Cytokines Induce Nitric Oxide-mediated mtDNA Damage and Apoptosis in Oligodendrocytes

Nitric oxide (NO) that is produced by inducible NO synthase (iNOS) in glial cells is thought to contribute significantly to the pathogenesis of multiple sclerosis. Oligodendrocytes can be stimulated to express iNOS by inflammatory cytokines, which are known to accumulate in the multiple sclerotic brain. The potentially pathological levels of NO produced under these circumstances can target a wide spectrum of intracellular components. We hypothesized that one of the critical targets for damage that leads to disease is mtDNA. In this study, we found that cytokines, in particular a combination of tumor necrosis factor-α (50 ng/ml) and IFNγ (25 ng/ml), cause elevated NO production in primary cultures of rat oligodendrocytes. Western blot analysis revealed a strong enhancement of iNOS expression 48 h after cytokine treatment. Within the same time period, NO-mediated mtDNA damage was shown by Southern blot analysis and by ligation-mediated PCR. Targeting the DNA repair enzyme human 8-oxoguanine DNA glycosylase (hOGG1) to the mitochondria of oligodendrocytes had a protective effect against this cytokine-mediated mtDNA damage. Moreover, it was shown that mitochondrial transport sequence hOGG1-transfected oligodendrocytes had fewer apoptotic cells compared with cells containing vector only following treatment with the cytokines. Subsequent experiments revealed that targeting hOGG1 to mitochondria reduces the activation of caspase-9, showing that this recombinant protein works to reduce apoptosis that is occurring through a mitochondria-based pathway.

Oligodendrocytes, one of the primary types of glial cells, produce the unique lipid-rich myelin membrane that forms multilamellar spirally wrapped sheaths around neuronal axons in the central nervous system (CNS). 1 Alterations in oligodendrocyte function lead to disruption of myelinogenesis and can cause severe neurological deficits, such as those found in multiple sclerosis (MS) and in the extensive white matter degeneration observed after an ischemic insult to the CNS (1). In demyelinating diseases, such as MS, oligodendrocytes are under immune attack, both cell-mediated and humoral (1,2). Recent evidence suggests that oligodendrocytes, besides being targets, may also be sources of cytokines and nitric oxide (NO), under inflammatory conditions (3,4). Cytokines are important mediators in the inflammatory demyelination observed in human MS, as well as in animal models of MS, such as experimental allergic encephalomyelitis or Theiler virus infection. In these pathologies, proinflammatory cytokines released by endogenous cells or infiltrated macrophages and CD4 ϩ Th1 cells accumulate and exert pleiotropic effects on oligodendrocytes (5,6). It has been shown that two proinflammatory cytokines, tumor necrosis factor-␣ (TNF␣) and interferon-␥ (IFN␥), accumulate in the brain and cerebrospinal fluid of MS patients (7,8). Separate studies using transgenic mice have revealed the development of a spontaneous inflammatory disease with experimental allergic encephalomyelitis-like symptoms and demyelination as a direct consequence of CNS-specific expression of TNF␣ (9) and IFN␥ (10).
Among the cytotoxic effector molecules evoked by proinflammatory stimuli mentioned above, increasing evidence supports a role for NO in the tissue damage observed in demyelinating diseases. In agreement with this notion is the finding that there is significant inducible nitric-oxide synthase (iNOS) expression in these pathologies (11,12). Additionally, several reports have shown that oligodendroglial cells could be induced to express iNOS by proinflammatory cytokines and bacterial lipopolysaccharide and release micromolar concentrations of NO (3,14,15).
NO is a potentially toxic molecule that has been implicated in a wide range of diverse (patho)physiological processes in the CNS. It is known that the production of NO from L-arginine is important for intercellular signaling, nonspecific host defense, and in helping to kill tumors and invading pathogens. Massive NO formation, under certain conditions, can be cytotoxic and initiate apoptosis. In vitro studies have shown a direct toxicity of NO on oligodendrocytes, suggesting a role for NO in cell injury (16,17). NO cytotoxicity may result in the production of peroxynitrite, a highly reactive product of the reaction between NO and superoxide. Nitrosylation of proteins, peroxidation of lipids, S-nitrosylation of thiol groups, and inhibition of mitochondrial activity are among the cytotoxic effects resulting from exposure to NO/peroxynitrite and other reactive nitrogen species. Because NO is a highly reactive molecule that can interact with a variety of cellular components, it appears likely that damage to several key cellular constituents may lead to the final demise of the cell. One likely critical site for injury is DNA. NO can react with DNA via multiple pathways. Once produced, subsequent conversion of nitric oxide to nitrous anhydride can lead to the nitrosative deamination of DNA bases, such as guanine and cytosine. Complex oxidation chemistry can also occur causing DNA base and sugar-oxidative modifications (18). In a previous study from our laboratory (19), it was shown that, in pancreatic ␤-cell cultures, NO damages mtDNA to a greater extent than nuclear DNA. Additionally, Ballinger et al. (20) showed that, following exposure to NO and peroxynitrite in vascular endothelial and smooth muscle cells, mtDNA was preferentially damaged relative to the nuclear ␤-globin gene, resulting in decreased cellular ATP levels and mitochondrial redox function. Mutations and deletions in mtDNA, which could arise from unrepaired DNA damage, have been linked to a variety of diseases and aging. Moreover, alterations in the mitochondrial genome, which influence electron transport, could affect cells through the initiation of progressive cell death. Neurodegenerative diseases, including MS, are associated with a progressive loss of cells (21). Previously, we found a correlation between the induction of apoptosis by oxidative stress and repair of mtDNA in oligodendrocytes (22). In the present study, we investigated the mtDNA damage caused by reactive nitrogen species produced after iNOS induction by proinflammatory cytokines in rat oligodendrocytes. The results showed that there was significant mtDNA damage following exposure to cytokines. Additionally, we targeted the DNA repair enzyme 8-oxoguanine DNA glycosylase (hOGG1) with intrinsic AP lyase activity, which may be crucial for the repair of damage to the DNA sugar-phosphate backbone, to mitochondria of oligodendrocytes. Overexpression of hOGG1 in mitochondria of oligodendrocytes had a protective effect against NO-mediated mtDNA damage and increased cellular survival following treatment with cytokines.

EXPERIMENTAL PROCEDURES
Cell Culture, Transfection, and Treatment-Rat oligodendrocytes were harvested and cultured as previously described (23). Differentiation of oligodendrocytes was assessed by immunofluorescence microscopy, using antibodies to surface antigens (A2B5, O4, O1), in which sequential and partially overlapping expression defines early precursor, late precursor, and terminally differentiated oligodendrocyte stages, as described previously in detail (22). For several experiments, a vector containing a mitochondrial transport sequence (MTS) upstream of the sequence for hOGG1 was transfected into the cells. The construct has been described previously (22). Cells were transfected with a combination of 2 g of DNA/6 l of FuGENE 6 (Roche Diagnostics) according to the specification of the manufacturer. The highest levels of recombinant protein expression were observed 24 -72 h after transient transfection. The efficiency of transfection and the localization of the recombinant protein were determined as described previously (21). Transfected or untransfected cells were exposed to 50 ng/ml TNF␣ (R & D Systems, Minneapolis, MN) and 25 ng/ml IFN␥ (eBioscience, San Diego, CA) individually or in combination for 24, 48, or 72 h.
NO Determination-The intracellular production of NO was determined as the accumulation of nitrite (a stable oxidation product of NO) in the culture medium using the Griess method (24). Aliquots of 100 l of culture medium were mixed with an equal volume of the Griess reagent (0.1% (1-naphthyl)ethylenediamine dihydrochloride, 1% sulfanilamide, and 2.5% phosphoric acid). The absorbance at 550 nm was determined in a microplate reader, using sodium nitrite solutions (10 -100 M/l) as standards.
Assay for DNA Damage Detection by Southern Blot-As described in our previous publications (19,22,23,25), DNA damage was detected using a quantitative Southern blot technique in conjunction with either a mitochondrial-specific or an IgE-specific PCR-generated probe. The probe used to hybridize to mtDNA was generated via PCR from the mouse mtDNA sequence using the following primers: 5Ј-GCAGGAA-CAGGATGAACAGTCT-3Ј from the sense strand and 5Ј-GTATCGT-GAAGCACGATGTCAAGGGATGAG-3Ј from the antisense strand. The 745-bp PCR product recognizes a 10.8-kb restriction fragment when hybridized to rat DNA digested with BamHI. To detect the nuclear IgE gene, primers were selected as follows: 5Ј-ACATCACCAAGCCCACTG-TAGATCTACTCC-3Ј from the sense strand and 5Ј-CACGGTGATAT-TCTCCTCACTTTCCAGGTC-3Ј from the antisense strand. The 516-bp PCR product recognizes an 8-kb restriction fragment when hybridized to rat DNA digested with BamHI. Autoradiographic bands were densitometrically scanned. The break frequency was determined, as previously described (25), using the Poisson expression s ϭ Ϫln P 0 , where s is the number of breaks/fragment and P 0 is the fraction of fragments, free of breaks.
Assay for mtDNA Damage at the Nucleotide Level by Ligation-mediated PCR (LM-PCR)-DNA extraction and LM-PCR conditions have been previously described in detail (28). Briefly, untreated and alkalitreated DNA from each sample was subjected to ligation-mediated PCR. Primer extension and ligation were all performed using the procedure initially described by Pfeifer et al. (26,27) with certain modifications (28). After ligation, the reaction products were precipitated and then resuspended in H 2 O. PCR amplification was performed as previously described (26,27), the only exception being the use of 5 pmol/l primer and a longer oligonucleotide for the asymmetric linker. The increase in the amount of primer used in the LM-PCR was designed to compensate for the increased number of mtDNA templates present in cells compared with single-copy nuclear genes. After PCR, all samples were precipitated by cold ethanol. Pellets were air-dried and resuspended in a formamide dye solution. Samples were electrophoresed using a regular sequencing gel, blotted, and then UV light was cross-linked to the membrane, as previously described (28). Hybridization was performed with a single-strand PCR-generated probe using a primer for the appropriate strand. After exposure of the membrane to Kodak XAR film, the density of individual hybridization bands corresponding to nucleotide bases of interest was evaluated by densitometric scanning of the radioautographs.
Western Blot Analysis-Whole cell protein lysates, nuclear, mitochondrial, or cytosolic protein fractions, obtained as described previously (22), were used for Western blot assays. Protein concentration was determined using the Bio-Rad protein dye microassay according to the manufacturer's recommendation. SDS-polyacrylamide gel electrophoresis and transfer of separated proteins to polyvinylidene difluoridemembrane were performed by standard procedures. Blocking and antibody immunoblotting were performed in 6% nonfat dry milk, Trisbuffered saline (TBS)/0.1% Tween 20 (TBS-T). TBS-T and TBS were used for washing. The monoclonal anti-iNOS/NOS type II antibodies were from BD Biosciences; the polyclonal anti-hOGG1antibodies were from Novus Biologicals (Littleton, CO); anti-cytochrome c monoclonal antibody was purchased from Pharmingen; the anti-lamin B1 polyclonal antibodies were from Santa Cruz Biotechnology; the antiserum to caspase-9 was from Cell Signaling Technology (Beverly, MA); and the anti-actin antibody was from Sigma. The immune complexes formed by these antibodies were detected with horseradish peroxidase-conjugated anti-mouse IgG or anti-rabbit IgG antibodies (Promega, Madison, WI) using chemiluminescent reagents (SuperSignal; Pierce, Rockford, IL).
TUNEL Assay-The presence of DNA breaks were evaluated by a TUNEL (terminal deoxynucleotidyltransferase-mediated dUTP nickend labeling) procedure as described previously (29). For detection and quantitation of apoptotic cells, the DeadEnd fluorometric TUNEL system from Promega was used. It measures the fragmented DNA of apoptotic cells by catalytically incorporating fluorescein-12-dUTP at 3Ј-OH DNA ends using the enzyme terminal deoxynucleotidyltransferase. The assay was performed according to the manufacturer's procedure. Briefly, cells were grown on slides, fixed in 4% methanol-free formaldehyde, and permeabilized with 0.2% Triton X-100. Thereafter, DNA strand breaks were labeled with fluorescein-12-dUTP by the terminal deoxynucleotidyltransferase enzyme, and the nuclei were stained with DAPI. The number of apoptotic cells with green fluorescence (fluorescein-12-dUTP)/100 blue DAPI-stained nuclei was calculated by fluorescence microscopy.
Caspase-8 and -9 Activity Assay-After cytokine treatment, cells were rinsed with 1ϫ phosphate-buffered saline and collected via centrifugation at 200 ϫ g for 10 min. Caspase-8 and -9 activities were measured using colorimetric assay kits (R & D Systems, Minneapolis, MN). Briefly, cells were lysed on ice for 10 min, centrifuged at 10,000 ϫ g for 1 min, and 200 g of supernatant proteins were incubated with caspase-8-specific or caspase-9-specific substrates. The assay was read on a microplate reader using a 405-nm wavelength of light.
Data Analysis-All statistical analyses were performed using the Student's t test to compare individual means with significant differences at a confidence level of p Ͻ 0.05.

RESULTS
Several studies have reported that rat oligodendrocytes produce nitric oxide as a result of the induction of the type II iNOS gene (5,13). In agreement with these reports, our experiments utilizing the combination of TNF␣ (50 ng/ml) and IFN␥ (25 ng/ml) stimulated NO production in oligodendrocytes. As shown at Fig. 1A, NO production was determined as nitrite (a stable oxidation product of NO) released into the culture medium using a colorimetric method (see "Experimental Procedures"). To evaluate the expression of iNOS in cytokine-treated cultures, we carried out immunoblot analyses using antibodies specific for iNOS. The data illustrated in Fig. 1B confirm the induction of iNOS synthesis in response to the combination of TNF␣ and IFN␥. iNOS expression reached maximum levels 48 h after treatment. The cultures treated with TNF␣ and IFN␥ individually did not produce NO nor did they contain detectable immunoreactive iNOS protein (Fig. 1).
Previously, our laboratory has shown that endogenously produced NO can damage mtDNA in pancreatic ␤-cells (19). In the present work, we examined the relationship between cytokineinduced NO production and mtDNA damage in oligodendrocytes. Quantitative Southern blot analysis revealed damage to mtDNA in oligodendrocytes 48 h after TNF␣ and IFN␥ treatment (Fig. 2). mtDNA damage in oligodendrocytes 72 h after the introduction of the cytokines was less pronounced, suggesting the activation of mtDNA repair mechanisms in the remaining viable cells. Treatment with cytokines and the inhibitor of iNOS, aminoguanidine, significantly reduced mtDNA damage in the cells (Fig. 2). The number of DNA breaks in similarly sized fragments of nuclear DNA containing the IgE gene was ϳ10-fold less than that detected in mitochondrial DNA and falls in the range of background noise for quantitative Southern blot analysis (Fig. 3).
To identify the specific pattern of cytokine-mediated mtDNA damage at the nucleotide level, LM-PCR was performed on a 200-bp sequence from the heavy strand of mtDNA from treated oligodendrocytes. This sequence contained one of the break points for the 5-kb "common deletion" that accumulates in mtDNA with aging (30,31). Previously, we have shown that NO can modify DNA by causing the deamination of purines, whereas reactive oxygen species generated by the enzymatic reaction of xanthine oxidase and substrate hypoxanthine cause modifications in both purines and pyrimidines (28). The pattern of cytokine-induced mtDNA damage was compared with the pattern of damage produced by PAPA/NO, xanthine oxidase/hypoxanthine, and peroxynitrite (Fig. 4). Using a Maxim-Gilbert sequencing ladder to identify bases, it appeared that most of the damage caused by PAPA/NO involved guanines, with some adenines also being affected (Figs. 4 and 5). In comparison, peroxynitrite and cytokines preferentially modified guanine and adenine; however, some thymine and cytosine residues also were modified. Thymine and guanine were the preferred targets of xanthine oxidase/hypoxanthine. It is readily apparent from the lesion map in Fig. 5 that the pattern of damage was virtually identical for cytokines and peroxynitrite, suggesting that peroxynitrite is the oxidizing agent responsible for most of the cytokine-mediated mtDNA damage.
Recently, we reported (22) that targeting of hOGG1 to mitochondria in oligodendrocytes enhanced the repair of oxidative lesions in mtDNA. Oligodendrocytes were transiently transfected with a vector containing a MTS upstream of the sequence for human OGG1 or empty vector for control. Previously, localization of recombinant protein was confirmed by fluorescence microscopy (22). Fig. 6 shows that expression of the OGG protein is increased in the mitochondrial (but not nuclear) protein fraction of MTS-hOGG1-transfected cells as compared with vector-only-transfected cells. Anti-lamin B1 and anti-cytochrome c antibodies were used as loading controls as well as a control for the purity of fractions.
Quantitative Southern blot analysis showed a significant decrease in cytokine-induced mtDNA damage in MTS-hOGG1transfected oligodendrocytes compared with cells containing only vector (Fig. 7A). The break frequency/10.8-kb fragment is shown on Fig. 7B.
Next, we investigated whether the observed increase in mtDNA repair influences viability of the cells following TNF␣ and IFN␥ treatment. The DeadEnd fluorometric TUNEL system used for the specific detection and quantitation of apoptotic cells revealed that MTS-hOGG1-transfected oligodendrocyte cultures had fewer apoptotic cells (17% Ϯ5.2), compared with cultures containing vector-only cells (31% Ϯ3.6) after 48 h of cytokine treatment. These results show that targeting the repair enzyme hOGG1 to the mitochondria of oligodendrocytes significantly protects against the induction of apoptosis by 48 h after TNF␣ and IFN␥ treatment.
To determine the pathway of apoptosis affected by enhanced mtDNA repair, we examined the activation of both caspase-8 and -9 in MTS-hOGG1-and vector-transfected oligodendrocytes following cytokine treatment. The induction of apoptosis through the extrinsic death receptor mechanism resulted in the activation of caspase-8, whereas the intrinsic mitochondrial death signal led to the activation of caspase-9. Previously, it had been reported (32) that TNF␣ and IFN␥ induce the expression of the Fas (death-receptor) in primary oligodendrocyte cultures, and cytokine-induced apoptosis mainly goes through the extrinsic pathway. In our study, colorimetric activity assays based on cleavage of caspase-8 (33)-or caspase-9 (34)specific substrate were performed. For a positive control, Jur-kat cells were treated with camptothecin. Fig. 8 shows that 48 h after TNF␣ and IFN␥ treatment, both caspase-8 and -9 were activated in MTS-hOGG1 and vector cells. But intriguingly, following cytokine induction, caspase-9 activity in vectoronly-transfected oligodendrocytes was significantly elevated compared with MTS-hOGG1 transfectants.
Further, to confirm that overexpression of hOGG1 in mitochondria of oligodendrocytes led to a substantial inhibition of the activation of caspase-9 following exposure to cytokines, we performed Western blot analysis using an antibody against cleaved caspase-9. Cytosolic protein fractions of control and TNF␣ ϩ IFN␥-treated samples from MTS-hOGG1 and vectoronly-transfected oligodendrocytes were subjected to immunoblotting. A specific antibody that detects cleaved (activated) 40-kDa and 38-kDa subunits of caspase-9 was employed. As shown in Fig. 9, the bands corresponding to the cleaved subunits of caspase-9 were markedly increased in vector-only samples as compared with MTS-hOGG1 transfectants following cytokine induction. To ensure equal loading, the membranes were reblotted with an anti-actin antibody. DISCUSSION The results of the present investigation show that endogenous NO production following cytokine-induced iNOS expression can cause mtDNA damage in oligodendrocytes. Furthermore, targeting of the specific DNA repair protein (hOGG1) to mitochondria of oligodendrocytes decreases the sensitivity of mtDNA to NO-mediated damage and protects these cells against apoptosis.
In the CNS, astrocytes, microglia, and immune-derived cells all release inflammatory cytokines and cytotoxic mediators that may damage and destroy myelinating oligodendrocytes and their progenitors. A substantial body of evidence supports the involvement of cytokine-induced iNOS-mediated NO production in CNS-demyelinating diseases (35,36,11,12). It is now apparent that two proinflammatory cytokines, TNF␣ and IFN␥, known to accumulate in the brain and cerebrospinal fluid of MS patients, play prominent roles in the immunopathogenesis of the disease (7,8). In the present study, we observed that treatment of oligodendrocytes with the combination of TNF␣ and IFN␥ resulted in the expression of iNOS and the concomitant production of NO. This observation is in agreement with previous reports, which show that cultured rat oligodendrocytes express iNOS and release NO following activation with lipopolysaccharide/IFN␥ (5,14) and that the central Oligodendrocytes were transfected with a vector containing a MTS upstream of the sequence for human OGG1 (OGG) and empty pcDNA3.0neo (V) for control. Both cultures were treated with TNF␣ (50 ng/ml) and IFN␥ (25 ng/ml) for 24, 48, and 72 h. High molecular weight DNA was isolated and digested to completion with BamHI. Samples were exposed to 0.1 N NaOH prior to Southern blot analysis and hybridization with a PCR-generated mtDNA probe. A, representative autoradiograph from cells treated 48 h (c, untreated control; tr, cytokine treated). B, break frequency/10.8-kb mtDNA restriction fragment. Autoradiographic bands representing a 10.8-kb mtDNA fragment were densitometrically scanned. The break frequency was determined using the Poisson expression s ϭ Ϫln P 0, where s is the number of breaks/fragment and P 0 is the fraction of fragments, free of breaks. Data are means ϩ S.E. of three separate experiments. *, p Ͻ 0.05 when the data from MTS-OGG transfectants was compared with vector-only cells using Student's t test.
FIG . 8. Activation of caspase-8 and -9 after cytokine treatment. MTS-hOGG1 (OGG)-and vector-only (V)-transfected oligodendrocytes were treated with TNF␣ (50 ng/ml) and IFN␥ (25 ng/ml) for 48 h. The cells were rinsed with 1ϫ phosphate-buffered saline, collected via centrifugation, and lysed on ice for 10 min. The lysate was then centrifuged at 10,000 ϫ g for 1 min and 200 g of supernatant proteins incubated with caspase-9-or caspase-8-specific substrate. The assay was read on a microplate reader using 405-nm wavelength light. Data are means ϩ S.E. of three separate experiments. *, p Ͻ 0.05 when the data from MTS-OGG transfectants was compared with vector-only cells using Student's t test. For a positive control, Jurkat cells (Jur) treated with 10 M camptothecin for 5 h were used. glial four oligodendrocyte cell line expressed iNOS mRNA and protein after exposure to TNF␣/IFN␥ (13).
NO is a highly reactive, potentially toxic molecule that can interact with a variety of cellular components. One likely critical site for injury is DNA. In the present work, we found that exposure of oligodendrocytes to the combination of TNF␣ and IFN␥ induced mtDNA damage, which correlated with cytokineinduced NO production and the expression of iNOS. That this damage was due to the generation of NO is supported by the observation that it was attenuated by treatment with the inhibitor of iNOS, aminoguanidine. We were not able to detect appreciable DNA damage in similarly sized fragments of nuclear DNA. It has been reported that iNOS-mediated endogenous generation of NO in mammalian cells does not appreciably increase the steady-state level of nuclear DNA damage (37). However, studies from our laboratory and those of others have shown that mtDNA is vulnerable to NO-mediated damage (19,20,27,38). Although in this study we cannot rule out the damage to the nucleus at or below 1 adduct/100 kb, it is apparent that the cytokine-mediated DNA damage was much greater in mtDNA.
Damage to DNA can occur through reactions with nitrous anhydride (N 2 O 3 ), formed by the reaction of NO with molecular oxygen (O 2 ), or reactions with peroxynitrite (ONOO Ϫ ), formed by the combination of NO with superoxide (39). Peroxynitrite can oxidize and nitrate DNA and may cause single-strand breaks through attack on the sugar-phosphate backbone (18). N 2 O 3 can nitrosate amines to form N-nitrosamines, which, after metabolic activation, can alkylate purines. Additionally, N 2 O 3 can cause nitrosation of primary amines in DNA bases, which leads to the formation of diazonium ions and subsequent deamination (18). Deamination of cytosine, adenine, and guanine may cause mutations in cells (40). Furthermore, xanthine (produced by the deamination of guanine) and hypoxanthine (the product of the deamination of adenine) are unstable and can form abasic sites that are both toxic and mutagenic (40,41).
Previously, analysis of a 202-bp sequence of mtDNA by the technique of LM-PCR showed that exogenous NO generated by PAPA/NO damaged specific guanines and adenines (28). No damage was detected to any pyrimidines in the sequence evaluated. For comparison, studies with the reactive oxygen species generator xanthine oxidase/hypoxanthine showed that we were able to detect a reproducible pattern of oxidative lesions that encompassed both purines and pyrimidines (28). In this study, we compared the pattern of nucleotide damage caused by the cytokines with the patterns produced by NO (PAPA/NO), reac-tive oxygen species (xanthine oxidase/hypoxanthine), and peroxynitrite. The pattern of damaged mtDNA nucleotides produced by TNF␣ and IFN␥ is virtually identical to that seen with peroxynitrite, suggesting that this molecule is the predominant damaging species. The spectrum of peroxynitrite DNA damage tends to be much more complex than that caused by N 2 O 3 . The two main types of base modification attributed to ONOO Ϫ are oxidations and nitrations (42). Also, it has been shown that peroxynitrite can cause DNA strand breaks (43). Mitochondria constitute a primary site for the intracellular reaction of peroxynitrite with DNA. Although these organelles possess scavenging and repair systems for peroxynitrite-dependent oxidative modifications, our data reveal that they can be overwhelmed under enhanced cellular formation of NO, as well as under conditions that lead to augmented superoxide formation, such as through defects in electron transport chains (44).
Recently, we and others have reported (22,45,46,47) that targeting of bifunctional glycosylase/AP lyases to mitochondria of various cells enhances the repair of oxidative lesions in mtDNA. In the present study, we observed that oligodendrocytes transfected with hOGG1, containing a MTS, showed a significant decrease in mtDNA damage following cytokine exposure compared with the cells transfected with vector only. The attenuation of cytokine-induced mtDNA damage by overexpression of hOGG1 in mitochondria may be explained by the AP lyase activity of OGG1. This activity may be crucial for the repair of damage to the sugar-phosphate backbone, which comprises a significant proportion of the total lesions in DNA following oxidative stress (48). The importance of the protective role of AP lyase activity against oxidative damage is suggested by the results that the bacterial enzymes endonuclease III (Endo III) and endonuclease VIII (Endo VIII), which have AP lyase activity but glycosylase activity for lesions other than 8-oxoguanine, protected mtDNA in HeLa cells from oxidative damage caused by exposure to menadione (46).
It has been reported that oligodendrocyte progenitors are particularly vulnerable to the activation of the cell death program following exposure to the combination of TNF␣ and IFN␥ (32). Our results using TUNEL labeling showed that apoptosis was attenuated in MTS-hOGG1-transfected oligodendrocytes as compared with vector-only-containing cells 48 h following cytokine treatment. The finding that targeting hOGG1 to mitochondria of oligodendrocytes reduced peroxynitrite-induced apoptosis supports our previous report in which we found that overexpression of hOGG1 in oligodendrocytes protected cells against apoptosis caused by an oxidative insult (22). These findings allowed us to suggest that damage to mtDNA may be an early step in the induction of apoptosis. Our data employing a colorimetric activity assay revealed that 48 h after TNF␣ and IFN␥ treatment, both caspase-8 and -9 were activated in both MTS-hOGG1-and vector-containing cells. However, caspase-9 activity in MTS-hOGG1 transfectants was significantly diminished compared with the vector-only-transfected oligodendrocytes. The majority of oligodendrocytes express TNF receptors, which recognize their specific ligands, resulting in the activation of the respective death receptors, followed by activation of caspase-8 (49). An increased expression of the death receptor Fas in cytokine-treated cultures of oligodendrocytes has been shown by others (31). It has recently become apparent that apoptotic signals coming from activated receptors can be amplified by mitochondria-dependent apoptotic pathways. For instance, caspase-8 can cleave the Bcl-2 family member Bid, and the truncated Bid is able to translocate from the cytosol to mitochondria and induce cytochrome c release followed by caspase-9 activation and thus amplify the apoptotic signal (50). FIG. 9. Cleaved (activated) caspase-9 in cytokine-induced cultures. Western blots were performed on the cytosolic protein fraction from untreated cells and cells treated with TNF␣ (50 ng/ml) and IFN␥ (25 ng/ml). 48 h after treatment, the cells were lysed, the cytosolic fraction was isolated, and SDS-PAGE was performed. Polyclonal antibodies were used to recognize the 40-kDa (prodomain ϩ large fragment cleaved at Asp-368) and 38-kDa (prodomain ϩ fragment cleaved at Asp-353) subunits of caspase-9. To confirm equal loading, the membrane was stripped and reblotted with an anti-actin antibody (42 kDa Our data indicate that cytokine-mediated mtDNA damage may be an additional stimulus that can activate caspase-9 to augment the apoptotic signal from caspase-8. The exact mechanism by which mtDNA damage caused by reactive nitrogen species may enhance cytokine-induced apoptosis in oligodendrocytes remains to be fully elucidated. However, it appears reasonable to speculate that, because mitochondria have a central role in energy metabolism and redox regulation, the accumulation of unrepaired lesions in mtDNA would cause an alteration in the transcription of mtDNA and modify the flow of electrons by changing key electron transport complexes (51) leading to defects in oxidative phosphorylation and impaired ATP/ADP exchange. This mitochondrial dysfunction would ultimately lead to the activation of the mitochondrial portion of the programmed cell death pathway. Overexpression of the DNA repair enzyme hOGG1 in mitochondria decreased the ratio of unrepaired mtDNA damage and thus prevented apoptotic signaling by limiting the amount of mitochondrial dysfunction. Clearly, further experimentation is required to fully understand the precise role of unrepaired mtDNA damage in the induction of the apoptosis.
Inflammatory and demyelinating diseases of the CNS, such as MS, are associated with a progressive loss of cells. Cytokines, which are major effector molecules in many disease processes, stimulate iNOS-mediated excessive production of NO in oligodendrocytes. This can cause NO-dependent mtDNA damage, which can contribute to the induction of apoptotic cell death. Targeting of a specific DNA repair enzyme to mitochondria attenuates cytokine-mediated mtDNA damage and increases the survival of these cells. This may be a viable approach for the protection of oligodendrocytes from stress caused by reactive oxygen species and reactive nitrogen species and may be a feasible future gene therapy strategy for dealing with demyelinating diseases of the CNS.