Apoptosis in liver disease

Abstract

Abstract: The description of the morphological hallmarks of programmed cell death, apoptosis, in 1972 by Kerr, Wyllie and Currie started a field of research that revolutionized our understanding of cellular proliferation, tissue homeostasis and pathophysiology of many diseases. In the following years, a series of proteins involved in signaling and intracellular death pathways were identified and 30 years later the Noble Prize for physiology and medicine was awarded to S. Brenner, H. R. Horvitz and J. E. Sulston for their discoveries related to describing the mechanisms of cell death (apoptosis). The delineation of the signaling pathways that mediate apoptosis changed the paradigms of understanding in many liver diseases. The most detailed analyzed mode of apoptosis involves a cell surface-based receptor–ligand system. Death receptors are typically members of the tumor necrosis factor-receptor superfamily and comprise an intracellular death domain. Following ligand binding to the receptor, intracellular adapter molecules are recruited to the receptor and subsequently transmit the apoptotic signal. Intracellular organelle-dependent signaling occurs, and effector molecules then augment the receptor-initiated apoptosis process. Cell death and degradation follows the activation of a highly regulated set of cytosolic and nuclear proteases and DNAses. Receptor-independent activation of the apoptotic process can occur as part of the cytotoxicity related to UV radiation, chemotherapeuticals or other DNA-damaging agents through activation of intracellular sensors of cellular integrity, e.g. the tumor suppressor gene p53. In contrast to necrosis, apoptosis is not commonly accompanied by an inflammatory response that causes collateral cell damage. The apoptotic program is highly effective in eliciting cell death and thus must be tightly controlled. This is achieved through continuous integration of pro- and antiapoptotic signals at the individual cell level. Dysregulation of the apoptotic process, resulting in too much or too little cell death, has potentially devastating effects and has been implicated in many forms of liver disease like acute liver failure or hepatocellular carcinoma. This review will focus initially on recent progress in signaling events of hepatocellular apoptosis and subsequently discuss the consequences for the hepatic pathophysiology that involves disarrangement of hepatocellular apoptosis.

Mechanisms of apoptosis in the liver

Apoptosis signaling

Apoptosis is a highly preserved and controlled mechanism to achieve tissue homeostasis through targeted elimination of single cells without disrupting the biological functionality of the tissue. The morphological changes associated with apoptosis include nuclear condensation, cell shrinkage, and plasma membrane blebbing resulting in apoptotic bodies (see Fig. 1) (1). Rapid engulfment of these through neighboring cells or macrophages prevents an inflammatory response as observed with necrotic cell death in most cases. However, recent studies of cholestatic liver injury have indicated a role of resident macrophages in the liver, Kupffer cells, in the pathogenesis of a sustained inflammatory response following apoptotic cell death (2). Apoptosis can be initiated through an extrinsic or an intrinsic pathway depending on the initial site of activation of the cell death process. The extrinsic pathway is initiated after binding of trimeric ligand molecules to corresponding preassembled receptor multimers (see Fig. 2) (3, 4). Most prominent among the cytokines to induce hepatocellular apoptosis are the members of the tumor necrosis factor (TNF) ligand superfamily CD95L, TNF and TNF-related apoptosis-inducing ligand (TRAIL) (3). These cytokines exert a physiological function through their cognate receptors, namely the CD95 receptor (Apo1/Fas receptor), TNF receptor type 1 (TNF-R1, p55/65, CD120a) and type 2 (TNF-R2, p75/80, CD120b), TRAIL receptor type 1 and type 2 (5). Ligand–receptor interactions occur at the plasma membrane and result in conformational changes of the receptor, initiating the assembly of an early intracellular signaling complex, to which downstream signaling molecules are subsequently recruited (6). The intracellular domains of death receptors are devoid of intrinsic kinase activity and therefore depend on homophilic protein–protein interactions for the initiation of cell signaling (7). The apoptotic cell death pathway is activated following recruitment of the Fas-associated death domain (FADD/MORT1), which results in activation of the caspase (cysteine aspartate protease) family of enzymes (8). Caspases are capable of cleaving substrates after a loosely specific motif of four amino acids that contain aspartate in the first position. Crucial to their catalytic activity is the presence of a cysteine residue in the active center of the molecule (9). The family can be divided into upstream initiator caspases, such as caspase-8, -9, -10 and downstream effector caspases, caspase-3 and -6. The effector caspases are responsible for the cleavage of proteins whose functional loss induces apoptosis (10). Caspases are constitutively expressed as inactive zymogens or procaspases that require cleavage into smaller active subunits. A total of 13 caspases, 11 of which are expressed in humans, have been cloned, but not all of them participate in apoptosis (11). Upon recruitment and colocalization with FADD, high local concentrations of procaspases-8 undergo autoproteolytic cleavage, releasing activated caspase-8. This complex has been termed the death-inducing signaling complex (DISC), and the mode of activation is referred to as the induced proximity model of activation (12). The central role of the adapter molecules could be shown in mice expressing dominant-negative (dn) FADD protein. These mice are protected from liver injury following activation of CD95 and TNF signaling (13). Expression of the FLICE inhibitory proteins (FLIP-long and -short) can block apoptosis at this point and prevent caspase-8 activation by inhibition of the recruitment and processing of procaspases-8 at the level of the DISC (14). Recent studies have implied a dichotomous role for the long isoform of FLIP in the context of cell death. While the antiapoptotic function prevents execution of the cell death process, mice with defective FADD exhibited, in addition, an impaired proliferative response upon 2/3 hepatectomy, implying a potential role of the apoptotic pathways in liver regeneration (15). The intrinsic pathway of apoptosis involves mitochondria and caspase-9 activation. Cleavage of the Bcl-2 family member Bid by caspase-8 results in truncated Bid (tBid), which triggers oligomerization of the proapoptotic Bcl-2 family members Bax and Bak (16). These molecules then insert into the mitochondrial membrane, resulting in mitochondrial permeability transition (MPT) and release of mitochondrial proteins including cytochrome c (17). This function of Bid has been demonstrated to mediate hepatocyte death from TNF. Hepatocytes deficient for Bid show resistance to TNF and CD95-induced cell death associated with the prevention of mitochondrial depolarization and cytochrome c release (18). In vivo, Bid-deficient mice were partially protected from toxic liver injury (19). In accordance with these observations, inhibition of cytochrome c release from mitochondria by the MPT inhibitor cyclosporin A prevents hepatocyte apoptosis at a point downstream of the adapter molecule FADD, but upstream of caspase-3 activation (20). Following release into the cytosol, cytochrome c triggers formation of the apoptosome, a complex with apoptosis protease-activating factor-1 and procaspases-9. Caspase-9 becomes activated and in turn activates caspase-3, leading to degradation of structural proteins and resulting in apoptosis (21). Hepatocytes are dependent on the mitochondrial death pathway and have been termed type II cells. In contrast, type I cells generate high levels of caspase-8 activity that directly cleave caspase-3 (22). Accordingly, expression of the antiapoptotic factors Bcl-2 and Bcl-X_L that inhibit Bid and Bax activation prevents apoptosis in type II, but not type I cells (22). Supporting the concept of the hepatocyte as a type II cell is that in vivo Bcl-2 or Bcl-X_L overexpression is partially effective in preventing liver injury from TNF (23, 24). Another mechanisms involving the intrinsic pathway in hepatocytes is the activation of the lysosomal cysteine protease cathepsin B. In vitro and in vivo TNF-induced apoptosis is dependent on the release of cathepsin B from vesicles (25). The proapoptotic effect of cathepsin B is above the level of mitochondrial cytochrome c release as this process is blocked in TNF-treated cathepsin B null hepatocytes. Additional evidence of an effect of this protease on mitochondria is that in a cell-free system, cathepsin B induced mitochondrial cytochrome c release (26).

Figure 1. Apoptosis in HepG2 cells. Hepatoma cells incubated with agonistic CD95 antibody undergo apoptosis within 8 h of treatment. Morphologic cell changes include nuclear condensation, cell shrinkage and blebbing. Staining: Hoechst 33342.

Figure 2. Apoptosis receptor signaling pathways. Binding of corresponding ligand and receptor induces recruitment of proteins through homophilic interactions of conserved domains. In the death pathway (right-hand side of the figure), autoproteolytic activation of procaspases-8 results in the DISC. Activated caspase-8 cleaves Bid producing a truncated and active tBid. tBid in turn activates the proapoptotic Bcl-2 family members Bax and Bak. Their oligomerization and integration into the mitochondrial membrane result in cytochrome c release. In a complex consisting of cytochrome c and APAF-1, procaspases-3 is recruited and activated, resulting in apoptosis. Additionally, caspase-8 activation causes release of the lysosomal enzyme cathepsin B that acts in an undefined fashion to activate the mitochondrial death pathway. NF-κB-regulated gene products act to block the TNF apoptotic death pathway and TNF also mediates cellular proliferation involving the TNF-R1 (left-hand side of the figure). Activation of IKKs that phosphorylate IκB leads to proteasome-dependent degradation of phosphorylated IκB, releasing NF-κB heterodimers that translocate to the nucleus and activate genes that may be necessary for TNF-induced hepatocyte proliferation. Activation of the JNK/c-Jun/AP-1 pathway can promote hepatocyte proliferation through AP-1-dependent gene expression, but sustained AP-1 activation may alternatively serve to trigger the apoptotic death pathway. DISC, death-inducing signaling complex; tBid, truncated Bid; TNF, tumor necrosis factor; NF-κB, nuclear factor-κB; IKK, IκB kinase; APAF-1, apoptosis protease activating factor-1; AP-1, activator protein-1; JNK, c-Jun N-terminal kinase.

Recent findings related to the regulation of apoptosis signaling pathways have implied a central role of mitogen-activated protein kinases (MAPK) in apoptosis signaling through TNF and to a lesser extent through CD95 and TRAIL (27–29). A prominent member of the MAPK family is the c-Jun N-terminal kinase (JNK). Activated JNK phosphorylates substrates that include the activator protein-1 (AP-1) transcription factor subunit c-Jun, leading to increased AP-1 transcriptional activity (30). Transient JNK/AP-1 activation occurs after TNF stimulation and may in part mediate the proliferative effects of TNF (31); however, with nuclear factor-κB (NF-κB) inhibition, TNF induces a prolonged activation of JNK and AP-1 in hepatocytes and non-hepatic cells in association with cell death (32, 28). Death from TNF following NF-κB inactivation can be blocked in hepatocytes by inhibition of c-Jun, suggesting that an AP-1 gene product activates the TNF mitochondrial death pathway (28).

TNF, as opposed to other apoptosis-inducing ligands, mediates pleiotropic effects on tissue homeostasis, apoptosis and proliferation. The transcription factor NF-κB has been implicated in the modulation of apoptosis induced by TNF, and inhibition of NF-κB induces apoptosis in hepatocytes following TNF exposure (33). NF-κB activation is achieved through degradation of its inhibitor IκB following its phosphorylation through IκB kinases (IKKs) (34). TNF production in the liver can be observed after partial hepatectomy within 1 h and neutralizing antibodies to TNF inhibit the regenerative stimulus of this cytokine (35, 36). TNF-mediated regeneration involves activation of IL-6 and STAT3 (37, 38). The role of NF-κB in regeneration is less clear, and adenoviral inhibition NF-κB did not affect liver regeneration in vivo (39).

Apoptosis in the pathophysiology of liver disease

Overactivation of the fine-tuned apoptotic process can lead to significant hepatocellular damage, while inhibition of apoptosis can promote the proliferation and transformation of cells. The following sections will outline the relevance of the apoptosis signaling pathways mentioned above in the context of liver disease.

Toxic liver injury

Central to the understanding of hepatocellular injury from toxins are reactive oxygen species (ROS) and cytokines secreted from inflammatory cells. Until recently, liver injury was thought to result from the direct biochemical effects of toxins or biological agents on hepatocytes. However, it is now apparent that liver injury is triggered in large parts from inflammatory cell products. Among the proapoptotic cytokines, TNF is a prominent mediator of toxin-induced liver injury and was one of the first to be identified as a mediator of hepatocellular death following toxin exposure. TNF is produced as part of the liver's response to hepatotoxins such as carbon tetrachloride or the sensitizer galactosamine and this type of injury can be prevented by administration of neutralizing antibodies to TNF (40). Subsequent studies demonstrated that both Tnf- and Tnf-r1-null mice are also protected from carbon tetrachloride injury (41). Interference with the TNF signaling pathway below the TNF receptor through adenoviral inhibition of FADD function was able to reduce galactosamine-induced liver injury through blockage of caspase-3 activation (42, 13). Mechanistically, the inhibition of NF-κB or overactivation of JNK through hepatotoxins are possible mechanisms that might lead to increased sensitivity of hepatocytes toward TNF-induced apoptosis. Hepatocyte-specific impairment of subunits of the NF-κB inhibitory kinases (IKKs) was associated with massive hepatocyte apoptosis form TNF, supporting the notion that NF-κB is the key factor in regulating TNF-induced hepatocyte injury (43).

CD95 has been implicated in the pathophysiology of rapid, acute liver failure with high mortality. The pregnancy-related hemolysis, elevated liver enzymes and low platelet count (HELLP) syndrome is one of these conditions. Increased expression of CD95L was detected in the serum of patients with the HELLP syndrome. In particular, maternal placenta particles expressing CD95 that elicit apoptosis in hepatocytes could be identified as the cause of this type of liver injury (44). The importance of the CD95 signaling pathway in liver injury was shown in mice with hepatocyte-specific expression of a dnFADD protein. While control animals die from acute liver failure characterized by hepatocellular apoptosis, dnFADD mice were protected from this type of injury (13).

Non-alcoholic and alcoholic liver disease

Non-alcoholic fatty liver disease (NAFLD) is the most frequent cause of elevated liver enzymes in countries with increasing prevalence of obesity. The progression of NAFLD to non-alcoholic steatohepatitis (NASH) is associated with inflammation and worsening fibrosis, resulting in cirrhosis in up to 15–50% (45). While many of the aspects of the pathophysiology of this disease are currently unclear, induction of the prooxidant cytochrome P450 2E1 (CYP2E1) is observed both in animal models and patients with NASH (46, 47). Overexpression of CYP2E1 in hepatocytes caused activation of the MAPK ERK1/2 and sensitized these cells to TNF-induced apoptosis through prolonged activation of JNK (48, 49). The generation of ROS contributes to this activation, and antioxidants reduce both MAPK activation and TNF-induced cell death in CYP2E1 expressing hepatocytes (49, 48). Additionally, patients with histologically proven NASH exhibit increased expression of the CD95 and TNF receptors (50, 51). Thus, activation of MAPK signaling induced through oxidative stress and increased apoptosis receptor expression in NASH appear to contribute to the hepatocellular injury in this disease.

Subclinical insulin resistance and diabetes mellitus are highly associated with liver injury in patients with NASH (52). Recent studies have provided evidence that a possible crosslink between metabolic and inflammatory pathways exists that leads to increased activation of hepatocellular death pathways (53). Firstly, JNK activation, as observed with increased ROS generation through CYP2E1 induction in hepatocytes, causes insulin resistance in vitro and in vivo (54). These observations link inflammation and oxidative stress to insulin resistance. Secondly, insulin resistance has been shown to be associated with decreased secretion of the adipocytokine adiponectin, and patients with NAFLD have decreased levels of adiponectin (55). In vivo mice were protected from LPS-induced TNF-mediated liver injury through application of adiponectin, implying that adiponectin is capable of counterregulating TNF-induced liver injury (56). These findings are of particular relevance in the context of the increasing incidence of obesity in our society, and epidemiological observations show that patients with diabetes mellitus have a higher morbidity from liver than from cardiovascular disease (57). Future studies will have to clarify the interactions of metabolic and cell death signaling pathways in patients with NASH.

Alcoholic steatohepatitis (ASH), similar to NASH, does exhibit increased expression of CYP2E1, leading to enhanced formation of ROS and lipidperoxides (58). Consistent with the importance of ROS, antioxidants reduced acute hepatocellular injury in rats from ethanol through prevention of release of proapoptotic factors from mitochondria (59). In patients with ASH, increased expression of apoptosis receptors and increased amounts of circulating apoptosis ligands could be detected (60–62). Additionally, the severity of ASH can be correlated with the amount of apoptotic cells in liver biopsies from these patients (63).

Hepatocellular carcinoma (HCCs)

HCCs are characterized by increased proliferation and loss of physiological tissue homeostasis induced through mutations of tumor suppressor genes and defects in apoptosis signaling (64). The loss of the tumor suppressor gene p53 contributes to the decreased expression of CD95 and reduced sensitivity of HCC cells towards this apoptosis pathway (65). Additionally, other apoptosis receptors (TRAIL-1 and -2) and cell cycle regulatory BH3-only domain proteins are decreased and antiapoptotic proteins are increased in HCCs (66, 67). In primary HCCs, the mutation and loss of a single apoptosis signaling molecule was significantly associated with reduced rates of apoptosis and the grade of the tumor (68). All of these aberrations contribute to the overall apoptosis-resistant phenotype of HCCs. In hepatoma cells, microinjections of wild-type p53 and treatment with the chemotherapeutical bleomycin restored sensitivity towards CD95-induced apoptosis (65). In addition to alterations of apoptosis sensitivity, HCCs are capable of an effect described as immunevasion. In biopsies from patients with HCCs, tumorous cells were shown to upregulate CD95 ligand, thus inducing apoptosis in infiltrating, tumor-specific T-lymphocytes and generating an immuneprivileged environment in which tumor growth can proceed (69).

Chronic viral hepatitis

The pathogenesis of liver cell injury in acute and chronic viral hepatitis is poorly understood, but proinflammatory cytokines are thought to play a central role in modulating the cellular immune response, virus replication and liver cell injury. TNF can promote viral clearance in vitro by cytotoxic and non-cytotoxic effects (70, 71). Central to the proapoptotic effects of TNF in infected hepatocytes is the HBV protein X (HBx). Expression of HBx increases the susceptibility of hepatocytes to cell death from TNF through activation of MAPK kinase 1 and n-Myc (72). Hyperactivation of caspase-8 and -3 in response to TNF stimulation occurs from inactivation of the inhibitor protein of caspase-8, FLIP (73). However, HBx has also been reported to produce antiapoptotic effects and could be involved in the emergence of HCC in chronic HBV infection (74, 75). Similarly, contrasting results have been reported in studies of HCV core protein in human hepatoma cell lines. Overexpression of HCV core protein was demonstrated to inhibit TNF-induced apoptosis through NF-κB activation (76). In contrast, this protein has also been reported to enhance TNF-induced apoptosis without affecting NF-κB activation (77). Thus, further studies are needed to resolve the apparently conflicting reports on the roles of death receptor ligand-mediated cytotoxicity in viral hepatitis.

Perspectives on therapies in apoptosis and liver disease

Modulation of apoptosis receptor signaling pathways is potentially beneficial in the treatment of liver disease. The antitumor activity of cell death ligands derives directly from its cytotoxic effects on hepatoma and carcinoma cells. However, dysregulation of the apoptosis signaling pathways and induction of antiapoptotic proteins in HCC have so far led to disappointing results. Secondly, undirected activation of cell death signaling pathways in liver disease seems to be undesirable as induction of cell death in healthy liver parenchyma would prohibit the clinical use. To avoid the deleterious side effects of apoptosis ligands on healthy liver tissue, substances with specific antitumor activity have been investigated with the aim of reversing the overall phenotype of apoptosis resistance and restoring apoptosis sensitivity. The class of histone deacetylase inhibitors has been shown to sensitize hepatoma cells towards CD95 and TRAIL-induced apoptosis, possibly involving the downregulation of the antiapoptotic caspase-8 antagonist cellular FLIP (c-FLIP) (78, 79). Similar effects have been described with the use of chemotherapeuticals (80). The advancement of techniques concerning cell type-specific targeting will provide therapeutic tools to avoid the deleterious effects of potential therapies from injury to healthy liver tissue. Recently, one report in this journal examined the efficiency of adenoviral, tumor cell-specific delivery of TRAIL and found increased rates of apoptosis in this cell population (81).

Treatment of liver injury and acute liver failure form overactivation of apoptosis signaling pathways, on the other hand, would have to rely on inhibition of apoptosis signaling pathways. Investigations of the efficiency of direct interference with caspase-8 and -3 by small interfering RNA have been performed. In two independent studies, this approach was effective in preventing acute liver injury from ischemia/reperfusion or injection of agonistic CD95 antibodies as reflected in a reduction of the elevated serum transaminases and an increase in overall survival (82, 83). The complexity of apoptosis signaling pathways was recently demonstrated in mice with a defect in the cell death signaling pathway. Mice expressing a dnFADD protein had impaired liver regeneration and higher mortality following 2/3 hepatectomy (15). Thus, the mere inhibition of cell death signaling in the development of new therapeutically strategies could potentially have harmful effects by inhibition of the regenerative capacity of the liver.

Summary

Apoptosis contributes to the physiological tissue homeostasis through controlled elimination of cells. However, disarrangement of death receptor signaling pathways has been identified as a major contributor to the initiation and aggravation of acute and chronic liver injury. The physiologically finely tuned process of cell death by apoptosis can be both augmented to result in excessive cell death as observed with toxic, viral, alcoholic and non-alcoholic liver injury or suppressed leading to excessive cellular proliferation and possibly HCC. The complexity of apoptosis signaling is additionally increased by the fact that apoptosis signaling pathways are involved in both cell death and proliferation, with one signaling molecule possibly contributing to either cell function under different conditions. New therapeuticals aimed at the modulation of apoptosis pathways will have to be specifically targeted to one cell population to avoid side effects in surrounding, healthy tissues. Research related to cell type-specific activation or inhibition of apoptosis signaling pathways will potentially provide new strategies for the treatment of liver diseases.