|Home | About | Journals | Submit | Contact Us | Français|
Ethanol intoxication stimulates the production of proinflammatory cytokines, increases the formation of reactive oxygen species, and induces mitochondrial impairment. However, information is limited as to the exact sequence and components involved in ethanol-induced hepatotoxicity. Acute ethanol exposure enhances mitochondrial superoxide (O2.−) production and impairs mitochondrial Ca2+ handling. In turn, O2.− facilitates cytochrome c release and mitochondrial membrane potential loss that is not dependent upon H2O2 and divalent cations and requires Bak in a Bax-independent fashion. Furthermore, triggering of Bak's proapoptotic activity requires the cytosolic presence of Bid, a BH3-only protein that is processed by the initiator caspase-2. Together, these studies identify an O2.−-driven, caspase-initiated apoptotic pathway that selectively involves the Bcl-2 family proteins Bid and Bak. This pathway manifests itself during chronic ethanol consumption in aged animals and identifies caspase-2, Bid, and Bak as essential mediators of O2.−-induced apoptosis that may prove effective targets for the development of therapeutics to treat alcoholic liver disease.
Alcoholic liver disease leads to hepatocyte death, liver cirrhosis, and organ dysfunction and is a major cause of death in the United States (22, 23, 37). Hepatocyte injury is most prominent in the centrilobular region of the liver acinus and stems from both necrotic and apoptotic cell death (17, 18, 32). However, the mechanism(s) by which alcohol leads to cell death is unclear. A prominent feature of alcohol exposure is the production of reactive oxygen species (ROS). Phagocytic NAD(P)H oxidase, xanthine (X) oxidase (XO), and the mitochondrial respiratory chain are the major sources of ROS under both physiological and pathological conditions. In hepatocytes, XO and mitochondria are likely sources of ROS during oxidative stress (10). Specifically, ROS contribute to alterations in mitochondrial morphology and bioenergetics. However, little is known about how alcohol-induced ROS production directly affects mitochondrial function and hepatocyte injury.
ROS are important mediators of apoptosis in liver diseases and are generated by all mammalian cells as by-products of abnormal metabolism and in response to paracrine factors such as ethanol (EtOH). Pathologically, both apoptosis and necrosis are initiated by ROS (2, 8, 20, 45). This damage is especially important in vascular and epithelial cells, which subsequently initiate a series of local chemical reactions and genetic alterations that ultimately result in amplification of the initial ROS-mediated tissue damage and/or cytotoxicity (9, 14, 30, 31). It is estimated that normal levels of ROS, generated in the basal region of cellular tissues from approximately 5% of the metabolized oxygen, are efficiently detoxified by endogenous enzymatic free-radical scavengers such as superoxide (O2.−) dismutase (SOD), glutathione peroxidase, and catalase (Cat) (9, 45). However, under conditions associated with excess production of ROS, such as acute and chronic EtOH exposure, the rate of ROS generated by tissues can exceed the capacity of endogenous oxidant defense mechanisms to detoxify ROS and prevent deleterious radical-mediated reactions.
An important target of cellular ROS is the mitochondria, resulting in mitochondrial dysfunction and permeabilization of the outer mitochondrial membrane and leads to the release of apoptotic proteins, including cytochrome c (1, 11, 30, 48). Specifically, we have demonstrated that the oxidant O2.− can lead to mitochondrial depolarization by facilitating cytochrome c release (27, 30). Moreover, the exogenous addition of cytochrome c rescued the mitochondria from this O2.−-induced depolarization, indicating that the primary mechanism of mitochondrial dysfunction via O2.− is cytochrome c release. Yet, the mechanism whereby O2.− facilitates outer mitochondrial membrane permeabilization and cytochrome c release is not clear.
A potential pathway in which O2.− may target the mitochondria is via the multidomain proapoptotic Bcl-2 family of proteins. Recent studies to characterize animals deficient in the Bcl-2 family members Bax and Bak suggest that these two proteins play essential but redundant roles in initiating mitochondrial apoptotic events (6, 25, 44, 49, 51, 55). However, despite their high homology, Bax and Bak have distinct subcellular localization and functional regulation. Bax is largely a cytosolic protein that must undergo conformational change to initiate its translocation to mitochondria prior to the initiation of apoptosis (33). In contrast, Bak is a resident mitochondrial protein whose proapoptotic activity can be activated by proapoptotic BH3 peptides (50). Although controversy exists about the induction of apoptosis in the liver of mice following chronic consumption of EtOH via a Lieber-DeCarli liquid diet, recent investigations have demonstrated that chronic EtOH consumption increases the expression of antiapoptotic Bcl-2 and Bcl-xL proteins by an interleukin-6-dependent mechanism (19). Upregulation of proapoptotic Bax proteins has also been observed in patients with alcoholic liver disease, but the modification of BH3-only proteins has not been determined (19).
Here, we used a novel approach to identify the cytosolic components required for mitochondrial dysfunction during oxidative stress. The O2.−-mediated mitochondrial phase of apoptosis is mainly dependent on Bid and Bak but not Bax. Further, we demonstrate that mitochondrial membrane potential (ΔΨm) loss and the release of cytochrome c via this pathway can be antagonized by Bcl-xL. We also show that O2.−-evoked caspase-8 activation and Bid cleavage require caspase-2. Alcohol-induced O2.− generation sensitizes cells to proinflammatory cytokine-mediated apoptosis. Consistent with these findings, Bid processing has been manifested in aged, chronically alcohol-fed animals.
HepG2 cells were grown in minimal essential medium supplemented with 10% fetal bovine serum (FBS), 2 mM glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. Wild-type (WT), bax−/−, bak−/−, bax−/− bak−/−, bid−/− (54), and caspase-2−/− (c2−/−) mouse embryo fibroblasts (MEFs) (5) were grown in Dulbecco's modified Eagle's medium supplemented with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin. Mito-GFP (green fluorescent protein)-expressing murine endothelial cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% FBS, 100 U/ml penicillin/streptomycin, and endothelial cell growth supplement.
Permeabilized cells or isolated cytosol fractions were incubated with X (100 μM; Sigma) and XO (5 mU/ml; Sigma) for 10 min or treated with 2 μM potassium O2.− (14) for 15 min at 37°C. SOD (1,000 U/ml) and Cat (2,000 U/ml) were added to the reaction mixture to scavenge O2.−. Mitochondrial O2.− production and Ca2+ uptake were measured as previously described (15, 31).
MEFs (~7 × 106) were resuspended and permeabilized with 40 μg/ml digitonin in 1.6 ml of intracellular medium buffer (ICM; 120 mM KCl, 10 mM NaCl, 1 mM KH2PO4, 20 mM HEPES-Tris, pH 7.2) supplemented with protease inhibitors and phenylmethylsulfonyl fluoride (1 mM). To prevent transition metal-mediated conversion of O2.− into H2O2 and OH. and to lower the ambient Ca2+, we passed all of the buffers used for ΔΨm and biochemical assays through a metal-chelating Bio-Rad Chelex-100 cation-exchange column prior to performing the experiments. The Ca2+-free ICM was prepared with a Chelex column before the addition of protease inhibitors. Cells were permeabilized with digitonin (40 to 60 μg/ml) for 7 min in ICM, and the effectiveness of permeabilization was evaluated by trypan blue exclusion. Cells were centrifuged at 14,000 rpm for 10 min at 4°C, and the supernatant was collected as the cytosol.
HepG2 cells (~10 million) were rinsed with Ca2+-free extracellular medium (120 mM NaCl, 5 mM KCl, 1 mM KH2PO4, 0.2 mM MgCl2, 0.1 mM EGTA, 20 mM HEPES, pH 7.4) prior to permeabilization. Cells were resuspended in ICM supplemented with protease inhibitors and phenylmethylsulfonyl fluoride (1 mM) and permeabilized with 40 μg/ml digitonin. To prevent cellular energy depletion, the cells were exogenously supplemented with ATP and an ATP-regenerating system (creatine phosphate and creatine phosphokinase). After 15 min of permeabilization, cells were centrifuged at 14,000 rpm for 20 min at 4°C. The supernatant was collected as HepG2 cytosol.
Mice were injected with heparin (1.4 U/g intraperitoneally) 10 to 30 min before being anesthetized with pentobarbital sodium. Hearts were perfused and removed. The isolated hearts were placed in ice-cold mitochondrial isolation buffer at 4°C, and mitochondria were isolated as previously described (28). The mitochondrial pellet was suspended in ICM buffer supplemented with succinate, ATP, and an ATP-generating system for ΔΨm measurement.
Fluorescence was monitored in a multiwavelength excitation dual-wavelength emission fluorimeter (Delta RAM; PTI) with 490-nm excitation/535-nm emission for the monomeric form (JC-1; Molecular Probes) and 570-nm excitation/595-nm emission for the J aggregate of JC-1 (7). ΔΨm was measured as the ratio of the fluorescence of the J aggregate (aqueous phase) and monomer (membrane-bound) forms of JC-1. The ΔΨm of GFP-expressing MEFs was measured by tetramethylrhodamine ethyl ester (TMRE) to avoid GFP interference with ΔΨm measurement with JC-1. Cells were loaded with 1 μM TMRE in digitonin-containing ICM supplemented with succinate, ATP, and an ATP-regenerating system. After loading of the TMRE for 7 min, cells were centrifuged and resuspended in ICM supplemented with succinate, ATP, and an ATP-regenerating system. Fluorescence was monitored at 540-nm excitation and 580-nm emission.
Digitonin-permeabilized cells were centrifuged at 14,000 rpm for 10 min. The supernatant was collected as the cytosol fraction and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting with an anti-cytochrome c antibody (BD-Pharmingen). To determine activation of the O2.−-mediated apoptotic cascade, digitonin-permeabilized cells were treated with O2.− for 10 min and centrifuged and the supernatants were probed for caspase-2 (Santa Cruz), caspase-8 (R&D Systems), and Bid (BD-Transduction). Bid processing in rat liver samples was probed with an anti-Bid antibody (Imgenex).
O2.−-treated cytosol samples were filtered with a 0.45-μm-pore-size cellulose acetate membrane-containing syringeless filter device (Whatman). Gel filtration was performed at 4°C by fast protein liquid chromatography (FPLC), and the Superose 6 column was equilibrated with ICM buffer. Column calibration was carried out with a gel filtration protein standards kit (Amersham). The pretreated samples were directly loaded onto a Superose 6 FPLC column at a flow rate of 0.3 ml/min. Forty fractions were collected, and aliquots containing 150 ml of the 500-ml fractions were used to assay cytochrome c release from the mitochondrial fraction. For in vitro assay of cytochrome c release from mitochondria, HepG2 cells were permeabilized with ICM buffer (40 μg/ml digitonin) supplemented with succinate and an ATP-generating system for 10 min at 37°C. The permeabilized cells were centrifuged at 14,000 rpm for 10 min and washed three times with ICM buffer. An aliquot of each gel filtration fraction was incubated with the cytosol-free cell pellet (mitochondrial fraction) for 15 min. Following centrifugation, the supernatants were subjected to immunoblotting for cytochrome c.
WT and Bid−/− knockout (KO) MEFs cultured on 25-mm-diameter glass coverslips were loaded with the O2.− indicator dihydroethidine (10 μM), and images were collected with a Carl Zeiss 510 META NLO confocal system as previously described (15).
MEFs cultured on glass bottom dishes were treated with an O2.− generation system for the indicated time periods. After treatment, the cells were incubated with annexin V Alexa 488 conjugate for 15 min with 1× annexin binding buffer (Molecular Probes). During annexin binding, TMRE (50 nM) was also loaded (10 min) for ΔΨm measurement. To assess plasma membrane integrity, the nuclear dye TOTO-3 iodide (1:1,400; Molecular Probes) was also applied during the annexin and TMRE loading. In normal cells, TOTO-3 iodide is not able to permeate the plasma membrane but enters cells as they lose plasma membrane integrity. Both untreated and treated cells were visualized by confocal microscopy after staining. A Carl Zeiss 510 Meta confocal imaging system was used to acquire images.
All animal experiments were conducted in accordance with the National Institutes of Health guidelines and were approved by the Institutional Animal Care and Use Committee. Fisher brown rats (F344BNF1) were purchased from a National Institute on Aging facility. Rats were housed individually, randomly assigned to a group, and fed a nutritionally adequate liquid diet in which alcohol (EtOH) was provided at 0.5 or 5%. Control rats received the same diet but with maltose-dextrin isocalorically substituted for alcohol. These diets were originally formulated by Lieber and DeCarli (24) and were purchased from Bio-Serv (Frenchtown, NJ). Control and alcohol liquid diet feeding started at an age of 2 months for the young group and at 24 months for the aging group. After 24 weeks of feeding, the animals were divided into six groups as follows: (i) young, control diet-fed rats (control young, n = 4); (ii) young, 0.5-g/dl alcohol diet-fed rats (0.5% alcohol young, n = 4); (iii) young, 5-g/dl alcohol diet-fed rats (5% alcohol young, n = 4); (iv) aging, control diet-fed rats (control aging, n = 5); (v) aging, 0.5-g/dl alcohol diet-fed rats (0.5% alcohol aging, n = 5); (vi) aging, 5-g/dl alcohol diet-fed rats (5% alcohol aging, n = 5). Animals were euthanized, and livers were harvested and homogenized with a Polytron homogenizer.
Chronic EtOH exposure promotes the development of alcoholic liver disease, but the mechanisms underlying EtOH-induced hepatotoxicity remain poorly understood (23). An important site manifesting damage brought on by EtOH is the mitochondria. To evaluate the role of mitochondria in alcohol-induced cell injury, we examined the effects of EtOH on mitochondrial morphology and functional changes. Acute delivery of EtOH (50 mM) resulted in mitochondrial fragmentation from a filamentous to a globular morphology (Fig. (Fig.1A),1A), which may impact mitochondrial function and could constitute a major cause of EtOH-induced cell injury. A prominent feature of mitochondrial functional alterations is the production of ROS (29). To assess ROS production, endothelial cells expressing Mito-GFP were loaded with the mitochondrial O2.− indicator MitoSOX Red and treated with EtOH for 30 h. MitoSOX Red is nonfluorescent until oxidized by O2.−, and an increase in the fluorescence of MitoSOX Red indicates oxidation by mitochondrial O2.−. Accordingly, EtOH-fragmented mitochondria exhibit exaggerated O2.− production (Fig. 1B and C). Since ROS have been implicated in numerous pathologies and have been proposed to contribute to mitochondrial dysfunction (46), we next tested whether EtOH has an effect on mitochondrial Ca2+ handling. In contrast to the transient Ca2+ uptake in control cells (Fig. (Fig.1D),1D), EtOH treatment evoked a sustained elevation of mitochondrial Ca2+ levels in response to the G-protein-coupled receptor physiologic agonist bradykinin (10 nM) (Fig. (Fig.1E).1E). Elevated mitochondrial ROS and altered mitochondrial Ca2+ handling have been shown to directly facilitate mitochondrial dysfunction (31, 38). This supports the hypothesis that EtOH-mediated O2.− production is required for mitochondrial functional changes.
These data demonstrate that EtOH treatment induces mitochondrial O2.− production, which affects mitochondrial structure and function. However, O2.− can be dismutated into hydrogen peroxide and it is unclear which oxidant species is responsible for the effects of EtOH on mitochondrial function. To regulate the formation of O2.−, we used a system of X plus XO (X+XO) where O2.− is generated directly in the cytosol (30). This O2.− generation system was introduced into MEFs that had been permeabilized and loaded with the ΔΨm indicator JC-1 in the presence of oligomycin to inhibit the F0/F1 ATPase's ability to maintain ΔΨm. Following the introduction of X+XO into permeabilized cells, there is a rapid loss of ΔΨm (Fig. (Fig.2A).2A). To demonstrate the specificity of O2.−, pretreating the cells with the O2.− scavenger Mn(III)tetrakis (4-benzoic acid) porphyrin (MnTBAP) prevented ΔΨm loss (Fig. (Fig.2B2B).
One mechanism implicated in mitochondrial depolarization during drug-induced, EtOH-induced, or ischemic cell death is mediated by Ca2+, which induces a cyclophilin D-dependent opening of the mitochondrial permeability transition pore, resulting in depolarization and irreversible swelling of mitochondria (3, 4, 35, 41). However, chelators of Ca2+ (EGTA) (Fig. (Fig.2C)2C) and the other sulfhydryl group-reactive metals [N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine (TPEN)] (Fig. (Fig.2D)2D) fail to prevent O2.−-induced ΔΨm loss. In addition, O2.−-induced mitochondrial depolarization was not able to be mimicked by the addition of hydrogen peroxide (H2O2) at doses of up to 1 mM (Fig. (Fig.2E),2E), suggesting that it is O2.− itself that initiates the apoptotic response and that it is not the result of a more generalized form of cellular stress caused by ROS. In addition to cyclophilin D-dependent ΔΨm loss, mitochondrial permeabilization can occur through the action of proapoptotic Bcl-2 family proteins, which can be countered by the antiapoptotic Bcl-2 family member Bcl-xL. As an independent confirmation that the loss of ΔΨm and cytochrome c release resulted from the action of proapoptotic Bcl-2 family members, we found that addition of recombinant Bcl-xL to the permeabilized cells, followed by cellular exposure to O2.−, prevented both the loss of transmembrane potential and cytochrome c release in a dose-dependent fashion (Fig. 2F to H). Paracrine signaling-derived O2.− was previously shown to elicit IP3-linked Ca2+ signaling (31) without inducing endoplasmic reticulum (ER) stress (data not shown). Together, these data suggest that O2.−-induced loss of ΔΨm results from activation of the intrinsic apoptotic pathway.
Studies of bax−/− bak−/− double-KO (DKO) mice indicated that the proapoptotic Bcl-2 family members Bax and Bak are essential for mitochondrion-mediated cell death (25). Our data (Fig. 2F to H) demonstrate that antiapoptotic Bcl-xL prevented O2.−-induced ΔΨm loss and cytochrome c release, implying a role for proapoptotic Bcl-2 proteins. Therefore, we studied whether the O2.−-induced mitochondrial functional changes are blocked by either Bax or Bak. bax−/− bak−/− DKO cells were simultaneously permeabilized and loaded with JC-1. After addition of the mitochondrial uncoupler carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP; 1 μM), the JC-1 ratio displayed a rapid mitochondrial depolarization in both DKO and WT cells (Fig. (Fig.3A).3A). However, treatment of WT MEFs with O2.− at the indicated time periods triggered a rapid and large ΔΨm loss (Fig. (Fig.3B)3B) and cytochrome c release (Fig. (Fig.3D)3D) whereas bax−/− bak−/− DKO cells were completely resistant to O2.−, indicating that Bax and/or Bak are required for these events (Fig. 3C and D).
Since O2.−-induced ΔΨm loss and cytochrome c release were absent in bax−/− bak−/− DKO cells and antiapoptotic Bcl-xL blocked the O2.− effect, we sought to determine whether Bax and/or Bak is the crucial candidate for the mitochondrial changes observed following O2.− exposure. MEFs from bak−/− mice were resistant to mitochondrial depolarization when exposed to the X+XO O2.−-generating system (Fig. (Fig.4A).4A). In contrast, Bax-deficient cells displayed levels of depolarization and cytochrome c release equivalent to those of WT cells (Fig. 4A and B). These data demonstrate that Bak is required for O2.−-induced mitochondrial depolarization and cytochrome c release. O2.− production at the level of the mitochondrial electron transport chain has been thought to be a crucial causative factor in the initiation of apoptosis in various disease conditions and in EtOH toxicity (46). O2.− produced as a consequence of antioxidant depletion has also been shown to trigger mitochondrial dysfunction and apoptosis. Since Bak is the major proapoptotic Bcl-2 family member constitutively resident in mitochondria, we investigated its role in O2.−-dependent ΔΨm loss in vitro. Reexpression of Bak in bax−/− bak−/− cells restored the ability of O2.− to initiate apoptosis (Fig. (Fig.4D).4D). In the experiment whose results are shown in Fig. 4C to E, GFP alone or together with Bax or Bak was introduced into bax−/− bak−/− cells. To avoid interference with the GFP, the cationic potentiometric dye TMRE was used rather than JC-1 for the measurement of mitochondrial depolarization. In this experimental model, TMRE dissociates from the mitochondria into the buffer upon depolarization and is detected as an increase in fluorescence. Reintroduction of Bak was sufficient to reestablish the loss of ΔΨm, as indicated by TMRE unquenching following treatment with O2.−. In contrast, reintroduction of GFP alone or Bax-GFP failed to restore the sensitivity of bax−/− bak−/− cells to X+XO (Fig. 4C and E). Together, these data implicate Bak as necessary for O2.−-initiated mitochondrial depolarization and cytochrome c release. Finally, to confirm that these results were not a tissue culture phenomenon, mitochondria were isolated from the hearts of WT, bak−/−, and bax−/− animals. Mitochondria were loaded with JC-1 and used to assess the ability of O2.−-treated cytosol to induce depolarization in different mitochondrial preparations over time. O2.−-treated cytosol induced a mitochondrial depolarization in both WT and bax−/− deficient mitochondria. In contrast, mitochondria isolated from bak−/− mice were resistant to O2.−-induced depolarization (Fig. (Fig.5).5). Remarkably, buffer supplemented with the O2.−-generating system alone did not elicit ΔΨm loss, pointing to the existence of an-as-yet-unidentified cytosolic factor that interacts with Bak to elicit its proapoptotic effects. Overall, our results clearly demonstrate that Bak, but not Bax, is essential for O2.−-induced ΔΨm loss and cytochrome c release.
Although Bak is a resident mitochondrial protein that can initiate cytochrome c release, its proapoptotic activity is believed to be initiated through the binding of BH3-only-containing proteins that act as apoptotic activators. This suggests that O2.− may be responsible for generating an active BH3 protein in the cytosol. We purified cytosol from HepG2 cells and treated it with X+XO to determine whether cytosol treatment with O2.− could induce apoptotic activity. This O2.−-treated cytosol was then incubated with mitochondria and tested for cytochrome c release activity. Cytosol pretreated with O2.− facilitated cytochrome c release from mitochondria (Fig. (Fig.6A).6A). Once the cytosol was treated with O2.−, addition of ROS scavengers to the cytosol prior to incubation with the mitochondria had no effect on the cytochrome c-releasing activity (data not shown), suggesting that it was not O2.− itself that was directly inducing Bak-dependent cytochrome c release from mitochondria but rather that O2.− was activating a cytosolic protein to initiate its Bak-dependent mitochondrial permeabilization. To determine the potential molecular weight of this substance, O2.−-treated, organelle-free HepG2 cytosol in which residual ROS was quenched by SOD plus Cat was loaded onto a Superose 6 gel filtration column and the flowthrough was collected. An aliquot of each fraction was incubated with mitochondria, and the cytochrome c release activity of the fractions was determined. High levels of cytochrome c release activity were observed in a series of low-molecular-weight fractions, suggesting that this activation of Bak was not due to an oligomeric complex but instead was potentially the result of a low-molecular-weight intermediate (Fig. (Fig.6B6B).
One low-molecular-weight BH3-only protein known to be expressed as a zymogen in the cytosol is the Bcl-2 family member Bid. Accordingly, bid−/− deficient MEFs were found to be insensitive to O2.−-induced mitochondrial depolarization and cytochrome c release (Fig. (Fig.6C).6C). Because the inability of bid−/− cells to respond was due to the absence of Bid in the cytosol, we next assessed whether reintroduction of Bid into the cytosol can restore O2.−-induced mitochondrial depolarization. Rapid mitochondrial depolarization was observed if a mitochondrial pellet from bid−/− cells was incubated with a WT cytosol exposed to O2.−. In contrast, a WT mitochondrial pellet incubated with O2.−-treated cytosol isolated from a bid−/− cell failed to induce mitochondrial depolarization (Fig. (Fig.6D).6D). The addition of full-length recombinant Bid to Bid-deficient cytosol restored the cytosol's ability to induce mitochondrial depolarization following O2.− exposure (Fig. (Fig.6E).6E). However, recombinant Bid along with O2.− did not directly induce depolarization of purified mitochondria (Fig. (Fig.6F),6F), indicating that an upstream player is required for Bid cleavage in the mechanism of O2.−-induced ΔΨm loss.
During apoptotic events, activation of Bid occurs via caspase cleavage. Bid has been reported to be a molecular intermediate in death receptor-initiated mitochondrial depolarization and cytochrome c release following its cleavage by caspase-8 (21, 26). In addition, caspase-2 has been reported to be capable of cleaving Bid into its proapoptotic form in response to activation by chemotherapeutic agents (39, 52). To investigate a potential role for caspases as the upstream mediators of O2.−-induced Bid activation, X+XO was introduced into digitonin-permeabilized cells along with either the O2.− scavenger MnTBAP or the pan-caspase inhibitor zVAD-fmk. Both drugs were able to prevent mitochondrial depolarization and cytochrome c release in response to O2.− (Fig. (Fig.7A).7A). O2.− treatment resulted in the generation of a 15-kDa form of Bid, which was prevented by either the O2.− scavenger MnTBAP or zVAD-fmk (Fig. (Fig.7B7B).
To identify which caspases were activated in response to O2.− treatment of the cytosol, cytosol from WT and bid−/− cells was prepared. Quantitative processing of both caspase-2 and caspase-8 to their enzymatically activated forms was observed following O2.− treatment of the cytosol of both WT and bid−/− MEFs (Fig. (Fig.7C).7C). In principle, the activation of either caspase would be sufficient to lead to the cleavage of Bid observed in WT cells, as both have been reported to utilize Bid as a substrate in their apoptotic cascade (12). Unlike caspase-8, a fraction of caspase-2 is normally localized to mitochondria, where O2.− generation occurs during EtOH exposure. In addition, procaspase-2 exists in an inactive, sulfhydryl-linked dimer that might be susceptible to oxidant-induced regulation (42). Therefore, caspase-2-deficient MEFs were tested for whether or not caspase-2 is required for the Bid cleavage activity in O2.−-treated cytosol. The O2.−-induced cleavage of both caspase-8 and Bid observed in WT cells was eliminated in caspase-2-deficient MEFs (Fig. (Fig.7C).7C). In addition, caspase-2-deficient cells failed to undergo O2.−-induced mitochondrial depolarization (Fig. (Fig.7D).7D). In contrast, caspase-8-deficient cells exhibited normal caspase-2 cleavage and substantial Bid cleavage (data not shown). Thus, O2.−-induced activation of the intrinsic apoptotic pathway depends on caspase-2.
Surprisingly, there does not appear to be any redundancy in the caspase and Bcl-2 family members that participate in the initiation of this apoptotic cascade. Caspase-2 is a required initiator, and Bid is the required caspase substrate responsible for inducing mitochondrial cytochrome c release in a Bak-dependent fashion. Together, these data demonstrate that individual caspases and Bcl-2 family members can make up dedicated sensing pathways that initiate apoptosis in response to the presence of toxic intracellular mediators.
Thus far, we have demonstrated a novel mechanism of O2.−-mediated mitochondrial membrane permeabilization that relies on the caspase-2-mediated generation of tBid and subsequent activation of Bak, resulting in mitochondrion-initiated apoptosis. To assess cell viability in intact cells, WT, bax−/− bak−/−, bak−/−, bax−/−, bid−/−, and c2−/− MEFs were treated with O2.− (X+XO) for 12 h, at which time they were incubated with TMRE to determine ΔΨm, along with the cell death markers annexin V and TOTO-3. An early event in the apoptotic cascade is externalization of inner membrane leaflet phosphatidylserine, which will then bind annexin V. As apoptosis progresses, plasma membrane integrity is compromised, allowing the indicator TOTO-3, which is normally unable to permeate the plasma membrane, to enter the cell. In response to O2.−, both WT and bax−/− cells exhibited a complete loss of ΔΨm, as well as positive annexin V and TOTO-3 staining, indicative of apoptosis. Conversely, bax−/− bak−/−, bak−/−, bid−/−, and c2−/− MEFs maintained an intact ΔΨm and were resistant to O2.−-induced apoptosis (Fig. (Fig.8A).8A). Cell viability was also preserved in bax−/− bak−/− (Fig. (Fig.8B),8B), bak−/− (Fig. (Fig.8C),8C), bid−/− (Fig. (Fig.8E),8E), and c2−/− (Fig. (Fig.8F)8F) MEFs, in contrast to bax−/− (Fig. (Fig.8D)8D) and WT (Fig. 8B to F) MEFs. This enhanced resistance to cell death was also present at increasing doses of O2.− (Fig. (Fig.8G).8G). To investigate the potential gain-of-function effects associated with Bid and C2 on O2.−-mediated apoptosis, WT and D59E mutant Bid and C2 WT plasmids were re-expressed in Bid- and C2-deficient MEFs, respectively (47). WT Bid- and C2-overexpressing cells, but not D59E mutant Bid-overexpressing cells, showed increased cell death (Fig. (Fig.8H).8H). To elucidate whether both extracellular and intracellular ROS converge on a similar pathway to trigger cell death via Bid, caspase-2, and Bak, we used the intracellular O2.−-generating agent menadione. Similar to paracrine signaling-derived ROS, bid−/−, c2−/−, and bak−/− deficient MEFs are resistant to menadione-induced cell death (Fig. (Fig.8I),8I), indicating that O2.− released from the mitochondria via the voltage-dependent anion channel (13) can activate a common O2.−-dependent apoptotic pathway (see Fig. S1 and S2 in the supplemental material).
It is well established that chronic EtOH consumption elicits signs of oxidative stress in both humans and rodents (34). In fact, chronic alcoholism in humans is associated with increased protein oxidative damage, as well as a depletion of small-molecule antioxidants in the liver (34). To test the significance of this selective role of ROS, the functional role of O2.− anion in tumor necrosis factor alpha (TNF-α)-stimulated cell death was determined in WT and bid−/− EtOH-treated cells. Both WT and bid−/− cells exposed to EtOH generated enhanced, but similar, levels of O2.− production over untreated cells (Fig. 9A and B). Although both cell types generate equal levels of ROS following EtOH exposure, WT but not bid−/− cells demonstrated increased sensitivity to TNF-α/cycloheximide-induced cell death (Fig. (Fig.9C).9C). Sensitization from TNF-α/cycloheximide-induced cell death in EtOH-fed WT MEFs was reversed by adenovirus-mediated gene transfer of mitochondrial manganese SOD (Fig. (Fig.9C).9C). These data suggest that enhanced ROS production in response to EtOH ingestion sensitizes cells to inflammation-mediated apoptosis. To translate our in vitro findings to an in vivo setting, we used the Lieber-DeCarli model of chronic EtOH consumption. Briefly, six groups of animals were studied: a young control group (2 months old) without chronic EtOH consumption (n = 4), a group of young animals fed a liquid diet containing either 0.5 or 5% EtOH for at least 24 weeks (n = 4), a group of control aged animals (24 months old) without chronic EtOH consumption (n = 5), and aged animals fed a liquid diet of either 0.5 or 5% EtOH for 24 weeks. Our in vitro findings demonstrated that Bid is requisite for O2.−-induced ΔΨm loss. We therefore probed for the generation of tBid in the above-mentioned six different animal groups. Liver samples from young animals subjected to two different doses of EtOH (0.5 and 5%) did not display any Bid cleavage (Fig. (Fig.9D).9D). Remarkably, four out of five liver samples from aged, chronically 5% EtOH-fed rats demonstrated substantial Bid cleavage, indicating that aged animal liver is more vulnerable to alcohol-induced damage (Fig. 9D and E). The fact that O2.−-induced apoptosis results from specific molecular mediators and the fact that EtOH exposure significantly triggered mitochondrial O2.− production and mitochondrial functional changes suggest that all three proteins in the O2.−-induced apoptosis pathway are potential molecular targets to prevent organ injury.
Using in vitro and in vivo models, we identified the intracellular components involved in the O2.−-evoked mitochondrial phase of apoptosis as it occurs during oxidative stresses such as that due to EtOH consumption. Further, we provided several lines of evidence that O2.−-mediated tBid generation induces selective activation of mitochondrial Bak, triggering cytochrome c release and ΔΨm loss that lead to apoptosis. ROS are a key factor in various diseases, including alcoholic diseases and ischemia/reperfusion, and trigger the decay of mitochondrial function through the cell death pathway. Although mitochondria play an important role during the onset of apoptosis, little is known about the organization of the upstream signaling pathways during alcohol-induced liver injury. Here we show that O2.− generation during oxidative stress promoted mitochondrial cytochrome c release and ΔΨm loss that are independent of other oxidants. The proapoptotic Bcl-2 family protein Bak, but not Bax, is essential for O2.−-evoked mitochondrial events and apoptosis. Oxidative stress induced by O2.− also activates caspase-2 and caspase-8, which induce Bid cleavage and subsequent activation of Bak. The antiapoptotic protein Bcl-xL and O2.− scavengers control O2.−-evoked cytochrome c release and ΔΨm loss. These results strongly indicate that O2.− mediates mitochondrial dysfunction through Bid and Bak activation controlled by Bcl-xL.
Previous work with cell-free systems has identified and characterized several Bcl-2 family proteins that are involved in apoptosis (36, 40). Although they possess multiple functions, an important role of Bcl-2 family proteins appears to be associated with mitochondrion-dependent apoptosis. In our study, cytochrome c release was noticed when O2.−-pretreated cytosol was incubated with mitochondria and occurred independently of other oxidants. Further, our fractionation study suggested that cytosolic components alone appear to be sufficient to evoke cytochrome c release following O2.− application. Although oxidative stress can regulate the antiapoptotic members Bcl-2, Bcl-xL, and Mcl-1 and the proapoptotic members Bid, Bad, Bak, and Bim, the mechanism through which O2.− modulates the proapoptotic Bcl-2 proteins is unknown. Since we observed a kinetic difference between tBid and O2.− (27, 30), we hypothesized that proapoptotic proteins could be altered following the application of O2.−. Previously, ROS have been shown to directly alter mitochondrial function and lead to apoptosis. Surprisingly, we found that bax−/− bak−/− DKO cells exposed to O2.− were resistant to mitochondrial and apoptotic events, indicating the involvement of Bcl-2 family proteins in oxidant stress. bax−/−, but not bak−/−, cells responded to O2.− similarly to WT cells, and activation of Bak was dependent upon the generation of tBid. In addition, animals chronically exposed to EtOH induction and oxidative stress demonstrated tBid cleavage, indicating the importance of this pathway in cell death. The in vivo and in vitro findings in our research strongly suggest that Bid and Bak are critical factors in O2.−-mediated apoptotic events at the mitochondrial level. In support, Wei et al. indicated that oligomerization of Bak triggered by tBid facilitated cytochrome c release by physical association (50). Enforced dimerization of Bax altered ΔΨm loss and subsequent ROS production, but the release of cytochrome c was not directly proportional to this phenomenon. However, since Bax is not responsible for O2.−-induced mitochondrial-phase events, it is likely that a cytosolic element is required for the activation of Bak during O2.− exposure. BH3 domain proteins such as Bid, Bad, Bim, Noxa, and Puma have been shown to absolutely require Bax or Bak to induce mitochondrial dysfunction and cell death in bax−/− bak−/− DKO cells. The BH3 protein Bid appears to be ubiquitously expressed, and generation of subnanomolar tBid concentrations is sufficient to trigger complete cytochrome c release and ΔΨm loss. Fas and TNF receptor 1 apoptotic death signals elicit cytochrome c release via caspase-8-processed Bid. Moreover, bid−/− deficient mice were resistant to Fas-mediated hepatocyte and ischemic neuronal death (53). Interestingly, our study shows that bid−/− KO cells are completely resistant to O2.−-induced mitochondrial events. Furthermore, our findings suggest that cytosolic Bid is a target for O2.− and that tBid selectively utilizes mitochondrial membrane-bound Bak. Fas- or TNF receptor-mediated caspase-8 activation cleaves Bid at position 59 (Asp) to generate tBid, and inhibition of Bid cleavage prevents apoptosis (21, 26, 47). In the present study, the potential role of Bid in O2.−-induced cell death suggests convergence at the Bid modification step. Although caspase-2 played a key role in apoptosis by trophic factor deprivation, β-amyloid cytotoxicity, and granzyme B, caspase-2-deficient mice lack a discernible phenotype. Nonetheless, caspase-2 is involved in stress and chemotherapeutic drug-induced apoptosis, most likely upstream of the mitochondria. Oxidative stress induced by O2.− also activates caspase-2 and Bid cleavage. This observation provides evidence for the involvement of caspase-2 and Bcl-2 family proteins in oxidative stress-induced apoptosis.
The most important implication of our studies is that Bid serves as a cytosolic factor for apoptosis during oxidative stress. The role of Bak and O2.−-modified Bid in mediating mitochondrial permeabilization is independent of Bax activity, as revealed by bak−/− and bid−/− KO cells. This suggests that the association of Bak and Bid plays a crucial role in mediating mitochondrial events. Consistent with this hypothesis, stable expression of Bak in bax−/− bak−/− DKO cells reestablished O2.−-mediated ΔΨm loss. Furthermore, cytosol exchange between WT and bid−/− KO cells reveals that WT cytosol containing Bid is sufficient for Bak-mediated ΔΨm loss. Previous reports suggest that Bax may interact with Bak in a Bid-independent fashion (7). Our results indicate that cross talk events involving the proapoptotic Bcl-2 proteins Bid and Bak may represent a vital component of O2.−-induced death during oxidative stress such as that due to chronic alcohol consumption. Studies have indicated that diverse classes of death agonists act at either the mitochondrial or the ER level and require Bax and Bak (43). Apoptotic signals delivered by lipid second messengers, H2O2, and intrinsic signal activators (staurosporine) require Ca2+ dynamics between the ER and mitochondria to evoke Bax- and Bak-induced apoptosis. Our experiments with Bax KO and Bax-rescued DKO cells establish a dispensable role for Bax in O2.−-mediated apoptosis. However, our data suggest that mitochondrial Bak is essential in mediating mitochondrial permeabilization and apoptosis and that the activation of Bid and Bak plays a pivotal role in the pathogenesis of alcohol injury. Finally, selective inhibitors of caspase-2, Bid, or Bak delivered intravenously may reduce cell death and restore organ function in cells susceptible to alcohol-mediated damage, such as those of the liver. TNF-α constitutes a major factor in the development of alcohol-induced liver injury (16). In the present study, alcohol-induced mitochondrial ROS production sensitized cells to TNF-α-induced cell death. Importantly, our results obtained with in vitro systems suggest that proinflammatory cytokine-induced cell death via caspase-2 cleavage of Bid is exaggerated in response to EtOH. In particular, caspase-2 is an attractive target for this type of therapeutic intervention as caspase-2 has a caspase cleavage site distinct from that of other initiator caspases. Proteins that contain thiol groups are particularly susceptible to oxidation and may represent important targets in oxidative signaling. Cysteine is an oxidative target because of the reactivity of the thiol group, which is susceptible to modification by free radicals. Oxidation of these critical thiol groups can increase or decrease the activity of these proteins. Since the caspase-2 processing site contains a cysteine residue at position 320, we hypothesized that oxidative modification or mutation of the cysteine residue at position 320 could alter caspase-2 cleavage. To test this, cells were transfected with either a WT or a C320A mutant caspase-2 plasmid and cytosolic fractions were treated with O2.− (X+XO). However, incubation with O2.− induced caspase-2 cleavage in both systems (data not shown), suggesting that O2.− triggers caspase-2 cleavage through an as-yet-unidentified mechanism. The results of this study suggest that mitochondrial Bak is essential for O2.−-processed Bid in mediating mitochondrial permeabilization and apoptosis and that the activation of Bid and Bak plays a pivotal role in the pathogenesis of organ injury during chronic alcohol exposure.
We thank Junying Yuan for caspase-2 KO MEFs, Xiao-ming Yin for Bid KO MEFs, and Andrew Gilmore and Scott Oakes for WT and D59E Bid plasmid constructs. We also thank Aron B. Fisher for helpful suggestions.
M.M. is supported by an American Heart Association-National Scientist Development grant and the National Institutes of Health (HL086699, 1S10RR022511). W.-X.Z. is supported by the Leukemia & Lymphoma Society and the Howard Temin Award (NCI). B.J.H. is supported by K99HL094536. P.P. is supported by the intramural NIH (NIAAA). C.B.T. is supported by grants from the NIH and NCI.
Published ahead of print on 30 March 2009.
†Supplemental material for this article may be found at http://mcb.asm.org/.