Recently, it has become apparent that mitochondrial DNA (mtDNA) damage can rapidly initiate apoptosis independent of mutations, although the mechanism involved remains unclear. To elucidate this mechanism, angiotensin II-mediated apoptosis was studied in cells that were transduced with a lentiviral vector to overexpress the DNA repair enzyme 8-oxoguanine glycosylase or were treated with inhibitors known to block angiotensin II-induced mtDNA damage. Cells exhibiting angiotensin II-induced mtDNA damage showed two phases of superoxide generation, the first derived from NAD(P)H oxidase and the second of mitochondrial origin, whereas cells prevented from experiencing mtDNA damage importantly exhibited only the first phase. Furthermore, cells with mtDNA damage demonstrated impairments in mitochondrial protein expression, cellular respiration, and complex 1 activity before the onset of the second phase of oxidation. After the second phase, the mitochondrial membrane potential collapsed, cytochrome c was released, and the cells underwent apoptosis, all of which were prevented by disrupting mtDNA damage. Collectively, these data reveal a novel mechanism of apoptosis that is initiated when mtDNA damage triggers mitochondrial superoxide generation and ultimately the activation of the mitochondrial permeability transition. This novel mechanism may play an important pathological role.
- angiotensin II
- mitochondrial permeability transition pore
- NADPH oxidase
oxidative stress is associated with a variety of pathological conditions, including cancer (12), diabetes (24), certain neurodegenerative diseases (5), ischemic heart disease (1), heart failure (13), and aging (31). It is now well established that a major target of oxidative damage is mitochondrial DNA (mtDNA), which undergoes base oxidation, deoxyribose damage, N-glycosidic bond cleavage (abasic site formation), and strand scission (4, 20, 32). It is therefore not surprising that oxidative mtDNA damage has been implicated as a causative factor in each of these conditions.
Interestingly, the majority of research investigating the role of mtDNA damage in disease has focused on the gradual accumulation of mutagenic oxidative base modifications in these pathologies. Indeed, several specific mtDNA point mutations and deletions have been linked to various types of cancers (19), diabetes (7), neurodegenerative disorders (33), and neuromuscular disorders (21). Furthermore, transgenic mice that accumulate excessive mtDNA damage have been shown to develop the metabolic syndrome (35), dilated cardiomyopathy and heart failure (17, 37), and die prematurely (15, 17, 34).
However, it is becoming increasingly clear that not all oxidative damage to mtDNA results in nucleotide substitution or mutation. Rather, single strand breaks as well as abasic sites are common forms of oxidative mtDNA damage. Recently, it has been estimated that, under normal physiological conditions, ∼9,000 abasic sites form each day (38), any one of which, according to our hypothesis, can be cytotoxic and result in acute cell death. However, despite an association between the many diseases linked to oxidative stress and acute cell death, the role of single strand breaks and abasic sites in this process has received little attention.
We have previously suggested that oxidative mtDNA damage plays a key role in the mechanism of apoptosis induced by angiotensin II (10). Moreover, Rachek et al. (23) have demonstrated that mtDNA repair enhances survival in cells subjected to oxidative stress. While it is logical to conclude that the gradual accumulation of mutagenic lesions can cause tissue damage, we propose that a small number of single strand breaks and abasic sites within the mitochondrial genome can rapidly induce apoptosis long before mutations can accumulate. However, the mechanism by which these acute lesions can cause apoptosis and tissue dysfunction remains unclear.
Because postmitotic tissues, as well as tissues with high-energy requirements, contain more mitochondria and are thought to be especially prone to the damaging effects of oxidative species and mtDNA damage (2), the mechanism underlying the effects of these cytotoxic lesions was studied in cultured neonatal cardiomyocytes subjected to angiotensin II. Notably, angiotensin II is known to increase reactive oxygen species and promote apoptosis in cardiac myocytes (8, 10, 29). Moreover, angiotensin II levels increase during heart failure, one of the many mtDNA damage-linked diseases in which acute cell death is thought to play a role. The present findings uncover a novel mechanism that links oxidative mtDNA damage to mitochondrial dysfunction and apoptosis independent of mutations.
MATERIALS AND METHODS
Cardiomyocyte preparation and incubation conditions.
Neonatal cardiomyocytes were prepared from 3-day-old Wistar rats as described previously (10). The use of animals for this study was approved by the Institutional Animal Care and Use Committee (Protocol 04023). After dissociation, the cells were preplated on plastic culture dishes for 90 min at 37°C to allow for nonmyocyte attachment. The unattached cells were then suspended in minimum essential medium (MEM) containing 10% newborn calf serum and 0.1 mM 5′-bromo-2′-deoxyuridine. The cells were incubated overnight to allow for attachment of viable cardiomyocytes, cultured in MEM for 3 days, and then harvested. Before analysis, cells were subjected to 0 nM (control) or 1 nM angiotensin II for various periods of time. Fifteen minutes before and throughout, angiotensin II treatment, or 40 min after angiotensin II treatment, some control and angiotensin II-treated cells were exposed to either 10 μM diphenyleneiodonium (DPI), an inhibitor of NAD(P)H oxidase, 0.5 mM apocynin, an inhibitor of NAD(P)H oxidase, or 100 μM NG-monomethyl-l-arginine (l-NMMA), an inhibitor of nitric oxide synthase (NOS) or 200 nM cyclosporin A, an inhibitor of the mitochondrial permeability transition pore (MPT). The effect of DPI and apocynin were identical; therefore, only the apocynin data are shown.
Lentiviral vector generation and overexpression of human 8-oxoguanine DNA glycosylase.
cDNA encoding either nuclear or mitochondrially targeted human 8-oxoguanine DNA glycosylase was inserted into a lentiviral vector for delivery to cells, with expression driven from a human cytomegalovirus promoter (as described previously for neonatal cardiomyocytes, 25) using standard recombinant DNA methods. The vector was packaged by transient transfection of the vector plasmid and three helper plasmids (encoding the viral Gag-Pol, Rev and the envelope protein VSVG) into 293 human embryonic kidney cells (HEK293) using the calcium phosphate method. Virus secreted into the cell culture media was collected and concentrated by ultracentrifugation to a typical titer of 109 transducing units per milliliter, as assessed by quantitative PCR of HEK293 cell genomic DNA following infection. Forty eight hours before analysis, virus was incubated with the cardiomyocyte cultures at a multiplicity of infection of 5:1. Control cells were incubated with either empty lentiviral vector or lentiviral vector containing cDNA for green fluorescent protein.
Annexin V and propidium iodide procedures.
To assess the number of apoptotic cells in control and human 8-oxoguanine overexpressing cells, the TACS Annexin V-FITC kit was used. After 4 h of exposure to control medium or medium supplemented with angiotensin II in the presence or absence of cyclosporin A (200 nM), the culture medium was removed from the cells. The cells were first rinsed with cold 1× PBS buffer and then placed in 100 μl of annexin V incubation reagent consisting of 10 μl of 10× binding buffer, 10 μl of propidium iodide, 1 μl of annexin V conjugate, and 79 μl of distilled water. After a 15-min incubation in the dark with the annexin V reagent, the cells were washed for 2 min with an excess of 1× binding buffer [100 mM HEPES (pH 7.4), 1.5 M NaCl, 50 mM KCl, 10 mM MgCl2, and 18 mM CaCl2]. Annexin V and propidium iodide staining were observed by fluorescence microscopy using an Olympus IX 70 inverted microscope. Five separate fields in microscope were examined for bright green (apoptotic) and red (necrotic) nuclei.
Western blot analysis.
Cellular content of ND5, Ogg1, succinate dehydrogenase, inducible NOS (iNOS), β-actin, cytochrome c oxidase (cox3), and cytochrome c were determined by Western blot analysis. The cells were washed twice in buffer containing 250 mmol/l sucrose, 10 mmol/l triethanolamine, pH 7.6, at room temperature. Cells were then lysed in ice-cold isolation buffer (pH 7.25 at 4°C) with the following composition: 150 mmol/l mannitol, 2 mmol/l sucrose, 5 mmol/l HEPES, 1 mmol/l EDTA, and protein protease inhibitors (a 1:100 dilution of protease inhibitor cocktail set III, 1% phenylmethysulfonyl fluoride, and 100 mmol/l sodium orthovanadate). The homogenate was centrifuged (Micromax RF, Thermo) at 800 g for 6 min at 4°C, and the supernatant was retrieved and centrifuged again at 12,000 g for 15 min. The pellet was defined as the mitochondrial fraction and used to determine ND5, Ogg1, cox3, and succinate dehydrogenase content. The supernatant was defined as the cytosolic fraction and used to determine iNOS, β-actin, and cytochrome c content. Protein concentration was determined by the Lowry method using bovine serum albumin as a standard. During Western blot analysis, the samples were subjected to 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The proteins were then transferred to nitrocellulose membranes, where they were blocked. After incubation with the appropriate antibody, the membranes were washed and then incubated with a secondary antibody. The Western blots were detected by the enhanced chemiluminescence reaction.
Quantitative real time PCR method of measuring levels of ND5 mRNA.
Total RNA was isolated from cells using TRIzol LS Reagent (Invitrogen) according to the protocol. Quantitative real time PCR was then performed using the iScript One-Step RT-PCR Kit with SYBR Green (Bio-Rad, Hercules, CA) according to the protocol. Primer sequences for ND5 were 5′-acgcctagcattaggaagca-3 (forward) and 5′-ttctagggcaatggcaaatc-3′ (reverse), and primer sequences used for β-actin were 5′-gtgacgttgacatccgtaaa-3 (forward) and 5′-ctcaggaggagcaatgatct-3′ (reverse).
Detection of mitochondrial DNA damage using quantitative Southern blots.
Isolated neonatal cardiomyocytes were exposed to 1 nM angiotensin for 20 min, 40 min, or 1 h. Forty eight hours before exposure to 1 nM angiotensin II, some cardiomyocytes were transfected with either human 8-oxoguanine glycosylase or green fluorescent protein. After the cells were lysed, high-molecular-weight DNA was phenol extracted, precipitated with ammonium acetate, and then exposed to 2 volumes of cold ethanol. The DNA samples were resuspended in water and then digested with the restriction endonuclease BamHI (10 U/μg DNA) at the same time it was treated with DNAse-free RNase (∼1.0 /ml) for 12–16 h at 37°C. After digestion, the samples were precipitated with ammonium acetate and resuspended in Tris-EDTA buffer (10 mM Tris and 1 mM EDTA; pH 8.0). The DNA was precisely quantified using a Hoefer TKO 100 Mini-Fluorometer and a TKO standards kit (Hoefer Scientific Instruments; San Francisco, CA). Samples containing 5 μg DNA were heated at 65°C for 20 min and then cooled at room temperature for an additional 20 min. After NaOH was added to a final concentration of 0.1 N, samples were incubated for 15 min at 37°C. This produced single-strand breaks at any abasic site in the DNA. Samples were then combined with 5 μl of loading dye, loaded onto a 0.6% alkaline agarose gel, and electrophoresed at 30 V (1.5 V/cm gel length) for ∼16 h in an alkaline buffer consisting of 23 mM NaOH and 1 mM EDTA. The gels were stained with ethidium bromide to confirm equal loading. After the gels were washed, DNA was transferred to a Zeta-Probe GT nylon membrane (Bio-Rad). The membranes were cross-linked and hybridized with a 32P-labeled rat mitochondria DNA-specific PCR-generated probe. Hybridization and subsequent washes were performed according to the manufacturer's recommendations. DNA damage was assessed as the following: break frequency = −ln (band intensity after 1 nM angiotensin II)/(band intensity of control).
Measurement of oxygen consumption.
Neonatal cardiomyocytes were incubated with angiotensin II (1 nM) for 0 (control), 60, 90, 120, 150, 180, or 240 min. Approximately 1 × 107 cells were then resuspended in control medium (1 ml) lacking angiotensin II and analyzed in a gas-tight vessel maintained at 37°C, equipped with a Clark-type oxygen electrode connected to a chart recorder, and calibrated. Uncoupled respiration was determined by adding carbonyl cyanide m-chlorophenylhydrazone (CCCP). To further evaluate flux through the electron transport chain, cells were permeabilized with digitonin (60 μg/mg protein) and resuspended in respiration buffer (1 ml) before analysis. Cells were then exposed to glutamate (10 mM) + malate (10 mM), and the response to ADP (0.5 mM) was monitored.
Detection of superoxide using dihydroethidium.
Neonatal cardiomyocytes were incubated with dihydroethidium (DHE, 5 μM), which is oxidized to fluorescent ethidium in the presence of superoxide. After 1 h, the dye-containing medium was removed and replaced with fresh medium. Cells were then incubated for an additional 30 min before analysis. Changes in DHE fluorescence were monitored by confocal microscopy at 15-min intervals with at least 10 cells monitored per sample. There were no changes in fluorescence in cells not incubated with angiotensin II.
Detection of mitochondrial membrane potential using DiOC6(3).
Neonatal cardiomyocytes were incubated with 3,3′-dihexyloxacarbocyanine iodide [DiOC6(3), 5 nM], which is taken up into the mitochondria of healthy cells where it fluoresces. After 1 h, the dye-containing medium was removed and replaced with fresh medium. Cells were then incubated for an additional 30 min before analysis. Changes in DiOC6(3) fluorescence were monitored by confocal microscopy at 15-min intervals, with at least 10 cells monitored per sample. To ensure mitochondrial-specific staining, DiOC6(3) was colocalized with MitoTracker Red.
Detection of nitric oxide using DAF-FM diacetate.
Neonatal cardiomyocytes were incubated with 4-amino-5-methylamino-2′,7′-difluorofluorescein diacetate (DAF-FM, 5 μM), a nitric oxide-sensitive dye. After 30 min, the dye-containing medium was removed and replaced with fresh medium. Cells were then incubated for an additional 30 min before analysis. Changes in fluorescence were monitored utilizing a Nikon Eclipse TE2000-U inverted fluorescence microscope at 15-min intervals with at least 10 cells monitored per sample. There were no changes in fluorescence in cells not incubated with angiotensin II.
It has been proposed that reactive species contribute to the toxicity of angiotensin II; however, neither the reactive species involved in the toxicity nor the mechanism of toxicity has been ascertained. To evaluate the role of reactive oxygen species, angiotensin II-induced superoxide generation (ethidium fluorescence) by the rat neonatal cardiomyocyte was monitored. Figure 1 shows that addition of angiotensin II (1 nM) to the medium initiated a rapid and robust increase in superoxide levels that reached a steady-state maximum (6.11 ± 0.78-fold) within 2–3 h (Fig. 1A).
Seshiah et al. (29) have previously shown that angiotensin II activates NAD(P)H oxidase in vascular smooth muscle cells, a major source of superoxide; however, the importance of NAD(P)H oxidase in the angiotensin II-treated cardiomyocyte has not been established. Therefore, to evaluate the role of NAD(P)H oxidase in the rapid generation of superoxide by the angiotensin II-treated cardiomyocyte, the effect of apocynin (0.5 mM), a nonflavoprotein inhibitor of NAD(P)H oxidase, was examined. Pretreatment with apocynin completely blocked angiotensin II-induced superoxide generation, whereas inhibition of NAD(P)H oxidase 20 min after angiotensin II addition abruptly halted the increase in ethidium fluorescence, eventually restoring basal levels (Fig. 1B). These data support the view that the rapid increase in superoxide generation was caused by the activation of NAD(P)H oxidase. Interestingly, however, when treatment with either NAD(P)H oxidase inhibitor was delayed 40 min after exposure to angiotensin II, the cells exhibited two peaks of ethidium fluorescence. The initial peak preceded the addition of the NAD(P)H oxidase inhibitor, whereas the second distinct phase of oxidant production was observed ∼80–100 min after the addition of the inhibitor (Fig. 1C). Initiation of the second oxidative phase coincided with the loss of mitochondrial membrane potential, suggesting that the second phase, but not the first phase, was responsible for the decrease in mitochondrial membrane potential (Fig. 1C).
NAD(P)H oxidase and the mitochondrial electron transport chain represent the major sources of superoxide in the myocyte (18). To determine whether the mitochondrial electron transport chain was the source of the second oxidative phase, cells were loaded with the mitochondrial-specific, superoxide-dependent dye MitoSox. In support of a mitochondrial origin, angiotensin II-induced increases in MitoSox fluorescence coincided with phase 2 oxidation (Fig. 1, A and C). Conversely, inhibition of the mitochondrial electron transport chain blocked phase 2 oxidation (data not shown). Together, these findings suggest that treatment with angiotensin II yields two overlapping phases of superoxide generation that can be resolved into distinct peaks via an NAD(P)H oxidase inhibitor.
Grishko et al. (10) had previously introduced the concept that both oxidative and nitrosative stress contribute to angiotensin II toxicity. To begin addressing whether nitric oxide and nitrosative stress could contribute to phase 2 oxidation, the time course of iNOS activation was examined. As seen in Fig. 2, iNOS is upregulated within 30 min of angiotensin II treatment. Although the levels of iNOS continued to increase between 30 and 60 min, Figs. 1 and 2 reveal that the generation of both superoxide [from NAD(P)H oxidase] and nitric oxide (from iNOS) preceded the onset of phase 2 oxidation.
To determine whether nitric oxide levels increased following exposure to angiotensin II, cells were loaded with the nitric oxide-sensitive dye DAF-FM (5 μM), and fluorescence was monitored over time. After 1 h of angiotensin II (1 nM) treatment, nitric oxide levels were increased by 41.13 ± 6.45% (Fig. 2, B and C). To assess the involvement of nitric oxide in the death cascade, l-NMMA, (100 μM) was used to block angiotensin II-induced NOS activation. As seen in Fig. 3A, NOS inhibition had little effect on angiotensin II-mediated activation of NAD(P)H oxidase but completely abolished phase 2 superoxide generation and the collapse of the mitochondrial membrane potential (Fig. 3B). This effect was not seen if addition of the NOS inhibitor was delayed 40 min following angiotensin II exposure (Fig. 3, C and D). Although the NOS inhibitor had no effect on superoxide formation by NAD(P)H oxidase, it prevented the formation of nitric oxide and another toxic substance, peroxynitrite, which is produced when nitric oxide reacts with superoxide. These findings demonstrated that both NAD(P)H oxidase-dependent superoxide generation and nitric oxide generation are required to initiate phase 2 oxidation.
We have previously shown that peroxynitrite is the putative oxidizing agent responsible for angiotensin II-mediated mtDNA damage (10). Based on Southern blot analysis, which measures strand breaks and abasic sites, angiotensin II-induced mtDNA damage was evident within 40 min of angiotensin II exposure (Fig. 4) and involved an estimated break frequency of 0.21 ± 0.03 breaks/10,000 bases. After 1 h of angiotensin II treatment, the estimated break frequency increased to 0.57 ± 0.04 breaks/10,000 bases, with a similar degree of damage seen after 2 and 4 h. However, little, if any, mtDNA damage was detected before the activation of both NAD(P)H oxidase and iNOS (within 20 min of angiotensin II addition), an observation consistent with our previous finding that either an NAD(P)H oxidase inhibitor or a NOS inhibitor is capable of preventing angiotensin II-mediated mtDNA damage (10). These data therefore suggest that peroxynitrite-mediated mtDNA damage follows phase 1 but precedes phase 2 oxidation.
Mitochondrial DNA encodes for 13 proteins that make up specific subunits of the electron transport chain. Since most of these subunits, including ND5, are regulated by a single promoter, any break in the DNA backbone could have a dramatic impact on the expression of the affected proteins. This conclusion is supported by the observation that despite a DNA break frequency of only 0.6, ND5 content began to decline within 120 min of angiotensin II addition and plunged 50% in the next 2 h (Fig. 5A). Similar declines in protein content are noted for two other mitochondrially encoded proteins ND1 and ND6 (data not shown). The importance of mtDNA damage in the observed decline in ND5 content is borne out by quantitative PCR measurements of ND5 mRNA levels, which plunged ∼30% following exposure to angiotensin II (Fig. 5B). The angiotensin II-mediated decline in ND5 protein levels was blocked by the pretreatment of the cells with the NAD(P)H oxidase inhibitor apocynin but not when the addition of the inhibitor was delayed by 40 min (Fig. 5A).
Under normal aerobic conditions, flux through the electron transport chain is efficient, and few electrons are diverted to other acceptors. However, when flux through the electron transport chain slows, more electrons are diverted to oxygen to form superoxide (14, 17, 39). To confirm that electron transport flux was reduced in angiotensin II-treated cells, oxygen consumption of the isolated cardiomyocyte was measured. The basal rate of oxygen consumption was suppressed by 19.25 ± 2.63% and 33.5 ± 1.6% after 2 and 4 h of angiotensin II exposure, respectively (Fig. 6A), with a slightly greater decline noted in state 3 respiration (Fig. 6B). The 44% decrease in complex 1 (NADH CoQ reductase) activity in sonicated mitochondria isolated from angiotensin II-treated cells supported a key role for complex 1 in the decline in respiration (Table 1). This view was further supported by the finding that the uncoupler of oxidative phosphorylation CCCP increased oxygen consumption more in control cells than in angiotensin II-treated cells (Fig. 6A). Similarly, angiotensin II caused a comparable suppression of respiration in the presence of the oxidative phosphorylation inhibitor oligomycin (data not shown), supporting the notion that angiotensin II primarily suppresses flux through complex 1 rather than inhibiting all steps of ATP synthesis.
The mitochondrial matrix contains very active antioxidant enzymes, such as Mn2+ superoxide dismutase and catalase, which can prevent oxidative stress in the face of increased superoxide production. To confirm that mitochondrial superoxide generation in the angiotensin II-treated cell overwhelms the antioxidant defenses of the mitochondria, the activity of two citric acid cycle enzymes, aconitase and succinate dehydrogenase, were examined. Both aconitase and succinate dehydrogenase are encoded by the nucleus, but only aconitase is inactivated by superoxide and peroxynitrite (26). As predicted, angiotensin II had no appreciable effect on succinate dehydrogenase activity but reduced aconitase activity by 66.25 ± 7.79% after 4 h.
To evaluate the effect of mtDNA damage on the cardiomyocytes, cells were infected with a high titer lentiviral vector containing cDNA encoding for either a mitochondrially or nuclear targeted DNA repair enzyme human 8-oxoguanine DNA glycosylase. Forty eight hours following transduction, the levels of 8-oxoguanine glycosylase in the mitochondria were significantly elevated, indicating successful overexpression of the mitochondrial-targeted enzyme (Fig. 7). As observed in Fig. 4, these cells exhibited a significant reduction in the frequency of angiotensin II-induced mtDNA strand breaks (from 0.5 to ∼0.2 lesions/10 kbase). Importantly, this improvement in mtDNA was associated with an elevation in ND5 levels (Fig. 8), elimination of phase 2 oxidation despite the presence of a prominent phase 1 oxidative peak (Fig. 9A), and prevention of the decline in the mitochondrial membrane potential (Fig. 9B).
A decline in the mitochondrial membrane potential of cells undergoing apoptosis is generally viewed as an indication of the MPT (11). Figure 9C shows that angiotensin II promoted substantial release of cytochrome c from the mitochondria, another characteristic of the MPT. In stark contrast, overexpression of mitochondrial 8-oxoguanine glycosylase led to a substantial decrease in cytochrome c release, collapse of the mitochondrial membrane potential, and apoptosis after 4 h of angiotensin II treatment (Fig. 9, C and D). Indeed, overexpression of mitochondrial 8-oxoguanine glycosylase was nearly as effective in reducing angiotensin II-mediated apoptosis as treatment with cyclosporin A, which is known to block cell death arising from the MPT (11). In contrast, overexpression of 8-oxoguanine glycosylase in the nucleus, which specifically repair nuclear DNA, did not decrease the extent of angiotensin II-mediated apoptosis.
The present study presents a novel mechanism that links oxidative mtDNA damage to apoptosis. As summarized in Fig. 10, a key element of this pathway is that it does not appear to require the longer term accumulation of mtDNA mutations. Rather, the data indicate that AT1 receptor activation induces apoptosis by promoting the formation of reactive species in the cytosol, which induce single strand breaks or abasic sites within the mitochondrial genome that block mitochondrial transcription before mutations can develop.
There is precedence for the notion that severe DNA damage can lead to cell death without forming mutations. Lankin and Jackson (16) have shown that if nuclear DNA damage is severe, p53 will facilitate cell death. Grishko et al. (10) previously demonstrated that angiotensin II-mediated activation of p53 is associated with the upregulation of the pro-apoptotic factor Bax and the downregulation of the anti-apoptotic factor Bcl-2. A similar change in the Bcl-2 family members in H9c2 cardiomyocytes was reported by Qin et al. (22), although they traced the elevation in the Bax/Bcl-2 ratio to the activation of p38 MAP kinase. In each case, however, the increase in the Bax/Bcl-2 ratio was modest, suggesting that neither pathway is likely to be the primary cause of angiotensin II-mediated apoptosis. This conclusion is supported by the observation that overexpression of the DNA repair enzyme 8-oxoguanine glycosylase in the nucleus does not render the cell resistant to angiotensin II-induced apoptosis (Fig. 9D). The dramatic cytoprotective activity of mitochondrially targeted 8-oxoguanine glycosylase, but not its nuclearly targeted counterpart, strongly suggests that the mechanism underlying mtDNA damage-mediated cell death is independent of the nuclear p53 mechanism of cell death. Rather, it depends on a novel mitochondrial pathway of cell death.
The observation that overexpression of 8-oxoguanine glycosylase in the mitochondria, which reduces the number of abasic sites and strand breaks, also provides important insight into the cytoprotective mechanism of the repair enzyme. Through a process of base excision repair, 8-oxoguanine glycosylase recognizes and removes 8-oxoguanine, a major oxidative base lesion in mtDNA (27). More importantly, the repair enzyme also exhibits AP lyase activity, which cleaves the DNA phosphate backbone 3′ of the lesion to help repair abasic sites and some DNA strand breaks. It has been documented that peroxynitrite can directly initiate DNA strand breaks and abasic sites (4, 20, 32), both which interfere with DNA transcription. Moreover, the reaction of the oxidant directly with DNA bases and deoxyribose causes further DNA damage (4, 20, 32). Absent this repair, the immediate effect of this damage would be a decline in the expression of mitochondrially encoded proteins, resulting in the slowing of electron transport chain flux and promotion of mitochondrial superoxide generation.
Mitochondrial polymerase DNA polymerase gamma will frequently miscode adenine opposite 8-oxoguanine during replication (27). Although this reaction would lead to a mutation, several lines of evidence suggest that it is not a cause of pathology in the angiotensin II-treated cell. First, a variety of mutagenic base modifications are alkali labile, therefore, NaOH treatment should provide a substantially distinct pattern of angiotensin II-induced mtDNA damage compared with cells not treated with NaOH. Yet, as demonstrated by Grishko et al. (10), the LM-PCR-based pattern of angiotensin II-induced mtDNA damage was similar whether or not the mtDNA was first treated with NaOH. Second, the formation of mutations is a slow process compared with the relatively rapid response to DNA strand breaks and abasic sites. Indeed, the effects of a mutation would only be noted after sufficient mtDNA damage has occurred, the damaged DNA is replicated [a process that requires a minimum of 60 min to complete, (9)], the mutated DNA is transcribed, and the modified protein has time to modulate mitochondrial function. By contrast, Fig. 5 shows that ND5 mRNA levels decrease within ∼30 min of the initial sign of mtDNA damage, a time course inconsistent with gradual accumulation of mutations. Third, no new protein bands were detected by the ND5 antibody following angiotensin II exposure (data not shown).
Unlike mutagenic lesions, abasic sites and single strand breaks create a block to replication and transcription (3, 28). Since transcription of most of the mitochondrially encoded polypeptides is driven by a single promoter (30), a single abasic site or single strand break could prevent the expression of all polypeptides located downstream from the damage, resulting in an immediate and dramatic decrease in the expression of mitochondrial proteins that make up the electron transport chain. One of these proteins is ND5, which combines with nuclear encoded and mitochondrially encoded polypeptides to form complex 1 of the electron transport chain (6). As seen in Figs. 4 and 5, the progressive reduction in ND5 mRNA and protein levels rapidly followed the onset of mtDNA damage.
We propose that oxidation of mtDNA is a key irreversible step in the signaling pathway. In support of this notion, Fig. 9 reveals that either overexpression of 8-oxoguanine glycosylase in the mitochondria or treatment with a NOS inhibitor, conditions that minimize mtDNA damage, abolishes phase 2 oxidation. Phase 2 oxidation is also abolished by pretreatment with an NAD(P)H oxidase inhibitor. In all of these conditions in which mtDNA damage was limited, the downstream expression of the mitochondrial proteins was robust and electron transport was sufficient to prevent the diversion of electrons to oxygen. However, prevention of further mtDNA damage after the initial insult by delayed addition of either a NOS inhibitor or an NAD(P)H oxidase inhibitor did not rescue the decline in ND5 levels or the appearance of phase 2 oxidation. Therefore, only the initial mtDNA damage is necessary to trigger the entire apoptotic pathway.
A key element of this novel death cascade is the existence of two distinct sources of superoxide that uniquely link the cytosol and mitochondria. Cytosolic NAD(P)H oxidase participates in the death cascade by providing one source of superoxide (phase 1 oxidation). As seen in Fig. 1A, superoxide levels increase rapidly following exposure to angiotensin II. Inhibition of phase 1 superoxide generation prevents mitochondrial DNA damage and angiotensin II-mediated apoptosis (10). Before cell death, however, superoxide combines with nitric oxide to form peroxynitrite. Although peroxynitrite is known to be the oxidizing agent responsible for angiotensin II-induced mtDNA damage (10), it is not certain whether peroxynitrite is formed in the cytosol, as proposed in Fig. 10. Since mitochondria contain their own NOS (mtNOS) (36), it is possible that superoxide generated in the cytosol could combine with locally produced nitric oxide in the mitochondria to induce mtDNA damage. Under these circumstances, exposure to l-NMMA would still be predicted to provide protection consistent with that observed in Fig. 3 by inhibiting mtNOS, as opposed to iNOS in the cytosol. However, the notion that cytosolic superoxide can enter the mitochondria, against a strong concentration gradient, is unlikely. Conversely, cytosolic peroxynitrite is known to readily enter the mitochondria. Thus it seems more logical to propose that peroxynitrite is forming in the cytosol. Indeed, as seen in Fig. 2, there is significant nitric oxide in the cytosol, which increases following exposure to angiotensin II. Therefore, there appears to be a sufficient stimulation of both superoxide and nitric oxide generation to facilitate the formation of peroxynitrite, the oxidant involved in mtDNA damage.
Phase 2 oxidation arises from the impairment of flux through the electron transport chain, as evidenced by decreased ND 5 levels, complex 1 activity, and oxygen consumption. Abolishing phase 2 oxidation irrespective of the mode of inhibition [through 8-oxoguanine glycosylase overexpression, NAD(P)H oxidase inhibition, and iNOS inhibition] invariably leads to cell survival. Indeed, disruption at an early stage in the pathway is equally effective in preventing death as the disruption of phase 2 oxidation using the antioxidant Tiron, suggesting that death occurs as the antioxidant defenses are overwhelmed. Evidence that angiotensin II increases MitoSox fluorescence and promotes the loss of aconitase activity suggests that the mitochondrial antioxidant defenses are overwhelmed by oxidants generated within the mitochondria. According to Halestrap et al. (11), one of the major consequences of mitochondrial oxidative stress is the oxidation of key residues on adenine nucleotide translocase, which facilitates the interaction of this translocase with cyclophilin D and promotes MPT pore formation. The observation that cyclosporin A blocks angiotensin II-mediated apoptosis (Fig. 9D) supports our view that phase 2 oxidation triggers the MPT. This conclusion is also borne out by the observation that angiotensin II stimulates the release of cytochrome c from the mitochondria and promotes the collapse of the mitochondrial membrane potential (Fig. 9, B and C), both characteristic features of the MPT.
The novel mechanism described in the present study has direct relevance to heart failure and diabetes, diseases in which angiotensin II plays a central role (8). However, the pathway appears to have wider implications. Druzhyna et al. (23) found that the oxidant menadione is also capable of initiating an acute apoptotic cascade, although the steps involved in that death pathway were not determined. Since oxidative stress is a key feature of cancer, certain neurodegenerative diseases and aging, the novel death cascade initiated by mtDNA damage may be important in a variety of pathological conditions.
This study was supported in part by American Heart Association Grant 0755288B.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society