Toxic bile acids facilitate Fas and tumor necrosis factor-associated apoptosis-inducing ligand (TRAIL) death-receptor oligomerization and activation. Bile acid modulation of death-receptor signaling is multifactorial and includes trafficking of Fas to the cell surface, enhancing TRAIL-R2/DR5 expression, and suppression of function of cFLIP, an antiapoptotic protein modulating death-receptor function. Because bile acid-associated death receptor-mediated apoptosis is a common mechanism for cholestatic hepatocyte injury, inhibition of death receptors and their cascades may prove useful in attenuating liver injury during cholestasis.
- tumor necrosis factor-associated apoptosis-inducing ligand
bile acids are amphipatic molecules synthesized from cholesterol within hepatocytes. After their synthesis, these compounds are secreted into bile, enter the small intestinal lumen, are reabsorbed in the terminal ileum, and are efficiently transported from the portal vein back into the hepatocyte. Bile acid homeostasis is tightly regulated in health, and their cellular and tissue concentrations are restricted. However, when their biliary secretion is impaired by pathophysiological processes, hepatocytes are exposed to elevated concentrations of bile acids. The elevated hepatic bile acid concentrations trigger cell death and subsequent fibrosis. Thus the mechanisms responsible for bile acid-induced liver injury are of biomedical importance. Current information suggests that bile acids induce hepatic injury by death receptor-mediated processes.
DEATH-RECEPTOR SIGNALING IN HEPATOCYTES
Apoptosis can be initiated by triggering cell surface death receptors, which belong to the TNF or nerve growth factor receptor superfamily. Currently, six death receptors are known, including Fas (also called CD95/APO-1), which binds Fas ligand; TNF receptor 1 (TNF-R1; also called p55/CD120a), which binds TNF-α; tumor necrosis factor-associated apoptosis-inducing ligand [TRAIL-R1; also called death receptor (DR) 4] and TRAIL-R2 [also called DR5/apolipoprotein (APO)-2/KILLERTRICK2], which bind TRAIL; DR3 (also called APO-3/TRAMP/WSL-1/LARD), which binds Apo-3 ligand/TWEAK; and DR6, which contains a death domain (DD) but has no identified ligand. Of these receptors, Fas, TRAIL-R1 and -R2, and TNF-R1 are richly expressed in liver cells.
Ligation of the death receptors with their cognate ligands facilitates receptor oligomerization; however, death-receptor oligomerization may also occur by a ligand-independent mechanism (1, 2). Both ligand-dependent and -independent receptor oligomerization signal apoptosis by recruitment of Fas-associated DD (FADD) protein to the oligomerized receptors via homotypic interactions between the DD of both proteins. FADD contains a death effector domain (DED) that promotes recruitment of the so-called initiator caspases (procaspase-8 and -10) by homotypic interactions between the DEDs, which are also present on procaspase-8/10, resulting in formation of the death-inducing signaling complex (DISC) (Fig.1). The induced proximity of the initiator caspases is postulated to cause their activation by autocatalytic processes. The activation of initiator caspases at the DISC is believed to be a necessary step in death receptor-mediated apoptosis.
The cytotoxic signals downstream of the DISC differ between cell types and have been characterized as “type I” and “type II” cellular responses (19). In type I cells, activated initiator caspases cleave and thereby activate the effector caspases (caspase-3/6/7), which are responsible for the execution of the cell death program. Other cell types such as hepatocytes are so-called type II cells. In these cells, sufficient activation of the effector caspases requires a mitochondrial amplification pathway. Activated initiator caspases cleave Bid, a proapoptotic member of the Bcl-2 family, which, in turn, induces mitochondrial dysfunction (Fig. 1). Truncated Bid (tBid) transmigrates to mitochondria, where it induces release of proapoptotic mitochondrial factors including apoptosis-inducing factor, second mitochondria-derived activator of caspases/direct inhibitors of apoptosis (IAP) binding protein with low isoelectric point, Omi/HtrA2, and cytochromec. Once in the cytosol, cytochrome c binds to the cofactor apoptosis-activating factor 1 facilitating its binding to procaspase-9, forming a complex referred to as the apoptosome. Through an energy-requiring reaction, procaspase-9 is processed to the mature enzyme, which, in turn, activates caspase-3, starting a caspase cascade downstream of the mitochondria. In contrast to type I cells, the mitochondrial pathway appears to be essential for the execution of the apoptotic program in type II cells. For example, hepatocytes from Bid-knockout mice are resistant to Fas-induced apoptosis (25).
BILE ACID-MEDIATED DEATH-RECEPTOR ACTIVATION
Current concepts suggest that bile acid-associated hepatocyte apoptosis is death-receptor dependent (5, 15, 20). Bile acid cytotoxicity, to date, has been shown to occur by Fas and TRAIL-R2/DR5 death receptors (5, 9). In contrast, bile acids do not appear to enhance TNF-α/TNF-R1 cytotoxicity (unpublished observation).
Micromolar concentrations of glycochenodeoxycholate induce Fas aggregation on the plasma membrane via a Fas ligand-independent mechanism (5, 17). Observations in bile duct-ligated mice, a model of extrahepatic cholestasis with elevated serum and hepatic bile acid concentration, also demonstrate that hepatocellular apoptosis and fibrosis are diminished in bile duct-ligated Fas-deficient lpr mice compared with wild-type animals (4,15). Thus Fas ligand-independent Fas death-receptor activation appears to be a critical pathway for bile acid-associated hepatocyte apoptosis.
Toxic bile acids appear to promote Fas activation by altering cellular trafficking of this death receptor (Fig.2). A mechanism for the posttranscriptional regulation of Fas receptors is to sequester this death receptor within intracellular pools, especially vesicles associated with the Golgi complex and the transGolgi network (3,20). On stimulation, these Fas-containing vesicles can be shuttled to the plasma membrane, presumably by a microtuble transport pathway, to initiate cell death signals. During bile acid-induced hepatocyte apoptosis, a Golgi-associated and microtubule-dependent pathway has been implicated in the trafficking of Fas to the cell surface (20). Both PKC and JNK kinase have been implicated in this vesicle trafficking (8). The increased density of cell surface Fas promotes their oligomerization, initiating a death-signaling pathway in a ligand-independent mechanism. The increased density of cell surface Fas also sensitizes hepatocytes to Fas agonist-induced cell death (20). Thus toxic bile acids promote both ligand-independent and -dependent hepatocyte apoptosis by Fas.
In vivo, genetic deletion of Fas only reduces hepatocyte apoptosis by 50%, suggesting alternative pathway(s) for bile acid-triggered cytotoxicity (15). This Fas-independent apoptosis also appears to be death-receptor dependent, because it is prevented by inhibiting death-receptor signaling. Indeed dominant-negative FADD, selective caspase-8 inhibitors such as IETD-cho or cytokine response-modifier A (CrmA), and depletion of Bid using specific antisense oligodeoxynucleotides all prevent both Fas-dependent and -independent apoptosis (9, 11). In a Fas-deficient cell line, enhanced expression and oligomerization of TRAIL-R2/DR5 have been observed following treatment with a toxic bile acid (9). Importantly, these observations have also been confirmed in the in vivo bile duct-ligated mouse model (10). Enhanced TRAIL-R2 expression in this model sensitizes the animal to hepatotoxicity by exogenous TRAIL administration. Because TRAIL is synthesized by activated Kupffer cells (6), it is likely that TRAIL also potentiates cholestatic liver injury due to Kupffer cell-generated TRAIL, which causes the hepatocyte apoptosis. Thus, although the role of TRAIL-mediated apoptosis in health is controversial (7), TRAIL is hepatotoxic in cholestasis.
BILE ACID MODULATION OF DEATH- RECEPTOR SIGNALING
Bile acids not only initiate ligand-independent death-receptor oligomerization but also modulate the signaling pathway, resulting in a strong sensitization of the cells to death receptor-mediated apoptosis. Current information suggests that death receptor-mediated apoptosis is regulated at the cell surface by the level of the receptor expression, at the DISC by expression of procaspase-8, FADD, or apoptosis inhibitor cellular FLICE inhibitory protein (cFLIP; also called Flame-1/I-FLICE/Casper/CASH/MRIT/CLARP/Usurpin), and at mitochondria by the expression of proapoptotic or antiapoptotic Bcl-2 family proteins such as Bcl-2, Bcl-xL, Bid, or Bax, which regulate mitochondrial cytochrome c release.
At the DISC level, bile acids do not alter expression of procaspases, FADD, or cFLIP. However, bile acids appear to prevent cFLIP recruitment to the DISC and thereby enhance initiator caspase activation (12). cFLIP exists as two different isoforms, cFLIP long (cFLIP-L; 55 kDa) and cFLIP short (cFLIP-s; 26 kDa). cFLIP-L is structurally similar to procaspase-8 in that it contains two DEDs and a caspaselike domain. However, the caspaselike domain lacks the critical active site cysteine residue essential for catalytic activity. cFLIP-s is also composed of two DEDs but lacks the entire caspaselike domain. Both cFLIP-L and cFLIP-s bind to FADD within the DISC via DED-DED homotypic interactions and inhibit activation of the initiator caspases in the DISC. We have recently shown that cFLIP is a phosphoprotein and that toxic bile acids potentiate phosphorylation of both cFLIP-L and -s by a PKC-dependent mechanism (12). The phosphorylation of cFLIP decreases binding of cFLIP to FADD, resulting in reduced recruitment to the DISC. The altered composition of the DISC (decreased cFLIP and increased caspase-8/10) results in enhanced activation of the initiator caspases.
At the mitochondrial level, the potential mitochondrial toxicity of bile acids may also enhance mitochondrial cytochrome crelease. Direct toxicity of bile acids on isolated hepatocyte mitochondria has been demonstrated (14). Recent observations suggest that treatment of cells with hydrophobic bile acids augments mitochondrial dysfunction with cytochrome crelease (18). Thus bile acids may potentiate death receptor-mediated apoptosis at multiple steps in the proapoptotic signaling cascade.
CELLULAR CONSEQUENCES OF DEATH- RECEPTOR ACTIVATION
Apoptosis signaling following death-receptor activation leads to mitochondrial dysfunction in hepatocytes. The mitochondria-dependent pathway of apoptosis is regulated by the Bcl-2 family proteins. One subfamily includes Bcl-2, Bcl-xL, Bcl-w, and Mcl-1 and promotes cell survival. The proapoptotic subfamily can be divided into those proteins containing three Bcl-2 homology domains (BH3) such as Bax, Bak, and Bok and those consisting of only one BH3 region like Bad, Bid, Bik, Bim, Bmf, and Hrk. The BH3-only protein, such as Bid, promotes translocation to and/or membrane insertion into mitochondria of Bax and Bak. Bid activation of Bax and/or Bak at the level of the mitochondria is essential for cytochrome crelease in hepatocytes following Fas activation. Bax has been implicated as an upstream mediator of bile acid-induced mitochondrial cytochrome c release (18). Rodrigues et al. (18) demonstrated that Bax translocates to mitochondria during deoxycholate-induced apoptosis in rat hepatocytes. In these studies, mitochondrial Bax was accompanied by the mitochondrial permeability transition (MPT) and cytochrome c release. Bile acid-mediated death-receptor signaling likely promotes Bax-dependent mitochondrial dysfunction via Bid-induced activation of Bax (Fig. 1).
Apoptotic mitochondrial dysfunction is accompanied by impairment of the electron transport chain and secondary oxidative stress. Oxidative stress is thought to occur when a cell, tissue, or organism is exposed to excess oxidant generation, in particular to superoxide anion (O ) and its metabolites (Fig. 1). It has been postulated that hydrophobic bile acids retained in the hepatocyte during cholestasis initiate the generation of reactive oxygen metabolites from mitochondria, leading to lipid peroxidation and loss of cell viability (21). The mitochondrial oxidative stress can trigger the MPT, resulting in exaggerated mitochondrial cytochromec release and apoptosis. Yerushalmi et al. (24) reported that relatively low concentrations of bile acids induced oxidative stress in hepatocytes before the onset of apoptosis, and a significant linear correlation between apoptosis and the magnitude of oxidative stress was observed (24). These alterations were attenuated by caspase-8 inhibitor IETD-cho, suggesting that mitochondrial oxidative stress is death-receptor dependent. When the mitochondrial dysfunction is severe, sustained mitochondrial permeabilization and massive cytochromec release may cause loss of ATP synthesis, likely causing necrosis rather than apoptosis. Indeed, Sokol et al. (21) have reported that bile acids induce oxidative stress in rat hepatic mitochondria, leading to hepatocyte necrosis. Nonetheless, it is likely that Bid/Bax activation following death-receptor signaling contributes to these observations.
POTENTIAL MOLECULAR TARGETS TO PREVENT BILE ACID-MEDIATED DEATH-RECEPTOR SIGNALING
Because hepatocytes express multiple death receptors, targeting a certain death receptor or ligand appears to be insufficient to prevent liver injury. Indeed, bile acid-induced activation of death-receptor signaling is observed by alternative activation of TRAIL-R2/DR5 in the absence of Fas (9, 11). Therefore, proteins that participate in common pathways during death-receptor signaling would be the best therapeutic targets.
Because FADD appears to be essential in death-receptor signaling, prevention of FADD expression or use of a dominant-negative FADD should block bile acid-induced death-receptor signaling. Indeed, dominant-negative FADD prevents bile acid-induced hepatocyte apoptosis in vitro by both Fas and TRAIL-R2 pathways (5,9). A recent preliminary report also suggests that dominant-negative FADD attenuates liver injury in the bile duct-ligated mouse (16). Selective gene silencing, using RNA interference technology, may be a potential therapeutic approach to decrease hepatic FADD expression.
Initiator caspases may also be potential molecular targets, because the selective caspase 8 inhibitors, IETD-cho, or the pox virus protein CrmA prevent bile acid-associated hepatocyte injury (5,9). Although caspase-10 is implicated in death-receptor signaling in the absence of caspase-8, the ability of caspase-10 to substitute for caspase-8 in mammalian cells remains controversial (22). A pan-caspase inhibitor, IDN-6536, has been developed and shown to attenuate elevated serum transaminases in humans (23). Although the long-term safety of inhibiting caspases remains to be determined, such a drug may prove useful in human cholestatic liver diseases (23).
Because Bid is essential for death receptor-mediated apoptosis in hepatocytes, we examined the role of Bid in bile acid-mediated apoptosis in primary mouse hepatocytes and bile duct-ligated mice. Inhibition of Bid expression using antisense technology but not scrambled antisense oligonucleotides markedly attenuated bile acid-induced hepatocyte apoptosis (10). More importantly, Bid antisense treatment of bile duct-ligated mice also reduced liver injury as assessed by serum alanine aminotransferase values. Thus Bid would also appear to be an attractive pharmacological target for the treatment of cholestatic liver diseases.
Certain isoforms of PKC have been shown to be activated by bile acids and participate in bile acid cytotoxicity (13). Bile acid-stimulated PKC activity appears to be instrumental in shuttling of Fas-containing vesicles to the plasma membrane and in cFLIP phosphorylation (12, 20). Indeed, PKC inhibition reduces bile acid-associated death receptor-mediated hepatocyte apoptosis (12, 20). The role and safety of PKC inhibition in cholestatic conditions require further studies.
Current studies demonstrate the importance of death receptor-mediated apoptosis in bile acid cytotoxicity. Bile acids promote death-receptor oligomerization by either ligand-dependent or -independent mechanisms. Toxic bile acids increase cell surface TRAIL-R2/DR5 or Fas by induction of their mRNA and/or enhancing cell surface trafficking to plasma membrane, respectively. The increased death-receptor protein levels on the plasma membrane result in ligand-independent death-receptor oligomerization or sensitization to ligand-mediated apoptosis. Oligomerized death receptors lead to DISC formation composed of the receptors, FADD, the initiator caspase (caspase- 8/10), and the apoptosis inhibitor cFLIP. Bile acids stimulate cFLIP phosphorylation by a PKC-dependent mechanism. The phosphorylation reduces cFLIP recruitment to the DISC, resulting in enhanced caspase- 8/10 activation. Initiator caspase-dependent Bid cleavage is also crucial for hepatocyte apoptosis. Thus inhibition of the death-receptor signaling would appear to be crucial for the treatment of cholestatic liver diseases. Selection of the most effective molecular targets (e.g., caspase-8/10, FADD, Bid, or PKC) will require further studies and will ultimately be based on effectiveness and safety.
We thank S. Erickson and E. Bungum for secretarial assistance.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-41876 and by the Mayo Foundation.
Address for reprint requests and other correspondence: G. J. Gores, Professor of Medicine, Mayo Medical School, Clinic, and Foundation, 200 First St. SW, Rochester, MN 55905 (E-mail:).
- Copyright © 2003 the American Physiological Society