Mitochondrial function and antioxidative defence in human muscle: effects of endurance training and oxidative stress

Abstract

1. The influence of endurance training on oxidative phosphorylation and the susceptibility of mitochondrial oxidative function to reactive oxygen species (ROS) was investigated in skeletal muscle of four men and four women. Mitochondria were isolated from muscle biopsies taken before and after 6 weeks of endurance training. Mitochondrial respiration was measured before and after exposure of mitochondria to exogenous ROS (H₂O₂+ FeCl₂).

2. Endurance training increased peak pulmonary O₂ uptake (V̇_O2,peak) by 24 % and maximal ADP-stimulated mitochondrial oxygen consumption (state 3) by 40 % (P < 0.05). Respiration in the absence of ADP (state 4), the respiratory control ratio (RCR = state 3/state 4) and the ratio between added ADP and consumed oxygen (P/O) remained unchanged by the training programme.

3. Exposure to ROS reduced state 3 respiration but the effect was not significantly different between pre- and post-training samples. State 4 oxygen consumption increased after exposure to ROS both before (+189 %, P < 0.05) and after training (+243 %, P < 0.05) and the effect was significantly higher after training (P < 0.05, pre- vs. post-training). The augmented state 4 respiration could in part be attenuated by atractyloside, which indicates that ADP/ATP translocase was affected by ROS. The P/O ratio in ROS-treated mitochondria was significantly lower (P < 0.05) compared to control conditions, both before (−18.6 ± 2.2 %) and after training (−18.5 ± 1.1 %).

4. Muscle activities of superoxide dismutase (mitochondrial and cytosolic), glutathione peroxidase and muscle glutathione status were unaffected by training. There was a positive correlation between muscle superoxide dismutase activity and age (r= 0.75; P < 0.05; range of age 20–37 years), which may reflect an adaptation to increased generation of ROS in senescent muscle. The muscle glutathione pool was more reduced in subjects with high activity of glutathione peroxidase (r= 0.81; P < 0.05).

5. The influence of short-term training on mitochondrial oxygen consumption has for the first time been investigated in human skeletal muscle. The results showed that maximal mitochondrial oxidative power is increased after endurance training but that the efficiency of energy transfer (P/O ratio) remained unchanged. Antioxidative defence was unchanged after training when expressed relative to muscle weight. Although this corresponds to a reduced antioxidant protection per individual mitochondrion, the sensitivity of aerobic energy transfer to ROS was unchanged. However, the augmented ROS-induced non-coupled respiration after training indicates an increased susceptibility of mitochondrial membrane proton conductance to oxidative stress.

An organism’s or a particular tissue’s capacity for aerobic energy production is generally adapted to the energy requirement. Physical exercise increases skeletal muscle energy demand manyfold. Regularly performed exercise elicits a rapid upregulation of human muscle capacity for oxidative energy production, as indicated by measurements of mitochondrial density (Turner et al. 1997) and oxidative enzyme activities (Morgan et al. 1971). Recently, by exploiting the high sensitivity of bioluminescence, the mitochondrial ATP production rate in isolated mitochondria (MAPR) was shown to increase after endurance training (Wibom et al. 1992; Berthon et al. 1995; Starritt et al. 1999). However, the information that can be obtained from MAPR is limited to ATP production. Measurement of mitochondrial respiration can provide additional information of both quantitative (maximal oxidative power) and qualitative (the efficiency of oxidative energy transfer (P/O ratio) and the degree of respiratory control) aspects of mitochondrial function. The effect of training on mitochondrial respiratory function in human skeletal muscle has not been investigated yet and the question of whether the efficiency of mitochondrial energy production is altered by endurance training remains unanswered.

Energy production via the electron transport chain with oxygen as the terminal electron acceptor, while in many aspects superior to the anaerobic pathways, is associated with the production of highly reactive univalent oxygen species and their derivatives (ROS) (Boveris & Chance, 1973; Sjodin et al. 1996; Leeuwenburgh et al. 1999). Flavins, coenzyme Q (CoQ), cytochromes b and non-haem iron proteins in the early part of the mitochondrial electron transport chain are the main targets for single-electron oxidation by O₂ in which superoxide radicals (O₂⁻.) are produced (Halliwell, 1994). Superoxide radicals can be further reduced through the dismutation reaction, which generates hydrogen peroxide (H₂O₂) (Halliwell, 1994). The generation of superoxide radicals and hydrogen peroxide by animal mitochondria has been demonstrated by a variety of techniques and is a well established physiological phenomenon (Boveris & Chance, 1973; Halliwell, 1994). Hydrogen peroxide can be non-enzymatically reduced to highly reactive hydroxyl radicals (OH.) via either the Fenton reaction or the Haber-Weiss reaction (Sjodin et al. 1996). Recently, direct evidence that heart muscle mitochondria produce hydroxyl radicals in vivo was provided by using isotope dilution gas chromatography mass spectrometry (Leeuwenburgh et al. 1999). It seems likely that in cells containing a large volume of mitochondria, such as myocytes, electron transport chain activity is the major source of ROS. While the reactivity of different ROS varies, some of them, such as hydroxyl radicals, may react with almost all molecules in living cells, interfering with a number of cellular processes, including mitochondrial biogenesis (Chandwaney et al. 1998) and energy metabolism (Chandwaney et al. 1998; Miranda et al. 1999). Due to the high reactivity of ROS, cellular structures located in the vicinity of free-radical production sites become primary targets for ROS attacks. Therefore, mitochondrial DNA and lipids and proteins in the mitochondrial membrane are, presumably, the main targets for ROS in eukaryotic cells. Recent observations by Leeuwenburgh et al. (1999) that the mitochondrial but not the cytosolic proteins of heart muscle were oxidised by hydroxyl radicals during exercise in rats support this view.

A training-induced augmentation of aerobic power in skeletal muscle is associated with an increased density of the electron transport chain and, consequently, greater potential for free-radical generation. Myocytes can increase their oxidative power and thus their capacity to produce ROS manyfold. An increased ROS production of this magnitude may seriously distort the cellular homeostasis if it is not matched by a proportional increase in the capacity of the cellular antioxidative defence system. Different scavenger systems including glutathione and the enzymes of glutathione and hydroperoxide metabolism are capable of converting ROS to more stable species and play a key role in this elaborate system.

Our knowledge about adaptation of the muscle antioxidative defence system to endurance training in humans is limited. The effect of endurance training on antioxidative defence in human muscle has only been investigated in one study (Tiidus et al. 1996). The results of this study demonstrated that 8 weeks of aerobic training increased whole-body pulmonary O₂ uptake (V̇_O2,max) and the capacity for flux through the citric acid cycle in skeletal muscle without altering muscle antioxidative enzymes activities or glutathione status (Tiidus et al. 1996). In addition, the relationship between aerobic training status and muscle antioxidative capacity was investigated in two cross-sectional studies. However, the results of these studies are conflicting. Jenkins and co-workers (1984) found a linear relationship between muscle superoxide dismutase (SOD) activity and training status expressed as V̇_O2,max. In contrast, Hellsten et al. (1996) found no significant relationship between the activities of SOD, glutathione peroxidase (GPX) or glutathione reductase and the activities of mitochondrial and other metabolic enzymes in human skeletal muscle. Since mitochondria are both the main source and the main target for ROS in skeletal muscle, the effect of endurance training on specific mitochondrial antioxidative defence systems (e.g. mitochondrial superoxide dismutase) is of particular interest. The specific adaptation of the ROS-scavenger systems of mitochondrial origin to endurance exercise training has not been investigated in humans.

Although the response of muscle antioxidative defence mechanisms to training differs widely between studies involving animals, the general consensus is that the training-induced increase in antioxidant status in animal muscle is lower than the relative increase in oxidative enzymes activities (Powers et al. 1999). A mismatch between oxidative and antioxidative potential would result in a decreased ratio of antioxidants to mitochondial mass. The physiological consequence would be increased susceptibility of mitochondrial function to oxidative stress. The influence of training on the susceptibility of integrated oxidative function to ROS has not been studied in human muscle mitochondria. A recent study in rat cardiac muscle demonstrated that the inhibition of mitochondrial respiration by exogenous ROS was more pronounced in mitochondria from trained animals compared to untrained controls (Leichtweis et al. 1997). However, in another study in rats, muscle mitochondria from trained animals showed a greater resistance to the ROS-induced inhibition of respiratory function than those from untrained rats (Chandwaney et al. 1998).

The purpose of the present study was to investigate the influence of endurance training on: (i) quantitative and qualitative aspects of mitochondrial oxidative function, (ii) the adaptation of human muscle antioxidative defence systems, and (iii) the susceptibility of human skeletal muscle mitochondria to oxidative stress and to elucidate its mechanism(s).

METHODS

Subjects

Eight motivated healthy volunteers (four females and four males) participated in the study. The subjects’ mean (range) age, weight and height were, respectively, 26 (20–37) years, 70.4 (56.5–98) kg and 1.76 (1.65–1.87) m. The subjects had not done any regular physical training for the 6 months preceding the study, although they were recreationally active. The subjects’ peak pulmonary oxygen uptake (V̇_O2,peak) was 2.7 (2.0–3.6) l min⁻¹ corresponding to 38.5 (27.8–46.9) ml min⁻¹ (kg body mass)⁻¹. The subjects were informed concerning the procedure and risks involved in the experiment before giving their written consent. The experimental design of the study was approved by the Ethics Committee of the Karolinska Institutet, Stockholm, Sweden and conformed to the Declaration of Helsinki.

Performance tests

Incremental ergometer tests for estimation of V̇_O2,peak and lactate threshold (LT) were conducted before and 2 days after the 6 week training period. For the female subjects the time of the tests was matched to their menstrual cycle. Subjects cycled (Monark 829 E cycle ergometer, Varberg, Sweden) at three or more submaximal work rates (4 min at each work rate and with a power step of 25 W) and (after about 5 min of low intensity cycling) at a supramaximal work rate until exhaustion. Expired gases were collected in Douglas bags and analysed for oxygen and carbon dioxide concentrations using a Beckman S-3A and LB-2 analyser, respectively (Beckman Instruments, Fullerton, CA, USA). Heart rate during the tests was recorded continuously (Polar Sport Tester 3000, Polar Electro, Kempele, Finland). V̇_O2,peak was defined as the highest V̇_O2 recorded during the test. Blood was sampled from a finger capillary at the end of each work period and analysed for lactate using YSI 2300 STAT lactate analyser (YSI, Yellow Springs, OH, USA). LT was defined as the interpolated workload corresponding to a blood lactate concentration of 4 mmol l⁻¹.

Endurance training

Subjects attended four training sessions per week over a period of 6 weeks. Training sessions consisted of 30 min of cycle exercise (80 r.p.m.) at a constant workload estimated to correspond to 70 % of V̇_O2,peak, followed by five 2 min bouts of exercise at 100 % of V̇_O2,peak, interspersed with 4 min periods of exercise at 50 % of V̇_O2,peak. Each training session was preceded by a standardised warm-up, which consisted of 5 min cycling at 50 W. The workloads were increased every second week, assuming an increase in V̇_O2,peak of 2.5 % per week. Heart rate was monitored continuously during each training session.

Muscle biopsy sampling

Muscle biopsies were taken from the lateral aspect of the quadriceps femoris muscle 48 h following the performance tests. After local skin anaesthesia (1–2 ml of 20 mg ml⁻¹ Carbocain (mepivacaine); Astra, Södertälje, Sweden) incisions were made (one on each leg) through the subcutaneous tissue and fascia at about one-third of the distance from the upper margin of the patella to the anterior superior iliac spine. Biopsies were taken at a depth of 2–3 cm using a Bergström needle with suction. Biopsies from both legs were combined. Each muscle sample was divided into two portions. One portion (20–25 mg) was quenched in liquid nitrogen and stored at −70°C until determination of muscle enzyme activities and glutathione content. The other portion (72–202 mg) was used for preparation of isolated mitochondria, as previously described (Tonkonogi & Sahlin, 1997). Briefly, the muscle specimen was minced with scissors and mitochondria were isolated by proteinase treatment (Nagarse EC 3.4.21.62), followed by homogenisation and subsequent differential centrifugation. The final mitochondrial pellet was resuspended (0.4 μl mg⁻¹ initial muscle) in a medium consisting of (in mmol l⁻¹): 225 mannitol, 75 sucrose, 10 Tris, 0.1 EDTA, pH 7.40. An aliquot of the suspension (10 μl) was taken for measurements of mitochondrial enzyme activities.

Capillary blood was sampled in parallel with the muscle biopsies and analysed for haemoglobin concentration and haematocrit using standard techniques.

Measurement of mitochondrial respiratory activity

The respiration rates of isolated mitochondria were measured with a Clark-type electrode (Hansatech DW 1, Norfolk, UK) in a water jacketed glass chamber of 0.3 ml capacity. A temperature of 25°C was maintained in the chamber. The measurements were carried out in a reaction medium containing (in mmol l⁻¹): 225 mannitol, 75 sucrose, 10 Tris, 10 KCl, 10 K₂HPO₄, 0.1 EDTA, 5 pyruvate, and 2 malate, pH 7.40. The solubility of oxygen in the medium was considered to be equal to 237.5 μmol l⁻¹.

The mitochondrial suspension was added to the reaction medium and state 3 respiration was initiated by the addition of ADP (final concentration 270 μmol l⁻¹). When all of the ADP added had been phosphorylated to ATP, the respiratory rate returned to state 4. The respiratory control ratio (RCR) was calculated as the ratio of the respiratory rate in state 3 to that in state 4. The ratio between phosphorylated ADP added and oxygen consumed (P/O ratio) was determined by the method of Chance & Williams (1955). By measuring the citrate synthase (CS) activity both in isolated mitochondria and in whole-muscle homogenate, it was possible to estimate the percentage of mitochondria freed from the muscle and calculate mitochondrial oxygen consumption per weight of muscle.

To study the susceptibility of muscle mitochondria to ROS, an aliquot of mitochondrial suspension was incubated in the presence of 100 μmol l⁻¹ H₂O₂ and 1 mmol l⁻¹ FeCl₂ for 30 min at 0°C. This treatment exposes mitochondria to both hydrogen peroxide and hydroxyl radicals (OH.) generated by the Fenton reaction. Control mitochondria in these experiments were incubated for the same period of time and at the same mitochondrial protein concentration and temperature but without ROS treatment. Oxidative phosphorylation in control and ROS-treated mitochondria was then evaluated as described above.

To study possible effects of ROS treatment on different mitochondrial membrane compounds, ROS-treated mitochondria respiring in state 4 were subjected to successive additions of atractyloside (50 μmol l⁻¹), GDP (1 mmol l⁻¹), oligomycin (1.5 μg ml⁻¹), and cyclosporin A (10 μmol l⁻¹), which inhibit, respectively,. ADP/ATP translocase (AAT), uncoupling proteins 2 and 3, F₁-ATPase, and the permeability transition pore. Addition of these inhibitors allowed the relative contribution of the proteins to proton leak through the mitochondrial membrane and thus to state 4 respiration to be evaluated. The concentrations of atractyloside and oligomycin required to completely inhibit AAT and F₁-ATPase were determined in a separate experiment by dose-response analysis performed with mitochondria isolated from rat (n= 6; male Sprague-Dawley rats killed by an overdose of pentobarbital sodium) and human (n= 2) skeletal muscle. The concentrations used were sufficient to completely inhibit AAT and F₁-ATPase, as indicated by a lack of increase in the respiratory rate after further addition of ADP to mitochondria respiring in state 4. At the concentrations used, none of the inhibitors exhibited any effects on state 4 oxygen consumption during control conditions.

Enzyme assays

Muscle activities of metabolic and antioxidant enzymes were determined in freeze-dried muscle samples. The specimens were dissected free from solid, non-muscle constituents, powdered and homogenised in ice-cold buffer (6.25 mg ml⁻¹) composed of Triton X-100 (0.05 % v/v), 50 mmol l⁻¹ KH₂PO₄ and 1 mmol l⁻¹ EDTA, pH 7.40. Enzyme activities in isolated mitochondria were measured after disruption of mitochondria by freeze-thawing and dilution (× 5) in the Triton solution described above.

The activity of CS (EC 4.1.3.7) in whole muscle and isolated mitochondria was determined as described by Alp et al. (1976). β-Hydroxyacyl-CoA dehydrogenase (HAD, EC 1.1.1.35) activity in muscle was measured by the method of Bass et al. (1969). The activity of 6-phosphofructokinase (PFK, EC 2.7.1.11) was assayed according to Opie & Newsholme (1967). Glutathione peroxidase (GPX, EC1.11.1.9) activity was measured by the method of Flohé & Günzler (1984), using H₂O₂ as substrate. Superoxide dismutase (SOD, EC 1.15.1.1) was assayed spectrophotometrically by monitoring the rate of acetylated cytochrome c reduction by superoxide radicals generated by the xanthine-xanthine oxidase system (Flohé & Otting, 1984). One activity unit of SOD is defined as that amount of enzyme which inhibits the rate of acetylated cytochrome c reduction by 50 %. To distinguish mangano-SOD (MnSOD), exclusively located in mitochondrial matrix, from cuprozinc-SOD (CuZnSOD), which is primarily located in the cytosol, the isolated mitochondria were assayed after incubation with 2 mmol l⁻¹ KCN for 30 min. At this concentration cyanide inhibits the CuZn isoform of the enzyme, but does not affect the Mn isoform (Geller & Winge, 1984). By using CS as a mitochondrial marker, it was possible to calculate MnSOD activity relative to the weight of muscle. All enzyme activity measurements were carried out at 30°C. Values were converted from dry weight to wet weight by assuming a water content of 77 %.

Glutathione assays

For determination of reduced (GSH) and total glutathione content freeze-dried muscle samples were trimmed free of visible blood and connective tissue and powdered. For GSH determination approximately 2 mg dry muscle was homogenised (150 μl mg⁻¹) in a solution consisting of 20 mmol l⁻¹N-ethylmorpholine, 10 % (v/v) acetonitrile and 50 mmol l⁻¹ monobromobimane (derivatisation agent), pH 8.0. The homogenate was incubated at room temperature for 10 min. The derivatisation reaction was then stopped by addition of sulphosalicylic acid to a final concentration of 10 %. The derivatised homogenate was centrifuged at 14 000 g, at 4°C for 15 min. The supernatant obtained was stored at −70°C until analysis of the glutathione thiol-bimane adducts by HPLC followed by fluorometric detection. The chromatographic conditions were as described by Luo et al. (1995). The preparation of freeze-dried samples and determination of muscle total concentration of the peptide pool of glutathione were performed as described by Luo et al. (1995) for frozen muscle specimens. The muscle concentration of oxidised glutathione (GSSG) was calculated as ([total glutathione]–[GSH])/2. All values were converted from dry weight to wet weight by assuming a water content of 77 %.

The determination of glutathione content has not been previously performed on freeze-dried muscle samples. The effect of freeze-drying was, therefore, evaluated in a preliminary experiment. Human muscle samples were obtained from four subjects, divided into two portions and frozen. One portion of each sample was freeze-dried prior to determination of glutathione content, whereas the second part was directly transferred to the homogenisation solution. The glutathione content in the samples was then assayed as described above. Neither the concentrations of GSH or total glutathione, nor the ratio of GSH to total glutathione were affected by the freeze-drying procedure. The freeze-drying of muscle samples has the advantage that the samples can easily be trimmed free from contaminating non-muscle compounds.

Data analysis

All values reported are means ±s.e.m.a Differences between means were tested for statistical significance by ANOVA with a repeated-measures design, followed by the Student-Newman-Keul’s post hoc test. Bivariate correlation coefficients were computed on the data and Fisher’s r to z test was used to determine their statistical significance. Significance was accepted at the 5 % level.

RESULTS

The 6 weeks of endurance training increased V̇_O2,peak by 24 % from 2.69 ± 0.21 to 3.34 ± 0.30 l min⁻¹ (P < 0.05). During incremental exercise, blood lactate concentration reached 4 mmol l⁻¹ (LT) at 121 ± 13 W before the training period and 185 ± 15 W (P < 0.05) after it. Blood haemoglobin concentration and haematocrit were 141 ± 2 g l⁻¹ and 41.4 ± 0.7 %, respectively, prior to training, and were not significantly changed after training.

Training significantly increased the activity of the mitochondrial enzymes. Muscle CS activity underwent a sharp increase from 21.1 ± 1.0 to 31.0 ± 1.5 mmol min⁻¹ (kg wet wt)⁻¹ (P < 0.05) in response to the training programme. The increase in CS activity was significantly correlated to LT (r= 0.76; P < 0.05). HAD activity increased by 19 %, from 8.45 ± 0.46 to 10.05 ± 0.63 mmol min⁻¹ (kg wet wt)⁻¹ (P < 0.05) in response to training. The training programme did not affect muscle activity of PFK, a marker of glycolytic activity (72.6 ± 4.3 and 74.2 ± 4.6 mmol min⁻¹ (kg wet wt)⁻¹ in pre- and post-training samples, respectively; n.s.).

The mitochondrial yields, calculated from the fraction of muscle CS recovered in the isolated mitochondria, were 18.8 ± 2.0 and 18.4 ± 1.4 % in muscle samples taken pre- and post-training, respectively (n.s.). Prior to training, state 3 respiration was 1.39 ± 0.10 mmol min⁻¹ (kg wet wt)⁻¹, which is similar to that previously measured in isolated mitochondria (Tonkonogi & Sahlin, 1997) and skinned fibres (Tonkonogi et al. 1998) from human vastus lateralis muscle. The training programme increased state 3 respiration by 40 % (Fig. 1A). However, the ratio between state 3 oxygen consumption and muscle CS activity was not changed. State 4 oxygen consumption of isolated mitochondria was not affected by training (Fig. 1B). RCR (state 3/state 4) was 8.7 ± 0.8 before training and 10.0 ± 0.4 afterwards (n.s). Prior to the training period, the P/O ratio of isolated mitochondria incubated under control conditions was 2.72 ± 0.05, which is similar to that previously observed in human skeletal muscle mitochondria oxidising pyruvate + malate (Tonkonogi & Sahlin, 1997). After the training period, the P/O ratio was 2.64 ± 0.03, not significantly different from pre-training.

Figure 1. Effect of endurance training on mitochondrial oxygen consumption rates.

State 3 (A) and subsequent state 4 (B) oxygen consumption rates of isolated human skeletal muscle mitochondria before and after 6 weeks of endurance training. Values are means ±s.e.m.a, n= 8. *P < 0.05 post- vs. pre-training. †P < 0.05 ROS-treated vs. control.

State 3 respiration was lower after ROS treatment (vs. control), but the difference was significant only after training (−9.1 %vs. control, P < 0.05; Fig. 1A). The relative change in state 3 respiration induced by exposure to ROS was not significantly different in post- and pre-training periods. State 4 respiration showed a large increase in response to ROS treatment both before and after the training programme (Fig. 1B). The increase in state 4 respiration induced by ROS treatment was significantly higher in post- than in pre-training samples (243 vs. 189 %; P < 0.05). Addition of 50 μmol l⁻¹ atractyloside, a specific inhibitor of AAT, reduced state 4 respiration in ROS-treated mitochondria by 39.9 ± 1.6 (P < 0.05) and 36.9 ± 1.6 % (P < 0.05) before and after training, respectively. Consecutive additions of GDP, oligomycin and cyclosporin A had no significant effect on the mitochondrial oxygen consumption rate.

Exposure to ROS significantly decreased the P/O ratio of mitochondria isolated from pre- and post-training muscle samples by 18.6 ± 2.2 and 18.5 ± 1.1 %, respectively. RCR was markedly reduced by ROS treatment both before and after training (32.1 ± 2.1 and 27.0 ± 1.6 % of control values, respectively, P < 0.05). The decrease in RCR was mainly caused by the large increase in state 4 respiration following ROS treatment.

Neither skeletal muscle SOD nor GPX activities, expressed relative to muscle weight, were affected by endurance training (Table 1). When related to muscle CS activity (a marker of mitochondrial mass), the activity of antioxidative enzymes decreased significantly after training (Fig. 2). There was a significant positive correlation between GPX activity and total muscle SOD activity (r= 0.78; P < 0.05) and the activity of the cytosolic isoform (CuZnSOD, r= 0.77; P < 0.05), but not between GPX activity and the activity of the mitochondrial isoform (MnSOD). The total muscle SOD activity showed a significant correlation to the age of the subjects (Fig. 3).

Table 1.

Antioxidant enzymes activities and glutathione status in human skeletal muscle before and after 6 weeks of endurance training

	Pretraining	Posttraining	Significance
SOD
Mitochondrial	373 ± 28	386 ± 33	n.s.
Cytosolic	1003 ± 23	910 ± 42	n.s.
Total	1376 ± 16	1296 ± 62	n.s.
GPX	964 ± 46	1003 ± 84	n.s.
GSH	1576 ± 146	1352 ± 95	n.s.
GSSG	97 ± 45	146 ± 38	n.s.
GSH+GSSG	1770 ± 113	1645 ± 71	n.s.

Values are means ±s.e.m. (30 °C); n= 8. SOD, superoxide dismutase (kU (kg wet wt)⁻¹); GPX, glutathione peroxidase (U (kg wet wt)⁻¹); GSH, reduced glutathione (μmol (kg wet wt)⁻¹); GSSG, glutathione disulfide (μmol (kg wet wt)⁻¹); GSH+2GSSG, total glutathione (μmol (kg wet wt)⁻¹).

Figure 2. Effect of endurance training on antioxidant enzyme activity.

Ratio between activities of antioxidative enzymes and citrate synthase (mitochondrial marker enzyme) in human skeletal muscle before and after 6 weeks of endurance training. Values are means ±s.e.m.a*P < 0.05 post- vs. pre-training.

Figure 3. Relationship between age and total superoxide dismutase activity in human skeletal muscle.

○, female subjects; •, male subjects. Values are means of pre- and post-training samples (y= 12.927x+ 998.51; r= 0.75; P < 0.05; n= 8).

Skeletal muscle contents of GSH, GSSG and total glutathione (Table 1) were similar to those previously observed in human skeletal muscle (Luo et al. 1995). Muscle glutathione status was unaffected by training. Linear regression analysis demonstrated a negative correlation between muscle GSSG concentration and GPX activity (r=−0.81; P < 0.05).

DISCUSSION

The results demonstrate that maximal muscle oxidative power increased by 40 % after the 6 week programme of endurance training. These data obtained by measurements of oxygen consumption in isolated mitochondria confirm earlier observations on mitochondrial ATP production measured using a bioluminometric technique, where 6 weeks of training resulted in a 43 % increase in MAPR with pyruvate + malate as substrates (Wibom et al. 1992) (calculated by us from their Table 2). An important novel finding in the present study was that the mitochondrial P/O ratio remained unchanged after 6 weeks of training. Recently, Schrauwen and co-workers (1999) reported that the expression of uncoupling protein 3 (UCP3) mRNA per mitochondrion was lower in skeletal muscle from trained subjects than in that from untrained subjects, and that UCP3 expression correlated negatively with mechanical efficiency. Based on these results, the authors suggested that endurance training and/or physical fitness reduces the expression of UCP3 (a mitochondrial protein that has been shown to uncouple respiration from ATP synthesis), allowing for a higher efficiency of mitochondrial energy production in trained subjects (Schrauwen et al. 1999). The data in the present study obtained by direct measurements of mitochondrial P/O ratio argue against this theory and demonstrate that the efficiency of mitochondrial oxidative energy transfer in human muscle is unchanged by endurance training.

Despite enhanced muscle oxidative power, the activities of antioxidative enzymes and glutathione status remained unchanged after the training programme. These results are consistent with previous data (Tiidus et al. 1996), where aerobic cycle training increased CS activity in the vastus lateralis muscle of healthy humans, but had no significant effects on muscle antioxidative enzyme activities, muscle vitamin E level or glutathione status. These results appear to be in conflict with the findings from a cross-sectional study (Jenkins et al. 1984) where the activity of muscle scavenger enzymes (SOD and catalase) were positively correlated to aerobic training status (V̇_O2,max). It is possible that the results of Jenkins et al. (1984) are a consequence of inter-individual fibre type differences, since subjects with a high whole-body V̇_O2,max also have a higher proportion of oxidative muscle fibres (Bergh et al. 1978). The fibre-type specific distribution of antioxidative enzyme activities has not been established in human muscle, but if it is similar to that in animal muscle (Hollander et al. 1999; Powers et al. 1999) it may offer an explanation for the positive relationship between training status and antioxidative enzyme activities observed in the study of Jenkins et al. (1984). Another possibility is that the training periods used in the current study (6 weeks) and in the study of Tiidus et al. (1996) (8 weeks) were too short to induce an upregulation of antioxidative status in skeletal muscle.

Since the mitochondrial electron transport chain is considered to be the major source of ROS in skeletal muscle, a delayed (or non-existent) adaptation of antioxidative systems to endurance training, along with enhanced mitochondrial capacity, may cause an imbalance between generation and removal of ROS. This imbalance may be potentially harmful for the cell and could result in oxidative damage of vital cellular components. However, it is possible that such perturbation of cellular homeostasis is a prerequisite for adaptation to training. It was suggested by Davies et al. (1982) that free radical-induced damage to mitochondrial membranes may stimulate mitochondrial biogenesis. Recently, a study by Miranda et al. (1999) provided evidence that endogenous ROS enhance the expression of nuclear mitochondrial biogenesis genes Nuclear Respiratory Factor-1 and Mitochondrial Transcription Factor A.

In contrast to the lack of adaptation of human muscle antioxidative defence to aerobic training, high-intensity sprint training has been shown to improve the level of antioxidant protection in the muscle (Hellsten et al. 1996). High-intensity sprint cycle exercise is associated with pronounced energetic stress accompanied by adenine nucleotide degradation to hypoxanthine. During this condition of metabolic stress, xanthine dehydrogenase may be converted to the oxidase form, which uses molecular oxygen instead of NAD⁺ as an electron acceptor and generates superoxide radicals (Kuppusamy & Zweier, 1989). These events may increase the rate of ROS generation more than the enzyme activities during exercise at moderate intensity and may explain the different responses of the antioxidative systems to the respective training protocols. It seems likely that the cellular localisation of oxidative stress during high-intensity exercise is different from that during aerobic exercise, due to the different localisation in muscle of xanthine oxidase (endothelial cells) and the electron transport chain (mitochondria). The adaptability of antioxidants located near different sources of ROS generation in muscle may also differ. The observed increase in the muscle antioxidative defence mechanism after high-intensity exercise training may be restricted to the vicinity of xanthine oxidase and may not necessarily provide better protection of muscle mitochondrial function against free radicals generated by the electron transport chain. This idea is consistent with previous observations of marked differences in the antioxidative capacity between various tissue compartments and organelles of an individual cell (Reid, 1996).

Our observation that endurance training did not change the antioxidative status of mitochondria and whole muscle despite an increase in the oxidative potential indicates that after aerobic training every individual mitochondrion is less protected against harmful effects of free radicals. However, the relative impairment of state 3 respiration induced by exposure to ROS was not significantly different before or after exercise, which indicates that the susceptibility of the mitochondrial respiratory complexes, and thus maximal muscle oxidative power, to free radicals remained unchanged. In contrast, the ROS-mediated increase in state 4 oxygen consumption was more pronounced after the training period. State 4 respiration is due to back-leakage of protons through the inner membrane of the mitochondrion and is controlled mainly by the proton permeability of the inner membrane. Therefore, the finding that ROS treatment caused a larger increase in state 4 respiration after training than before it indicates that the mitochondrial membrane is more sensitive to oxidative damage after endurance training, probably due to the reduced antioxidative protection of each individual mitochondrion.

The results in the present study demonstrate that the increase in state 4 respiration induced by ROS treatment was reduced by addition of atractyloside, a specific inhibitor of ADP/ATP translocase. This effect indicates that AAT was affected by oxidative stress and that part of the increased state 4 oxygen consumption following ROS treatment is caused by proton leakage through the AAT. Recent studies on rat cardiac and housefly flight muscles demonstrated that even mild oxidative stress produces striking changes in mitochondrial AAT, manifested as a gradual increase of the apparent molecular mass by up to 1.2 kDa, a progressive loss of lysine, cysteine, arginine and valine residues, an increase in carbonyl content, and a loss of AAT exchange activity (Giron-Calle et al. 1994; Giron-Calle & Schmid, 1996; Yan & Sohal, 1998). It has been reported that carboxyatractyloside inhibited the peroxidation-induced molecular mass shift of AAT (Giron-Calle & Schmid, 1996). Furthermore, it was demonstrated on flight muscle mitochondria that AAT was the only mitochondrial membrane protein exhibiting an increase in carbonyl content, adducts of the lipid peroxidation products, and a loss in its functional activity following exposure to oxidative stress (Yan & Sohal, 1998). Our results are compatible with these reports and indicate that, in human skeletal muscle mitochondria, AAT is one of the most ROS-sensitive compounds. Like AAT, uncoupling proteins 2 and 3 (Fleury et al. 1997; Gong et al. 1997), F₁-ATPase (Novgorodov et al. 1990) and the non-specific protein-aqueous channel known as the permeability transition pore and recently characterised in isolated skeletal muscle mitochondria (Fontaine et al. 1998) can, under certain conditions, participate in proton transport through the mitochondrial membrane. Sequential additions of the inhibitors of these membrane compounds (GDP, oligomycin and cyclosporin A) had no effect on state 4 respiration of ROS-treated mitochondria, which may indicate that the H⁺ leak through these structures was not affected by exposure to ROS.

The most prominent effects of ROS treatment on the function of isolated mitochondria were the reduction in the P/O ratio and the increase in state 4 oxygen consumption. These changes indicate that less ATP is produced per oxygen molecule consumed (i.e. the efficiency of aerobic energy production is reduced). A large number of studies have shown that whole-body oxygen consumption increases progressively during prolonged exercise at constant workloads exceeding 60 % of V̇_O2,max. The mechanism behind this phenomenon is not known, but hypotheses such as decreased mechanical efficiency (altered recruitment pattern of fibres and muscles; Coyle et al. 1992) or decreased chemical efficiency of oxidative phosphorylation (Tonkonogi et al. 1998) have been proposed. During exercise, generation of ROS is increased (Davies et al. 1982; Jackson et al. 1985; Leeuwenburgh et al. 1999) and it is possible that ROS generation during prolonged exercise may exceed the capacity of local antioxidative defence mechanisms, which can result in an oxidative attack on mitochondria. This may, as indicated by the results in the present study, cause an uncoupling of oxidative phosphorylation and thus contribute to an upward drift in V̇_O2 during exercise. The increased uncoupling previously observed after prolonged exercise (Tonkonogi et al. 1998) is consistent with this hypothesis.

An interesting finding in this study was that muscle SOD activity was positively correlated to the age of the subjects. This finding is consistent with previous reports on the effects of ageing on the muscle antioxidative defence mechanism in animals. Many studies on rat skeletal muscle have demonstrated that muscle from aged rats has a significantly higher total activity of antioxidative enzymes than that from young rats (Ji, 1993; Leeuwenburgh et al. 1994; Chandwaney et al. 1998). It has been demonstrated that generation of free radicals and oxidative tissue damage is more pronounced in muscle from aged experimental animals (Ji, 1993; Bejma & Ji, 1999) and therefore it was suggested that enhanced activity of antioxidative enzymes might be a compensatory response to the higher oxidative stress imposed on the aged muscle (Ji, 1993). In humans, ageing has been shown to be associated with increased oxidative damage of DNA, lipids and proteins in skeletal muscle (Mecocci et al. 1999). It is therefore reasonable to assume that the positive relationship between muscle SOD activity and age observed in the current study reflects the adaptation of the human muscle antioxidative defence mechanism to an age-dependent increase in oxidative stress. It is also possible that age-related changes in the composition of human muscle fibre types may contribute to the apparent upregulation of muscle total antioxidative enzyme activity. It is well documented that ageing is associated with a significant reduction in type II fibre size, whereas type I muscle fibres seem to be resistant to age-associated atrophy (Rogers & Evans, 1993). Since there is no significant alteration in the percentages of fibre types with age (Rogers & Evans, 1993), there will be an increase in the proportion of type I fibre area in the muscle, and, consequently, the whole muscle will become more oxidative. As mentioned previously, evidence from the literature suggests that the activity of antioxidative enzymes is greater in oxidative muscle fibres than in glycolytic fibres (Hollander et al. 1999; Powers et al. 1999). Thus, aged muscle with a larger relative type I fibre area will exhibit greater activity of antioxidative enzymes. The role of age-related changes in the human muscle antioxidative defence mechanism remains unclear and deserves further studies.

In summary, the influence of endurance training on mitochondrial oxygen consumption has for the first time been investigated in isolated human skeletal muscle mitochondria. It is concluded that short-term training increases maximal mitochondrial oxidative power, whereas the efficiency of aerobic energy production remains unchanged. Muscle antioxidative defence systems were unaffected by training, but there was an increase in mitochondrial content. This results in a reduced protection of each individual mitochondrion against toxic oxygen species. This phenomenon may underlie the increased susceptibility of the mitochondrial membrane to oxidative stress, indicated by the larger increase in proton conductance induced by exposure to ROS after the endurance training period. The sensitivity of maximal muscle oxidative power to ROS was unaffected by endurance training. The finding that the muscle superoxide dismutase activity is upregulated with age indicates that the antioxidative defence system in human muscle undergoes significant alteration during ageing.