Contractions of skeletal muscles produce increases in concentrations of superoxide anions and activity of hydroxyl radicals in the extracellular space. The sources of these reactive oxygen species are not clear. We tested the hypothesis that, after a demanding isometric contraction protocol, the major source of superoxide and hydroxyl radical activity in the extracellular space of muscles is mitochondrial generation of superoxide anions and that, with a reduction in MnSOD activity, concentration of superoxide anions in the extracellular space is unchanged but concentration of hydroxyl radicals is decreased. For gastrocnemius muscles from adult (6–8 mo old) wild-type (Sod2+/+) mice and knockout mice heterozygous for the MnSOD gene (Sod2+/-), concentrations of superoxide anions and hydroxyl radical activity were measured in the extracellular space by microdialysis. A 15-min protocol of 180 isometric contractions induced a rapid, equivalent increase in reduction of cytochrome c as an index of superoxide anion concentrations in the extracellular space of Sod2+/+ and Sod2+/- mice, whereas hydroxyl radical activity measured by formation of 2,3-dihydroxybenzoate from salicylate increased only in the extracellular space of muscles of Sod2+/+ mice. The lack of a difference in increase in superoxide anion concentration in the extracellular space of Sod2+/+ and Sod2+/- mice after the contraction protocol supported the hypothesis that superoxide anions were not directly derived from mitochondria. In contrast, the data obtained suggest that the increase in hydroxyl radical concentration in the extracellular space of muscles from wild-type mice after the contraction protocol most likely results from degradation of hydrogen peroxide generated by MnSOD activity.
- hydroxyl radicals
during contractions, skeletal muscle generates reactive oxygen (ROS) and reactive nitrogen species. Early studies demonstrated with electron spin resonance (ESR) techniques that skeletal muscles contained free radical species and muscle contractions increased the magnitude of these species (7, 21). Subsequent studies showed that contractions of rat diaphragm strips in vitro (37) and hindlimb muscles of mice in vivo (27) caused an increase in superoxide anions detected in the extracellular environment. Other work suggested that the activity of hydroxyl radicals was increased in the circulation of cats during and after contractions of the triceps surae muscle (32), and contractions of extensor digitorum longus muscles of rats in vitro resulted in a large increase in the NO concentration in the surrounding medium (2). After a variety of contraction protocols, increased concentrations of the products of the reactions of ROS with lipids, proteins, and DNA in tissues and blood cells of animals and humans have been reported (38, 39).
The sources of the superoxide anion and hydroxyl radicals generated during and after contractions are unclear, although NO appears to be generated by a type I (neuronal type) NO synthase (NOS) located in the plasma membrane (18, 35). In contracting skeletal muscle, mitochondria are considered to be the primary site of superoxide anion radical generation (7, 37). Muscle fibers contain two forms of superoxide dismutase that reduce superoxide toxicity (15). In the mitochondrion, the superoxide anion is dismutated to hydrogen peroxide by MnSOD (SOD2) and in the cytosol by CuZnSOD (SOD1). During contractile activity the increased oxygen consumption by the contracting skeletal muscles inevitably leads to an increase in superoxide anion radical production (25). Superoxide anions are charged and highly reactive, and simple diffusion from the mitochondria across both mitochondrial and plasma membranes to the extracellular space seems highly unlikely. There is evidence that superoxide can cross membranes in the protonated form (15), although this may be relatively insignificant at physiological pH (14). Additionally, superoxide may be released from mitochondria and pass through other membranes via anion channels (17, 33, 43), but whether this contributes to extracellular levels of superoxide is unknown. There are other potential, primarily nonmitochondrial, sites for superoxide generation including xanthine oxidase enzymes, prostanoid metabolism, and membrane-bound NAD(P)H oxidoreductases, and, although the importance of some of these other sites has been questioned (20, 27), skeletal muscle myotubes grown in culture contain relatively few mitochondria, yet when electrically stimulated, the myotubes are a major source of extracellular superoxide generation (27).
O'Neill et al. (32) reported an increase in formation of ortho-, meta-, and para-tyrosines from phenylalanine in the venous outflow from the triceps surae muscle of cats during muscle contraction. They concluded that this was due to an increased extracellular hydroxyl radical activity and that the formation of this species was catalyzed by iron because the rise was preventable by treatment of animals with the iron chelator Desferal. The substrate for this iron-dependent reaction is likely to be hydrogen peroxide, normally formed by dismutation of superoxide anions and released from mitochondria during increased activity (15). In contrast to the superoxide anions, hydrogen peroxide can readily diffuse across both mitochondrial and plasma membranes (15).
The phenotype of homozygous knockout Sod2-/- is lethal, with mice surviving 1–18 days (19, 24). In contrast, knockout mice heterozygous for MnSOD (Sod2+/- mice) show an ∼50% reduction in MnSOD activity in all tissues yet survive without any obvious phenotypic defects (45). Adult Sod2+/- mice show no changes in the activity of other major antioxidant defense enzymes, although an increase in mitochondrial oxidative damage is evident as well as a decrease in mitochondrial function (45, 47).
We (27, 34) previously used microdialysis techniques to examine the superoxide anion content and hydroxyl radical activity in skeletal muscle extracellular space, and these techniques were used to compare extracellular superoxide anion and hydroxyl radical activity in the interstitial fluid of the gastrocnemius (GTN) muscles of wild-type mice (Sod2+/+) and knockout mice heterozygous for MnSOD (Sod2+/-) in vivo both with and without a period of repetitive isometric contractions. We tested the hypothesis that, after a demanding isometric contraction protocol, a major source of ROS in the extracellular space of muscles is the mitochondrial generation of superoxide anions and that with a reduction in MnSOD activity the concentration of superoxide anions in the extracellular space would be unchanged, because mitochondrial superoxide does not significantly influence extracellular superoxide levels, but the concentration of hydroxyl radicals would be decreased due to a reduction in the hydrogen peroxide released into the extracellular space from the mitochondria.
MATERIALS AND METHODS
The studies were carried out on adult (6–8 mo old) male C57BL/6 Sod2+/- and Sod2+/+ mice. At this age, body mass and force development of wild-type C57BL/6 mice have reached maturity and body and muscle mass have stabilized (3). The Sod2+/- mice were characterized by Williams et al. (47) and Van Remmen et al. (45). The superoxide anion content and hydroxyl radical activity in the interstitial fluid from quiescent muscles were obtained by sampling the extracellular space with microdialysis probes placed in the GTN muscles for 15-min periods before the contraction protocol. For muscle samples from quiescent mice, mice were euthanized by overdose of anesthesia and the GTN muscles were removed for biochemical analyses without undergoing the contraction protocol. These muscles were rapidly frozen in liquid nitrogen and stored at -70°C until being analyzed.
Contraction protocol. The adult mice were anesthetized with pentobarbital sodium, with an initial dose of 65 mg/100 g of body mass via an intraperitoneal injection. Supplemental doses were administered as required to maintain a depth of anesthesia sufficient to prevent response to tactile stimuli. The knee of one hindlimb was fixed to a base plate, and the hindlimb musculature was stimulated to contract by surface electrodes placed around the upper limb and the ankle to induce isometric contractions under in vivo conditions (27). Fiber length was set at the optimum length for force production. Maximum isometric tetanic contractions were produced by square wave pulses of 0.2-ms duration, a voltage slightly greater than that required to produce a maximum twitch (usually ∼70 V), and a frequency of 100 Hz (3). Maximum isometric contractions were held for 500 ms, with a contraction every 5 s for a total of 180 contractions during the 15 min of the contraction protocol. Mice remained under anesthesia until the end of the experiment and were then euthanized by overdose of pentobarbital sodium. The GTN muscles were rapidly dissected and prepared for histological examination as previously described (44).
Measurement of extracellular ROS. Microdialysis probes (MAB 3.8.10, Metalant, Stockholm, Sweden) with a molecular mass cutoff of 35 kDa were placed into the GTN muscles of both limbs of anesthetized animals with a 22-gauge plastic introducer. Probes were perfused with either 20 mM salicylate (31) in normal saline or 50 μM cytochrome c (27) in normal saline at a flow rate of 4 μl/min and allowed to stabilize for 30 min. Samples were collected from the probes over sequential 15-min periods. The microdialysis probe consists of a 10-mm tubular dialysis membrane that has a diameter of 0.5 mm with concentric inlet and outlet tubes. Perfusion of the probes occurs at a very slow rate such that small molecules reach equilibrium in the perfusate and extracellular space and are then collected via the outlet tube. The specificity of the analyses of ROS were examined by pretreatment of Sod2+/+ mice with either NG-nitro-l-arginine methyl ester (l-NAME; 50 mg/kg body wt given by intraperitoneal injection 30 min before study) as an inhibitor of NOS, or desferrioxamine mesylate (Desferal; 10 mg/kg body wt given by intravenous injection 30 min before study) as an iron chelator.
Hydroxyl and superoxide radical analyses. 2,3-Dihydroxybenzoic acid (2,3-DHB) generated from hydroxyl radical reaction with salicylate in the microdialysis fluids was measured as an index of hydroxyl radical activity by HPLC with electrochemical detection as previously described (34). Reduction of cytochrome c in the microdialysate was used as an index of superoxide anion radical concentration in microdialysates as previously described (27).
Muscle total SOD activity and MnSOD content. Total muscle SOD activity was measured according to the method of Crapo et al. (6). The MnSOD content of muscles was analyzed by Western blotting with an antibody against MnSOD (Stressgen, Victoria, BC, Canada). Briefly, the frozen muscle was ground under liquid nitrogen and the powdered sample was aliquoted. Powdered muscle was homogenized in a range of protease inhibitors, samples were then centrifuged at 4°C, and the supernatant was analyzed for total protein content. Fifty micrograms of total protein were separated on SDS-PAGE followed by Western blotting.
Muscle catalase activity. Muscle homogenates were analyzed for catalase activity by following the kinetic decomposition of hydrogen peroxide spectrophotometrically at 240 nm with the method described by Claiborne (5).
Total glutathione and protein thiol content. The automated glutathione recycling method described by Anderson (1) was used to assess the total glutathione content of samples with a 96-well plate reader (Benchmark, Bio-Rad). The protein thiol content of samples was analyzed by the method of Di Monte et al. (10), adapted for use on a 96-well plate reader.
Malondialdehyde content. The malondialdehyde (MDA) content of the sample was measured as an index of lipid peroxidation with the HPLC-based method of Chirico (4).
Protein determination. The protein content of samples was determined with use of a bicinchoninic acid (BCA) protein assay kit (Sigma, Poole, UK) based on the method of Smith et al. (40).
Statistical analyses. Data are presented as means ± SE of values from four to six animals for each experimental condition. Data were initially analyzed by ANOVA for repeated measures. Where significance was indicated, means were compared with the Bonferroni correction.
Comparison of quiescent GTN muscles from Sod2+/- mice with those from Sod2+/+ mice. For quiescent GTN muscles from Sod2+/- mice compared with those from Sod2+/+ mice, the reduction in the MnSOD protein content was ∼65% (Fig. 1, Table 1). The reduction in MnSOD produced a reduction in total muscle SOD activity. No significant changes in muscle catalase activity were seen, although the values tended to be higher in the Sod2+/- than the Sod2+/+ mice (Table 1). The GTN muscles of Sod2+/- compared with Sod2+/+ mice showed no differences in glutathione, total protein thiol, or MDA content (Table 1).
Effect of 180 isometric contractions on generation of force. For both Sod2+/+ and Sod2+/- mice, the metabolically demanding protocol of 180 isometric contractions induced a loss of ∼50% of the initial maximum force developed by the GTN muscle within the first 3 min of commencing the repetitive contractions (Fig. 2). The loss of force production was sustained throughout the remaining 12 min of the contractile protocol. No difference in force production was observed between the Sod2+/+ and Sod2+/- mice at any time during the 15-min contractile protocol. Furthermore, no evidence of any gross changes in muscle fiber structure was observed for histological cross sections of GTN muscles from Sod2+/- or Sod2+/+ mice stained with hematoxylin and eosin (Fig. 3).
Effect of contraction protocol on reduction of cytochrome c in extracellular space. The microdialysis probes in the quiescent GTN muscles of both Sod2+/- and Sod2+/+ mice detected ∼0.5 nmol of superoxide/15 min (Fig. 4). Data are presented as superoxide equivalents, and no significant differences were found between the superoxide anion contents in the extracellular space of GTN muscles of Sod2+/- and Sod2+/+ mice at specific time points. ANOVA revealed that overall the Sod2+/+ mice tended to have higher values throughout (P = 0.071). Both groups of mice showed a significant increase in reduction of cytochrome c during the contraction protocol compared with the 15-min period immediately before the contractions. Pretreatment of Sod2+/+ mice with the NOS inhibitor l-NAME resulted in a greater reduction in cytochrome c in the microdialysates from the GTN muscle during the four 15-min collections before contractions. Values were 166 ± 15% of those in untreated mice. During the contractile activity, cytochrome c reduction in l-NAME-treated animals was unchanged at 115 ± 16% of that observed from the contracting muscle of the untreated group. In contrast, desferrioxamine mesylate treatment had no significant effect on the cytochrome c reduction, which was 109 ± 9% of the untreated value at rest and 101 ± 15% of that from the untreated group during contractile activity.
Effect of contraction protocol on formation of 2,3-DHB in extracellular space. 2,3-DHB was detected in all microdialysis fluids from GTN muscles. In both Sod2+/- and Sod2+/+ mice, the initial rate of generation of 2,3-DHB was ∼160 pmol/15 min (Fig. 5). During the period of contractile activity the 2,3-DHB content increased significantly in microdialysates from the Sod2+/+ mice, an increase not observed in the GTN muscles of the Sod2+/- mice, in which 2,3-DHB levels did not change significantly throughout the contraction protocol. Pretreatment of Sod2+/+ mice with the NOS inhibitor l-NAME did not significantly influence the 2,3-DHB generation in microdialysates from the GTN muscle during the four 15-min collections before contractions. Values were 98 ± 6% of those from untreated mice. During the contractile activity, the 2,3-DHB formation in muscle microdialysates from l-NAME-treated animals did not differ from that in untreated contracting muscle, although both l-NAME-treated and untreated mice showed a 50–70% increase compared with resting values. In contrast, desferrioxamine mesylate treatment significantly reduced 2,3-DHB generation to 28 ± 7% of the generation in untreated mice at rest and to 54 ± 12% of that in the untreated group during contractile activity.
In all tissues, including skeletal muscles, studied by us and by others (45–47), Sod2+/- mice displayed an ∼50% reduction in MnSOD activity, as well as a reduction in total skeletal muscle SOD activity. Despite these deficits in SOD activity, Sod2+/- mice survive without any visible morphological defects with as long a life span as Sod2+/+ mice (Van Remmen H, unpublished observations). Van Remmen and colleagues (45) reported that the decrease in MnSOD activity increased the ROS load within the mitochondria, leading to increased ROS damage to mitochondrial DNA and to mitochondrial proteins that caused a decrease in mitochondrial function. The Sod2+/- mice examined in the current study showed a similar 50% decrease in MnSOD content, resulting in a decrease in total muscle SOD activity.
Both the morphological appearance and contractile characteristics are consistent with the conclusion that the heterozygous knockout of MnSOD had no major effect on skeletal muscle integrity or function. Reid and coworkers (36, 37) and Lawler and Hu (23) reported that ROS decreased muscle force production by diaphragm muscles administered isometric contraction protocols in vitro. In the present study, the force response expressed as a percentage of the maximum force of Sod2+/- compared with Sod2+/+ mice followed a similar pattern in response to the metabolically demanding 180-isometric contraction protocol. No differences were observed between the two groups in the pattern or magnitude of force loss, and the heterozygous knockout of MnSOD did not have any impact on force production before, during, or after the contraction protocol.
Microdialysis techniques were used in the current study to assess the superoxide content and hydroxyl radical activity in skeletal muscle interstitial space. These techniques were used previously in studies of ROS in brain interstitial fluid (31) as well as skeletal muscle interstitial fluid in the mouse (27) and rat (34). We used the formation of 2,3-DHB from salicylate in the microdialysates as a measure of hydroxyl radical activity as previously reported (11, 31, 34). Formation of 2,3-DHB from salicylate in biological systems has been claimed to occur specifically through hydroxyl radical-mediated hydroxylation (13) in a similar manner to formation of ortho-, meta-, and para-tyrosines from phenylalanine (41). In vitro studies have indicated that reaction with peroxynitrite also leads to hydroxylation of salicylate and phenylalanine (16, 30), and it is currently unclear whether these reactions involve intermediary generation of hydroxyl radicals (16, 22, 26, 30). O'Neill and colleagues (32) reported that phenylalanine hydroxylation in muscle extracellular fluid was prevented by use of the iron chelator desferrioxamine mesylate, and our data confirm that a similar inhibition occurred in 2,3-DHB generation from salicylate in microdialysates from skeletal muscle. Desferrioxamine can reduce hydroxyl radical formation by iron chelation, but it has also been reported that this substance can scavenge peroxynitrite (9). To clarify the particular species involved in 2,3-DHB formation from salicylate in muscle microdialysates, Sod2+/+ mice were treated with l-NAME before study. No significant effect of this treatment on 2,3-DHB generation was seen in control Sod2+/+ mice at rest, and l-NAME did not influence the increase in 2,3-DHB formation seen during contractile activity. Thus, although peroxynitrite can induce formation of 2,3-DHB from salicylate in some systems, our data are more compatible with the major role being played by hydroxyl radicals in muscle microdialysates.
Monitoring the reduction of cytochrome c is acknowledged to lack specificity as a measure of the superoxide content of biological fluids (29), and in a previous publication we reported (27) that levels of cytochrome c reduction in mouse muscle microdialysates were reduced by 50% on addition of purified SOD. We now consider that that at least a portion of the reduction of cytochrome c occurs outside the microdialysis probe. Cytochrome c has a molecular mass of ∼12 kDa, and the dialysis membrane cut-off is 35 kDa; hence there will be substantial diffusion of the cytochrome c out of the probe, and because of the slow flow rate there will also be diffusion back into the probe. In this case, addition of high-molecular-mass purified SODs to the microdialysis fluid could not prevent the reduction of cytochrome c by superoxide that occurs outside the probe. Murrant and Reid (29) also report that superoxide and nitric oxide are the major reductants of cytochrome c in muscle extracellular space. To determine the likely contribution of NO to cytochrome c reduction in muscle microdialysates, Sod2+/+ mice were treated with l-NAME before initiation of the experiment. The reduction of cytochrome c was significantly increased in the treated animals compared with untreated animals, indicating that NO was unlikely to make a significant contribution to the reduction of cytochrome c in this system. The increase in cytochrome c reduction seen in the l-NAME-treated mice may be attributed to an increased bioavailability of superoxide in the absence of NO generated by NOS enzymes or to a reduction in peroxynitrite-mediated oxidation of cytochrome c as previously suggested (42). It has also been suggested that small molecules such as ascorbate or glutathione can reduce cytochrome c in vivo, but in unpublished studies we have observed that reduction of microdialysate cytochrome c stopped immediately on death of the mouse, suggesting a dependence of this reduction on metabolic activity that is not compatible with this hypothesis. Overall, therefore, the data are consistent with superoxide playing a substantial role in the reduction of cytochrome c in microdialysates, but the lack of complete suppression of the reduction by exogenous SOD means that we cannot define precise levels from the current data.
For contracting skeletal muscle, mitochondria are considered to be the major source of superoxide generation (7). During contractile activity the increased oxygen consumption by the active skeletal muscle leads to an increase in superoxide anion production due to electron delocalization from the electron transport chain (25). This is an attractive hypothesis, but it does not readily account for the increased extracellular release of superoxide anions reported from contracting muscles (27, 37) or myotubes (27). Although simple diffusion of superoxide from the mitochondria across both mitochondrial and plasma membranes to the extracellular space seems unlikely, the protonated form may cross membranes, and there is evidence that superoxide does cross mitochondrial and potentially other membranes through anion channels (17, 33, 43). However, the present data support the hypothesis that mitochondrial superoxide does not contribute to the superoxide anion content detected in the extracellular space of skeletal muscle. Overall, the superoxide release from Sod2+/- mice was lower than from Sod2+/+ mice, although values at each specific time point were not significantly different. Furthermore, the increase in superoxide anion content during contractions was equivalent for both groups of mice despite the greater rise in mitochondrial superoxide anions that must occur in the Sod2+/- compared with Sod2+/+ mice.
A number of alternative hypotheses can be formulated regarding the source of the superoxide anions released into extracellular space of contracting muscle fibers. One potential mechanism for the generation of extracellular superoxide anions is the plasma membrane oxidoreductase postulated by de Grey (8) and partially identified by Morré et al. (28). In their model, intracellular NADH drives one-electron reduction of external molecular oxygen to produce superoxide anions via a plasma membrane oxidoreductase. This NADH-driven superoxide anion generation has been proposed to be particularly important to augment survival of muscle fibers with defective mitochondria in aging mammals, allowing the fibers to utilize exclusively glycolytic metabolism. Alternatively, this mechanism of superoxide anion generation could play a role to maintain NADH/NAD+ homeostasis in skeletal muscle during the early, relatively anaerobic stages of demanding contraction protocols.
Microdialysates from Sod2+/- and Sod2+/+ mice contained similar amounts of 2,3-DHB at rest, but the Sod2+/+ mice showed a rise during and immediately after contractions, whereas no rise was observed for the Sod2+/- mice. It was previously concluded that the generation of 2,3-DHB in skeletal muscle microdialysates was primarily due to hydroxyl radical activity and that the rise in extracellular hydroxyl radical activity that accompanies contractile activity of skeletal muscles was catalyzed by iron and preventable by treatment with the iron chelator desferrioxamine (32). Apparently, the hydroxyl radicals were generated in the extracellular space from hydrogen peroxide released from the contracting muscle fibers. Furthermore, the lack of an increase in hydroxyl radical activity in the interstitial fluid of muscles of Sod2+/- mice after the contraction protocol implies that the substrate for this iron-dependent reaction is likely to be hydrogen peroxide. Normally, hydrogen peroxide diffuses readily from mitochondria (15). Our data support the possibility that a reduction in the release of hydrogen peroxide from contracting muscle cells occurs in Sod2+/- mice. There are various possible mechanisms by which this may occur. We favor the hypothesis that mitochondrial generation of hydrogen peroxide is reduced because of a reduction in MnSOD activity in the Sod2+/- mice. Alternatively, the actual amount of MnSOD present may influence superoxide generation from the electron transport chain. Both Forman and Azzi (12) and Reid and coworkers (37) have argued that the apparent rate of superoxide produced and/or detected may be increased by the presence of increased amounts of agents or scavengers that trap superoxide, potentially inducing a net increase in generation or flux of superoxide through a mass effect. In this case it is possible that the amount of MnSOD present might influence the rate of superoxide generation in wild-type compared with heterozygous MnSOD knockout mice.
It is also possible that relatively small changes in the activity of hydrogen peroxide-metabolizing enzymes or glutathione content may have influenced the amount of hydrogen peroxide that crosses the plasma membrane in Sod2+/- mice. The mean muscle catalase activities were approximately doubled in Sod2+/- mice, and the mean muscle glutathione content increased by ∼50% (Table 1), although these changes were not statistically significant. It is also possible, but less likely, that a change in the “free” iron available in the extracellular space had occurred in Sod2+/- mice, and such changes would also modify the generation of hydroxyl radicals.
In conclusion, the comparative studies of Sod2+/- and Sod2+/+ mice support the hypothesis that the contraction-induced increase of hydroxyl radical activity in the extracellular space of skeletal muscles is derived from the iron-catalyzed degradation of hydrogen peroxide generated by MnSOD activity in the mitochondria. In contrast, the increased superoxide anion radical concentration detected in the extracellular space of muscle fibers after a demanding contraction protocol is unaffected in Sod2+/- compared with Sod2+/+ mice, and thus these data do not support the mitochondria as a major contributor to superoxide anion levels in muscle extracellular space.
Expert technical assistance was provided by Tatania Kostrominova, Cheryl Hassett, and Carol Davis.
Present address of T. T. Huang: Department of Neurology and Neurological Sciences, Stanford University and Geriatric Research, Education and Clinical Center, Palo Alto Department of Veterans Affairs Health System, 3801 Miranda Ave., Palo Alto, CA 94304.
We thank the US National Institute on Aging (Grants AG-20591 and AG-16998) for financial support. A. McArdle and M. J. Jackson also thank the Wellcome Trust and Research into Ageing for additional financial support.
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 © 2004 the American Physiological Society