Increased cytotoxicity of 3-morpholinosydnonimine to HepG2 cells in the presence of superoxide dismutase. Role of hydrogen peroxide and iron.

3-Morpholinosydnonimine (SIN-1) is widely used to generate nitric oxide (NO) and superoxide radical (O). The effect of SOD on the toxicity of SIN-1 is complex, depending on what is the ultimate species responsible for toxicity. SIN-1 (<1 mM) was only slightly toxic to HepG2 cells. Copper, zinc superoxide dismutase (Cu,Zn-SOD) or manganese superoxide dismutase (Mn-SOD) increased the toxicity of SIN-1. Catalase abolished, while sodium azide potentiated, this toxicity, suggesting a key role for H2O2 in the overall mechanism. Depletion of GSH from the HepG2 cells also potentiated the toxicity of SIN-1 plus SOD. Although Me2SO, sodium formate, and mannitol had no protective effect, iron chelators, thiourea and urate protected the cells against the SIN-1 plus Cu,Zn-SOD-mediated cytotoxicity. The cytotoxic effect of Cu,Zn-SOD but not Mn-SOD, showed a biphasic dose response being most pronounced at lower concentrations (10-100 units/ml). In the presence of SIN-1, Mn-SOD increased accumulation of H2O2 in a concentration-dependent manner. In contrast, Cu,Zn-SOD increased H2O2 accumulation from SIN-1 at low but not high concentrations of the enzyme, suggesting that high concentrations of the Cu,Zn-SOD interacted with the H2O2. EPR spin trapping studies demonstrated the formation of hydroxyl radical from the decomposition of H2O2 by high concentrations of the Cu,Zn-SOD. The cytotoxic effect of the NO donors SNAP and DEA/NO was only slightly enhanced by SOD; catalase had no effect. Thus, the oxidants responsible for the toxicity of SIN-1 and SNAP or DEA/NO to HepG2 cells under these conditions are different, with H2O2 derived from O dismutation playing a major role with SIN-1. These results suggest that the potentiation of SIN-1 toxicity by SOD is due to enhanced production of H2O2, followed by site-specific damage of critical cellular sites by a transition metal-catalyzed reaction. These results also emphasize that the role of SOD as a protectant against oxidant damage is complex and dependent, in part, on the subsequent fate and reactivity of the generated H2O2.

Nitric oxide (NO ⅐ ) 1 and other reactive nitrogen oxidative metabolites (NO x ) are produced by a variety of mammalian cells. These agents may be important for host defense, but under certain conditions, NO ⅐ and NO x may cause tissue damage by still to be clarified mechanisms. Both NO ⅐ and superoxide anion radical (O 2 . ) are known to be generated by macrophages, neutrophils, and endothelial cells (1)(2)(3). The production of these two radicals under physiological conditions can lead to the formation of peroxynitrite (ONOO Ϫ ) and other reactive species that are cytotoxic (4 -8). The cytotoxicity of ONOO Ϫ may be mediated by the ability of this compound to initiate lipid peroxidation (9,10), cause oxidation of protein and non-protein sulfhydryls (11), produce nitration of tyrosine residues in proteins (5), or react with sugars (4) and DNA (12). Superoxide dismutase (SOD) present in the cytoplasm and mitochondria within cells and in the extracellular space dismutates two O 2 . into H 2 O 2 and O 2 (13). The protective effect of SOD against oxygen-derived free radicals in vivo, in intact cells and in vitro studies is well documented (14,15). However, in some studies a bell-shaped dose-response curve for the protective effect of SOD was observed. At low concentrations, SOD was usually protective, but at very high concentrations, its protective effect decreased or was reversed such that SOD potentiated toxicity (16 -18). SIN-1 (3-morpholinosydnonimine), the active metabolite of the vasodilatatory drug molsidomine, is frequently used as a model compound for a continuous release of O 2 . , NO ⅐ and/or NO x , and other potent oxidants such as ONOO Ϫ and hydroxyl radical ( ⅐ OH) (16,19,20). These SIN-1-derived strong oxidants degrade deoxoyribose (6), oxidize low density lipoproteins (21), mediate loss of microsomal ␣-tocopherol (22), damage surfactant protein A (23), and inhibit glyceraldehyde-3-phosphate dehydrogenase (24) and hepatic gluoneogenesis (25). Cytotoxic effects of SIN-1 have been demonstrated in a variety of cell lines (26 -28). For example, the LD 50 for SIN-1 cytotoxicity against Escherichia coli was 0.5 mM (29), and 1 mM SIN-1 was very toxic to neurons (30). SOD had a protective effect against SIN-1 toxicity to cultured neurons (30) and to E. coli (29), but had no protective effect with rat hepatoma cells (28), and SOD actually increased SIN-1 toxicity to Leishmania major (31). The effect of SOD on the toxicity of SIN-1 is complex, e.g. if ONOO Ϫ is the cause of toxicity, SOD will be protective since dismutation of O 2 . will prevent formation of ONOO Ϫ . However, if NO ⅐ itself is the toxic agent, SOD can potentiate NO ⅐ toxicity * This work was supported by Grants AA-03312 and AA-09460 from the National Institute on Alcohol Abuse and Alcoholism. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  , EDTA, and DTPA were from Sigma. Sodium azide was from Aldrich. 5,5-Dimethylpyrroline-N-oxide (DMPO) was from Sigma and was purified before use by charcoal treatment to remove paramagnetic impurities.
Cell Cultures-The human hepatoma cell line HepG2 (American Type Culture Collection, HB8065) was used for these studies. Cells were cultured in minimum essential medium (MEM, Life Technologies, Inc.) supplemented with 1% PSN antibiotic mixture (Life Technologies, Inc.), and, 10% fetal calf serum (Sigma) in 25-cm 2 plastic flasks (Corning) at 37°C in a humidified atmosphere of 95% air and 5% CO 2 . For cytotoxicity experiments, cells were plated at 4 ϫ 10 4 cells/well/0.5 ml of tissue culture medium on 24-well tissue culture plates. After 2 h of incubation and attachment, SIN-1, SOD, as well as other compounds were added and the cells were incubated for the indicated time periods at 37°C. Stock solutions of all compounds were prepared in MEM medium immediately before use. The preparation of solutions and incubation with the HepG2 cells were carried out under protection from light. Cell viability (cytotoxicity) was determined using the Cell Titer 96 non-radioactive cell proliferation/cytotoxicity assay kit (Promega), which is based on the cellular conversion of a tetrazolium salt into a formazan product that can be detected spectrophotometrically (32). GSH levels were determined using the GSH-400 kit assay (Cayman Chemical, Ann Arbor, MI).
Determination of H 2 O 2 -H 2 O 2 generation at different concentrations of SIN-1, in the absence or presence of SOD, was determined by two techniques after a 24-h incubation in tissue culture medium at 37°C in 95% air and 5% CO 2 in the absence of cells. In the first method, the H 2 O 2 content was measured by formation of formaldehyde from the oxidation of methanol by the catalase-H 2 O 2 compound I complex (33). Incubations contained 100 mM methanol, 1200 units/ml catalase, SIN-1, and tissue culture medium. Reactions were terminated after a 1-h incubation at 37°C by the addition of 20% trichloroacetic acid. The generation of formaldehyde was determined by the Nash reaction (34), and calculation of H 2 O 2 concentration was carried out as described previously (35). In the second method, the formation of H 2 O 2 was determined by the horseradish peroxidase-catalyzed reaction of H 2 O 2 with 4-amino-antipyrine plus 3,5-dichloro-2-hydroxybenzenesulfonic acid. After a 1-h incubation of SIN-1 with the above compounds in tissue culture medium at 37°C, the resulting product was measured spectrophotometrically at 546 nm (28,36).
EPR Measurements-The production of ⅐ OH from the SOD-catalyzed decomposition of H 2 O 2 was determined by spin trapping with DMPO. MEM containing 100 or 1000 units/ml Cu,Zn-SOD or Mn-SOD and 100 mM DMPO was incubated with 80 M H 2 O 2 . The samples were filled in a Varian flat cell and measured within 1 min after addition of H 2 O 2 , using a Varian EPR 200 spectrometer operating in the X-band mode. Instrument parameters are indicated in the legends to figures.
In separate experiments evaluating iron-catalyzed decomposition of H 2 O 2 , 1 mM SIN-1 was co-incubated with either Cu,Zn-SOD (100 or 1000 units/ml) or Mn-SOD (100 or 1000 units/ml) for 24 h at 37°C in the absence of cells. At the end of the incubation period, 100 mM DMPO and 600 mM Me 2 SO were added and decomposition of H 2 O 2 was initiated by the addition of 40 M FeSO 4 . In the presence of Me 2 SO, ⅐ OH radicals are converted to CH 3 radicals, which were spin-trapped with DMPO (DMPO/CH 3 adducts), thus providing a characteristic OH radical fingerprint (37). The samples were transferred to a EPR flat cell and measured immediately after the addition of the last compound.

Cytotoxicity of SIN-1 to HepG2
Cells-The cytotoxic effect of SIN-1 in the absence or presence of Cu,Zn-SOD to the human hepatoma HepG2 cell line was studied at different SIN-1 and Cu,Zn-SOD concentrations over a 24-h time period (Fig. 1). SIN-1 alone at concentrations of 0.1 or 0.316 mM was not toxic over this time course; some toxicity was observed at 1 mM SIN-1. At the very high concentration of 3 mM, SIN-1 toxicity was pronounced and increased during incubation. The addition of Cu,Zn-SOD potentiated the cytotoxic effect of SIN-1; this is most notable at SIN-1 concentrations of 0.316 and 1 mM (Fig. 1, B and C). Pronounced cytotoxicity was found in the presence of Cu,Zn-SOD plus SIN-1 at concentrations where SIN-1 alone had no effect (0.316 mM) or only a minor effect (1 mM). The potentiation of the cytotoxic effect of SIN-1 by Cu,Zn-SOD increased during incubation. A surprising observation was the finding that in the presence of 0.316 mM SIN-1, the potentiating cytotoxic effect of Cu,Zn-SOD was pronounced at 10 and 100 units of SOD per ml, but decreased as the units of added Cu,Zn-SOD were elevated to 1000 units/ml. In other experiments, 0.316 or 1 mM SIN-1 in the absence and presence of 100 units of Cu,Zn-SOD was preincubated in culture medium in the absence of HepG2 cells for 3 h. The HepG2 cells were added, and cell viability was compared to that of a system in which SIN-1, SOD, and cells were added together at the same time. At all time points tested (2-18 h), the medium "conditioned" by preincubating SIN-1 and Cu,Zn-SOD was considerably more toxic, which suggests accumulation of a stable toxic metabolite produced from the incubation of SIN-1 in the presence of SOD.
A more detailed study of the concentration dependence of the  Fig. 2. The potentiation of SIN-1 cytotoxicity by Cu,Zn-SOD was biphasic; toxicity increased in a concentration-dependent manner up to 100 units of Cu,Zn-SOD/ml but decreased at higher concentrations ( Fig. 2A). Boiled Cu,Zn-SOD did not potentiate the cytotoxicity of SIN-1, and Cu,Zn-SOD in the absence of SIN-1 had no cytotoxic effect ( Fig. 2A). Another type of superoxide dismutase, Mn-SOD, also potentiated the cytotoxicity of SIN-1, but in contrast to the Cu,Zn-SOD, the potentiation of SIN-1 cytotoxicity by the Mn-SOD did not decrease at higher concentrations of this isoform (Fig. 2B).
Since SOD increases the rate of production of H 2 O 2 by catalyzing O 2 . dismutation, the effect of catalase on the SIN-1 plus SOD toxicity was determined. Catalase (200 units/ml) abolished the SOD-mediated cytotoxic effect of SIN-1 at all Cu,Zn-SOD concentrations tested ( Fig. 2A). The possible involvement of intracellular H 2 O 2 and catalase in the cytotoxic effect of SOD plus SIN-1 was studied using sodium azide, an inhibitor of the intracellular catalase. Sodium azide slightly increased the toxicity of SIN-1 in the absence of SOD and further potentiated the cytotoxic effect produced by the combination of SIN-1 plus Cu,Zn-SOD ( Fig. 2A).
Since H 2 O 2 can be a precursor for formation of ⅐ OH, the possible involvement of ⅐ OH-like species in the SOD-mediated cytotoxic effect of SIN-1 was evaluated using ⅐ OH scavengers. Viability of the HepG2 cells was lowered by 10% by 0.632 mM SIN-1 and by 90% in the presence of SIN-1 plus 100 units/ml Cu,Zn-SOD (Fig. 3). Thiourea and uric acid were very effective in protecting the HepG2 cells against this SOD-mediated cytotoxicity. However, other ⅐ OH scavengers tested, including Me 2 SO, sodium formate, and mannitol, had no protective effect ( Fig. 3). Similarly, ethanol (10 -100 mM) and benzoic acid (0.5-10 mM) also had no protective effect (data not shown).
Formation of ⅐ OH and other potent oxidizing species from H 2 O 2 requires catalysis by transition metals such as iron. The role of transition metals in the SOD plus SIN-1 cytotoxicity was studied using the iron chelators deferoxamine and 2,2Ј-dipyridyl. As shown in Fig. 4, the cytotoxicity of SOD plus SIN-1 was reduced in a concentration-dependent manner by co-incubation of the samples with these two iron chelators. In contrast, EDTA and DTPA had only a small protective effect; the viability of the cells exposed to the SIN-1 plus SOD combination increased from 20% in the absence of chelator to 34 and 39% in the presence of 1 mM EDTA or 1 mM DTPA, respectively (data not shown).
The role of intracellular glutathione in the protection of HepG2 cells against the toxicity of SIN-1 itself or the potentiation of SIN-1 toxicity by SOD was studied using BSO, which inhibits GSH synthesis (38). GSH functions to help remove H 2 O 2 via glutathione peroxidase activity; GSH can also scavenge peroxynitrite (29). BSO at a concentration (0.1 mM) that depleted cellular GSH by 85 to 90% increased the cytotoxic effect of SIN-1 itself from about 5% loss of viability in the absence of BSO treatment to about 50% loss of viability after BSO treatment (Fig. 5). In addition, the BSO treatment also increased the potentiation of SIN-1 cytotoxicity produced by Cu,Zn-SOD (Fig. 5), analogous to the potentiation produced by azide ( Fig. 2A).
Production of H 2 O 2 -The protective effects of catalase and GSH against SIN-1 and especially SIN-1 plus SOD toxicity suggests that H 2 O 2 was involved in the overall mechanism leading to toxicity. The increase in SIN-1 toxicity produced by addition of Cu,Zn-SOD or Mn-SOD could reflect elevated production of H 2 O 2 in the presence of these enzymes. Accumulation of H 2 O 2 from SIN-1 was therefore determined. Measurement of H 2 O 2 by two independent methods showed agreement between the concentrations of accumulated H 2 O 2 by the various reaction systems and the cytotoxicity to the HepG2 cells. Mn-SOD (Fig. 6); this accumulation of H 2 O 2 is consistent with the enhancement of SIN-1 toxicity produced in the presence of these concentrations of SOD. Of interest was the observation that levels of H 2 O 2 produced from SIN-1 were actually lower in the presence of higher concentrations of the Cu,Zn-SOD, but not the Mn-SOD (Fig. 6). The Cu,Zn-SOD potentiation of SIN-1 toxicity was also lower at elevated concentrations of the Cu,Zn-SOD (Fig. 2B). The possibility that catalase was present as a contaminant in the Cu,Zn-SOD preparation, thereby decreasing H 2 O 2 accumulation and SIN-1 toxicity was ruled out by demonstrating that very high concentrations of azide (5 mM), which should completely inhibit any catalase that could be present, did not affect the accumulation of H 2 O 2 produced from SIN-1 in the presence of low and high concentrations of the Cu,Zn-SOD (data not shown).
Spin Trapping of ⅐ OH-The availability of H 2 O 2 accumulated after a 24-h incubation in the SIN-1 plus SOD systems for iron (Fe(II))-catalyzed decomposition was demonstrated by spin trapping with DMPO (Fig. 7). The products of the Fenton reaction ( ⅐ OH radical or Fe(IV) species) were converted by reaction with Me 2 SO to methyl radicals which were then spintrapped with DMPO. Only low levels of ⅐ OH radicals (DMPO/ CH 3 adducts; a N ϭ 16.24 G; a H ϭ 23.19 G) could be detected (Fig. 7) upon addition of Fe(II) to the samples which were incubated with either SIN-1 alone (trace E) or SIN-1 plus 1000 units/ml Cu,Zn-SOD (trace B). Formation of ⅐ OH was significantly higher in the samples incubated with SIN-1 plus 100 units/ml Cu,Zn-SOD, 100 units/ml Mn-SOD, or 1000 units/ml Mn-SOD (Fig. 7, traces A, C, and D). Production of ⅐ OH (DMPO/ CH 3 adduct) therefore paralleled accumulation of H 2 O 2 (Fig. 6).
The mechanism of the observed pattern of H 2 O 2 accumulation in the presence of Cu,Zn-SOD and Mn-SOD was studied. A strong DMPO/OH signal (a N ϭ a H ϭ 14.9 G) was observed when 50 M H 2 O 2 (comparable concentrations were produced by the SIN-1 plus SOD system during a 24-h incubation) was added to the medium containing 1000 units/ml Cu,Zn-SOD and 100 mM DMPO (Fig. 8B). The intensity of the DMPO/OH signal was considerably smaller when 100 units/ml Cu,Zn-SOD was used (Fig. 8A), and no DMPO/OH signals could be detected when Mn-SOD (100 or 1000 units/ml) was used (Fig. 8, C and  D). The Cu,Zn-SOD concentration-dependent formation of DMPO/OH adducts is shown in Fig. 9. Formate, but not Me 2 SO, inhibited the DMPO/OH signal produced in the Cu,Zn-SOD/H 2 O 2 system (data not shown), consistent with the decomposition of H 2 O 2 within the channel of Cu,Zn-SOD in the active site of the enzyme (39 (Fig. 10). . and NO ⅐ in order to mimic the release of these agents by macrophages, neutrophils, and endothelial cells (6, 19 -22). The generation of these two radicals under physiological conditions can lead to the formation of ONOO Ϫ and other reactive species (4,5,8). Superoxide dismutase (SOD), ubiquitously present in living aerobic organisms, generally has protective effects against oxygen-derived free radicals (14,15). However, in some studies, a bell-shaped dose-response curve of the protective effect of SOD was observed (16 -18). An increase in SOD activity has been postulated to enhance lipid peroxidation (40,41) and cellular injury (42,43 (26 -28). SIN-1 induced neuronal cell death in a dose-dependent manner; however, the neurons were completely protected if SOD was present in the reaction system (30). Brunelli et al. (29) found that SIN-1 killed E. coli with an LD 50 of 0.5 mM. Cu,Zn-SOD (50 -400 units/ml) provided substantial, but not complete, protection against this SIN-1 killing. Cu,Zn-SOD and catalase together completely protected E. coli against SIN-1 toxicity. Assreuy et al. (31) recently reported that SIN-1 (0.3 mM) was efficient in killing L. major, and this cytotoxicity was enhanced by the Cu,Zn-SOD (500 units/ml). The authors suggested that the killing of L. major by SIN-1 was dependent only on NO ⅐ , and the enhancement of toxicity by SOD was due to scavenging of O 2 . , thereby increasing the half-life of the NO ⅐ .
Ioannidis and DeGroot (28) found that the cytotoxicity of SIN-1 to Fu5 hepatoma cells was not affected by SOD, but catalase diminished cell damage; the authors suggested a cooperative toxic action between H 2 O 2 and the tumoricidal activity of NO ⅐ in the Fu5 cells. Siegfried et al. (47) reported that SIN-1 had significant cardioprotective effects in a myocardial ischemiareperfusion model, perhaps via quenching of O 2 . by NO produced from the SIN-1. We found that Cu,Zn-SOD did not protect against the cytotoxicity to human hepatoma HepG2 cells caused by a high concentration of (3 mM) SIN-1. Actually, both Cu,Zn-SOD and Mn-SOD potentiated SIN-1-mediated cytotoxicity against the HepG2 cells, resulting in toxicity at SIN-1 concentrations that alone had no, or only a minor, cytotoxic effect. Lack of potentiation by boiled Cu,Zn-SOD indicated that catalytically active enzyme was required for potentiating this cytotoxic effect of SIN-1. Catalase abolished the SIN-1 plus SOD-mediated cytotoxic effect at all Cu,Zn-SOD concentrations evaluated (Fig. 2), whereas boiled catalase had no effect. This suggests that H 2 O 2 , produced by a SOD-catalyzed dismutation of O 2 . , is involved in the overall cytotoxic effect of SIN-1 plus SOD. The further potentiation of the SIN-1 plus Cu,Zn-SOD cytotoxicity by sodium azide, an inhibitor of intracellular catalase, is in accord with this suggestion, as are assays of H 2 O 2 accumulation produced by SIN-1 in the absence and presence of SOD. Alternatively, catalase may be protective by reaction with NO ⅐ thereby decreasing levels of NO ⅐ and ONOO Ϫ . In preliminary experiments, we found that catalase had no effect on assays of NO x formation from SIN-1 as detected with the Griess reagent. It is likely that Cu,Zn-SOD and Mn-SOD are increasing dismutation of O 2 . produced from SIN-1 to H 2 O 2 in the culture medium rather than in the HepG2 cells themselves. For example, the conditioned medium containing SIN-1 plus Cu,Zn-SOD was cytotoxic when added later to HepG2 cells indicating formation of a stable toxic product, probably H 2 O 2 . This raises the question as to whether the SIN-1 plus SOD, i.e. H 2 O 2 -mediated, toxicity is occurring extra-or intracellularly.
The toxicity of H 2 O 2 produced by the SIN-1/SOD system appears to be intracellular and mediated by transition metalcatalyzed decomposition of H 2 O 2 at or near critical cellular sites, based upon the following results. 1) The presence of sodium azide, an inhibitor of intracellular catalase, increases cell killing by the SIN-1/SOD combination (Fig. 2); 2) depletion of intracellular glutathione by BSO also enhanced the observed toxicity of SIN-1/SOD (Fig. 5); 3) penetrable metal chelators such as deferoxamine and 2,2Ј-dipyridyl were very effective in reducing the SIN-1/SOD-mediated cytotoxicity, while EDTA and DTPA, which do not readily cross cellular membranes and hence act to chelate iron extracellularly (48), provided only a small protection. The inhibition of SIN-1/SOD toxicity by metal chelators would argue against a significant role for ONOO Ϫ or singlet oxygen, which may be produced from the interaction of H 2 O 2 with NO ⅐ (49) or H 2 O 2 with ONOO Ϫ (50) in the mechanism of SIN-1 toxicity. The possibility that the cytotoxicity was mediated by the reaction of NO ⅐ with H 2 O 2 is not likely since no reaction between H 2 O 2 and NO ⅐ was found at physiological pH (51). The requirement for H 2 O 2 and iron in the overall mechanism by which SIN-1/SOD causes toxicity to the HepG2 cells would suggest that ⅐ OH-like species formed via a Fenton-type reaction may be the oxidants responsible for the cytotoxicity. Indeed, thiourea and urate, potent ⅐ OH scavengers, were effective in protecting the HepG2 cells against SIN-1/SOD toxicity. However, other ⅐ OH scavengers such as Me 2 SO, formate, mannitol, benzoate, and ethanol did not afford any protection; Me 2 SO, ethanol, and formate are permeable agents and should be available to scavenge ⅐ OH that is accessible. The inability of ⅐ OH scavengers to provide protection against oxidant injury has sometimes been explained by site-specific damage produced when H 2 O 2 is decomposed by transition metals bound to critical cellular targets (52)(53). The ability of thiourea and urate to protect against the SIN-1/SOD cytotoxicity may reflect the ability of these agents to chelate metals (54) or react with H 2 O 2 (55,56).
The effectiveness of Cu,Zn-SOD to potentiate SIN-1 toxicity displayed an unusual SOD concentration dependence in that lower concentrations of SOD enhanced the SIN-1 toxicity to a greater extent than higher concentrations. The Mn-SOD did not display this biphasic response. These effects can probably be related to the levels of H 2 O 2 that accumulate in the reaction system, since H 2 O 2 plays a central role in the cytotoxicity of SIN-1/SOD. Highest levels of H 2 O 2 are found when SIN-1 is incubated with the lower concentration of the Cu,Zn-SOD or with low and high concentrations of the Mn-SOD. The lowest levels of H 2 O 2 are found when SIN-1 is incubated with the high concentration of Cu,Zn-SOD; these levels are even lower than those produced by SIN-1 alone, in the absence of SOD. These results suggest that high concentrations of the Cu,Zn-SOD, but not the Mn-SOD, can decompose H 2 O 2 . The ability of Cu,Zn-SOD but not Mn-SOD to decompose H 2 O 2 in a Fenton-type reaction was originally reported by Yim et al. (39) . This raises the question as to why ⅐ OH formed by the decomposition of H 2 O 2 by high concentrations of Cu,Zn-SOD, did not contribute to the cytotoxicity of SIN-1/SOD, i.e. less toxicity is observed under conditions in which ⅐ OH is being produced, and Mn-SOD which does not decompose H 2 O 2 to ⅐ OH was equally cytotoxic at concentrations of 100 units/ml and more cytotoxic at 1000 units/ml than Cu,Zn-SOD at the same concentrations. Most likely the ⅐ OH produced by this reaction is formed deep within the channel of the Cu,Zn-SOD (39) and reacts with the enzyme before escaping into the bulk solution to damage the cells. It must also be considered that although levels of H 2 O 2 found in the presence of SIN-1 plus the high concentration of Cu,Zn-SOD are lower than that produced by SIN-1 alone, toxicity is observed in the former but not the latter case (Fig. 1, B and C), which may actually be a reflection of sufficient concentrations of H 2 O 2 crossing the cellular membrane before decomposition by the extracellular Cu,Zn-SOD occurs or there is a contribution toward overall toxicity by ⅐ OH produced by the SIN-1 plus Cu,Zn-SOD combination. SOD is generally thought to play a protective role in biological systems against oxidant action. It is now becoming apparent that the net effect of SOD is complex and often depends on the fate of the generated H 2 O 2 and the availability of H 2 O 2 degradation enzymes such as catalase and glutathione peroxidase. SIN-1 is the active metabolite of molsidomine, an antianginal drug (59), which is used for the treatment of ischemia reperfusion injury (60) and human erectile dysfunction (61). It is also used for the production of O 2 . and NO ⅐ in a variety of experimental systems. The potentiation of SIN-1 cytotoxicity by SOD therefore requires caution in the use of this agent and in considering the therapeutic efficacy of SOD in systems that generate O 2 . and NO ⅐ . Our results indicate that the observed potentiation of the cytotoxicity of SIN-1 by SOD is linked to the formation of H 2 O 2 by SOD-catalyzed dismutation of O 2 . . The largely extracellularly produced H 2 O 2 then diffuses inside the cells, where it exerts its toxic effect via reaction with transition metals bound to critical cellular sites. The Cu,Zn-SOD did not significantly increase or decrease the cytotoxicity of SNAP or DEA/NO (sources of NO ⅐ but not O 2 . ), nor did catalase protect HepG2 cells against the cytotoxicity of SNAP or DEA/NO (Fig.  10), which suggests no significant role for NO ⅐ in the observed cytotoxic effects of SIN-1 at the concentrations utilized (e.g. 0.316 or 1 mM) and under these reaction conditions. A role for NO ⅐ in SIN-1 toxicity may occur under different conditions or concentrations, e.g. in the potent toxic effects of higher concentrations of SIN-1 (e.g. 3 mM), found in the absence of SOD (Fig. 1D).