The apoptotic regulatory protein ARC (apoptosis repressor with caspase recruitment domain) prevents oxidant stress-mediated cell death by preserving mitochondrial function.

ARC is an apoptotic regulatory protein expressed almost exclusively in myogenic cells. It contains a caspase recruitment domain (CARD) through which it has been shown to block the activation of some initiator caspases. Because ARC also blocks caspase-independent events associated with apoptosis, such as hypoxia-induced cytochrome c release, we examined its role in cell death triggered by exposure to hydrogen peroxide (H(2)O(2)) in the myogenic cell line, H9c2. Cell death in this model was caspase-independent and characterized by dose-dependent reduction in ARC expression accompanied by disruption of the mitochondrial membrane potential (Delta psi(m)) and loss of plasma membrane integrity, typical of necrotic cell death. Ectopic expression of ARC prevented both H(2)O(2)-induced mitochondrial dysfunction and cell death without affecting the stress kinase response, suggesting that ARCs protective effects were downstream of early signaling events and not due to quenching of H(2)O(2). ARC was also effective in blocking H(2)O(2)-induced loss of membrane integrity and/or disruption of Delta psi(m) in two human cell lines in which it is not normally expressed. These results demonstrate that, in addition to its ability to block caspase-dependent and -independent events in apoptosis, ARC also prevents necrosis-like cell death via the preservation of mitochondrial function.

domains are present on the cytoplasmic domains of many death receptors, such as Fas/CD95 and TNFR1, where they recruit adaptor molecules, such as TRADD and FADD, to the deathinducing signaling complex (1). Adaptor molecules then recruit additional proteins, including initiator caspases which become activated by autoprocessing within the death-inducing signaling complex (5). CARD motifs are found on a majority of initiator caspases (the notable exception being caspase 8) (2), on daptor proteins that interact with and activate these caspases, such as CRADD/RAIDD and Apaf-1 (6,7), on the IAP family members, cIAP-1 and cIAP-2 (8), and on a growing number of regulatory proteins that modulate intracellular kinase activity and NF-B activation and influence pro-survival and pro-death intercellular signaling, such as RICK, CARDIAK, and bcl-10/mE10/CIPER (9 -13).
ARC (apoptosis repressor with CARD) is a CARD protein that is expressed almost exclusively in myogenic tissue, e.g. the heart and skeletal muscle. ARC selectively interacts with the initiator caspases 2 and 8 and significantly attenuates death receptor-induced apoptosis dependent on the activation of these caspases (14). In the H9c2 rat embryonic heart cell line, ARC expression was shown to decrease in association with activation of hypoxia-induced apoptosis. Ectopic expression of ARC prevented apoptosis, protection being mediated in all or part through the blockade of hypoxia-induced cytochrome c release from the mitochondria (15). Since small peptide broadspectrum inhibitors of caspases had no effect on this release, these results suggested that, in addition to its potential ability to directly prevent the activation of specific initiator caspases, ARC is also capable of blocking caspase-independent events critical for the execution of apoptosis.
In this study, we have examined the role of ARC in hydrogen peroxide (H 2 O 2 )-induced cell death in the rat embryonic cell line, H9c2. Cell death mediated by reactive oxygen species is a significant component of ischemia-reperfusion injury in tissues, such as the heart (16,17). While ischemia causes some cell death on its own, the reintroduction of oxygen during reperfusion is associated with accelerated apoptotic cell death and necrosis, resulting, in part, from a burst of free radical production. H 2 O 2 and its reactive by-products are potential mediators of cell death induced by diverse stimuli (18,19). Under the conditions of the experiments reported in this study, H 2 O 2 caused a dose-dependent dissipation of the mitochondrial membrane potential, plummeting intracellular ATP levels, and loss of plasma membrane integrity linked to the early changes in mitochondrial integrity. These changes were accompanied by reduced ARC protein levels and were unaffected by a broad spectrum caspase inhibitor. Moreover, H9c2 cell lines that lacked endogenous ARC expression (H9-cells) were more sen-sitive to death-inducing effects of H 2 O 2 . Maintaining ARC levels in these cells as well as the parental H9c2 cells by driving exogenous ARC expression from a constitutively active promoter prevented the dissipation of the mitochondrial membrane potential and the resulting cell death. H 2 O 2 -induced changes in Erk1/2 phosphorylation were unaffected by enforced ARC expression, suggesting that ARCs protective effects were not due to the direct or indirect quenching of H 2 O 2 but to blocking downstream effects of H 2 O 2 within the cell. We also show that ARC provides equivalent protective effects in two nonmyogenic cell lines in which H 2 O 2 also causes necrosis-like cell death. These results demonstrate that, in addition to its ability to block caspase-dependent and -independent events in apoptosis, ARC can also prevent cellular changes associated with necrotic cell death through its ability to prevent mitochondrial dysfunction.

EXPERIMENTAL PROCEDURES
Materials-A rabbit polyclonal antibody to ARC was generated as previously described (15). To remove antibodies directed against the N-terminal CARD, which ARC shares with the nucleolar protein, Nop30 (20), the antisera was affinity purified using the C-terminal proline-glutamic acid (P/E)-rich domain of ARC immobilized on nickelsaturated nitriloacetic acid (Ni-NTA)-agarose matrix (Qiagen). The Cterminal domain of rat ARC was cloned in-frame into the bacterial expression vector, pTrcHis-TOPO (Invitrogen), to generate a 6Xhistagged protein. Expression of the protein in bacteria was induced by isopropyl-1-thio-␤-D-galactopyranoside and a guanidinium hydrochloride lysate prepared and loaded onto the Ni-NTA column according to the manufacturers instructions (Qiagen). After washing the column to elute nonspecific binding and equilibrating it in 150 mM NaCl, 50 mM Tris-HCl, pH 7.4, the Ni-NTA-agarose matrix with the bound 6XHistagged C-terminal domain of rat ARC was incubated with antisera and washed extensively in equilibration buffer. Antibodies were then eluted in 4 M MgCl 2 (21) and dialyzed overnight in the cold against two changes of phosphate-buffered saline, pH 7.4.
Cell Culture-The H9c2 embryonal rat heart-derived cell line (21) was obtained from the American Type Culture Collection (ATCC) (CRL 1446) and cultured in Dulbecco's modified Earle's medium containing 4500 mg/liter D-glucose and 110 mg/ml pyruvate, supplemented with 10% heat-inactivated fetal calf serum and penicillin (100 units/ml)/ streptomycin (100 g/ml). Cultures were routinely monitored for the presence of ARC and the myogenic determination factor, myoD. A subline of H9c2 lacking detectable ARC expression (referred to as H9Ϫ) was generated by maintaining cell cultures at high density for several weeks and then isolating clonal cell lines by limiting dilution. Cell lines in which ARC was stably expressed were generated by transfecting H9Ϫ cells with either an empty expression vector (pcDNA3.1(ϩ)hygro) or one containing the human ARC cDNA (pcDNA3.1ϩhygro-hARC). Transfection was performed using Fugene (Roche Molecular Biochemicals) at a DNA:Fugene ratio of 1:6. Hygromycin-resistant clones were selected in media containing 100 g/ml hygromycin. MCF-7 and HeLa cell lines were obtained from ATCC and grown in Dulbecco's modified Earle's medium containing 10% fetal calf serum and penicillin/streptomycin.

Measurement of Cell Death and Mitochondrial Membrane
Potential-Cell viability and cytotoxicity was measured using calcein-AM (500 nM), a membrane permeant dye that in live cells is cleaved by esterases to yield cytoplasmic green fluorescence and ethidium homodimer-1 (EtD-1 (1 M)), a red nucleic acid dye to which live cells are impermeant, but that accumulates in the nuclei of membrane-compromised cells (Molecular Probes L-3224). Staining was performed on live cells in phosphate-buffered saline for 15 min. Cell death was measured by determining the percent of EtD-1 positive cells to total cell number (EtD-1 and calcein-AM positive cells) in at least 8 fields from each experiment and then averaged over a total of three to six experiments.
Mitochondrial membrane potential was determined using Mito-TrackerCMXRos (Molecular Probes M-7512). At the end of the experiment, medium was removed and Dulbecco's modified Earle's medium with 50 nM MitoTrackerCMXRos was added to live cells for 15 min at 37°C. Dissipation of mitochondrial membrane potential was measured by determining the percent of MitoTracker negative cells in at least 8 fields from each experiment and then averaged over a total of three to six experiments.
Adenoviruses and Adenoviral Infections-A recombinant replication-defective adenovirus encoding human ARC was constructed using the pAdTrack shuttle vector (22). This contains two separate cytomegalovirus promoters to drive either transgene or green fluorescent protein (GFP) expression. Bacterial recombination between the shuttle vector and the replication-defective adenoviral genome, virus amplication, and virus purification were performed by the Gene Therapy Center of the University of North Carolina at Chapel Hill. Control virus expressed only GFP. Viral infections were performed on cells in suspension with the optimal amount of virus/cell determined empirically using the GFP marker to achieve 85-90% infection. Infection was performed in serumfree media for 2 h at 37°C, after which serum-containing media was added and the cells plated onto tissue culture plastic. Experimental manipulations of the cells were performed 48 h after infection. Cytochrome c Release-Cells were seeded at a density of 1 ϫ 10 6 cells/100-mm dish and then treated 48 h later with 400 M peroxide in serum-free Dulbecco's modified Earle's medium for 15-60 min. The cells from 3 dishes were collected for each time point by trypsinization, washed twice in fractionation buffer (250 mM sucrose, 20 mM HEPES, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA), and resuspended in 0.15 ml of fractionation buffer supplemented with a protease inhibitor mixture (Sigma P-8340). The cells were homogenized on ice using 100 passes of a tight fitting Dounce homogenizer. Unbroken cells and nuclei were spun down at 2,500 rpm for 10 min. The supernatant was spun at 14,000 rpm for 25 min to separate the "cytosolic" (supernatant) and "heavy membrane" fractions (pellet). Cytosolic protein (25 g) was loaded onto 16% Tris glycine gels (Novex), electrophoretically transferred to a polyvinyldine fluoride membrane, and cytochrome c release detected by Western blotting as described below.

H 2 O 2 Causes Mitochondrial Membrane Potential (⌬ m ) Dissipation and Loss of Plasma Membrane Integrity in H9c2s-
Exposing rat embryonic cardiac H9c2 cells to relatively high doses of H 2 O 2 caused disruption of the mitochondrial membrane potential (⌬ m ) and rapid loss of plasma membrane integrity, events characteristic of necrosis. Fig. 1 shows the results of staining untreated and H 2 O 2 -treated H9c2 cells with markers of ⌬ m (MitoTrackerCMX-Ros), loss of plasma membrane integrity (ethidium dimer-1 (EtD-1)), and cell viability (calcein-AM). In untreated cultures, the vast majority of cells were alive, impermeant to EtD-1, trapping calcein-AM within the cell, and displaying an intact ⌬ m (Fig. 1, B, C, and A, respectively). After exposure to 400 M H 2 O 2 for 12 h, however, most cells were dead, showing loss of plasma membrane integrity, as indicated by uptake of EtD-1 and failure to accumulate calcein-AM (Fig. 1, E and F), and disruption of the ⌬ m (Fig.  1D, respectively). Fig. 2A shows the dose response for ⌬ m dissipation and the appearance of dead (EtD-1ϩ) cells after 12 h of treatment. Both serum starvation alone and in combination with a low dose of H 2 O 2 (50 M) caused some ⌬ m dissipation but without increased cell death. As the concentration of H 2 O 2 increased, however, both the percentage of cells undergoing ⌬ m dissipation and EtD-1 uptake increased. At 400 M H 2 O 2 , all of the cells showed a dramatic reduction in ⌬ m with 85-90% of these now permeable to EtD-1. Fig. 2B shows the time course of the percent of total cells undergoing ⌬ m disruption and cell death at 400 M H 2 O 2 . ⌬ m dissipation occurred within the first hour of exposure and was apparent in nearly all of the cells by 2 h. Over this period, cellular ATP levels dropped to Ͻ10% (data not shown). The appearance of dead (EtD-1 positive) cells lagged behind the changes in ⌬ m suggesting that cell death as measured by EtD-1 uptake may be dependent on mitochondrial dysfunction measured here as a disruption in ⌬ m . To directly test this, we first pretreated cells with cyclosporin A to determine if the disruption of the ⌬ m could be blocked by this mitochondrial permeability transition (MPT) inhibitor. Fig. 3A shows that pretreatment with 500 nM cyclosporin A markedly prevented the H 2 O 2 -induced ⌬ m dissipation seen after 2 h. Similar results were seen with bongkrekic acid (data not shown). Fig. 3B shows that CsA not only prevented ⌬ m dissipation but also the H 2 O 2 -induced increase in EtD-1 positive cells, indicating that the loss of plasma membrane integrity was dependent, in large part, on either ⌬ m dissipation or MPT.
shows that H 2 O 2 caused a rapid release of cytochrome c from the mitochondria into the cytosol. While increases in cytosolic cytochrome c can trigger activation of caspase 9 and more distal caspase associated with apoptosis, such as caspase 3, we failed to observe caspase 3 cleavage (a measure of its activation) at any of the H 2 O 2 doses tested (Fig. 4B). Caspase 3 cleavage, however, was readily detectable when cells were incubated with staurosporine and was blocked by preincubating the cells with the broad spectrum inhibitor of caspases, zVAD-fmk, indicating that it was due to processing by upstream caspases. To exclude the possibility that apoptosis may have occurred earlier than any of the times examined or over such an accelerated time frame that it was missed, we pretreated the cells with zVAD-fmk before exposure to H 2 O 2 . Fig. 4C shows that this pretreatment had no effect on either ⌬ m dissipation or cell death. Together these results demonstrate that the necrosislike cell death we observed with high doses of H 2 O 2 is not a caspase-dependent event or an event secondary to caspase-dependent processes.
Down-regulation of the Muscle-specific CARD protein, ARC, Links H 2 O 2 to Cell Death in H9c2s-Because ARC protein levels in H9c2s decrease in parallel with the development of apoptosis caused by hypoxia (12), we examined whether ARC protein levels were also altered in response to H 2 O 2 exposure and, if so, whether these changes were linked to the onset of cell death. Fig. 5A shows a Western blot for ARC expression in various rat tissues and in H9c2 cells. ARC was detected in adult and embryonic heart but not in lung or liver, consistent with previously published Northern blotting data showing that expression of ARC mRNA is restricted to myogenic tissues and the testis (14). The protein detected by our antibody migrates at a slower mobility (between 30 and 36 kDa) than predicted from the primary amino acid sequence of ARC (ϳ25 K d ), due to ARCs proline-rich C-terminal domain. In addition, the protein detected in H9c2 cells (a rat myogenic cell line) migrates faster than the protein in normal rat myogenic tissues. A comparison of the DNA sequence of the ARC cDNAs isolated from H9c2 cells and the rat heart revealed the presence of a 12-amino acid deletion in the proline-rich C-terminal domain of H9c2 ARC. Whether this deletion is of any functional significance is not known, although similar deletions in this region are seen in human and mouse ARCs from both cultured cells and intact tissues. Fig. 5B shows that increasing doses of H 2 O 2 resulted in progressive loss of ARC from the cell. This dose-dependent decrease in expression paralleled the dose-dependent loss in ⌬ m and plasma membrane integrity. As a control for these measurements, Fig. 5B also shows that hypoxia caused ARC down-regulation as described previously (15), while staurosporine at doses sufficient to cause apoptosis did not affect ARC levels, indicating that not all cell death stimuli may alter ARC expression.
To determine the functional significance of ARCs down-regulation in H 2 O 2 -induced cell death, we infected H9c2 cells with a replication-defective recombinant adenovirus encoding the cDNAs for GFP and human ARC in separate expression cassettes (AdTrack.ARC). Infected H9c2 cells expressed human ARC protein and this expression, unlike the endogenous protein, was unaffected by H 2 O 2 (Fig. 5C). Fig. 6A shows a representative experiment examining the effect of exogenous ARC expression on ⌬ m dissipation caused by 200 M H 2 O 2 . In addition to CMXRos, unfixed cells were also incubated with EtD-1. As shown above, H 2 O 2 caused dissipation of the ⌬ m in the vast majority of cells. In addition, many of these cells now accumulated EtD-1 in their nuclei (Fig. 6, white arrows), indicating loss of plasma membrane integrity. zVAD-fmk or the control AdTrack virus expressing only GFP (AdTrack) had no effect on these changes induced by H 2 O 2 . In cells infected with AdTrack.ARC, however, the ⌬ m remained intact. Fig. 6B shows the combined results of this and several other experiments. Infection with AdTrack.ARC reduced ⌬ m dissipation from Ͼ95% to less than 25% of the cells and the loss of plasma membrane integrity from ϳ70% to Ͻ10% of the cells.
The Protective Effects of Endogenous ARC-The experiments described in Fig. 6

FIG. 5. Endogenous ARC expression is linked to cell death. A,
Western blot of 10 g of total cell extracts from adult heart, liver, lung, embryonic heart, and H9c2s probed for ARC. B, upper panel, Western blot of 10 g of total cell extract from control and treated H9c2s for ARC. Peroxide at doses Ͼ100 M and hypoxia cause large reductions in ARC expression. Staurosporine (100 nM) has no effect on ARC expression. Lower panel, same extracts reprobed for equal loading using an antibody against total Erk. C, H9c2 cells infected with AdTrack.ARC and treated 36 -48 h later show no reduction in exogenous ARC expression. do this, we took advantage of a subline of H9c2s that does not express ARC. This ARC-deficient subline was isolated and subcloned in the laboratory as described under "Experimental Procedures" and is referred to as H9Ϫ. A Western blot of whole cell extracts from the parental H9c2 and H9Ϫ cell lines confirms that H9Ϫ cells do not express ARC (lanes 1 and 2 in Fig.  7A). The bottom panel shows the same extracts probed for Erk1/2 expression to demonstrate that approximately equal amounts of extract were loaded for each sample. Fig. 7B compares the response of H9c2 and H9Ϫ cells to increasing doses of H 2 O 2 in terms of EtD-1 staining. The dose-response curve is shifted to the left for H9Ϫ cells, indicating that they are more sensitive than the parental H9c2 cells to death-inducing effects of H 2 O 2 .
Since it is possible that changes in gene expression in addition to the loss of ARC expression may contribute to the increased sensitivity of H9Ϫ cells to H 2 O 2 , we introduced the human ARC cDNA driven by the cytomegalovirus promoter into H9Ϫ cells. A number of stable cell lines were isolated that expressed the transgene (lanes 3-6 in Fig. 7A). Unlike AdTrack.ARC-infected H9c2 cells, in which transgene expression was 20 -30 times that of the endogenous gene, the level of transgene expression in these stable cell lines was only 2-4 times that in the parental H9c2 cell line (lane 1). Like the recombinant adenovirally infected cells, however, transgene expression was unaltered by H 2 O 2 in these cell lines (Fig. 8A). Fig. 8B displays the results of a number of experiments examining the effect of H 2 O 2 in these cell lines. In all cases, the cell lines were now totally resistant to H 2 O 2 in terms of ⌬ m dissipation and loss of plasma membrane integrity (EtD-1 uptake), strongly suggesting that the lack of ARC expression was an important factor contributing to the increased sensitivity of H9Ϫ cells to H 2 O 2 . In addition, these results show that massive overexpression of ARC seen in the adenovirally infected cells was not necessary to provide a protective effect. In fact, the protection afforded by the H9Ϫ/hARC cell lines was greater than that seen following adenoviral delivery of ARC, presumably because of the slight cytotoxicity associated with viral infection of these cells.
One possible mechanism through which ARC could protect cells from oxidative damage is through either the direct quenching of H 2 O 2 or indirect quenching through the increased expression or activation of anti-oxidant defenses. To test this, we examined the effect of enforced ARC expression on H 2 O 2induced stress kinase activation. In Fig. 8C, the H 2 O 2 -induced activation of p38 and Erk1/2 was examined in H9Ϫ, parental H9c2, and three of the H9Ϫ/hARC cell lines. (No changes in JNK1/2 activation were observed under the conditions of our experiments.) Increased p38 activation as measured by changes in its phosphorylation state was evident in H9Ϫ cells and, to a much lesser degree, in one of the H9Ϫ/hARC cell lines. In contrast, activation (i.e. phosphorylation) of Erk1/2 in response to H 2 O 2 was seen in all the different cell lines, despite difference in ARC expression. Cells that express ARC tended to have increased basal levels of Erk phosphorylation, but in all cases, peak expression was comparable, occurring between 10 and 15 min after stimulation. These results demonstrate that in ARC-expressing cells, the stress kinase response involving Erk1/2 phosphorylation is unaltered, suggesting that ARC does not directly or indirectly quench H 2 O 2 . The activation of p38 in H9Ϫ cells may reflect additional changes in gene expression unrelated to its lack of ARC expression or to its increased sensitivity to oxidant stress.

ARC Prevents H 2 O 2 -induced Cell Death in Nonmyogenic Cell
Lines-To determine if the protective effects of ARC against necrosis-like cell death involved in myogenic-specific signaling pathways, we tested whether forced expression of ARC could attenuate the cytotoxic effects of H 2 O 2 on the MCF-7 and HeLa human cell lines. Fig. 9A shows that increasing doses of H 2 O 2 caused complete ⌬ m dissipation followed by loss of plasma membrane integrity in MCF-7 cells. These changes were completely prevented by prior infection with the recombinant replication-defective adenovirus expressing human ARC (AdTrack.ARC). Fig. 9B shows that ⌬ m dissipation also increases with increasing H 2 O 2 levels in another human cell line, HeLa cells. While infection with AdTrack.ARC was not as effective in HeLa as in MCF-7 cells, it still significantly attenuated ⌬ m disruption. EtD-1 uptake was not measured in these cells since H 2 O 2 causes the cells to detach from the dish. These results demonstrate that while ARC expression is mostly confined to myogenic cells it engages pro-survival signaling pathways present in nonmyogenic cell types. DISCUSSION We have shown in this study that the muscle-specific repressor of apoptosis, known as ARC, provides an essential link between oxidant stress and cell death in the rat myogenic cell line, H9c2. We showed that H 2 O 2 caused a dose-dependent dissipation of the ⌬ m and loss of plasma membrane integrity (Figs. 1 and 2) that was accompanied by reductions in ARC protein levels (Fig. 5). We also showed that H9c2 sublines that are ARC deficient (H9Ϫ) are more sensitive to the death-inducing effects of H 2 O 2 (Fig. 7). Enforced expression of ARC by either adenoviral-mediated gene transfer (Fig. 6) or the creation of stably integrated cell lines expressing ARC (Fig. 9) dramatically increased the resistance of parental H9c2 and H9Ϫ cells to H 2 O 2 . Although adenoviral gene transfer resulted in large overexpression of the protein, it is especially noteworthy that expression of the transfected ARC transgene in the H9Ϫ cells was only 2-4 times that of the endogenous gene in H9c2 cells (Fig. 7), indicating that protection does not depend on levels of ARC which are unlikely to be achieved under any physiological or pathophysiological conditions. We also showed that while ARC blocks the mitochondrial disrupting or cell death-inducing effects of H 2 O 2, it does not block H 2 O 2 -induced Erk1/2 activation (Fig. 8), demonstrating that ARC interferes only with selective aspects of H 2 O 2 -induced signaling. It is, therefore, unlikely that ARC blocks cell death by either directly quenching H 2 O 2 or activating antioxidant defenses to do the same. ARC is a protein composed principally of two domains: a C-terminal domain consisting of multiple proline/glutamate repeats and an N-terminal CARD. CARD proteins have been linked to caspase activation and inhibition and, in some cases, NF-B regulation (6 -13). In the case of ARC, previous studies have established that it can bind to the initiator caspases 2 and 8, but not caspase 9, and that it can inhibit caspase 8 activity, presumably by blocking its activation (14). But ARC has also been shown to block cytochrome c release and hence, the downstream caspases activated by the cytochrome c⅐(d)ATP⅐Apaf-1⅐caspase-9 complex. Notably, this inhibition was mediated by a caspase-independent effect of ARC (15). Most other CARD  proteins potentiate apoptosis, including RAIDD/CRADD (6), bcl-10 (9 -12), Apaf-1 (7), and CARD4/Nod1 (23,24), although the inhibitors of apoptosis, cIAP-1 and -2, contain a CARD (8). The CARD domains of cIAP-1 and -2, however, are not required for their direct interaction with caspases and the blocking of apoptosis (25), but may function to disrupt other CARD-dependent apoptotic events (26). ARCs ability to block not only the activation of initiator caspases 2 and 8 and to modulate caspase-independent mitochondrial events associated with caspase-dependent apoptosis and caspase-independent cell death distinguishes it from the IAPs.
The early loss in both plasma membrane integrity and cellular ATP levels we observed in this study are hallmarks of necrotic cell death. Because loss of plasma membrane integrity can also occur secondary to apoptosis, inhibitors of apoptosis, such as ARC, could prevent the secondary appearance of the necrosis-like phenotype by preventing primary apoptosis. This scenario, however, is inconsistent with the time course of plasma membrane disruption and the almost complete loss of cellular ATP, as well as with the fact that no evidence for caspase activation was obtained after H 2 O 2 treatment and that broad spectrum caspase inhibitors had no effect on loss of plasma membrane integrity or dissipation of the ⌬ m in response to H 2 O 2 (Figs. 4 and 6).
Understanding the mechanism(s) through which H 2 O 2 causes cell death is likely to provide insight into the mechanism of ARCs protective effects. Our data indicate that H 2 O 2induced cell death in H9c2 cells is likely to start with disruption of the mitochondria, possibly caused by the MPT. In this study, we measured the dissipation of the ⌬ m as an indicator of mitochondrial dysfunction. ⌬ m dissipation is often used as a indirect marker of MPT, but not all changes in ⌬ m are caused by MPT. To investigate the potential role of the MPT, we pretreated cells with cyclosporin A (CsA), an inhibitor of cyclophilin D, which is an inner membrane protein and a component of the mitochondrial permeability pore complex (27,28). (Similar results were seen in cells pretreated with bongkrekic acid, a ligand for another member of the mitochondrial permeability, the adenine nucleotide translocator.) The fact that CsA prevented both H 2 O 2 -induced ⌬ m dissipation and loss of plasma membrane integrity (Fig. 3) indicates that 1) ⌬ m dissipation is secondary to MPT, and that 2) events triggered by either MPT and/or ⌬ m dissipation cause cell death. H 2 O 2 s ability to induce MPT may be due to a direct oxidization of thiol groups on the pore complex (29,30) or on calcium transport systems on the mitochondria, the endoplasmic reticulum, or the plasma membrane (31,32).
Based on our results, the likely sequence of events occurring during H 2 O 2 -induced cell death is that H 2 O 2 activates the MPT, which in turn leads to cytochrome c release, disruption of the ⌬ m and the uncoupling of oxidative phosphorylation, the likely cause for plummeting cellular ATP levels. In the absence of ATP, cytochrome c fails to activate caspase 9 since Apaf-1 requires ATP as a cofactor (7) and there is, therefore, no apoptotic component to the process of cell death. Instead, the loss of ATP likely contributes to the rapid loss of plasma membrane integrity and cell death via a necrosis-like mechanism. The fact that enforced expression of ARC also prevents H 2 O 2 -induced ⌬ m dissipation (Figs. 6, 9, and 10) as well as cytochrome c release 2 suggests that ARC blocks cell death by either stabilizing the ⌬ m in the presence of the MPT or directly preventing the MPT. That mitochondria may be an important target for ARCs preventative actions is consistent with previous studies showing that ARC translocates from the cytosol to the mito-chondrial membrane fraction in response to an apoptotic stimulus (hypoxia) and prevents the release of cytochrome c (15), an event which is not caspase-dependent and not predictable based on the original description of ARC as an inhibitor of selective initiator caspases.
ARCs ability to block both apoptotic and necrosis-like cells is reminiscent in some respects of that of the anti-apoptotic members of the bcl-2 family. Thus, enforced bcl-2 expression completely blocks both staurosporine-induced apoptotic and necrotic cell death in Jurkat cells (33), while other investigators have shown that bcl-2 retards necrotic cell death in PC12 caused by respiratory chain poisons such cyanide, rotenone, or antimycin A and partially protects Rat-1 fibroblasts from high dose hydrogen peroxide (34,35). Given the importance of the mitochondria as gatekeepers for both necrosis and apoptosis (28,36,37), the shared ability of bcl-2 and ARC to block both apoptotic and necrotic-like cell death may be linked to their constitutive or transient localization to the mitochondrial membrane fraction.
In summary, we have presented data demonstrating that ARC links oxidative stress to cell death in H9c2 cells. H 2 O 2 causes necrotic-like cell death that is caspase-independent, triggered by disruption of mitochondrial function, and accompanied by down-regulation of ARC expression. Enforced expression of ARC prevents cell death by preserving mitochondrial function. While ARC expression is, for the most part, limited to myogenic cells and tissues, it engages pro-survival signaling pathways common to nonmyogenic cells as well. These results identify a novel function for the CARD protein, ARC, and suggest that it may be particularly effective in counteracting both apoptotic and necrotic-like forms of cell death to provide protection against cellular damage accumulated during ischemia and in response to the reperfusion of ischemic heart tissue.