Role of Neutrophils in the Pathogenesis of Acute Inflammatory Liver Injury

Abstract

Polymorphonuclear leukocytes (neutrophils) are essential in the defense against invading microorganisms, tissue trauma or any inciting inflammatory signals. Hepatic infiltration of neutrophils is an acute response to recent or ongoing liver injury, hepatic stress or unknown systemic inflammatory signals. Once neutrophils reach the liver, they can cause mild-to-severe tissue damage and consequent liver failure. For neutrophils to appear in the liver, neutrophils have to undergo systemic activation (priming) by inflammatory mediators such as cytokines, chemokines, complement factors, immune complexes, opsonized particles and other biologically active molecules, e.g., platelet activating factor. Neutrophils accumulated in the hepatic microvasculature (sinusoids and postsinusoidal venules) can extravasate (transmigrate) into the hepatic parenchyma if they receive a signal from distressed cells. Transmigration can be mediated by a chemokine gradient established towards the hepatic parenchyma and generally involves orchestration by adhesion molecules on neutrophils (β₂ integrins) and on endothelial cells (intracellular adhesion molecules, ICAM-1). After transmigration, neutrophils adhere to distressed hepatocytes through their β₂ integrins and ICAM-1 expressed on hepatocytes. Neutrophil contact with hepatocytes mediate oxidative killing of hepatocytes by initiation of respiratory burst and neutrophil degranulation leading to hepatocellular oncotic necrosis. Neutrophil-mediated liver injury has been demonstrated in a variety of diseases and chemical/drug toxicities. Relevant examples are discussed in this review.

Introduction

The liver plays an important role in the innate immune response, providing the first line of defense against microbes and toxins crossing the intestinal barrier (Janeway, 2002). Kupffer cells, the resident macrophages in the liver, are critical for the rapid clearance of microorganisms from the systemic circulation (Gregory and Wing, 2002; Nagy, 2003). Although Kupffer cells by themselves are highly phagocytic and are able to remove microorganisms, they also facilitate generation of inflammatory response leading to recruitment of inflammatory cells such as neutrophils, monocytes, T and B lymphocytes, as well as natural killer (NK) cells, to the liver. Mediators of neutrophilic inflammation generated by Kupffer cells include tumor necrosis factor α (TNF-α), interleukins (IL-1, and IL-6), chemokines and reactive oxygen species (ROS) (Sweet and Hume, 1996). These inflammatory mediators will then activate neutrophils in the hepatic microvasculature (sinusoids and postsinusoidal venules) leading to a variety of events culminating in hepatocellular death (Figure 1). Ischemia-reperfusion injury (IRI), toxicity of ethanol (ETOH), acetaminophen (APAP), and alpha-naphthyl-isothiocyanate (ANIT) are examples, which are discussed in this review, where neutrophils contribute to or even dominate the injury mechanism.

General Mechanisms of Neutrophil-Mediated Liver Pathology

Liver injury mediated by neutrophils has been reported in a number of experimental animal models such as ischemia-reperfusion injury (Jaeschke et al., 1990), endotoxemia (Jaeschke et al., 1991), alcoholic hepatitis (Bautista, 2002; Jaeschke, 2002; Ramaiah et al., 2004), obstructive cholestasis (Gujral et al., 2003), ANIT toxicity (Hill et al., 1999; Xu et al., 2004) and APAP toxicity (Liu et al., 2006). Drug/chemical-related liver inflammation can result from chemical/drug-induced oxidant stress and tissue injury and/or by the local upregulation of inflammatory mediators. Whatever the stimulus, neutrophil invasion often exacerbates the liver damage caused by the drug/chemical. To understand the mechanism of liver injury, it is important to review the pathophysiology of neutrophil recruitment into hepatic vasculature, which includes neutrophil activation within the hepatic vasculature, neutrophil transmigration or extravasation and adherence of neutrophils to hepatocytes causing a respiratory burst and oxidant stress-mediated killing of hepatocytes.

Neutrophil Accumulation in the Liver

Neutrophils accumulate within the hepatic microvasculature, which includes sinusoids and in postsinusoidal venules, before transmigration. A variety of inflammatory mediators such as TNF-α, IL-1, CXC chemokines [(e.g., IL-8, Macrophage inflammatory protein-2 (MIP-2), keratinocyte-derived chemokine (KC), cytokine-induced neutrophil chemoattractant-1 (CINC-1)], and platelet activating factor (PAF)), can mediate neutrophil accumulation within the hepatic microvasculature (Schlayer et al., 1988; Bautista and Spitzer, 1992; Witthaut et al., 1994; Essani et al., 1995; Bajt et al., 2001). These mediators up-regulate the expression of Mac-1 (CD11b/CD18), a member of the β₂ integrin family of adhesion molecules, on neutrophils resulting in priming and activation of these phagocytes (Jaeschke, 2006). In contrast to adhesion in postsinusoidal venules, neutrophil accumulation within sinusoids does not require β₂ integrin adhesion molecules (Jaeschke, 2000, 2003). However, neutrophil activation/accumulation within sinusoids and postsinusoidal venules typically does not cause liver damage (Jaeschke, 2006). In general, extravasation into the parenchyma is necessary for neutrophils to cause tissue damage (Chosay et al., 1997).

Neutrophil Extravasation into Liver Parenchyma

Neutrophil extravasation from the hepatic microcirculation (typically sinusoids) into the parenchyma is facilitated by β₂ integrins (Jaeschke and Hasegawa, 2006). Integrins are constitutively expressed in low affinity state (e.g., CD11a/CD18, CD11b/CD18) and can be activated to a high affinity state on the cell surface, which enables the neutrophil to firmly adhere to the sinusoidal endothelium by binding to endothelial ICAM-1 and to extravasate if a chemotactic signal is received. Chemokines represent a large family of chemotactic peptides with a broad range of cellular targets and can be produced by hepatocytes, sinusoidal endothelium, cholangiocytes, Kupffer cells and stellate cells. Chemotactic signals include CXC chemokines such as IL-8, MIP-2, KC or CINC-1, which are potent chemoattractant for neutrophils (Bajt et al., 2001; Maher et al., 1997; Okaya and Lentsch, 2003). Osteopontin (OPN), also known as secreted phosphoprotein-1, localized within the biliary epithelial cells is another potential chemokine that can possibly mediate neutrophil chemotaxis (Apte et al., 2005; Banerjee et al., 2006a, 2006b).

It has been reported that increased production of CXC chemokines within hepatocytes can cause neutrophil infiltration, extravasation, and consequent neutrophil-dependent cytotoxicity (Colletti et al., 1996; Maher et al., 1997; Lentsch et al., 1998). However, CXC chemokine formation does not always lead to neutrophil extravasation (Simonet et al., 1994; Dorman et al., 2005). For CXC chemokines to have an effect, the mediators need to be produced in the right location to establish a chemotactic gradient, at the right time and in sufficient quantities relative to other potential chemoattractants.

Neutrophils do not attack healthy cells but respond to distressed or dying cells (Jaeschke, 2006). Thus, apoptosis is known to trigger neutrophilic inflammation (Figure 1, Jaeschke et al., 1998; Lawson et al., 1998; Faouzi et al., 2001). Blocking apoptosis by pancaspase inhibitors effectively prevented neutrophil extravasation into the parenchyma, suggesting a role of apoptosis in mediating neutrophil extravasation (Jaeschke et al., 1998). Although it has been shown that apoptotic hepatocytes can generate enough CXC chemokines to trigger neutrophil infiltration into liver (Faouzi et al., 2001), in the presence of other potent proinflammatory mediators, this effect appears to be of limited relevance (Ito et al., 2006). Direct cell contact of the neutrophil with apoptotic hepatocytes through endothelial cell gaps may have induced extravasation (Ito et al., 2006). Necrosis can also incite additional neutrophilic infiltration either by cell content release through complement activation or by reactive oxygen species (ROS) generation. Lipid peroxidation products including lipid aldehydes are potent chemotactic factors for neutrophils (Jaeschke, 2000).

In addition to mediators originating from the injured cells, the activated neutrophils can also generate potent chemotactic factors such as leukotriene B₄, which can recruit more neutrophils into the inflammatory site as seen during ischemia-reperfusion injury (Jaeschke, 2003).

Adherence of Neutrophils, Respiratory Burst and Cytotoxicity

Once neutrophils are in the parenchyma, they use β₂ integrins to make contact with their target by utilizing ICAM-1 on hepatocytes (Jaeschke, 2006). The engagement of neutrophil β₂ integrins, in particular CD11b/CD18, triggers a long-lasting oxidant stress through NADPH oxidase in close proximity to the target. Superoxide generated by NADPH oxidase dismutates to oxygen and hydrogen peroxide, which is a highly diffusible oxidant (Jaeschke, 2003; Gujral et al., 2004b). Myeloperoxidase released from the primary granules of neutrophils can generate hypochlorous acid resulting in the formation of chloramines (Bilzer and Lauterburg, 1991). Both species are highly toxic to cells. There is direct evidence for an intracellular oxidant stress in hepatocytes induced by neutrophil-derived hydrogen peroxide (Jaeschke et al., 1999) and hypochlorous acid (Gujral et al., 2004b). The fact that deficiency of glutathione peroxidase aggravated neutrophil-induced liver injury (Jaeschke et al., 1999) and that inhibition of NADPH oxidase protected against neutrophil cytotoxicity (Gujral et al., 2004b) suggests that these oxidants are the dominant mediators of the injury mechanism in vivo. In addition to ROS, neutrophil-derived serine proteases are known to contribute to hepatocyte damage (Jaeschke and Smith, 1997a, 1997b).

TNF-α, IL-1, and IL-18 proforms are processed by neutrophil proteinase-3 (a neutrophilic protease) to generate their active forms; inhibition of serine proteinase-3 is known to eliminate toxic liver injury mediated by galactosamine and endotoxin (Niehorster et al., 1991). Similarly elastase can process the membrane-bound protransforming growth factor into the soluble active form, which can then activate its receptor (Wiedow and Meyer-Hoffert, 2005). Elastase can also induce the formation of neutrophil- and monocyte-derived chemokines in Kupffer cells (Yamaguchi et al., 2000). Thus, neutrophil serine proteases can process cytokines, chemokines, growth factors, and their receptors, resulting in an aggravated inflammatory response and increased neutrophil cytotoxicity. In addition, neutrophil-derived proteases are able to directly cause hepatocellular injury (Ho et al., 1996). The direct cytotoxic effects of proteases may contribute to the overall injury process during a more prolonged neutrophilic hepatitis.

In addition to the described mechanisms for the amplification of the inflammatory response, an even more complex picture begins to emerge. Damaged cells release cell contents into the tissue and blood. Among others, liver cells can release calpains, which are calcium-regulated cytosolic cysteine proteases that exists in several isoforms; activation of these enzymes has been implicated in liver damage mainly by progression of liver injury (Limaye et al., 2003, 2006; Mehendale and Limaye, 2005). Experimental intervention with cal-pain inhibitors substantially mitigates progression of liver injury initiated by toxicants (Rose et al., 2006). Similarly, the release of deoxyribonuclease 1 during cell necrosis has been suggested to promote the expansion of the injury (Napirei et al., 2006). On the other hand, it was reported that necrotic cells release of high mobility group box-1 (HMGB-1) protein, which can bind to the toll-like receptor-4 on macrophages and promote cytokine formation, which causes neutrophil recruitment into the liver (Scaffidi et al., 2002). Thus, necrotic cell death resulting from killing by neutrophils or by direct toxicity of drugs/chemicals can directly and indirectly amplify the neutrophilic inflammatory response by a complex network of mediators and mechanisms.

Specific Examples of Neutrophil-Mediated Liver Toxicity

Ischemia-Reperfusion-Mediated Liver Injury

During transplantation, liver surgery (Pringle maneuver), and hemorrhagic shock, the liver can be exposed to prolonged periods of ischemia, which results in cellular stress and cell death during ischemia itself and also during reperfusion. In fact, the inflammatory response triggered by the early injury is a critical factor in the organ damage during reperfusion (Jaeschke, 2003). The acute inflammatory response during reperfusion consists of 2 phases: 1. A Kupffer cell-mediated injury phase (0–6 h of reperfusion) where Kupffer cells generate reactive oxygen and aggravate the ischemic damage (Jaeschke and Farhood, 1991; Jaeschke et al., 1991a). In addition, activated Kupffer cells and infiltrating lymphocytes generate a number of cytokines, which further promote the inflammatory response (Colletti et al., 1990; Lentsch et al., 2003). 2. A neutrophil-induced injury phase (6–24 h of reperfusion), where neutrophils are fully activated, generate reactive oxygen and dominate the injury process (Jaeschke et al., 1990, 1992; Hasegawa et al., 2005). Oxidant stress and injury by neutrophils can be strongly attenuated by blocking antibodies directed against CD11b (Jaeschke et al., 1993b) or CD18 (Liu et al., 1995) and by inhibiting NADPH oxidase (Hasegawa et al., 2005). However, blocking ICAM-1 was markedly less effective in this model (Farhood et al., 1995). These findings suggested that the extravasation process was less dependent on integrin/ICAM-1 interactions.

The extensive endothelial cell damage during ischemia and reperfusion facilitates the extravasation process without the need for adhesion molecule involvement. On the other hand, the adherence to the target cells and most importantly the adherence-dependent oxidant stress depends strictly on CD11b/CD18. These observations during reperfusion are different compared to other models of neutrophil-induced injury such as obstructive cholestasis. Here, the extravasation process into the parenchyma is dependent on β₂ integrin/ICAM-1 interactions and the reduced oxidant stress and injury in animals deficient of these adhesion molecules is mainly caused by the prevention of neutrophil extravasation (Gujral et al., 2003, 2004a).

Neutrophil activation and recruitment during reperfusion is mediated by TNF-α and activated complement factors (Colletti et al., 1990; Jaeschke et al., 1993a). TNF-α also induces, through activation of the transcription factor NF-κB, the upregulation of ICAM-1 on liver cells (Colletti et al., 1998) and the formation of CXC chemokines (Colletti et al., 1996; Lentsch et al., 1998), which may trigger extravasation of neutrophils. Complement factors not only directly activate (upregulation of CD11b) and prime neutrophils for ROS formation (Jaeschke et al., 1993a) but also promote a Kupffer cell-induced oxidant stress and injury, which indirectly enhances the neutrophilic response (Jaeschke et al., 1993). More recently it was recognized that the early cell death during reperfusion, which is a dominant necrotic process (Gujral et al., 2001), releases proteins (such as HMGB-1), which can bind to toll-like receptor-4 on macrophages and trigger the production of proinflammatory cytokines (Tsung et al., 2005a, 2005b). These findings explain the strong correlation between the early injury at the beginning of reperfusion and the later neutrophil response. The release of HMGB-1 and other proteins from dying cells does not only trigger the initial cytokine formation but represents a continuous stimulus for the formation of proinflammatory mediators and the recruitment of neutrophils and thus propagation of the neutrophilic hepatitis.

Alcoholic Hepatitis

It is well known that continued ingestion of alcohol in humans results in neutrophilic steatohepatitis (Bautista, 2002; French, 2002; Jaeschke, 2002). In fact, pathological evaluations of liver tissue have consistently demonstrated the presence of neutrophils in the liver parenchyma. Neutrophil infiltration in the liver is known to significantly contribute to the pathologic findings noted in alcoholic liver injury (ALI) (Bautista, 1997, 2002; French, 2002). Alcoholic steatohepatitis (ASH) seldom (< 10% cases) reverts to normal hepatic histology, even when the precipitating condition is removed (Enomoto et al., 1999; French, 2002). Rather, patients with steatohepatitis often develop increased hepatic fibrosis and, with time, cirrhosis occurs in a substantial fraction (i.e., in almost 50%) of these individuals (Galambos, 1972; MacSween and Burt, 1986; Maher, 2002). Thus, liver-related morbidity and mortality occur in patients with steatohepatitis and this stage appears to represent a rate-limiting step in the progression to cirrhosis and clinical liver disease in patients with ALD (Galambos, 1972; Uesugi et al., 2001; Diehl, 2002; Nanji, 2002; Thurman, 2002).

The histomorphological pattern of alcoholic steatohepatitis in human alcoholics consists of the infiltration of polymorphonuclear neutrophils, hepatocyte degeneration, ballooning, and oncotic necrosis (Galambos, 1972; Bautista, 2002; Diehl, 2002). Similar pathologic changes are noted in rodent models (Apte et al., 2005; Ramaiah et al., 2005). There is no direct evidence that ethanol per se causes inflammation. The presence of fat in the liver seems to be a prerequisite to the development of neutrophilic inflammation, possibly because a fatty liver is more vulnerable to various hitherto unidentified factors that trigger inflammation (Day and James, 1998). For instance, there is evidence for an involvement of bacterial endotoxins and viral hepatitis (Thurman et al., 1998). Furthermore, oxidative stress induced either by dietary polyunsaturated fatty acids or by iron supplementation may aggravate the neutrophilic inflammation in experimental models (Nanji et al., 1994; Tsukamoto et al., 1995).

Substantial progress has been made in the general understanding of neutrophil-infiltrative mechanisms in the liver (Figure 2, Pennington et al., 1998; Lieber, 2000; Bautista, 2002; French, 2002; Jaeschke, 2002). However, the precise mechanisms of neutrophil-mediated injury in alcoholic hepatitis remain unclear. The role of inflammatory mediators TNF-α, complement factors, PAF, vasoconstrictors (endothelin-1), adhesion molecules (selectins, LFA-1/Mac-1, ICAM-1, VCAM-1), chemokines and cytokines have been reported in literature as possible mechanisms for neutrophil mediated liver damage (Bautista, 2002). Apoptosis of hepatocytes that occurs during alcoholic hepatitis (Casey et al., 2001; Murohisa et al., 2002) is another mechanism that has been reported for hepatic neutrophil infiltration (Ziol et al., 2001). The authors showed a strong correlation between apoptotic hepatocytes and neutrophils in a model of ASH, which is also the case in case of endotoxemia-induced liver failure (Jaeschke et al., 1998; Lawson et al., 1998). Interestingly, endotoxemia is a common feature of clinical hepatitis and chronic alcohol exposure in human and experimental models suggest variable levels of endotoxemia (Wagner and Roth, 1999; Lieber, 2000; Bautista, 2002; French, 2002).

Recent evidence suggests that a novel class of extracellular proteins called matricellular proteins (MCP) play a critical role in the pathogenesis of several inflammatory diseases (Kim et al., 1997; O’Regan and Berman, 2000; Francki et al., 2001; Sodek et al., 2002). Unlike extracellular matrix proteins MCPs are involved in cell-to-cell and cell-to-matrix communication rather than structural support. Osteopontin (OPN), is one of the major MCP involved in cell-to-matrix communication and may play a role in several inflammatory diseases including glomerular nephritis (O’Regan and Berman, 2000; Denhardt et al., 2001), inflammation during CCl₄-induced hepatotoxicity (Kawashima et al., 1999), in puromycin-induced nephrotoxicity (Denhardt et al., 2001), and nonalcoholic steatohepatitis (Sahai et al., 2004). OPN is also a known chemoattractant for macrophages and neutrophils (Giachelli et al., 1998; Denhardt et al., 2001). The role of OPN in hepatic neutrophil infiltration during ASH was recently tested in a rat model of ASH.

In this study it was identified that OPN (native and thrombin-cleaved form) was induced within the hepatocytes in the ASH rodent model and there was a strong correlation between cleaved form of OPN and hepatic neutrophil infiltration (Banerjee et al., 2006a). These studies implicate OPN as a new mediator in the pathogenesis of neutrophilic ASH. The mechanisms by which OPN signals to neutrophils remains to be studied but based on previous reports (Giachelli et al., 1998; Denhardt et al., 2001), it can be speculated that OPN directly activates neutrophils via α₉β₁integrins. Recently, we have shown that both un-cleaved and cleaved OPN upregulated CD11b/CD18, which correlates quantitatively with neutrophil infiltration (Apte et al., 2005; Banerjee et al., 2006a). Overall, there is substantial evidence that neutrophils contribute to the liver pathology in alcoholic steatohepatitis although the extent of investigation into neutrophil-mediated liver pathology during ALI is relatively less compared to other models of neutrophil-mediated liver damage.

Alpha Naphthylisothiocyanate (ANIT) Liver Toxicity

ANIT is a liver toxicant that damages bile epithelial cells and hepatocytes. ANIT is used experimentally in rodents as a model to investigate intrahepatic cholestasis. ANIT induces acute cholangitis, which upon prolonged exposure results in biliary hyperplasia and fibrosis (Kodali et al., 2006). Biochemical and histopathologic alterations following ANIT administration are well documented. However, the precise mechanism by which ANIT causes liver injury remains unknown. ANIT is detoxified by glutathione (GSH) conjugation within hepatocytes and the complex of ANIT-glutathione is secreted into bile (Kodali et al., 2006). However, the ANIT-GSH conjugate dissociates rapidly in bile leading to release of the parent compound and thus exposing biliary epithelial cells to toxic concentrations of ANIT (Dietrich et al., 2001). Since there is continuous recycling of the compound by repetitive conjugation and disassociation, ANIT excretion is delayed thereby causing hepatobiliary toxicity. In addition, hepatic GSH levels are also depleted. Concurrent to recycling of the compound and exposing the biliary epithelial cells and hepatocytes to toxic concentrations of ANIT, there is a marked neutrophilic inflammatory response (Dahm et al., 1991; Xu et al., 2004). Neutrophils typically invade the bile ducts at the portal areas causing marked necrosis of the biliary epithelium. Thus, neutrophils play an important role in ANIT-mediated liver toxicity.

There are several studies that have investigated the mechanisms behind neutrophil infiltration into the liver during ANIT toxicity (Dahm et al., 1991; Jean and Roth, 1995; Hill et al., 1999; Kodali et al., 2006). Adhesion molecules such as β₂(CD18)integrins and chemokines have been the major focus of investigation of ANIT-mediated neutrophil infiltration (Xu et al., 2004). Using CD18-deficient mice, it was shown that ANIT induced neutrophil inflammation in via both CD18-dependent and -independent mechanisms (Kodali et al., 2006). It was suggested that CD18 positive neutrophils augment ANIT toxicity worsening hepatocellular necrosis but with little impact on the ANIT-mediated biliary epithelial cell damage and cholestasis (Kodali et al., 2006).

Chemoattractants such as PAF, leukotriene B₄ and lipid peroxidation products, e.g., 4-hydroxynonenal, have been investigated as possible candidates in ANIT-mediated neutrophil infiltration (Evans et al., 1991; Henderson, 1994; Nitti et al., 2002). However, since blockade of these mediators by specific inhibitors did not result in abrogation of ANIT-induced neutrophilic inflammation, it appears that PAF and leukotrienes are not likely involved in neutrophil infiltration (Bailie et al., 1995, 1996). CXC chemokines are other potential candidates that were investigated for ANIT-mediated inflammation. CXC chemokines (MIP-2, KC, IL-8) have been implicated as mediators of hepatic inflammation in a model obstructive cholestasis (Saito and Maher, 2000). Similarly, ANIT induced MIP-2 formation in periportal hepatocytes resulted in increased neutrophilic inflammation in the liver. A study in mice with targeted disruption of CXCR2, a receptor for MIP-2 and KC, showed a 50% reduction of hepatic neutrophil infiltration in response to ANIT (Xu et al., 2004). However, blockade of CXC chemokines did not have any effect on ANIT-induced hepatocellular injury, cholestasis, or fibrosis.

Based on this it was suggested that ANIT toxicity can occur even in the absence of chemokine-dependent neutrophil inflammation. However, it has to be kept in mind that only a part of the total hepatic neutrophils actually transmigrate and contribute to the injury (Essani et al., 1995). In addition, severe neutropenia as well as deficiency of CD18 has been demonstrated to protect against the liver cell injury but not cholestasis in this model (Kodali et al., 2006). These results are somewhat different from studies of Dahm et al. (1991), who showed that ANIT toxicity (parenchymal cell and cholangiocyte injury) in rats can be completely blocked when neutrophils were depleted. The discrepancy of these studies may be attributed to the amount of neutrophils actually required to produce the full spectrum of ANIT-liver injury and that CXC chemokines may not be completely responsible for neutrophil infiltration in this model. However, both studies clearly highlight the importance of neutrophils in ANIT-mediated hepatocellular injury.

Acetaminophen Liver Toxicity

An acute overdose of acetaminophen (APAP) can cause serious liver damage in experimental animals and in humans. APAP-induced liver toxicity is currently the most frequent cause of drug-induced liver failure in the United States (Lee, 2004). The mechanism of APAP toxicity is initiated by metabolic activation. Approximately 90% of a therapeutic dose of APAP is conjugated with glucuronic acid or sulfate within the hepatocytes and excreted into bile or plasma. The remaining ~5–10% of the APAP dose is metabolized by cytochrome P450 (CYP), particularly CYP2E1 (Nelson and Bruschi, 2003). The product of this reaction is a reactive metabolite, N-acetyl-parabenzoquinone imine (NAPQI) (Nelson and Bruschi, 2003), which is detoxified by glutathione (GSH). Once hepatocellular GSH levels are depleted, NAPQI reacts with sulfhydryl groups of proteins leading to formation of APAP protein adducts (Jollow et al., 1973). APAP protein adducts, in particular in mitochondria, appear to be critical for the subsequent mitochondrial oxidant stress and the amplification of the intracellular signaling mechanisms leading to cell necrosis (Jaeschke and Bajt, 2006).

APAP liver damage is known to activate neutrophils leading to accumulation of these cells in the hepatic vasculature (Lawson et al., 2000). Inflammatory cytokines and CXC chemokines, which can systemically activate neutrophils and trigger their recruitment into the liver (Witthaut et al., 1994; Bajt et al., 2001), are also generated during APAP hepatotoxicity (Takada et al., 1995; Horbach et al., 1997; Lawson et al., 2000). However, the role of neutrophils in the pathophysiology of APAP hepatotoxicity still remains controversial (Jaeschke, 2005).

Following APAP administration, a significant number of neutrophils are recruited into the liver closely following the development of cell injury 4–24 hours after drug treatment (Lawson et al., 2000). However, several antibodies against neutrophil β₂-integrins did not protect against APAP-induced liver injury (Jaeschke et al., 1991; Lawson et al., 2000). In addition, animals deficient in gp91phox, an essential protein of the activated NADPH oxidase complex in phagocytes, APAP caused a similar oxidant stress, peroxynitrite formation and liver injury as in wild-type animals (James et al., 2003). In support of these findings, a chemical inhibitor of NADPH oxidase did not protect against APAP hepatotoxicity (Cover et al., 2006). In a rat model, an inhibitor of hemeoxygenase 1 (HO-1) attenuated the inflammatory response including hepatic neutrophil infiltration, but had no effect on liver injury (Bauer et al., 2000).

Similarly, treatment with pentoxifyline, which is known to improve microcirculatory perfusion, did not affect liver injury (Welty et al., 1993). These studies suggest that neutrophil accumulation within sinusoids is not a relevant factor in APAP-induced liver injury. In addition, primary cultured mouse hepatocytes can be destroyed by APAP through the same mechanisms as hepatocytes in vivo in the absence of any inflammatory cells (Shen et al., 1992; Bajt et al., 2004; Kon et al., 2004) Thus, none of these studies support a relevant role of neutrophils in the mechanism of APAP induced liver injury. In contrast, these studies suggest that the major function of recruited neutrophils is to remove damaged cells or cell debris.

On the other hand, there are few studies that suggest that neutrophils may be involved in APAP-induced hepatotoxicity. IFN-γ gene knockout mice and NK and NKT cell-depleted animals show less liver injury and a lower number of neutrophils in the liver after APAP overdose (Ishida et al., 2002; Liu et al., 2004). However, these studies only show a correlation. Direct support for a potential role of neutrophils is derived from a study in rats, in which pretreatment with a neutrophil antiserum significantly attenuated APAP hepatotoxicity (Smith et al., 1998). However, other possible mechanisms of protection of the crude antiserum were not investigated in this study such as the likely contamination of antiserum with endotoxin, which could potentially inhibit APAP bioactivation and protect independent of the neutrophilic inflammation. Also, Liu et al. (2006) showed that pretreatment with a neutropenia-inducing anti-Gr-1 monoclonal antibody for 24 hours attenuated hepatic neutrophil accumulation and liver injury after 500 mg/kg of APAP.

However, causing neutropenia by treatment with the same antibody after APAP bioactivation did not protect (Cover et al., 2006). A recent preliminary study investigating the reasons for these opposite results with the same reagent indicated that causing neutropenia 24 hours before APAP treatment results in an acute phase response with upregulation of protective genes such as metallothionein (Jaeschke and Liu, 2007). Since metallothionein induction alone can protect against APAP hepatotoxicity (Liu et al., 1999), it remains unclear if pretreatment with the anti-Gr-1 antibody reduced APAP hepatotoxicity by causing neutropenia, induction of metallothionein, induction of other stress proteins or a combination of these mechanisms (Jaeschke and Liu, 2007).

Liu et al. (2006) also showed protection against APAP liver toxicity in ICAM-1-deficient mice. However, in our hands ICAM-1-deficient mice were not protected (Cover et al., 2006), a result that is consistent with the previously reported ineffectiveness of an anti-CD18 antibody (Lawson et al., 2000). Thus, the predominant experimental evidence does not support a hepatocyte damaging effect for neutrophils in the pathogenesis of APAP-mediated hepatotoxicity in the currently used rodent models. However, the presence of activated neutrophils in the liver during APAP-induced liver injury leaves the possibility open that conditions may develop where neutrophils aggravate APAP hepatotoxicity.

Summary

Neutrophil-mediated liver damage is mediated by activation/priming and subsequent accumulation within the hepatic vasculature most notably in the sinusoids. Neutrophil activation is mediated by proinflammatory mediators such as cytokines, CXC chemokines, complement factors and a lesser known recently reported protein, osteopontin. Accumulated neutrophils within the sinusoids sense the chemotactic signal from the hepatic parenchyma leading to neutrophil extravasation, which is mostly dependent on neutrophil integrins and endothelial adhesion molecules.

Next, the neutrophils undergo adhesion to hepatocytes via hepatocyte ICAM-1 and neutrophil Mac-1 (CD11b/CD18), which triggers formation of reactive oxygen by NADPH oxidase and release of proteases through degranulation. Neutrophil-derived reactive oxygen species diffuse into target cells and induce an intracellular oxidant stress, which eventually leads to the death of hepatocytes by oncotic necrosis. Characteristic for a neutrophilic hepatitis is that once the process is initiated, it can become a self-aggravating mechanism. The mediators responsible for the promotion of the inflammatory response are not always identified. Neutrophilic proteases can play an important role by facilitating production of inflammatory cytokines and chemokines. Products derived from dying cells can either be chemotactic for neutrophils directly (e.g., LPO products), trigger pro-inflammatory cytokine production by macrophages (e.g., HMGB-1) or cause direct cell damage (e.g., calpain release). In summary, there is extensive experimental evidence to indicate that neutrophils are important determinants of liver damage. Understanding the precise signals that regulate neutrophil infiltration into liver parenchyma and the mechanisms of neutrophil-induced cell damage can reveal therapeutic intervention strategies, which may significantly reduce acute liver failure during neutrophilic hepatitis and facilitate patient recovery when exposed to liver toxic stresses and chemicals.