FIB4, ELF, and HEPA score. Recently, a new scoring system, NIS4, has been reported. Study indicated that NIS4 is superior to other scoring system. Fecal microbiome may also be used as a non-invasive biomarker. Recently, Rohit Lumber's group has shown 19 gut microbiome signatures can predict nastyrosis. Sensitivity and specificity are pretty high. Also, they reported, when treated with FGF-19 analog, a rare gut bacteria, Valonila, was increased and correlated with drug response. The study suggested Valonila as a biomarker for drug response. Hyaluronan has been used for biomarker for liver fibrosis. Here, you can see blood hyaluron levels are increased during fibrosis progression. And fibrosis scoring system such as ELF score, HEPA score, and fibrometer contain hyaluronan. We can tell that hyaluronan is one of the best biomarkers for liver fibrosis. However, we don't know about hyaluronan beyond the biomarker. There are different sizes of hyaluronan. High molecular weight hyaluronan is protective, but low molecular weight is pro-inflammatory and pro-fibrogenic. Hyaluronan is produced as high molecular weight form through membrane-bound hyaluronan synthesis, HUS-1, HUS-2, and HUS-3. Upon inflammation, high molecular weight form degrades into low molecular weight form and induce inflammation and fibrosis through hyaluronan receptors, CD44, TR4, and TR2. Hyaluronan is systemically circulating and turnover is fast. One-third of total body hyaluronan is turned over daily. Interestingly, more than 90% of hyaluronan is taken up and degraded in liver endothelial cells. Importantly, in cirrhosis, liver endothelial cells are dysfunctional and cannot take up hyaluronan, and then hyaluronan accumulates in the blood. However, it has not been studied whether hyaluronan is overproduced in liver fibrosis. To study whether hyaluronan is overproduced, we measured HUS-1, HUS-2, and HUS-3 expression in liver fibrosis. Here, we used hepatitis B patient liver biopsy samples. Only HUS-2 is increased during the progression of liver fibrosis, and there is a strong correlation between liver HUS-2 expression and blood hyaluronan levels, but not with HUS-1 and HUS-3, indicating that HUS-2 is associated with hyaluronan production in liver fibrosis. This is similar with NAFLD patients. HUS-2 expression is upregulated in NAFLD livers with fibrosis compared with those without fibrosis. Hyaluronan staining shows more hyaluronan deposition in advanced NASH fibrosis. Then we moved to the mechanistic study. We found that activated cell cells express HUS-2, but not other cell types. Based on this information, we moved to functional study and generated cell cell-specific HUS-2 knockout mice. Here, we used choline-deficient high-fat diet model to induce NASH fibrosis. In knockout mice, serum and liver hyaluronan levels are reduced. So in liver fibrosis, not only clearance problem by endothelial dysfunction, overproduction of hyaluronan by HUS-2 is associated with blood and liver hyaluronan levels. Also, knockout mice show fibrosis is reduced, indicating that stellar cell-derived HUS-2 and hyaluronan are crucial for liver fibrosis. In clinical practice, only total hyaluronan is measured. We found in the blood of patient with fibrosis and in the mouth fibrotic livers, low molecular weight form is a dominant form of hyaluronan in liver fibrosis. Because low molecular weight hyaluronan is profibrogenic, these data suggest that low molecular weight hyaluronan can be a good biomarker for liver fibrosis. Finally, we tested whether hyaluronan can be a molecular target for treating NASH fibrosis. 4-methylambidiferone, 4-MU, also known as hymechromone, is an approved drug for biliary spasm in Europe and Asia. 4-MU also inhibits hyaluronan synthesis. 4-MU treatment dramatically inhibited fibrosis development, suggesting that inhibiting hyaluronan synthesis can be a treatment strategy for NASH fibrosis. In liver fibrosis, once sterocells are activated, sterocells produce hyaluronan, which is converted into low molecular weight form, which further activates sterocells to produce collagen. Fragmented type 3 collagen, Pro-C3, and low molecular weight hyaluronan can be measured in the blood. MiR-192 is a TGF-beta downstream target and can be detected in the blood. These can be biomarkers for liver fibrosis. Future directions. We need to study whether low molecular weight hyaluronan is useful for not only staging fibrosis progression, but also assessing fibrosis regression after treatment. Also using low molecular weight form, we may consider creating new scoring system. For treatment option, 4-MU is effective in mouse NASH fibrosis. We need further preclinical and clinical studies. We may want to consider other 4-MU derivatives or generating new drugs to target HS synthesis. Key takeaways. Fibrosis is a key determinant of the prognosis of NAFLD. We need to develop more useful non-invasive biomarkers for staging fibrosis and assessing treatment efficacy. Most of biomarkers are for liver injury, not specific to fibrosis. Microbiome is interesting, but we need more study. Collagen fragments, hyaluronan, MiR-192 could be more specific biomarkers for fibrosis. Low molecular weight hyaluronan could be a sensitive biomarker for fibrosis. Blocking HS synthesis could be a treatment strategy for NASH fibrosis. Lastly, I want to thank all laboratory members and CDER-SINI members, and also outside collaborators. I also want to thank NIH, as well as Karsh Family Foundation for funding support. Thank you for your attention.

liver fibrosis

NASH

molecular pathways

repair processes

immune-mediated regulation

liver macrophages

hepatocyte cholesterol test pathway

pro-fibrogenic state

cholesterol metabolism

fibrosis progression

EHBP1

hyaluronan

biomarkers