Correspondence Dr Qin Ning, Department of Infectious Disease, Institute of Infectious Disease, Tongji Hospital of Tongji Medical College, Huazhong University of Science and Technology, 1095 Jie Fang Avenue, Wuhan 430030, China Tel: +86 27 8366 2391 Fax: +86 27 8366 2391 e-mail: firstname.lastname@example.org
Acute liver failure (ALF) is a syndrome of diverse aetiology, including hepatic encephalopathy, renal, cardiac and pulmonary failures, which result in a rapid loss of hepatic function. The mechanisms of liver injury contributing to ALF can be summarized into two categories: direct damage and immune-mediated liver injury. This review summarizes current concepts of immune-mediated liver injury from both clinical studies and animal models. We highlight immune responses of ALF from the liver injury perspective, which combines a variety of molecular and cellular mechanisms, particularly, the contribution of cytokines and the innate immune system. Hepatic and circulating inflammatory cytokines play a significant role in the pathophysiology of ALF including hepatocyte necrosis, extrahepatic complications and hepatocyte regeneration. Overproduction of cytokines, if unchecked, is hazardous to the host and may cause severe outcomes. Measuring pro-inflammatory cytokines in ALF may be of value for predictors of outcome. Innate and adaptive immune systems both involved in ALF contribute to immune-mediated liver injury. The innate immune response is activated much more rapidly compared with adaptive immunity, particularly in acute liver injury where the host has little time to trigger an effective adaptive immune response. From this point of view, the innate immune system may make a more profound contribution than the adaptive immune system. Furthermore, immune responses crosstalk with other physiological or pathophysiological factors, for example, coagulation factors which in turn determine the outcome of ALF and these are discussed.
Acute liver failure (ALF) is a syndrome of diverse aetiologies in which patients without liver disease recognized previously sustain a liver injury that results in a rapid loss of hepatic function. The period between the initial symptoms and the manifestations of hepatic encephalopathy is crucial for the prognosis of these patients. Therefore, several groups have included the time window between the appearance of symptoms and the onset of encephalopathy in their definitions. The most recent definition uses the term ALF as hyperacute, acute and subacute liver failure; fulminant or sub-fulminant liver failure. APASL consensus and Chinese consensus suggest that liver failure can develop as ALF (in the absence of any pre-existing liver disease), acute-on-chronic liver failure (ACLF) (an acute deterioration of known or unknown chronic liver disease) or a chronic decompensation of an end-stage liver disease (1, 2).
The aetiology of ALF is the most important determinant of outcomes. Differences in aetiology, outcomes and management are well understood. In the East and the developing world, ALF is mainly because of viral infections, primarily hepatitis B, but also A and especially in developing countries, E, as well as other non-hepatotropic viruses. In principle, four different mechanisms are primarily responsible: infectious (mainly viral), drugs/toxins/chemicals, cardiovascular and metabolic (3).
Clinically, patients with ALF develop a cascade of complications, including hepatic encephalopathy, and renal, cardiac and pulmonary failures, which are often presaged by a systemic inflammatory response syndrome. Although management strategies have been developed for ALF during recent decades, such as liver support systems that partially substitute for functions like detoxification and metabolic homeostasis (4), liver transplantation remains the most promising treatment for ALF. However, considering the diverse aetiologies and the shortage of donor organs, efforts have been dedicated to understand the massive immune-mediated liver injury and devise alternative strategies for regulating the imbalanced immune system.
Mechanisms of acute liver failure: perspectives from liver injury
The liver injury mechanisms that contribute to ALF can be classified into two groups. Firstly, pathogens and toxic substances directly damage cellular organelles or trigger cell signalling cascade pathways, which disturb intracellular homeostasis. Intervention of these pathways could protect against ALF (5–8). Secondly, to control liver injury progression, the immune system must distinguish between evasion and tissue damage, both of which elicit similar inflammatory immune responses. This underlines that the immune response can be a double-edged sword for the host (Fig. 1). The immune system's ability to distinguish between internal and external damage is likely to have a substantial impact on therapeutic approaches. All of these mechanisms converge on liver cell death pathways, including apoptosis, autophagy, necrosis and necroptosis (also known as programmed necrosis), combined with other physiological or pathophysiological factors that determine the outcome of ALF, which clinically manifests as organ failure.
During these processes, three factors primarily determine the prognosis for liver failure: the metabolic consequences resulting from liver failure, the release of mediators and toxic metabolites and the capacity of the remaining hepatocytes to restore the liver mass (9). Direct mechanisms of liver injury and liver cell death have been excellently reviewed by several groups (10–12). However, the immune mechanisms of liver injury have not been thoroughly scrutinized. Immune-mediated liver injury mechanisms that contribute to ALF involve both the innate and adaptive immune systems. Additional factors that influence the immune response in liver injury have also been reviewed.
Immunological mechanisms of acute liver failure
Distinctive characteristics of liver immunology
The liver receives blood from both the systemic circulation and the intestine, which is rich in microbial products, environmental toxins and external antigens. Increasing evidence suggests that the liver plays a key role in innate immune defences against pathogens and acts as an immune organ. The liver is rich in innate immune cells, such as Kupffer cells (KCs), natural killer cells (NK), natural killer T cells (NKT) and γδT cells and liver endothelial cells. The proportion of unique non-parenchymal cells provides a distinctive local immune environment. Also, the liver is rich in immunosuppressive cytokines, including interleukin (IL)-10, and several liver cell subsets express the inhibitory programmed death ligand 1(PD-L1), the consequence of which is that many encounters between T cells and liver antigen-presenting cells (APCs) result in immune tolerance (13–15).
Inflammatory cytokine cascades and ‘sepsis-like’ immune paralysis
The presence of opposing systemic pro- and anti-inflammatory profiles that result in organ failure and immune dysfunction are well recognized in septic shock. Studies have demonstrated that ALF, including ACLF, display ‘sepsis-like’ immune paralysis (16, 17), considering the features of systemic inflammation, cellular immune depression and progression to multiple organ dysfunction. There is an increasing evidence that activation of systemic immune responses plays a pivotal role in ALF pathogenesis and outcomes. One study demonstrated that cellular immune function was severely depressed in patients with ACLF and subjects with severe sepsis or septic shock, in which ex vivo tumour necrosis factor (TNF)-α secretion and monocyte human leucocyte antigen (HLA) DR expressions were significantly reduced (16).
Cytokines play a significant role in the pathophysiology of ALF including hepatocellular death, extrahepatic complications and hepatocyte regeneration. Cytokines-mediated liver injuries are closely related to hepatocyte proliferation and regeneration (18, 19). Hepatic and circulating inflammatory cytokines mediate a positive feedback loop on the innate immune system; overproduction of cytokines, if unchecked, is hazardous to the host and may cause severe outcomes. Suppressor of cytokine signalling (SOCS) family proteins SOCS1 and SOCS3, which negatively feedback for cytokine signalling, through the activation of the signal transducer and activator of transcription (STAT) and nuclear factor-κB(NF-κB)-mediated pathways, has been shown to be rapidly induced during liver injury (20–21). In Concanavalin A (ConA)-induced fulminant hepatitis, a study indicates that SOCS1 plays important negative roles in fulminant hepatitis and the prevention of hepatocyte apoptosis mediated by Fas and TNF-α (22). Recent data from animal models and humans indicate that TNF-α plays an essential role in the pathogenesis of ALF. TNF-α in the acute phase response is significantly elevated in the livers and sera of ALF patients. IL-6 and IL-10 ameliorate acute liver injury in murine models of ALF through their ability to down-regulate TNF-α production (23–25) and HBV-related ALF (26). In a fulminant hepatitis mouse model, TNF-α produced by intrahepatic non-Ag-specific inflammatory cells is critical in the development of lethal necroinflammatory liver disease (27).
Recent studies have shown imbalanced intrahepatic expressions of IL-12, interferon (IFN)-γ and IL-10 in fulminant hepatitis B and ACLF (28). Intrahepatic pro-inflammatory IFN-γ and TNF-α expressions were markedly upregulated in ACLF compared with chronic hepatitis B (CHB) or normal controls. IFN-γ overexpression was significantly correlated with increased CD4 and CD8 T-cell accumulations. TNF-α upregulation was also significantly correlated with increased KCs (29).
Many studies have evaluated circulating inflammatory cytokines in ALF, including TNF-α, IL-1, IL-6 and IL-8. Given the parallels between the levels of circulating inflammatory cytokines and the degree of systemic inflammation in ALF, a few studies have suggested that measuring pro-inflammatory cytokines may be of value for discriminating between survivors and non-survivors. This indicates the value of soluble effectors of immune response as predictors of outcome. Their inclusion into prognostic scoring systems should be evaluated (30).
Death receptor and perforin/granzyme-mediated immune responses
Death receptors are transmembrane proteins with three domains: an extracelluar ligand-interacting domain, a transmembrane domain and an intracellular death domain. Death receptors associated with the liver include Fas (CD95), tumour necrosis factor receptor 1 (TNF-R1), tumour necrosis factor-related apoptosis inducing ligand receptors 1 and 2 (TRAIL-R) and death receptors 3 and 6. Death receptors' engagement by their corresponding ligands, including Fas ligand (Fas-L), TNF-α and TRAIL trigger intracellular signalling pathways.
Hepatocytes, cholangiocytes, sinusoidal endothelial cells, stellate cells and Kupffer cells can express Fas (CD95) (31–33). Fas is activated when bound to membrane-bound FasL or soluble FasL (34). Fas is constitutively expressed on hepatocyte cell membranes and is over-expressed during liver damage. FasL is expressed on cytotoxic T lymphocytes, NK cells, NKT cells and hepatic macrophages (35). In animal experiments, antibody to Fas induced fulminant liver failure because of massive hepatocyte loss. In human ALF, hepatocytes strongly express Fas, Fas-L is upregulated on infiltrating lymphocytes and serum soluble Fas-L is markedly elevated (36). Several growth factors including hepatocyte growth factor (HGF) have been found to prevent apoptosis in animal study (37, 38).
Tumour necrosis factor receptor 1 and TNFR2 are both expressed on hepatocytes, although only TNFR1 expresses a death domain and initiates an apoptotic programme (39). TNFR1 can activate both pro- and anti-apoptotic signalling pathways following interaction with its hepatocyte receptor (TNF receptor) (40). The pathogenic role of TNF-α has been extensively documented in both experimental models of and in human ALF (41). The liver injury caused by TNF-α is often a part of multiple organ failure, as seen in sepsis and ALF or conditions like alcoholism, ischaemia/reperfusion and viral hepatitis (42–46).
Tumour necrosis factor-related apoptosis inducing ligand receptors 1 and 2 can induce apoptosis via caspase activation. Although TRAIL-mediated apoptosis is a well-characterized pathway of apoptotic cell death, there is little evidence that it plays a significant role in cell death in ALF. TRAIL receptor 1 and TRAIL receptor 2 are of intense interest and emerging significance in the pathogenesis of chronic liver diseases (47, 48).
Cytotoxic T lymphocytes (CTL) and NK cells together with their granule components, perforin and granzyme, play an important role in the defence against hepatic infections caused by different pathogens. Moreover, it has been shown in vitro that hepatocytes can initiate cell death in a perforin-dependent manner. For example, hepatocellular apoptosis in d-galactosamine/lipopolysaccharide (d-Gal/LPS)-associated liver failure is mediated not only by TNF-α-dependent Fas/FasL cytotoxicity but also involves a perforin/granzyme cell death pathway (49). The combination of perforin and granzymes significantly increases the lytic ability of CTLs (50). Target cells may repair perforin pores by endocytosing membrane fragments, although they simultaneously incorporate granzymes that induce DNA fragmentation.
Innate immunity and adaptive immunity
The innate immune system is the first line of defence against initial environmental challenges and injury. In the liver, it is a complex system that includes NK cells, NKT cells, KCs, as well as neutrophils, eosinophils and complement components (Table 1). The innate immune response is activated much more rapidly compared with adaptive immunity, particularly in acute liver injury where the host has little time to trigger an effective adaptive immune response. From this perspective, the innate immune system may make a more profound contribution than the adaptive immune system.
Table 1. Role of different cell types in liver injury
The complement system is a biochemical cascade that plays an important role in the pathogenesis of various diseases. Activation of this cascade triggers a wide range of cellular responses from apoptosis (cell death) to opsonization (antigen/antibody binding). Evidence suggests that the complement system is also involved in the pathogenesis of various liver disorders. Recent investigations have demonstrated that the complement participates in the pathogenesis of ethanol-induced liver injury, fulminant hepatitis and hepatic ischaemia–reperfusion injury (51–53). Complement is also involved with liver regeneration after partial hepatectomy or toxic injury. C3 or C5 deficiency results in diminished liver regeneration, accompanied by transient or fatal liver failure after partial hepatectomy (54). C5-deficient mice develop severely defective liver regeneration and persistent parenchymal necrosis after exposure to CCl4 (55). Both C3 and C5 participate in liver regeneration after liver injury or loss of tissue by enhancing STAT-3 and NF-κB, the two important signals for initiating the regenerative response (51).
Monocytes, macrophages and Kupffer cells
Monocytes and macrophages are central to systemic inflammatory response syndrome (SIRS) by secreting large quantities of pro-inflammatory cytokines and are responsible for antigen presentation through their surface expressions of HLA class II molecules. In addition to their activation after exposure to non-specific microbial stimuli, monocytes can, in turn, trigger adaptive immune responses (30). A recent study showed that the level of HLA-DR expression on monocytes in ACLF was decreased, and that the toll-like receptor (TLR) 4 expression on monocytes for patients with cirrhosis and ACLF was higher than for healthy controls (56).
Tissue-specific macrophages, KCs in the liver, have the capacity for phagocytosis and can release a broad array of cytokines that critically determine the subsequent reactions of other immune cells and hepatocytes, as well as the degree of organ damage. Upon receiving inflammatory signals, monocytes and macrophages are rapidly recruited to the liver, and these cells have functional profiles similar to KCs (57).
Specific mechanisms of liver injury induced by neutrophils were recently summarized (58, 59). Neutrophil-mediated liver damage is mediated by activation/priming and the subsequent accumulation of inflammatory mediators in the liver vasculature. Accumulated neutrophils sense chemotactic signals from the hepatic parenchyma, which leads to neutrophil extravasation and contact with hepatocytes. This triggers complete neutrophil activation with prolonged adherence-dependent oxidative stress and degranulation. The oxidants diffuse into hepatocytes and trigger intracellular oxidative stress, mitochondrial dysfunction and, eventually, cause oncotic necrotic cell death.
Hepatic infiltration by neutrophils is an early response to tissue injury, cellular stress or systemic inflammation. Neutrophils can kill hepatocytes in vivo within 1 h. This correlates with the appearance of intracellular oxidative stress and mitochondrial dysfunction (60). Neutrophils also express FasL and can kill hepatocytes that express Fas receptors via apoptosis (61). However, there are multiple possibilities for the involvement of the FasL/Fas system in the mechanisms of neutrophil-mediated liver injury, which requires additional study (59).
Natural killer and natural killer T cells
Natural killer cells represent a population of lymphocytes with potent cytolytic activities that are exerted through the death receptor and perforin/granzyme pathways. Hepatic NK cells are regulated by KC-derived cytokines, such as IL-12 and IL-18, as well as NKT cell-derived IL-4, and NK cells produce large amounts of IFN-γ upon activation (62). Hepatic NK cells modulate T cell responses in the liver, promote intracellular changes in endothelial cells and hepatocytes and can even directly promote hepatocyte death or cell lysis. During CCl4-induced liver fibrosis, NK cells can kill activated stellate cells that have increased natural killer group 2D (NKG2D) ligand retinoic acid early inducible gene (Rae) 1 (63). In addition, NK cells that express TRAIL show strong cytotoxic effects against primary hepatocytes in liver injury (64). Recent work from our group has shown that NK cells play a vital role in murine hepatitis virus strain 3 (MHV-3)-induced ALF (65). Following MHV-3 infection, the number of NK cells in the livers of Balb/cJ mice markedly increased, peaked at 48 h post-infection and remained at a high level until sacrifice. Expression of CD69, cytotoxic activity and intracellular IFN-γ and TNF-α production by liver NK cells at 48 h post-infection were all significantly up-regulated. Depletion of NK cells 24 h post MHV-3 infection increased the survival of mice from 0/18 (0%) to 4/18 (22.2%). Highly activated liver NK cells were cytotoxic for MHV-3-infected hepatocytes and this effect was markedly inhibited by anti-FasL plus anti-NKG2D monoclonal antibodies.
Natural killer T cells express surface markers of both T and NK cells, which are found at unusually high frequencies in the liver. From studies of experimental liver injury in mice after administering the plant-derived lectin ConA, NKT cells have been identified as critical elements that promote acute liver damage by releasing IL-4 and IFN-γ, and by the direct induction of Fas-mediated hepatocyte apoptosis (66). Further, NKT cells are involved in antiviral defence mechanisms (e.g. during CHB virus infection), as suggested by studies with transgenic mice (67).
Dendritic cells are extremely important in viral hepatitis-induced liver injury. In order to control rapidly replicating cytopathic virus infections, the immune system must respond rapidly to control viral replication and dissemination before tissue damage and inflammation endanger host survival (68). Secretion of type I IFNs is an essential component of the innate immune response against viruses. Nonetheless, before effective adaptive immune responses are elicited, type I IFN-mediated innate immune responses are critical for host survival during the early phase of infection. The first wave of type I IFNs is produced almost exclusively by plasmacytoid dendritic cells, leading to virus containment and disease prevention (69).
T cells and B cells
In idiosyncratic drug-induced liver injury, the drug or its metabolite covalently binds to cellular proteins to form hapten-carrier conjugates. This phenomenon, known as ‘haptenization’, proposes that drugs act as haptens and covalently bind to endogenous proteins to form drug–protein adducts. Hapten-like autoimmunity may involve the phagocytosis, processing and presentation of antigens (altered proteins) by APCs, activation of T-helper cells, induction of hapten-specific cytotoxic T cells and the production of autoantibodies by B cells against target antigens in the liver, including the drug (hapten), part of the carrier protein or both (70). T cells infiltrating the liver immediately after partial hepatectomy and gut-derived antigens are indispensable for liver necrosis and may thus provide therapeutic targets to ameliorate liver damage following partial hepatectomy (71). HBV-specific CD8+ T-cell responses are believed to be of considerable importance in viral control and immune-mediated liver damage (72). The development of methods such as MHC/peptide tetramer staining, intracellular cytokine staining and Elispot, which are able to quantify virus-specific CD8+ cells directly ex vivo, has permitted a more accurate analysis of HBV-specific CD8+T cells during the different phases of HBV infection. During acute HBV infection, virus-specific CD8+ T-cell responses are often readily detectable and are multispecific (73–75), while during chronic HBV infection these responses are generally weak and can become exhausted (76–78). HBV-specific CD8+ T cells may also directly kill virus-infected hepatocytes and contribute to liver pathology (72, 79–80). One study compared the transcriptional profiles of CD4+ T cells of healthy controls and patients infected by acute hepatitis B (AVH-B) or CHB using a custom expression array comprising 350 genes related to CD4+ T cell and Treg functions (81). The fingerprints enabled a clear discrimination between patients suffering from AVH-B or CHB infection. The observed profiles suggested the accumulation of effector T cells with a potential role in necro-inflammation during the acute stage. Subsequent down-regulated effector functions may result in suppressed CD4+ effector T cells and viral persistence during the chronic infection stage. In a mouse model of fulminant autoimmune hepatitis, liver damage is driven by CD4+ T cell production of IFN-γ, and independent of both CD8+ T cells and the Fas ligand/Fas pathway (82). A study investigated the role of STAT3-in Con A-induced T cell-mediated hepatitis. Myeloid STAT3 activation inhibits T cell-mediated hepatitis via the suppression of a Th1 cytokine (IFN-γ) in a STAT1-dependent manner, whereas STAT3 activation in T cells promotes T-cell hepatitis to a lesser extent, via an induction of IL-17 (83). The frequency of circulating Th17 cells increased with disease progression from CHB to ACLF patients vs healthy control. Th17 cells were also found to be largely accumulated in the livers of CHB patients. The increases in circulating and intrahepatic Th17 cells positively correlated with plasma viral load, serum alanine aminotransferase levels and histological activity index. In addition, the concentration of serum Th17-associated cytokines was also increased in CHB and ACLF patients (84).
Regulatory T cells also play a crucial role in the control of immune responses to hepatic immune responses. CD4+CD25+ Treg play an important role for limiting liver injury in Con A-induced hepatitis via a TGF-β-dependent pathway (85), and IL-10 is crucial for tolerance induction in Con A-induced hepatitis and is mainly expressed by CD4+CD25+ Treg and KC (86). Another study demonstrated that fibrinogen-like protein 2 (Fgl2) was an important effector cytokine for Treg that contributed to susceptibility to MHV-3-induced fulminant hepatitis (87).
Liver sinusoidal endothelial cells
Liver sinusoidal endothelial cells not only constitute a physical barrier between the sinusoidal lumen and parenchyma but also actively participate in acute liver injury for both hepatic and systemic inflammatory conditions. A distinct function of these cells is their ability to clear pro-inflammatory substances, such as lipopolysaccharide (LPS), from sinusoidal blood without inducing widespread inflammation. This unique ability is mediated through an enhanced production of anti-inflammatory mediators and reduced expressions of adhesion and antigen presentation/co-stimulator molecules during exposure to pro-inflammatory mediators, such as LPS (30). These cells also express MHC class I and class II and costimulatory molecules characterized by active stimulatory APCs (88). However, they respond to TLR4 ligation with the secretion of IL-10, to which they also respond by downregulating their APC functions (89), and their main effect appears to be the induction of T cell tolerance (13). Liver sinusoidal endothelial cell lectin, a recently identified member of the DC-SIGN family, specifically recognizes activated T cells and negatively regulates their immune responses (90). In warm hepatic ischemia/reperfusion injury, hepatic sinusoidal endothelial cells play a pivotal role as antigen-presenting cells by expressing B7-1 and B7-2 (91). In LPS and β-galactosamine (GA) or carbon tetrachloride-induced ALF, IL-1-induced matrix metalloproteinases by HSCs within the space of Disse and thereafter extracellular matrix degradation may provoke the collapse of sinusoids, leading to parenchymal cell death and loss of liver functions (92).
Eosinophils are observed in several liver diseases, but their contributions in the pathogenesis of these disorders remain poorly investigated. In Con A-induced immune-mediated liver injury, liver eosinophil infiltration and IL-5 production have been observed. Liver injury was dramatically reduced in IL-5-deficient or eosinophil-depleted mice. Activated NKT cells also produce IL-5, a critical cytokine for eosinophil maturation and function. These results highlight the pathologic role of IL-5 and eosinophils in experimental immune-mediated hepatitis (93). Fulminant hepatic failure patients showed a high number of intrahepatic eosinophils concomitant with an increased expression of IL-6. Besides, the IL-6-positive eosinophils associated with the lack of IL-5 (94).
Crosstalk with other physiological and pathophysiological mediators
Coagulation pathway and platelets
Activation of the coagulation cascade is an integral component of host inflammation. Intimate interactions exist between inflammatory cascades and coagulation: not only is coagulation activated by many bioactive substances, including endotoxin, cytokines, bacterial products and viruses, but experimental evidence from animal models also indicates that the coagulation cascade plays crucial roles in the outcomes of septic and inflammatory insults. Infusing neutralizing antibodies for these potent coagulants either ameliorates or prevents these inflammatory conditions, supporting their role in the pathogenesis of these diseases (95, 96). Activated protein C, a potent anticoagulant serine protease, has been shown to have cell-protective properties by virtue of its anti-inflammatory and anti-apoptotic activities (97).
In mouse models of acute viral hepatitis, platelet depletion reduces intrahepatic accumulation of virus-specific CTLs and organ damage. Also, anticoagulant treatment that prevents intrahepatic fibrin deposition without reducing platelet counts does not avert liver injury (98). In reperfusion using the cold ischaemic rat liver model, sinusoidal endothelial cell apoptosis is a central feature of reperfusion injury. One study demonstrated that platelets caused sinusoidal endothelial cell apoptosis upon reperfusion of liver grafts and that preventing platelets adhesion was protective (99).
Fibrinogen-like protein 2
Fibrin deposition and thrombosis within the microvasculature are now appreciated as playing a pivotal role in the hepatocellular injury observed in viral hepatitis. Importantly, the pathways by which fibrin generation is elicited in viral hepatitis may be mechanistically distinct from the classical pathways of coagulation induced by mechanical trauma or LPS. Activated endothelial cells and macrophages express distinct cell-surface procoagulants, including a novel pro-thrombinase, Fgl2/fibroleukin, which is important for both the initiation and localization of fibrin deposition. There is a critical role for Fgl2/fibroleukin in the pathophysiology of fulminant viral hepatitis in mice and severe hepatitis B or acute-on-chronic liver failure. Human fgl2 (hfgl2) was detected in 21 of 23 patients (91.30%) with severe ACLF because of hepatitis B (100–102). Hepatitis B virus induced hfgl2 transcription is dependent on the c-Ets-2 and MAPK signalling pathways (103). There was a positive correlation between hfgl2 expression and the severity of liver disease, as indicated by the bilirubin levels. Measuring the hfgl2/fibroleukin expression in peripheral blood mononuclear cells (PBMC) may be a useful marker for monitoring the severity of acute-on-CHB and both hfgl2 and c-Ets-2 as potential targets for therapeutic intervention.
In acute liver injury, fibronectin (FN) is deposited at sites of hepatocellular necrosis. In contrast to quiescent hepatic macrophages, liver inflammatory mononuclear phagocytes synthesize abundant amounts of FN (104). Plasma fibronectin is a glycoprotein synthesized in the liver that aids with the clearance of circulating microbes and microbial products via their opsonization and subsequent uptake and clearance by Kupffer and other cells of the reticuloendothelial system (RES). The circulating levels of this opsonic molecule are markedly reduced in ALF. Studies in humans have shown a correlation between low serum fibronectin concentrations and impaired Kupffer cell functions, as documented by the low systemic clearance of micro-aggregated albumin and by an increased predisposition to infection and a higher mortality rate (30).
The early stages of alcohol-induced liver injury involve chronic inflammation. A recent study suggested that enhanced LPS-induced liver injury caused by ethanol was mediated, at least in part, by fibrin accumulation in the liver, mediated by inhibiting fibrinolysis by plasminogen activator inhibitor-1. These results also support the hypothesis that fibrin accumulation may play a critical role in the development of early alcohol-induced liver injury (105).
Neurohumoural regulatory system: sympathetic nerve system, oestrogen and thyroid hormone
During the past decade, the nervous system has become increasingly recognized as a significant modulator of inflammatory processes and immune functions (106, 107). Hepatic sympathetic nerves not only control various physiological functions but also contribute to liver injury. Studies suggest that nor epinephrine released from hepatic sympathetic nerves plays a critical role in protecting the liver from Fas-mediated fulminant hepatitis, possibly via mechanisms that include anti-apoptotic proteins and IL-6 (108). Modulation of autonomic nervous system functions may open novel therapeutic strategies for immune and inflammatory liver diseases (106). In addition, direct electrical stimulation of the peripheral vagus nerve in vivo during lethal endotoxaemia in rats inhibited liver TNF synthesis, attenuated peak serum TNF amounts and prevented the development of shock (109).
Host hormones also play important roles in liver diseases (110–113). Females are generally considered to be more susceptible to alcohol-induced liver injury than males. In animal models, ethanol increased liver weight and fat accumulation, an effect that was minimized by ovariectomy and partially reversed by oestrogen replacement. Infiltrating leucocytes are increased by ethanol (114). These effects were significantly blunted by ovariectomy and reversed by oestrogen replacement, possibly associated with an oxygen radical-mediated mechanism.
Recently, studies reported that thyroid hormone-induced calorigenesis upregulated the hepatic expression of mediators promoting cell protection in partial hepatic ischaemia–reperfusion. Triiodothyronine [T (3)] administration to rats induced significant depletion of reduced liver glutathione (GSH), with higher protein oxidation, O2 consumption and KC function (115). In a surgical liver devascularization model, Serum thyroxin [T (4)] and T (3) levels markedly decreased, whereas free-triiodothyronine and thyroxin-stimulating hormone levels did not change. T (4) and T (3) levels correlated with the degree of liver failure and with malondialdehyde (MDA) and interleukin-6 levels. The downregulation of T (4) and T (3) levels during ALF appears to correlate well with the severity of disease. This downregulation related to inflammation and oxidative stress and resulted in changes in myocardial thyroid receptors (116).
A variety of animal models have been used to study the mechanisms of acute liver injury or failure, such as those induced by chemical substances (ConA, PolyI:C, D-GaLN, a-Galcer, LPS, APAP, CCl4, CpG, DMSO, CD40L, amanitin, azoxymethane, thioacetamide and carrageenan), metabolic substances (ethanol, high fat, sulphatide and NAD), infectious pathogens and surgery models. Viral hepatitis is a major cause of ALF in Asian countries; however, the use of viruses to develop animal models of ALF has generally been unsuccessful. This review focuses on virus-induced ALF animal models. For other models, readers are referred to recent excellent reviews (117, 118). The standards for ALF animal models are summarized in Table 2.
Table 2. Requirement for model of acute liver failure
Hepatic failure should be potentially reversible to allow survival if an effective treatment is utilized
Nearly universal mortality without treatment
Death from liver failure
Selective liver damage should result in death
Time to death should be long enough to allow initiation and assessment of treatment
Minimal hazard to personnel
Toxins should present minimal risk for laboratory personnel
The species utilized should have metabolic and physiological properties similar to those of humans
Murine hepatitis virus strain 3-induced acute liver failure model
MHV-3 infection in BALB/cJ mice that caused ALF is characterized by macrophage activation and a marked production of pro-inflammatory mediators and massive hepatocellular necrosis and apoptosis (95). Following MHV-3 infection, the virus replicates predominantly, but not exclusively, in the liver, and within 24–48 h, evidence of macrophage activation and sinusoidal thrombosis is observed. Infiltrating mononuclear cells and neutrophils are observed in areas of hepatic necrosis. The viral load itself does not appear to be responsible for the development of massive liver injury. It has been shown that, following MHV-3 infection, macrophages and endothelial cells express the pro-thrombinase fgl2, which cleaves pro-thrombin to the active moiety thrombin. This results in fibrin deposition in the sinusoids, disturbances of hepatic microcirculation and hepatocyte necrosis.
Our recent work also found that increased killing of liver NK cells by Fas/FasL and NKG2D/NKG2DL contributed to hepatocyte necrosis in virus-induced liver failure (65). Similar phenomena have also been observed in the liver and PBMCs from patients with HBV-induced ACLF (Fig. 2). In a mouse model of MHV-3-induced hepatitis, we found that telbivudine treatment enhanced survival and improved clinical conditions and histological lesions. This was because of telbivudine preserved T-helper 1 cytokine production that enhanced the ability of T cells to proliferate and secrete cytokines (119).
Transgenic mouse model of acute liver failure
Transgenic expression of HBsAg in inbred mice led to a fatal necro-inflammatory liver disease model for examining HBV-induced fulminant hepatitis (118, 120). Injecting MHC class I molecules restricted HBsAg-specific spleen cells, and transferring cloned HBsAg-specific CTLs into HBsAg-transgenic mice induced an acute necro-inflammatory liver disease (121). The development of disease is divided into three stages, ranging from single cell necrosis to the massive destruction of most hepatocytes.
In stage one, within 1 h of injecting CTLs, a few hepatocytes undergo apoptosis as a result of direct interactions with the specific CTLs. The second stage, occurring at 12–24 h post-injection of CTLs, is characterized by increased hepatocellular apoptosis and the formation of necro-inflammatory foci. Non-HBsAg-specific inflammatory cells, especially radiosensitive mononuclear cells and neutrophils, are pivotal for this injury. The third stage, at 24–72 h post-injection, is characterized by massive liver necrosis, inflammatory cell infiltration and hyperplasia of sinusoid lining cells (KCs). This histopathological feature matches the pathogenic liver changes observed in patients who have HBV-induced fulminant hepatitis.
Non-HBsAg-specific inflammatory cells, typically of macrophage origin, play a critical role in the massive hepatocellular necrosis. Prior administration of antibodies to IFN-γ reduces cell death by over 97%, demonstrating the importance of this cytokine in the development of fulminant hepatitis, perhaps through the activation of macrophages. Although the transgenic HBsAg model is an elegant means of dissecting the pathogenic mechanisms of FHF, the model has limitations in that it differs markedly from the clinical situation, in which replicating viruses are present (95).
Rabbit haemorrhagic disease virus -induced ALF model
Rabbit haemorrhagic disease virus (RHDV) was reported recently as a new animal model of ALF that caused an acute and highly fatal disease in wild and domestic rabbits (122). Viral antigens could be found in hepatocytes at 12 h post-infection and at 36–48 h they were found in 60–80% of hepatocytes (123). Rabbits died within 36–54 h with clinical signs of progressive ALF and coma. This model reproduced representative biochemical and histological parameters and clinical signs of human ALF. This model may have advantages for collecting blood samples and makes monitoring easier for intracranial pressure and biochemical alterations that occur during the course of infection. Also, significantly increased inducible nitric oxide synthase and TNF-α activity are observed (124). Therefore, RHDV experimental infection induces an ALF in rabbits that has a number of physiological and biochemical features seen clinically in humans, is highly reproducible, has a long therapeutic window and generates intracranial hypertension and an associated encephalopathy. Furthermore, this model could provide a useful tool for the study of ALF and the evaluation of new liver support technologies for humans.
This review highlights the recent understanding of immune-mediated ALF, from animal models to studies involving samples from patients with ALF and HBV-induced ACLF. The importance of cytokines, the innate immune system and accompanying pathophysiological processes, particularly coagulation factors, have been discussed. Given the complexities of the various aetiologies and host immune responses, the challenge is to provide a rational immune basis for managing patients with ALF and ACLF, by either enhancing or suppressing the unbalanced immune response.
This work was supported by the National Key Basic Research Program of China (2007CB512900).