Physiological or programmed cell death typically occurs by what is now referred to as apoptosis. Defects in the physiological pathways of apoptosis that result in either too much or too little cell death play a key role in a litany of human diseases, including many liver disorders. Consequently, great interest has emerged in developing therapeutic strategies for modulating the critical factors that regulate life-and-death decisions.
Cell demise by apoptosis is marked by a well-defined sequence of morphological events, including chromatin condensation, cell shrinkage, and release of small membrane-bound apoptotic bodies. These remnants are typically engulfed by neighboring as well as circulating cells, such as macrophages and Kupffer cells of the liver. The morphological alterations are accompanied by biochemical changes, such as DNA fragmentation, and specific protein degradation by activation of intracellular proteases and endonucleases. In contrast to apoptosis, cells that die in response to tissue damage exhibit diverse morphological changes referred to as necrosis. Nevertheless, apoptosis and necrosis frequently represent alternate outcomes of the same cellular pathways to cell death.1
A remarkable feature of apoptosis is its highly regulated nature. Key insights into the molecular mechanisms regulating cell death come from genetic studies in Caenorhabditis elegans development.2 Specific mutations in the ced-3 or the ced-4 genes that resulted in loss of gene function were shown to significantly alter programmed cell death. In contrast, CED-4 is a protease-activating factor that targets the CED-3 precursor protein, which in turn initiates cell death. In ced-9 mutants, all cells die during embryonic development. These studies indicate that CED-3 and CED-4 are required for cell death to occur; CED-9 acts upstream of CED-3 to suppress apoptosis; and the apoptotic pathways can be activated in all cells.
Many of the morphological changes of apoptosis in vertebrates are caused by a family of cysteine proteases, or caspases which are specifically activated in apoptotic cells. In fact, the main effector protease in C. elegans is CED-3. Effector caspases are usually activated proteolytically by an upstream caspase; initiator caspases are activated through regulated protein-protein interactions. Indeed, protein-protein interactions are one of the major themes in apoptosis. Capase-8 and caspase-10 contain a death-effector domain (DED), whereas caspase-2 and caspase-9 contain a caspase activation and recruitment domain (CARD). A third protein interaction module is present in several upstream regulators of apoptosis, such as CD95 and the adapter molecule Fas-associated death domain (FADD). It is possible that DED and CARD are derived from a common ancestral domain.3
The precise balance of protein-protein interactions is also crucial to the activity of the well-characterized Bcl-2 family of apoptosis regulators. The members of this family can be either pro- or anti-apoptotic, can homodimerize, or form key heterodimers to regulate cell death.4–6 One key function of Bcl-2 family members is to regulate the release of pro-apoptotic factors, such as cytochrome c, from the mitochondrial intermembrane space into the cytosol.5, 6 Pro-apoptotic family members added to isolated mitochondria are sufficient to induce cytochrome c release, whereas overexpression of Bcl-2 family members will counteract such an effect. The mechanisms by which Bcl-2 family members regulate cytochrome c release are complex and continue to be unraveled.
Death receptor-ligand interactions are important initiators of apoptosis that are processed via the death receptor or extrinsic pathway (Fig. 1). Death receptors belong to the tumor necrosis factor (TNF)/nerve growth factor receptor superfamily. These transmembrane proteins contain an extracellular ligand-receptor interacting domain, a single transmembrane domain and an intracellular death domain. In the liver, the most relevant death receptors include Fas (CD95/Apo-1), tumor necrosis factor receptor 1 (TNFR1), tumor necrosis factor-related apoptosis inducing ligand (TRAIL) receptor 1 (TRAIL-R1/death receptor 4 or DR4), TRAIL receptor 2 (TRAIL-R2/DR5), death receptor 3 (DR3), and death receptor 6 (DR6). Engagement by their cognate ligands Fas (FasL/CD95L), TNF-α, and TRAIL brings three receptor molecules into close proximity resulting in receptor oligomerization and activation. This induces the recruitment of several adaptors molecules, thus triggering the apoptotic cascade. TNF-α, which is released from macrophages, triggers the cell death and tissue destruction seen in certain chronic inflammatory diseases. The FasL is a cell-surface protein expressed by activated natural killer cells and cytotoxic T lymphocytes, which usually triggers death of virus-infected cells, tumor cells, and foreign graft cells.
Activated TNFR1 first activates NF-κB and JNK pathways, and then induces apoptosis subsequent to internalization.7 TRAIL-R1 and TRAIL-R2 initiate apoptosis, while TRAIL-R3 and TRAIL-R4, which do not contain death domains, function as decoy receptors.8 Once activated, death receptors initiate cleavage of procaspase-8 to active caspase-8, which in turn directly activates caspase-3, or engage the mitochondrial pathway of apoptosis by cleavage and activation of the Bcl-2 family member, Bid. In Type 2 signaling cells, such as hepatocytes, truncated Bid induces mitochondrial permeabilization and release of mitochondrial amplifiers of caspase-8-dependent apoptosis, such as cytochrome c. Cytosolic cytochrome c binds to the Apaf-1 complex, which activates caspase-9 and then caspase-3 via an ATP-dependent mechanism.
Can modulation of apoptotic pathways be a therapeutic strategy? Among all the apoptosis-drug targets, strategies aimed at modulating caspase activity are at the forefront in blocking cell death for many human diseases. Suppression of caspase-8/10 by both pharmacological and genetic approaches ameliorates apoptosis in several models of hepatic disease.9–12 One of these caspase targets is a broad-spectrum inhibitor that forms irreversible adducts with the active site cysteine of several caspases, and is in clinical trials for acute liver injury and hepatitis. A selective and reversible caspase-1 inhibitor is intended to treat arthritis and other inflammatory diseases. Interestingly, a nitric oxide derivative of ursodeoxycholic acid has been shown to effectively protect against liver damage in murine models of autoimmune hepatitis induced by concanavalin A or Fas agonist antibody, by inhibiting caspase activity.13 In this case, cysteine S-nitrosylation by intracellular nitroxide formation was suggested as the main mechanism responsible for caspase inhibition, rather than by an effect of ursodeoxycholic acid. Finally, silencing of caspase gene expression by small interfering RNA (siRNA) protects livers from ischemia/reperfusion- and Fas-induced injury.10, 12 Many issues related to caspase targets and disease indications remain unsolved, including the merits of broad spectrum versus selective inhibitors, reversible versus irreversible drugs, acute versus chronic indications, caspase-dependent cell death, and physiological roles of caspases beyond apoptosis.
Hepatocyte apoptosis can occur through both death receptor-dependent and -independent processes, ultimately involving mitochondrial dysfunction. Several approaches have been proposed for exploiting Bcl-2 family members as therapeutic targets. For example, in cholestatic liver injury, the BH3-only protein Bid is essential for hepatocyte apoptosis.14 Efforts to attenuate Bid-induced apoptosis decrease hepatic fibrosis that may eventually be useful for liver cytoprotection strategies. Also, manipulation of BH3 proteins with pharmacologic molecules and the use of siRNAs for Bax/Bak are attractive strategies for therapeutic intervention. Finally, knowledge regarding protein modifications that suppress Bcl-2 function could help in creating gain-of-function mutants with enhanced cytoprotective properties, including the use of caspase- and kinase-resistant forms of Bcl-2.15, 16 Bcl-2 gene transfer into cells before transplantation might be used with advantage in cell-replacement therapies, an application likely to emerge from stem cell technology.
Other potential targets for modulating mitochondrial dysfunction that are relevant to the regulation of cell death have been described. For example, manipulation of components of the mitochondrial transition-pore complex may prevent downstream apoptotic events. Nonimmunosuppressive analogs of cyclosporine A have shown therapeutic promise.17 Ursodeoxycholic acid is an effective treatment against cholestatic liver diseases. This endogenous bile acid acts as a potent inhibitor of the classical mitochondrial pathway of apoptosis, in part, by directly stabilizing membranes and preventing the mitochondrial permeability transition.18, 19 Furthermore, as a cholesterol-derived molecule, ursodeoxycholic acid interacts with nuclear steroid receptors, such as the glucocorticoid receptor, suggesting that it may also modulate gene expression.20, 21 Ursodeoxycholic acid was also shown to partially prevent apoptosis, via the death receptor pathway, in primary mouse hepatocytes co-cultured with fibroblasts that express the FasL.22
Several DED-carrying proteins have been identified in humans, in addition to DED-containing caspase-8 and caspase-10. Of these, the adaptor protein FADD is the only one that contains both a DED and a death domain (DD), which represents a nonredundant link between these two protein families. Therefore, it is no surprise that FADD is an attractive antisense target for disrupting caspase activation that is initiated by the TNF family of death receptors.23–25 Other strategies can target events upstream of caspase activation. These include neutralization of soluble Fas and TNF-α, which prevents Fas-mediated hepatocyte apoptosis.26 Fas gene silencing protects mice from both fulminant hepatic failure in acute injury and liver fibrosis after chronic injury.27 Finally, recombinant ligands and agonistic antibodies that trigger apoptosis through TNFR and altered intracellular signaling events are being investigated as potential cancer therapeutics.28
In this issue of HEPATOLOGY, Descamps et al.29 created death domain-deficient (ΔDD) membrane-anchored receptors to specifically block death signaling to target cells. Both TNF-α and FasL are known to be inflammatory cytokines and bind to TNFR1 and Fas on the cell membrane, respectively. In fact, the recognition of hepatitis virus by cytotoxic T lymphocytes activates FasL/Fas and TNF-α/TNFR and induces apoptosis of liver cells. In an effort to reduce the potential side effects from soluble TNF receptors, or anti-TNF-α antibodies, the authors focused on the use of both truncated TNFR1 (TNFR1ΔDD) and Fas (FasΔDD), which they concluded could potentially prevent TNF-α- and Fas-induced apoptosis of hepatocytes by interfering with trimerization of the death receptor subunits. Both decoy receptors were delivered to the liver by either recombinant adenovirus or hydrodynamic administration of nonviral recombinant plasmids.
As proof-of-principle, TNFR1ΔDD or FasΔDD were initially expressed on the surface of HeLa cells and inhibited their death by TNF-α and FasL, respectively. Expression of the Fas-decoy receptors in mice not only prevented concanavalin A-induced acute hepatitis, but also death by anti-Fas monoclonal antibody-induced fulminant hepatitis. Moreover, mice expressing TNFR1 decoy receptors escaped endotoxin-induced death. Hydrodynamic delivery of Fas-decoy receptors to the liver protected mice against anti-Fas monoclonal antibody induced fulminant hepatitis. Notably, both hepatocyte FasΔDD and TNFR1ΔDD expression provided at least as effective protection as their soluble counterparts Fas-Fc and TNFR1-Fc, respectively. The authors concluded that the decoy receptors acted as dominant-negative factors exerting local inhibition, thus avoiding systemic effects. It was unclear, however, as to whether the antagonism was the result of competitive binding to ligand, or alteration in signal transduction via the endogenous intact receptor. The authors dutifully considered this possibility, as well as design of the most potent ΔDD receptor(s). It was assumed, but not confirmed, that the binding affinities to both decoy and wild-type receptors was identical. In addition, there was no discussion in the article regarding the potential role of membrane-anchored antagonists that remained soluble after administration either by adenovirus or hydrodynamic delivery. Finally, usage of the term “dominant-negative” implies that the ΔDD receptors are more (in)active on a per molecule basis than the endogenous respective receptors, which may not be the case.
The results from this study represent a significant advance in the field and certainly suggest that ΔDD decoys could potentially be used to treat autoimmune hepatitis, fulminant hepatitis, and liver failure due to septic shock. In fact, the potential of using nonsignaling membrane-anchored death receptors to inhibit apoptosis is quite exciting, particularly when one considers the adverse side effects associated with neutralizing antibodies to death ligands or soluble death receptors. The advantage of decoy receptors over RNA interference (RNAi) is not as obvious as the authors suggested. Granted, the off-target effects such as the interferon effect are well documented. However, a number of recent reports30 demonstrate the enormous progress that is being developed for therapeutic applications of the RNAi technology. While incomplete gene silencing and “off-target activity” might constitute a major drawback to their use, similar types of problems would almost assuredly be associated with delivery and expression of decoy receptors. The clinical relevance of the technology remains tenuous at best. Although the adenovirus appears to be an efficient system to deliver the decoy receptors to hepatocytes, the potential immune response cannot be ignored. The use of hydrodynamic delivery as a therapeutic approach to treating injury to the liver remains impractical, at least for the short term. With the development of new and more efficient delivery systems, the use of decoy receptors may ultimately find a place in the treatment of both acute and chronic liver disease. It remains to be seen, however, as to whether the effect is as potent under conditions in which the decoy receptors are administered after the injury, i.e. the more rigorous test of such a therapeutic agent. Those types of experiments will be critical in moving from bench to bedside. The road to the clinic will undoubtedly be significantly bumpier and more winding than what it was to generate these exciting data in animal models.
Understanding the molecular mechanisms of apoptosis has revealed potential strategies for therapeutic intervention in many human diseases in which cell survival and death are unbalanced. Several of these therapeutic approaches have even progressed to clinical testing in humans, and those results will be crucial in evaluating prospects of modulating apoptosis as a potential therapeutic strategy. Continued efforts in identifying target molecules involved in apoptosis, the way they function and the mechanisms of regulation, should offer new options for pharmacological and/or gene mediated therapies for patients with liver diseases. It is now time to add decoy receptors to the growing list of options.