Hi, my name is Wenzing Ding. I'm from University of Kansas Medical Center. Welcome to 2021 ASLD Hepatotoxicity SIG Symposium. We're very happy to have you to join us today. Hi everyone. I'm Lily Dara from the University of Southern California. Wenzing and I have an exciting lineup of speakers for you today. It's organized around the theme of blood coagulation and the role of anticoagulants in drug-induced liver injury. We'll start off by hearing a talk on novel anticoagulants by Dr. O'Leary, followed by hepatotoxicity from heparins by Dr. Ntagliato. We'll end the session with two excellent basic science talks. The first one from Dr. Lyondyke, who's going to talk about the one variant brand factor in hepatotoxicity, followed by Dr. Singsaju. She's going to share the story on chitinase-3-like protein in APAP. Please submit your questions in the chat box. We'll try our best to answer as many questions as we can during the meeting. Stay safe out there, and we hope to see you all in person next year. It's my privilege and honor to be with you today to discuss novel anticoagulants in liver disease and the risk of drug-induced liver injury. My name is Jacqueline O'Leary, and I'm Chief of Hepatology at the Dallas VA and Associate Professor of Medicine at UT Southwestern. These are my disclosures. We're going to be setting the stage for why DILI from anticoagulants is such an important topic, transitioning to DILI from adenosine diphosphate receptor blockers, and then spending most of our time discussing DILI from NOACs. Anticoagulation in cirrhosis has really changed for five main reasons. The first being that rebalanced coagulation is totally accepted in our cirrhotic patients, whereas a long time ago we used to think of them as hypocoagulable. But now we know that things really are rebalanced even in late-stage cirrhosis with decompensation. There are several factors that can be at play to create a hypocoagulable scenario, such as severe anemia, which impairs the red cell's ability to push platelets to the periphery of vessels to clot, severe renal dysfunction, certain medications, many of which we'll discuss, as well as sepsis can result in hypocoagulation. Although patients are having overall rebalanced coagulation, they do have an increased risk overall for VTE or venous thromboembolism, most often in their portal vein. Meta-analysis data has shown that this hazard ratio is approximately 1.7, especially in the setting of portal vein thrombosis, does improve recanalization in meta-analysis, as well as markedly decrease the risk for progression. As a result of several of these factors, a new review article in J-HEP has changed how we think about anticoagulation in portal vein thrombosis and recommended it for all patients who have no contraindications as well as acute portal vein thrombosis. The fourth reason is that anticoagulation actually lowers the risk for variceal hemorrhage in our patients with varices. This is surprising to some people because of the fact that anticoagulation lowers the risk for variceal hemorrhage. However, the risk for variceal hemorrhage is not dependent on clotting, but instead dependent on pressure. The risk for variceal hemorrhage is lower in meta-analysis in our cirrhotic patients with varices. The fifth reason is that in randomized controlled trials, enoxaparin has been shown to prevent not just portal vein thrombosis, but also decompensation as well as death. The enthusiasm for anticoagulation in patients with compensated and even decompensated cirrhosis has increased significantly, although there are exceptions, specifically platelets less than 50,000 and those patients with severe PHG, and more and more patients with compensated and decompensated anticoagulated for not just portal vein thrombosis, but also atrial fibrillation. Switching gears to adenosine diphosphate receptor blockers that are used instead of or in combination with aspirin, specifically clopidogrel, ticagrelor, presugrel, and triclopidine, there were 18 published cases of DILI from clopidogrel. About two-thirds were mixed pattern, about a quarter hepatocellular, and the minority cholestatic. The mechanism of clopidogrel-induced DILI is thought to be high activity of CYP2C19 and CYP2B6 because they generate higher levels of active metabolites, and pretreatment of one aminobenzotriazole, which is a nonspecific inhibitor of CYP450 activity, did block hepatotoxicity, confirming this mechanism of DILI. Ticagrelor, as opposed to clopidogrel, was studied first in the PLATO trial. Unlike clopidogrel, it is not required for activation. Mild hepatic impairment was shown to increase the risk for adverse events such as death at 3.1% versus 0.9%, and numerically SAEs at 20% versus 16%, as well as numerically AEs at 84% versus 81%. Ticagrelor is metabolized by CYP3A4, even though that is not required for it to be activated. There have been reports of mild transient liver enzyme elevations. However, only severe liver injuries have been reported as part of TTP, which rarely occurs, as well as hypersensitivity with rhabdomyolysis and secondary effects on the liver. Fortunately, there have been no independent reports of 2-DILI outside of the conglomeration of TTP or hypersensitivity in patients without hepatic impairment. However, patients with hepatic impairment, defined as child's B and C serotics, should not use Ticagrelor. Sucral versus clopidogrel has been studied in the TRITON trial. It is metabolized by CYP3A4 and 2B6 for activation like clopidogrel. Abnormal liver function is rare, and ALT greater than three times the upper limit of normal with bilirubin, the upper limit of normal only occurred in 0.12% of sucral patients versus 0.06% of patients on clopidogrel. Several reports of hypersensitivity with fever, eosinophilia, diarrhea, and abnormal LFTs have been reported. Fortunately, rapid reversal of this does happen with stopping of the medication, and one can switch safely to clopidogrel after these adverse events. Of note, increased GI bleeding risk has been shown when compared to use of clopidogrel. Ticlopidine is used for risk reduction of stroke in patients intolerant of aspirin. Ticlopidine undergoes extensive metabolism by several 450-cytochrome isoforms and carboxyesterases. RAD studies suggest that adducts cause hepatotoxicity after biliary excretion of glutathione-conjugated metabolites via MRP2-facilitated transport, and this results in a cholestatic pattern. The association of DILI is with HLA-A3303 as well as a polymorphism in CYP2B6 where there is a change from T to C, which makes patients more susceptible to Ticlopidin-induced cholestatic DILI with an odds ratio of 2 because of the high expression of the enzyme and therefore more rapid metabolism. Now switching gears to NOACs. The first NOAC to highlight that there may be a risk of DILI was exmagatin, which is an oral thrombin inhibitor. It was polled at stage development because of an 8% risk for DILI specifically in a hepatocellular pattern. The mechanism of DILI was elucidated by Lundgren and colleagues in 2007, and specifically the prodrug exmagatrin and its active metabolite meligatrin can associate with the active groove of HLA-DR7 and HLA-DQ2, and this association is key in activating and initiating DILI. Unfortunately, other NOACs that have made it through FDA approval have not had such significant risk, and in meta-analysis data of 29 phased through randomized controlled trials looking at 152,000 patients taking NOACs evaluating specifically for the risk of DILI, all trials were assessed as having low risk of bias, and NOACs were not shown to have an increased risk for DILI compared to other control groups, but this is of course critical. In specific sub-analysis of NOACs versus low molecular weight heparins, NOACs were actually associated with a lower risk of DILI than low molecular weight heparin. I won't go into this because you will be hearing more information later on in this seminar. When we look at the forest plots for apixaban, dibagatran, and direxaban, you can see that they all cross one despite having 22,000, 11,000 patients in the first two drugs, and doxaban and rivaroxaban similarly cross one. In the 62,000 patients analyzed here, the rate of DILI was different than patients who were on the placebo or other anticoagulation agent that was being analyzed. Administrative database from Canada, province Quebec, specifically looking at their health insurance data, they evaluated patients with a new diagnosis of non-valvular AFib over a four-year period, and they evaluated for the risk of acute liver injury. I will highlight that they only assessed this by hospitalization or death. If acute liver injury was diagnosed as an outpatient, they would have missed it. They did evaluate this in over 51,000 patients, and 4,000 patients did have prior liver disease and found no significant association between NOACs and drug-induced liver injury that resulted in death. When they broke it down by the different NOACs, specifically dibagatran, rivaroxaban, and apixaban, there was similarly no difference. In another Medline database search over a six-year period, a group published two additional new cases of rivaroxaban-induced DILI, in addition to 26 cases that they found in the literature. The median time of onset was 15 days. The average age was over 65 and 58 percent with the predominance of females, and hepatocellular pattern was more common at 42 percent, cholestatic at 27 percent, and mixed in 15. The group from Iceland, who does beautiful work, looked at the Iceland patient population for the country receiving outpatient oral anticoagulation over a 10-year period. This was roughly 15,000 patients. Patients with elevated liver enzymes were pulled, and the cases were reviewed for association to warfarin, rivaroxaban, apixaban, dibagatran, and endoxaban. Of these three cases, the first had a peak ALT of almost 300, the second of 2,200, and the third of almost 800. The latency period was a little over 300 days for the first case, over 500 days for the second, and eight for the third. All were hepatocellular in nature, and one was suspected drug-induced autoimmune hepatitis successfully treated with corticosteroids, and the patient did relapse after being taken off later. Overall, the risk of rivaroxaban-induced DILI was thought to be 1 in 1,100 patients. When choosing a NOAC in patients with cirrhosis, of note, no NOAC has been studied in patients with decompensated cirrhosis, and all are absolutely indicated in CTP-C. Dibagatran is least affected by hepatic metabolism, but does increase the risk for GI bleeding more so than the other NOACs, and so therefore is not favored. Apixaban, although hepatically metabolized, does not increase the drug exposure or AUC in CTP-B, and therefore may be considered as preferred. In conclusion, anticoagulation is not recommended in patients with compensated and decompensated cirrhosis. ADP receptor blockers, such as clopidogrel, has the lowest risk for DILI in that group. The pattern, when found, is often mixed, and the mechanism is high activity of CYP2C19 and CYP2B6. Ticagrelor has increased risk for TTP and hypersensitivity and cannot be used in decompensated cirrhosis, and Prasugrel has an increased risk for hypersensitivity and an increased risk for GI bleeding. NOACs have a low rate of DILI. The pattern is most often hepatocellular, although other patterns can be seen. Not all have the same risk, and likely Rivaroxaban has the slightly higher risk, 1,100. When choosing a NOAC for anticoagulation, never use it in CTP patients, CTPC patients, and Apixaban AUC may be the least affected in our CTP-B patients. Thank you so much for your attention. Hello. My name is Nicholas Intagliata. I'm from the University of Virginia, and I was invited by the Hepatotoxicity SIG to talk on hepatotoxicity from heparins. I'd like to thank the committee for inviting me, and thank you all for joining. So I wanted to start out, my background is in hepatology, and my interests are from a perspective of coagulation and liver disease. I wanted to start out with this quote from H.C. Hemker on an excellent review on the history of heparins. I'm not going to read it all, but of note, he recognizes that heparins have been around for a long time. They're used to treat a wide variety of diseases related to thrombosis. And thrombosis has this humble position in the pantheon of diseases, and it's so common, and so many people die from it, that the importance of these medications really cannot be overstated. This is a timeline of anticoagulation. Heparin was initially discovered over 100 years ago. Our first oral anticoagulant was discovered in 1940 with warfarin. It wasn't until 1970s and 1980s that low molecular weight heparin became available. Fonda-Paradox, which we will talk about a synthetic-derived low molecular weight heparin, became approved in 2001. And in the next talk, you'll hear about direct oral anticoagulants. These are the three classes of how we'll talk about these today. Unfractionated heparin, low molecular weight heparin, and Fonda-Paradox. Over on the right, you can see their mechanisms of action, which are similar, with the exception that unfractionated heparin can act indirectly both on factor Xa and thrombin. So starting with unfractionated heparin, as I said, it is the oldest anticoagulant and has been in use for many years. It's a naturally occurring polysaccharide derived from porcine mucosa, and previously also derived from bovine mucosa. It's important to note that the molecule is quite heterogeneous, and not all heparins are the same, and this can lead to unpredictability with dosing and side effects. It can be administered both subcutaneously and intravenously, and has variable binding of other proteins. Its metabolism is complex. It has a rapid saturable phase via the reticuloendothelial system, as well as a slow non-saturable phase via the kidneys. Here's a summary on unfractionated heparin from liver tox, and it's important to note the patient population at risk is very wide range. Many patients receive this medication for a wide variety of reasons. Its pattern of hepatotoxicity, it is typically associated with mild elevations in about 10% to 60% of patients with elevation in AST and ALT. Higher values, greater than five times the upper limit of normal, are much more rare, but do occur in patients receiving higher doses. Cholestatic liver injury is rare, and it is often self-limiting, and it arises within four to eight days of treatment. I'll come back to this theme again, and there are confounding associated clinical features. Other reasons that patients may have elevation of liver enzymes that are important to recognize, and its mechanism of the injury is unknown, but suspected to be a direct hepatotoxicity. The early reports of drug-induced liver injury with unfractionated heparin are from 1975, this paper entitled hypertransaminaseemia with heparin therapy. This looked at 14 consecutive inpatients receiving heparin, and found the majority of patients had some elevation in liver enzymes, with the range listed below. There's no real comment on clinical sequelae or further study, and subsequent reports were published over the course of the end of the decade. Saffel and colleagues in 1980 published this report that postulated and demonstrated a potential difference between porcine and bovine-derived heparin products, with a higher risk of elevation in liver enzymes from porcine-derived heparin. So Dukes and colleagues set out in 1984 to examine this question, and they randomized 86 patients to either receive bovine or porcine-derived heparin. The vast majority of patients had some increase in AST and ALT, some higher than others, as you can see here on the curve to the right. About 20% of the elevation resolved while continuing therapy, and all of it resolved after stopping it. They noted two significant risk factors that were associated, including gender, men, and baseline elevation in liver enzymes. There was no difference in animal source between porcine and bovine. More recent literature, there's really sparse data or reports implicating unfractionated heparin in liver injury. There's one case report to the right here that shows a mixed liver injury, thought secondary to unfractionated heparin. Overall, there are no convincing reports of severe liver injury to date, and we do have to recognize there are problems with confounders in this population. The bottom line is that mild elevations in AST and ALT can occur with unfractionated heparin and are self-limited, or resolve with discontinuation of therapy. So I'll come back to this randomized control trials, a pretty important paper, I think, in this field. And what Harrell and colleagues did in 2012 is randomized 48 healthy men to either receive unfractionated heparin or various types of low molecular weight heparin. The curve to the right shows you that the vast majority of patients receiving unfractionated heparin had some elevation in liver enzymes by day seven. It demonstrates really a clear relationship with unfractionated heparin and elevated liver enzymes that is apparently self-limited. Looking at the package insert, you can see that significant elevations occur in a high percentage of patients and healthy subjects who receive heparin. But they do note, again, this relationship of other factors that may be causing elevation in liver enzymes. Moving on to low molecular weight heparin. Remember, this is a purified fragment of heparin isolated by depolymerization. It is much more homogeneous and has predictable pharmacokinetics with less side effects. It has some limitations in use compared to unfractionated heparin that we are all familiar with. Looking at the summary from liver tox, it has a similar patient population at risk. Its pattern of hepatotoxicity is similar, typically associated with mild elevation in AST and ALT and 4 to 13% to patients. Again, dose-dependent effects that can arise three to seven days after starting therapy, and cholestatic injury is rare. It has similar patient population and confounding factors when trying to make a relationship between the elevation of liver enzymes and the drug injury. Its mechanism of injury is unknown but suspected to be a direct hepatotoxic effect on the liver. Looking at a randomized controlled trial that looked at logiparin or enoxaparin versus placebo in patients undergoing hip replacement. The authors note that overall, there were some increases in AST and ALT, however, not significant when compared to placebo. A more recent trial, a large double blind trial, examining different doses of enoxaparin for prevention of DBT in surgical patients. You can see here on the table to the right with the box that overall elevations in ALT were demonstrated greater than three times the upper limit of normal in 2 to 7% of patients, and this did have an apparent dose dependency. Coming back to the randomized control trial, you can see here the curve with unfractionated heparin to the top left compared to the three types of low molecular weight heparin. Again, all the curves look almost identical with a peak at seven days. The drugs were continued from day one to day five. What is unique about this paper is they explore possible underlying mechanisms. They evaluate molecules such as MIR-122 and GLDH, which are specific to hepatocytes. And note, they do become elevated as the drug is continued with some decrease and then another peak at day seven. The authors tested K18 fragments, which were absent, indicating that these molecules are likely leaked from necrotic hepatocyte death. And so they note there is this direct relationship with the early peak and potentially an indirect effect of the drug as well through secondary activation of the innate immune response. What is unclear is what stops this response and why do things go back to normal? Looking at the package inserts from inoxaparin and daltaparin here, respectively, you can see that ALT and AST are demonstrated to be elevated in clinical trials from about 4% to 6% of patients. They also note a similar warning of confounding comorbidities that could explain elevation of liver enzymes as well, and that all were fully reversible. I did found one case of serious drug-induced liver injury in liver tox. This is a 45-year-old man treated for a DBT. When converted to inoxaparin as ALT at about day seven, peaked at 770. This was stopped and eventually Fonda-Paranox was started with complete resolution. There are several case reports published looking at low molecular weight heparin and drug-induced liver injury listed here. These reports are fairly convincing. One actually had a re-challenge with demonstrating again increased in liver enzymes, as well as three that had biopsy supporting evidence. Moving on to Fonda-Paranox, this is the liver tox summary. Again, the patient population at risk will be similar, although there is more limited clinical use and experience with this drug. It does have a similar pattern of hepatotoxicity, although it seems to be rarer, so 1% to 3% with mild elevations in AST and ALT. A similar confounding relationships in the mechanism of injury is unknown, but suspected to be the same with direct hepatotoxicity to the liver. Looking at this study that examined Fonda-Paranox for the treatment of pulmonary embolism, compared two types of low molecular weight heparin to a small group that received Fonda-Paranox. You can see here that the risk of a mild elevation of liver enzymes, so using a very conservative definition of drug induced liver injury, that it was much more common with the two types of low molecular weight heparin when compared to Fonda-Paranox. The majority of patients were continued on anticoagulation and the liver dysfunction resolved. Multivariate analysis showed that type of anticoagulation and length of treatment were associated with risk. And also the presence of hepatitis B surface antigen, perhaps demonstrating that patients with chronic liver disease may be at increased risk. Looking at the data from clinical trials, the first bullet point looks at a trial using prophylaxis Fonda-Paranox and finds that increases in ALT and AST greater than three times the upper limit of normal, or up to 1.5% compared to placebo. Looking at another trial comparing Fonda-Paranox to enoxaparin and unfractionated heparin, you can see that Fonda-Paranox increase in AST and ALT was significantly lower than low molecular weight heparin and unfractionated heparin. Overall data from clinical trials I think is reassuring with Fonda-Paranox at very low rates of elevation liver enzymes. Similarly, looking at the package insert, elevations greater than three times the upper limit of normal were reported in 1.7 and 2.6% of patients, which seems lower than other trials in low molecular weight heparin. The elevations are reversible as well. I found one case report of a severe drug induced liver injury in a young child who was initially treated with enoxaparin and then switched to Fonda-Paranox. This result probably after stopping the agent. So we just talked a lot about how heparins can harm the liver, but what about helping the liver? And this, I think, comes out of this putative idea or role of thrombosis as a pro-inflammatory and pro-fibrotic condition in chronic liver disease, as put forth by Ian Walmas and his colleagues with the Parenchymal Extinction Hypothesis. Animal models using rats, models of cirrhosis have examined a low molecular weight heparin and demonstrated positive effects such as reduction of inflammation and fibrosis. One study published five years ago now wanted to look at low molecular weight heparin's effect on portal hypertension in a model of rats. It evaluated both acute and chronic effects, and although they did not find acute effects, demonstrated at week one and three, decreased portal pressure and hepatic vascular resistance, as demonstrated in this cartoon below, through possible mechanisms such as reduced fiber and deposition. One study published several years ago now looked at humans without portal vein thrombosis and randomized them to enoxaparin versus no treatment. They followed them over the course of two years and were treated for one year and found that patients that were treated with low molecular weight heparin did not develop portal vein thrombosis and were less likely to develop hepatic decompensation or death. This is a compelling study and exciting as a potential disease modifying agent. Potential mechanisms beyond the liver include potential effects on enteric intestinal microcirculation. So in conclusion, heparins are associated with a mild elevation in ALT and AST. And there are really no data to suggest these are associated with a significant risk of liver injury. The mechanism of elevation is unclear but seems direct injury with dose dependency. And there are animal models and some indirect human evidence that these medicines may actually be beneficial in chronic liver disease models. Thank you all. Hi, everyone. Today I'll be talking about some of our work focused on the role of von Willebrand factor in acetaminophen induced liver injury. But before I begin, I absolutely want to take the opportunity to thank the organizers for the invitation to share this work. Very much looking forward to the return to in person meetings, but I would welcome questions in the chat at the end of the session or during this session. And certainly don't hesitate to reach out by email. I'd like to begin by acknowledging the folks in the lab who do the work. The vast majority of the science that was done today was performed by Daphne Grunewald, who's a postdoc in the laboratory. And this has been a solid collaboration with Dr. Ton Lisman in the Netherlands. And I want to acknowledge science that was done several years ago by Bob Roth here at Michigan State University that really kicked off this whole field kind of looking at platelets in the acutely injured liver. And certainly want to acknowledge continued funding from NIH for supporting our science. So where we're headed today, just to give you a brief overview. So over the next 20 or so minutes, I want to begin by providing you with information on how to connect the hemostatic system, blood clotting and platelets with acute liver injury and repair. And we'll do that from the standpoint of both clinical and translational studies, as well as solid mechanistic studies in experimental settings. We'll focus on the role of platelets in acute liver injury, and I'll try to give you just a brief introduction to what the clinical and experimental evidence for that connection is. And we'll try to set the stage for how these studies can be a foundation for mechanistic studies drawn from experimental settings, in particular acetaminophen overdose in mice. I'll then move on to the bulk of the talk, which will focus on the role of this platelet adhesive protein called von Wildebrand factor in acetaminophen induced liver injury. Again, providing both a clinical view as well as experimental evidence on the role of von Wildebrand factor in liver injury. And I'll conclude by providing kind of a solid summary of what we've covered in the talk, but also try to provide a perspective on how I think we should move forward only insofar as collaboration across multiple fields is going to be required to take the next logical step. So we need to begin by talking about this delicate rebalance of coagulation and hemostasis that occurs in patients with liver disease. As you're well aware, the liver is the primary site of synthesis of pro-coagulant as well as anti-coagulant proteins, whereby the normal liver is producing these factors in a balance depicted here by the elephant on the pedestal. All of these pro and anti-coagulant proteins being in balance, it would take quite a bit of force to push this elephant one side or the other. So in other words, the system being in balance, it takes quite a bit to cause either bleeding or thrombosis. Whereas in the context of the diseased liver, the synthetic function of the liver is reduced and one is left with reduced production of both pro and anti-coagulant proteins, creating this delicately rebalanced state where the elephant is precariously balanced on the pedestal. And the amount of force of the event that it takes to push the system to thrombosis or bleeding is dramatically shifted. So this is a figure from a review that Anya Kopech and I wrote some years ago now, but I still think it holds true today. And that is the vast majority of studies in humans have focused on discovering the mechanistic basis for altered hemostasis. The liver is diseased or injured, what is changing in the hemostatic system? Whereas in the context of experimental models, the vast majority of the focus has been on what is the role for components of the hemostatic system in damaging the liver? What is the consequence for the pathogenesis of liver disease? And while there's some intersection between these two areas, this is predominantly the state of play as it is today as well. One particular area of overlap that is becoming increasingly exciting is the role of platelets, both in the context of how liver disease changes the activity of platelets and their regulators, as well as the role of platelets in promoting liver injury and chronic liver disease. So, here I'm presenting data from the acute liver failure study group from Dr. Straubitz documenting thrombocytopenia in patients with acute liver failure, and notably what you can observe is that in patients that go on to require a transplant or have a poor outcome, including death, there's a dramatic diminution and continued reduction in platelets after enrollment and study. Notably, blood platelet count also is significantly reduced after the challenge of mice with a hepatotoxic dose of acetaminophen. Here, data from Kazumi Okawa from Bob Roth's lab at Michigan State University published some years ago, where on the y-axis you see platelets and on the x-axis is the time after administration of acetaminophen, and you can notice a very rapid and persistent reduction in platelets in the blood. Where those platelets are going is the liver. Here, depicted in immunofluorescent labeling for alpha 2b beta 3 integrin on those platelets, otherwise known as CD41, labeled in red, and I think what you can appreciate is that the platelets are accumulating at the periphery of these areas of necrosis that are manifesting over time in the acetaminophen-treated mice. Notably, platelet depletion prior to administration of this toxic dose of acetaminophen caused a dramatic reduction in the area of hepatocellular necrosis 24 hours after acetaminophen administration, and that's also evident here by measuring serum ALT activity as an indicator of hepatotoxicity on the y-axis, and what you can appreciate is that the development of hepatotoxicity is significantly reduced in mice that had platelet depletion. So, this provides solid evidence to suggest that platelets are important for the pathogenesis of acute liver damage and outcome both in mice and humans, and we're left asking these next-level questions, including what factors are driving that initial hepatic platelet accumulation and what are the mechanisms whereby platelets are contributing to the progression of liver damage, and as we were thinking about this to try to think what we would do next, we decided to frame this question on observations that are made in patients with acute liver failure, and notably work from Tom Listman's group in collaboration with the acute liver failure study group had documented an increased plasma concentration of von Willebrand factor in patients with acute liver injury or acute liver failure. So, if you're not familiar, von Willebrand factor is a platelet adhesive protein. It circulates in plasma and is as well stored in endothelial cells and released upon various stimuli and is also stored in platelets. Von Willebrand factor plays a key role as this platelet adhesive protein by serving as a bridge between platelets and collagens that are revealed upon vascular injury, and notably von Willebrand factor contributes to normal hemostasis but can also contribute to pathological intravascular thrombosis, and perhaps we need to look no farther than the COVID-19 literature to see how VWF has emerged as a potential target to prevent pathologic intravascular thrombosis. With this in mind, that patients with acute liver injury and acute liver failure have elevated levels of plasma von Willebrand factor, the hypothesis had been put forward that perhaps inadequate regulation of von Willebrand factor in the context of liver injury was driving hepatic platelet accumulation or platelet hyperactivity, and this could potentially lead to liver damage, but this hadn't yet been tested in an experimental setting. So Daphna Huneveld, who's a postdoc in my lab, set out to test that hypothesis that von Willebrand factor contributes to hepatic platelet accumulation and liver injury after acetaminophen overdose in mice. These are some of the first data from these studies. Just to give you the background here, we used wild type or von Willebrand factor knockout mice. These are commercially available, were rederived from the Jackson Laboratory. These have been around for quite a while from Denise Wagner's group, generated these mice as a model of von Willebrand factor deficiency that's observed in humans. Daphna also developed a robust immunofluorescent labeling that allowed us to pick up von Willebrand factor in the liver as well as co-localize that with platelets that had accumulated in the liver marked by CD41 alpha 2B beta 3 integrin. So in liver sections from wild type mice, what you can appreciate is there are a few platelets in there that makes a lot of sense scattered throughout the lobule. VWF labeling is restricted to the large vessels, the portal vein and the central vein. This also makes sense insofar as those vascular endothelial cells are storing von Willebrand factor, as I noted earlier. Documenting the specificity of this stain when we stain livers from von Willebrand factor knockout mice. For von Willebrand factor, we don't see any green labeling. That makes sense. This is an antigen negative control showing specificity of the labeling, but you can still observe platelets scattered throughout the lobule as well as within the intima, for example, of the central vein. So 24 hours after challenge with a hepatotoxic dose of acetaminophen, in this case we use 300 mg per kg, what we observe is consistent with the prior studies of Dr. Roth insofar as there's a marked accumulation of platelets within those areas of necrosis, in particular the sinusoid surrounding those necrotic lesions, and there's dramatic colocalization with von Willebrand factor deposits within the injured liver 24 hours after challenge. Interestingly, our hypothesis had been that if we take away von Willebrand factor, we would see a dramatic diminution in the number of platelets accumulated in that injured liver, and that's simply not what we saw. While there seems to be a trend towards a reduction in platelet accumulation at this peak hepatotoxicity time point, there's still ample platelets that have accumulated in the liver indicated here by the fluorescent red staining in the von Willebrand factor knockout mice challenged with acetaminophen. So this suggests that von Willebrand factor is not a primary driving force for this early platelet accumulation in the acetaminophen-challenged liver, and I think that this sets the stage for our next talk from Dr. Ju, who will tell you probably the mechanism whereby that initial platelet accumulation occurs. Nonetheless, we were curious, and the primary motivation here was just to look at the role of von Willebrand factor over the full time course, examining both acute hepatotoxicity and repair of the injured liver, and lo and behold, what we found was that the stability of these VWF platelet aggregates, either that option or persistent platelet accumulation in the injured liver, was dramatically reduced in mice that lacked von Willebrand factor, depicted here between 24 and 48 hours, whereas there's clear retention of platelets within wild-type mice. In acetaminophen-challenged von Willebrand factor knockout mice, there's effectively a dramatic return to the normal number of platelets in that liver already by 48 hours, and that persists 72 hours after acetaminophen challenge, suggesting that von Willebrand factor deposits in the acetaminophen-injured liver and also those VWF deposits contribute to persistent platelet accumulation, alternatively stabilizing those platelet VWF aggregates. So what's then the impact of von Willebrand factor deficiency on acetaminophen-induced liver injury if those platelet aggregates have been destabilized? So to examine that, we simply quantified hepatocellular necrosis in wild-type and von Willebrand factor knockout mice challenged with acetaminophen, and while there was no difference in the amount of necrosis at 24 hours after acetaminophen challenge, the resolution and repair, the removal of this necrotic shrapnel was obviously accelerated in the von Willebrand factor knockout mice that were challenged with acetaminophen, and that can be observed here on the y-axis area of necrosis quantified from the H&E stained liver sections, and you can see this reduction in hepatic necrosis that occurred in von Willebrand factor knockout mice at both 48 and 72 hours. So that's all well and good, and what we wanted to do to confirm that this was a legitimate VWF effect was restore von Willebrand factor in those von Willebrand factor knockout mice, and that's because there's multiple reported phenotypes in these mice, and we simply wanted to provide that level of mechanistic clarity that von Willebrand factor was the bad actor here. So to evaluate that, we took von Willebrand factor knockout mice, and three hours after challenge with a hepatotoxic dose of acetaminophen, we administered either saline or a purified human von Willebrand factor concentrate called Humate P to restore von Willebrand factor levels to roughly physiological levels in the mouse, but again using this human von Willebrand factor, and then we examined liver necrosis 48 hours after acetaminophen challenge, and consistent with the hypothesis that von Willebrand factor inhibits liver repair, there was a restoration of hepatocellular necrosis at this 48-hour time point in von Willebrand factor knockout mice that had been repleted with this human von Willebrand factor, suggesting that indeed this is not some obtuse feature of the von Willebrand factor knockout mice. The observation, the impact on liver repair is a legit VWF effect. So to further sort of solidify that particular mechanism as well as begin to ask questions related to whether von Willebrand factor could be targeted as a, you know, a putative target to perhaps improve repair of the acetaminophen-injured liver, we sought to evaluate strategies that would allow us to do that in the context of wild-type mice, and lo and behold there are a number of emerging tools, and we'll talk about that in just a few minutes, but among the tools that have been around for quite a while include polyclonal antibodies developed from rabbits that target multiple epitopes in the von Willebrand factor protein, here showing an article from Dr. Ginsburg's laboratory in which they've mapped those VWF epitopes detected by one of these polyclonal antibodies that's commercially available from DACO, now Agilent, and what was nice about this particular von Willebrand factor polyclonal antibody was that it had at least been used in prior settings of liver pathology injury and regeneration and had documented effects on the accumulation of platelets within that tissue, here showing work from Don Lisman's laboratory where after partial hepatectomy there's a rapid accumulation of platelets within the liver sinusoids and administration of this particular polyclonal anti-VWF antibody caused a dramatic reduction in platelet accumulation within the liver, so these data gave us this proof-of-concept idea that this antibody could be used to block platelet accumulation within the liver and we could then use that tool to infer the impact of VWF platelet aggregates in liver repair. So this study was relatively simple as well. We simply took wild-type mice, these are C57 Black6J, and four hours after acetaminophen challenge, the mice were given either a control IgG antibody or anti-VWF antibody from DACO. Notably, we selected this four-hour time point afterwards to allow us to both avoid interference with that initial time of acetaminophen metabolism where NAPQI is generated as well as provide this initial proof-of-concept that in mice that had already developed a bit of hepatotoxicity that we could interfere with VWF platelet interactions and ultimately have downstream beneficial effects on the development of liver injury. So here we see hepatic necrosis in mice that were given acetaminophen as well as the control IgG quantified on the graph on the right, and lo and behold, when we administered the anti-VWF antibody, there was a significant reduction in hepatocellular necrosis 48 hours after acetaminophen challenge. So these results provide this concept that VWF promotes persistent hepatic platelet accumulation and that one of the downstream effects of this platelet accumulation is a delay in liver repair after acetaminophen challenge. And so this provides early evidence suggesting that blocking this interaction may actually accelerate liver repair. And while I don't have the opportunity to provide you with a full overview in the talk of all the data in the manuscript, I would certainly encourage you to go check out that paper that Daphna published I think now a year and a half ago or whatnot. So the obvious question with this is, this is great mouse doctoring and we've discovered these mechanisms and the intersection between VWF and platelets, but is there clinical or translational evidence for this? And we were absolutely thrilled to see a new publication in hepatology this year from Dr. Listman's group, again, in collaboration with the acute liver failure study group, in which they documented that not only are plasma von Willebrand factor levels increased in patients with acute liver injury and acute liver failure, they went so far as to document that in fact the elevated levels of von Willebrand factor were associated with a poor outcome in those patients with respect to a 21-day transplant-free survival, providing clinical evidence that in fact these changes in von Willebrand factor are deleterious in patients and these results are very much consistent with our mechanistic finding in the mice. So where do we go from this? I think there are a number of questions that remain unanswered, including defining the precise process is controlling the deposition, and I would contend durability, of von Willebrand factor deposits in the injured liver. We'd certainly like to uncover the mechanisms linking von Willebrand factor platelet aggregates to inhibition of liver repair, and those studies are ongoing in my laboratory. And I would very much suggest that we consider the question of whether emerging therapeutics that target von Willebrand factor directly could hold realistic promise as treatments in patients with acute liver injury. We absolutely need to keep in mind the delicate hemostatic rebalance in these patients, but targeting VWF may be realistic in this context. And we should also continue to ask the question of whether the mechanisms that we've discovered here in the context of acute hepatotoxicity are broadly applicable to chronic liver diseases as well, insofar as elevation of von Willebrand factor is also observed in patients with cirrhosis. So if you're curious about anything that I've talked about today, or this strikes you as something that you'd like to learn more about, I would absolutely refer you to a review article, and kind of a forward-looking review article written by Dr. Krooneveld in my laboratory and published recently in the Journal of Thrombosis and Hemostasis, in which she documents not only a lot of the biochemistry of von Willebrand factor and its role in the injured liver, but also raises some possibility of therapeutic agents that could be used to intervene. So with that, I'll close and I'll look forward to answering questions in the chat and to receiving emails and meeting with you all face-to-face in the future. Thank you so much. Hi, my name is Cynthia Zhu. I'd like to thank the organizers for inviting me to speak here. The title of my talk is chitinase 3-like-1 contributes to a simenophen-induced liver injury by promoting hepatic platelet recruitment. So chitinase 3-like-1, chitinase 3-like-1 in short, is a member of a family of enzymes called chitinase. So they degrade chitin through hydrolysis. Chitin is a glycopolymer found in fungal cell wall. So chitinase 3-like-1 is actually doesn't, is a member that doesn't have the chitinase enzymatic activity. But chitinase 3-like-1 is a very important protein because it has been associated with a battery of chronic inflammatory conditions. A GWAS study showed chitinase 3-like-1 is a risk factor for asthma development. More recently, chitinase 3-like-1 is considered a biomarker for Alzheimer's disease. Its level is increased in CSF and in astrocytes, and the levels correlate with the severity of the disease. Also, serum level of chitinase 3-like-1 is elevated, and it's an indicator of poor prognosis in many types of solid tumors. In terms of liver disease, serum level of chitinase 3-like-1 is elevated in liver fibrosis, NAFLD, alcoholic liver disease, and hepatocellular carcinoma. However, the function of chitinase 3-like-1 in liver diseases had not been well studied, and this area is actually a vacuum. We published the first study in 2018, where we found chitinase 3-like-1 has a role in triggering or promoting intrahepatic coagulation. We used a mouse model of concanavalin A-induced hepatitis. In this model, we found chitinase 3-like-1 promotes tissue factor production, which then triggers hypercoagulation in the liver and liver injury. So, after this study, we wondered if this is unique to con A-induced liver injury or if it also occurs in other models. So, we wanted to choose a model that occurs or an injury that occurs clinically. Based on the literature, a similar overdose-induced liver injury is accompanied by intrahepatic coagulation, and in fact, intrahepatic coagulation contributes to a simenophine-induced liver injury. There are many studies. I wanted to draw your attention to this publication in 2015 in Blood, where they showed platelets contribute to a simenophine-induced liver injury. They depleted platelets before they give the mice the simenophine, and they showed significant reduction of liver injury. We actually repeated the experiment, and we found the same thing, but we also tried to deplete platelets after simenophine, like three hours after simenophine treatment, and we were still able to see a dramatic reduction of ALT levels and liver histology. So, with that, we wanted to know if CHI3LAC1 plays a role in this simenophine-induced liver injury model. What we did first was we obtained liver and serum samples from patients with a simenophine overdose-induced liver failure. As you can see here, in patient samples, in patient's liver, we can detect a large number of CHI3LAC1-positive cells, but not in healthy livers. More dramatically, in serum of patients, we can see about 20-fold increase of CHI3LAC1 levels in comparing with serum samples from normal individuals. This also occurs in mice treated with a simenophine. Both the message and protein levels go up very rapidly after simenophine treatment. So, then we wondered if CHI3LAC1 has a role in this model. So, we treated the wild-type and CHI3LAC1 knockout mice with the simenophine. As you can see here, the knockout mice developed much more attenuated injury with much lower ALT levels at 6 and 24 hours, as well as smaller areas of liver necrosis, suggesting that CHI3LAC1 contributes to a simenophine-induced liver injury. So, to further confirm that, we injected the recombinant mouse CHI3LAC1 protein to knockout mice. And when we did that, we can see that liver injury was restored toward the wild-type level. ALT levels go up and necrotic areas also goes up. And what's really unexpected and amazing observation we found was in wild-type mice after simenophine treatment, very early on, like around…starting around two to three hours after simenophine, we can see accumulation of platelets in the liver, and they are in unique areas. They are around the central vein, and those are the areas that will be necrotic at a later time point. But we did not see any of that in CHI3LAC1 knockout mice. However, when we treated the knockout mice with the recombinant protein, we can recapitulate what we observed in wild-type mice. So, this really suggests that perhaps CHI3LAC1 is important in causing platelet accumulation. And we wanted to know how this protein does that. To answer that question, we asked two sub-questions. One is, because CHI3LAC1 is a soluble protein, we wanted to know what is its receptor, and we also wanted to know what is its target cell. We think that these will give us major clues in terms of how CHI3LAC1 promotes platelet accumulation. So, to address that first question, the receptor, we first did a screening, a proteomic screening, using liver tissues, and one candidate came out from our screening was CD44. And we did a series of experiments to see if CD44 indeed served as a receptor for CHI3LAC1. The first experiment was a pull-down assay. We used liver tissues from wild-type mice or CD44 knockout mice, and we used anti-CD44 antibody to pull down. And you can see here, the antibody can pull down CHI3LAC1 protein in wild-type liver but not in CD44 knockout. And Panel B is in vitro experiment, we transfected 293 cells with CD44, or the empty vector, and then we incubate the cells with His-tagged CHI3LAC1 protein. You can see here, only the cells that transfected with CD44, that expressed CD44, can bind to His-tagged CHI3LAC1. And Panel C is just incubating the recombinant CHI3LAC1 protein with CD44 protein. These are all human proteins, and then we use anti-CD44 antibody to pull down. You can see it can pull down CHI3LAC1 protein. And lastly, we did an interferometry study to measure the binding kinetics between CHI3LAC1 and CD44. We came up with a KD value of 250, and that is well within the range of receptor ligand binding. So, all of these experiments suggest perhaps CD44 act as a receptor for CHI3LAC1. If CD44 is the receptor mediating CHI3LAC1 function, we would expect that if we delete CD44, the mice should behave like CHI3LAC1 knockout mice with lower C-menifin-induced liver injury. So, this is the result comparing with wild-type mice shown in the filled circle. CD44 knockout mice had lower ALT levels, reduced liver necrosis, and much lower levels of fibrinogen and fibrin staining. And also, interestingly, we found that in terms of platelet accumulation, CD44 knockout mice behave very similarly to CHI3LAC1 knockout mice. There are almost no platelet accumulation in the liver, unlike the wild-type. Alternatively, we wanted to confirm the finding in CD44 knockout mice. Remember, I showed earlier that if we treat CHI3LAC1 knockout mice with the recombinant protein, we can restore liver injury to the wild-type level. So, we wondered if CHI3LAC1 protein can do that in CD44 knockout mice. And here is the result. The recombinant protein injection cannot change ALT levels, liver histology, or fibrinogen staining, suggesting that the recombinant protein needs CD44 receptor to function. And in the knockout mice, it doesn't do anything. So, all of these experiments suggest that CHI3LAC1 perhaps works through CD44, signals through CD44. All of these experiments were performed using CD44 knockout mice, but we wanted to use an alternative approach. So, here we used a CD44 blocking antibody. What we did was we treated CHI3LAC1 knockout mice with the recombinant protein. As I showed you earlier, this can restore liver injury to the wild-type level with high ALT levels, a dramatic increase of liver necrosis, and severe fibrinogen staining. However, if we block CD44 by the neutralizing antibody, the recombinant protein cannot change or cannot increase liver injury. So, all of these from different angles suggest that CHI3LAC1 functions through CD44. So, then our next question is, which cell that expressed CD44 and binds to CHI3LAC1? And as expected, there are a lot of cells in the liver that express CD44. It's a very common receptor, especially on immune cells. However, when we incubate these cells with CHI3LAC1, so this is CD44-positive cells. This peak represents cells that are not only CD44-positive, but also can bind to CHI3LAC1. And then we get on this peak, and using different cell surface markers to identify different cell types, we wanted to know which these cells are. And surprisingly, uniformly, all these cells express FRAT, suggesting there are liver macrophages. And so, with this revelation, we went back to the liver staining. So, we stained the tissues with macrophage markers, as well as platelet markers. And this, on the left, is human liver tissues. Brown staining represents macrophages, and the pink staining is for platelets. As you can see, in normal liver, there are a lot of macrophages, but in patients with a semen defense overdose, we see a lot of platelets accumulation, and these platelets, a lot of them overlap with macrophages. And this is mouse liver tissue with immunostaining of pink color representing macrophages, FRAT, and green representing platelets. As you can see, the white areas are areas where the platelets adhere to macrophages. So, these results really suggest that perhaps macrophages play an important role in platelet accumulation by adhering to the cells. If this is the case, we reasoned that if we deplete macrophages, we should reduce platelet accumulation. And here's the result. On the top panel, we treated mice with the empty liposomes containing PBS, and here you can see platelet accumulation, and they adhere to macrophages. However, when we treated mice with clodrinate-entrapped liposomes, we do not see any platelet accumulation. This suggests that macrophages are really important in recruiting and adhering to platelets. And our next question is, how do macrophages do this? What is the adhesion molecule that's involved? To address this question, we screened all the adhesion molecules that we can find from the literature that have been shown to play a role in platelet accumulation. And I wanted to draw your attention to this protein, podoplanin, because here we found that in both KY3LAC1 not clod and CD44 not clod mice, the expression is very, very low comparing with wild type. So, to further confirm, so this suggests that KY3LAC1 perhaps plays a role in the expression of this protein. To further confirm this, we isolated macrophages from the liver. And as you can see here, macrophages from KY3LAC1 not clod mice and CD44 not clod mice express very low levels of podoplanin comparing with wild type. However, if we treat macrophages with KY3LAC1 protein, we can induce the podoplanin expression in KY3LAC1 not clod cells. However, the protein cannot induce KY3LAC1 in CD44 not clod mice. So, this shows that KY3LAC1 can induce podoplanin expression in a CD44 dependent manner. Podoplanin, the only binding partner for podoplanin is Clk2, which is expressed by platelet. So, we did a series of experiments in vitro showing that podoplanin is important in mediating macrophage-induced platelet adhesion. And we wondered if this pathway is important in vivo. So, what we did was we treated the mice with anti-podoplanin antibody before we give them acetaminophen. As you can see here, after the antibody treatment, we cannot see any platelet accumulation in the liver, and liver injury is also dramatically reduced. Here is the ALT levels comparing with IgG-treated control mice. So, all of these data so far support our hypothesis that after acetaminophen treatment, KY3LAC1 is upregulated and released, and it binds to CD44 receptor on liver macrophages. And this KY3LAC1 CD44 signaling triggers upregulation of podoplanin, which through Clk2 on platelets leads to platelet adhesion and accumulation. And that results in intrahepatic coagulation and inflammation, and that contributes to hepatocyte injury. So, with this result, which really supports a critical role of KY3LAC1 in this acetaminophen-induced liver injury, we wondered if we can target this protein in treatment of overdose acetaminophen-induced liver injury. To do that, we generated a panel of monoclonal antibodies, and here on the top panel, you can see that we screened these antibodies, and we identified one clone that is very effective. When we treated mice with the anti-mouse KY3LAC1 antibody, we can abrogate the platelet accumulation in the liver. ALT is reduced significantly, and liver histology showed that necrosis area is reduced dramatically. So, this is a mouse monoclonal antibody against mouse protein. We also identified a clone of antibody that can effectively target the human protein. What we did was, we took advantage of our KY3LAC1 non-clonal mice, and the fact that the mouse and human protein share a large homology, about 80 percent. So, we inject the human protein to the KY3LAC1 non-clonal mice. As you can see here, this protein can induce platelet accumulation. It can increase ALT, and it can expand the areas of liver necrosis. However, if we treat the mice with the antibody against human KY3LAC1 protein, we can significantly reduce platelet accumulation, and ALT dropped dramatically, and liver histology showed necrotic area was diminished. So, with all these, we concluded that the KY3LAC1 plays a critical role in acetaminophen-induced liver injury through promoting platelet adhesion and recruitment into the liver. With that, I'd like to thank my team for their hard work, especially Dr. Sandy Shen, who is important and led this project, and also my collaborators from different institutions, Will Lee and Bing Gao, who helped us acquiring patient samples, and Dr. Zhiqiang An, who helped us generating the antibodies, and Feng, who helped us with intraviral microscopy studies. And I'd like to thank you for your attention.

proteins

von Willebrand factor

chitinase 3-like protein

CD44

drug-induced liver injury

anticoagulation

acetaminophen overdose

hepatic platelet recruitment

siminofin overdose

liver failure

podoplanin