Incremental knee extensor (KE) exercise performed at 25, 70, and 100% of single-leg maximal work rate (WRMAX) was combined with ex vivo electron paramagnetic resonance (EPR) spectroscopic detection of α-phenyl-tert-butylnitrone (PBN) adducts, lipid hydroperoxides (LH), and associated parameters in five males. Blood samples were taken from the femoral arterial and venous circulation that, when combined with measured changes in femoral venous blood flow, permitted a direct examination of oxidant exchange across a functionally isolated contracting muscle bed. KE exercise progressively increased the net outflow of LH and PBN adducts (100% > 70% > 25% WRMAX, P < 0.05) consistent with the generation of secondary, lipid-derived oxygen (O2)-centered alkoxyl and carbon-centered alkyl radicals. Radical outflow appeared to be more intimately associated with predicted decreases in intracellular Po2 (iPo2) as opposed to measured increases in leg O2 uptake, with greater outflow recorded between 25 and 70% WRMAX (P < 0.05 vs. 70–100% WRMAX). This bias was confirmed when radical venoarterial concentration differences were expressed relative to changes in the convective components of O2 extraction and flow (25–70% WRMAX P < 0.05 vs. 70–100% WRMAX, P > 0.05). Exercise also resulted in a net outflow of other potentially related redox-reactive parameters, including hydrogen ions, norepinephrine, myoglobin, lactate dehydrogenase, and uric acid, whereas exchange of lipid/lipoproteins, ascorbic acid, and selected lipid-soluble anti-oxidants was unremarkable. These findings provide direct evidence for an exercise intensity-dependent increase in free radical outflow across an active muscle bed that was associated with an increase in sarcolemmal membrane permeability. In addition to increased mitochondrial electron flux subsequent to an increase in O2 extraction and flow, exercise-induced free radical generation may also be regulated by changes in iPo2, hydrogen ion generation, norepinephrine autoxidation, peroxidation of damaged tissue, and xanthine oxidase activation.
- electron paramagnetic resonance
- lipid peroxidation
- mitochondrial redox
the molecular detection of free radical species during exercise is technically challenging because of their high reactivity and low steady-state concentration (19). As a consequence, investigators have traditionally relied on comparatively stable molecular “footprints” of oxidative damage to lipids, proteins, and DNA formed downstream of the primary production pathway (35).
However, the analytic techniques employed have been mostly indirect and the subject of much concern (37). Reactive intermediates exhibit markedly different thermodynamic and kinetic properties adding to the inconsistencies reported in the exercise science literature (4). Exercise models also deserve critical evaluation because they typically recruit heterogenous muscle groups characterized by different modes of contraction. Furthermore, blood sampling has usually been confined to the mixed venous circulation distal to the activated musculature of interest, which can also prove problematic.
Electron paramagnetic resonance (EPR) spectroscopy is considered the most sensitive, specific, and direct molecular technique for the detection and subsequent identification of free radicals sine qua non, although its application to the exercise environment has been limited (39). We (8) recently applied an ex vivo EPR spin-trapping technique to the functionally isolated single-leg knee extensor (KE) model to overcome some of these experimental limitations in an attempt to more accurately define the source and mechanisms associated with exercise-induced free radical generation. Through the insertion of a femoral arterial-venous catheter and simultaneous measurement of leg blood flow (LBF), we (8) documented, for the first time, direct analytic evidence for oxygen- and carbon-centered free radical outflow across a skeletal muscle bed during submaximal exercise. A schematic overview of the experimental techniques and exercise models traditionally employed and the methodological advances introduced by our laboratory is presented in Fig. 1.
However, whereas preliminary findings demonstrated a tentative association between radical outflow and hemodynamic parameters, it is equally plausible that alternative mechanisms may have contributed to the downstream generation of lipid-derived radicals. Potential mechanisms include altered lipid-substrate and hydrogen ion (H+) exchange, catecholamine auto-oxidation, molecular peroxidation of damaged skeletal tissue, and xanthine oxidase activation. Furthermore, to our knowledge, there are no published reports that have simultaneously examined the exchange of aqueous and lipid-soluble antioxidants to provide a more complete understanding of the dynamic interaction between pro-oxidant challenge and antioxidant defense during muscular work.
Using proton magnetic resonance spectroscopy to detect changes in myoglobin saturation using an identical experimental paradigm, Richardson et al. (49, 50) have consistently demonstrated a marked decrease in intracellular Po2 (iPo2) between the low-to-moderate intensity domains [25–70% maximal work rate (WRMAX)]. However, no changes were observed between the moderate-to-high intensity domains (70–100% WRMAX) subsequent to increased muscle O2 diffusional conductance, allowing an undeterred rise in muscle O2 flux. Thus the inclusion of an additional exercise increment in the present study afforded an opportunity to disassociate changes in O2 flux from changes in iPo2 and, for the first time, examine their respective contributions to radical exchange.
On the basis of existing knowledge, three experimental hypotheses were tested. First, that incremental exercise would be associated with an incremental increase in the net outflow (defined as the product of a positive venoarterial concentration difference and LBF) of lipid hydroperoxides (LH) and α-phenyl-tert-butylnitrone (PBN) adducts despite a net uptake or consumption of antioxidants. Second, a decrease in iPo2 would further compound radical outflow initiated by an increase in muscle O2 flux (measured as O2 uptake, V̇o2). Thus we anticipated comparatively greater radical outflow during the low-to-moderate exercise intensity transition compared with the moderate-to-high transition. Finally, we expected a comparable outflow in other redox-reactive parameters, including blood lipids, H+, catecholamines, biomarkers of tissue damage, and uric acid thus providing additional insight into potential sources and mechanisms associated with exercise-induced free radical generation.
MATERIALS AND METHODS
The study was designed with 90% power at the P < 0.05 level to detect biologically significant changes in oxidant exchange across the working leg based on a calculated critical difference of 121% for PBN adducts and previous findings (8). Five healthy men, aged 47 (mean) ± 22 (SD) yr (range: 22–67 yr) subsequently provided written informed consent according to the ethical requirements of the University of California San Diego, Human Protection Program. The heterogenous sample facilitated an examination of mechanisms pertinent to exercise-induced free radical generation across a broad spectrum of submaximal single-leg V̇o2 values. It was considered unethical to recruit additional subjects owing to the invasive nature and inherent risks associated with vascular catheterization. Subjects were asked to refrain from ingesting antioxidant vitamins, analgesics, and nonsteroidal anti-inflammatories 4 wk before the start of the investigative period. They ingested a high-carbohydrate meal the night before the study and arrived at the laboratory the following day after a 12-h overnight fast. All procedures conformed to the code of Ethics of the World Medical Association (Declaration of Helsinki).
Subjects performed two familiarization bouts on the dynamic KE apparatus before an incremental test to volitional exhaustion for the determination of WRMAX. Anecdotal subject reports, force tracings, electromyography, intramuscular temperature measurements, and T2-weighted magnetic resonance imaging have confirmed the isolation of mechanical work to the quadriceps muscle group (48). Initial WR was set at 3–10 W and increased by 3–6 W/min (at a cadence of 60 rpm) according to the subject’s predicted exercise capacity, resulting in a WRMAX of 37 ± 19 W.
Thirty minutes after vascular catheterization of the left femoral artery and vein, as previously described (50), each subject completed an incremental exercise test at 25% (low), 70% (moderate), and 100% (high-intensity domain) of WRMAX. Each increment was continued for 2–4 min to achieve steady-state pulmonary V̇o2. Arterial and venous blood were sampled simultaneously and immediately preceded by duplicate assessments of LBF via a constant-infusion thermodilution technique (2). Leg plasma flow (LPF) was calculated from LBF and venous hematocrit (Hct).
We specifically chose not to include resting data owing to the technical difficulties and subsequent inaccuracies associated with the measurement of resting flow as previously discussed (8). A complementary examination of related metabolic parameters was therefore not attempted due to our inability to document subsequent exchange.
O2 transport variables and lactate.
Samples of arterial and venous blood (2–3 ml) were presented anaerobically to an IL Synthesis blood gas analyzer (Clayton, NC) and IL 682 cooximeter (Clayton) for the determination of Po2, Pco2, H+, oxyhemoglobin saturation of arterial and venous blood (So2), hemoglobin (Hb), and Hct with an intrarun coefficient of variation (CV) of <5%. Blood O2 content (CO2) was calculated as 1.39 (Hb) × (So2/100) + 0.003 × Po2. O2 extraction was calculated as the difference between the femoral artery O2 content (CaO2) and femoral venous O2 content (CvO2) divided by CaO2 [×100 (%)]. O2 delivery was calculated as the product of LBF and CaO2 and muscle V̇o2 as the product of CaO2 − CvO2 difference and LBF. Whole blood lactate (intra-assay CV < 5%) was assayed enzymatically using a semiautomated analyzer (1500, Yellow Springs Instruments, Yellow Springs, OH). Net exchange (outflow or uptake) was calculated according to Fick’s principle by multiplying LPF or LBF by the venoarterial concentration difference for selected parameters. Changes in Hct and Hb concentrations were utilized to calculate relative shifts in arterial-venous plasma volume (PV) according to standard methods (25).
Indexes of Oxidative Stress
Ex vivo spin trapping with PBN (190 mM/l in 0.9% NaCl) was incorporated for ex vivo detection of secondary or tertiary lipid-derived radical species as previously described (4, 8). Briefly, PBN adducts were extracted from serum into toluene, vacuum degassed for two cycles, and analyzed at 21°C by using an EMX X-band EPR spectrometer fitted with an ER TM110 cavity (Bruker; Karlsruhe, Germany). The extraction efficiency for PBN was 85–90% as confirmed by UV spectroscopy. Spectrometer conditions were as follows: 20 mW experimentally validated nonsaturating incident microwave power, 0.5-G modulation, 1 × 105 receiver gain, 82-ms time constant, 3,450-G magnetic field center, and ±50 G scan width for 15 cumulative scans. EPR spectral parameters were obtained by using commercially available software (Bruker Win EPR System, Version 2.11) and filtered identically. Confirmation of magnetic flux density (g) values were established using the stable free radical diphenylpicryl hydrazyl. The average spectral peak-to-trough line height was considered a measure of the relative spin adduct concentration following conformation of peak-to-trough line-width conformity and double integration on selected samples. A comparison of the double integral of selected PBN adducts was also made with that of known concentrations of the stable 2,2,6,6-tetramethyl-1-piperidinyloxyl free radical dissolved in toluene. The concentration of spin-trapped radicals observed using the technique applied in the present study yielded values in the 10−5-to-10−6 M range (equivalent to ∼6 × 1017−18 spins).
Serum LH were determined by using the ferrous iron-xylenol orange (FOX) assay (61). Briefly, this assay incorporates the selective oxidation of ferrous to ferric ions by hydroperoxides. This reaction yields a blue-purple-colored complex due to the selective binding of xylenol orange to ferric ions. Catalase was added to prevent spontaneous LH generation during the ferrous oxidation step. Absorbance changes at 560 nm were monitored spectrophotometrically. The intra-assay and interassay CV was <2% and <4%, respectively.
Serum uric acid was measured by reflectance spectrophotometry using a Vitros 950 analyzer (Amersham; Little Chalfont, UK). The intra-assay and interassay CV values were both <2%.
For ascorbic acid measurements, plasma was stabilized and deproteinated by adding 900 μl of 5% metaphosphoric acid (Sigma Chemical; Dorset, UK) to 100 μl EDTA plasma. Ascorbic acid was subsequently assayed by fluorimetry based on the condensation of dehydroascorbic acid with 1,2-phenylenediamine (59). The intra-assay and interassay CV values were both <5%.
The plasma concentration of α-tocopherol, retinol, and the carotenoids lycopene, α-carotene, and β-carotene were determined by using a high-performance liquid chromatography method (21, 56). The intra-assay and interassay CV values were both <5%.
Serum concentrations of total cholesterol and triacylglycerol were determined via routine enzymatic techniques using an Olympus AU5200 automated analyzer and Olympus reagents (Southall; Middlesex, UK). The intra-assay and interassay CV values were <2% and <4%, respectively. The cholesterol content of serum high-density lipoprotein (HDL-C) was assayed enzymatically after chemical precipitation of other lipoproteins from serum with dextran sulfate and magnesium by using an ILab 600 analyzer (Instrumentation Laboratory, Warrington; Cheshire, UK). The intra-assay and interassay CV values were both <3%. Serum low-density lipoprotein cholesterol (LDL-C in mmol/l) was calculated as LDL-C = total cholesterol − triglycerides/2.2 − HDL-C (29). Enzymatic assays were used to analyze the plasma concentration of nonesterified fatty acids (Behring Diagnostics; La Jolla, CA) and glycerol (Wako Chemicals; Neuss, Germany). The intra-assay and interassay CV values for nonesterified fatty acids were <2% and 5% and 2% and 5% for glycerol.
Plasma concentrations of epinephrine, norepinephrine, and dopamine were measured by reverse-phase high-performance liquid chromatography using electrochemical detection (15). Norepinephrine spillover was calculated as previously described (51).
Serum total creatine phosphokinase (CPK) and LDH activities were determined via reflectance spectrophotometry using a Vitros 950 analyzer (Amersham). The intra-assay and interassay CV values were both <4%. Serum myoglobin was measured by using an automated chemiluminescence immunoassay (Bayer Centaur; Newbury, UK). The intra-assay and interassay CV values were <4% and 1%, respectively.
After application of repeated Shapiro-Wilk W-tests and confirmation of distribution normality, a two-way repeated-measures analysis of variance was incorporated to assess the effects of exercise intensity (25% vs. 70% vs. 100% WRMAX) and sample site (arterial vs. venous) on selected dependent variables. When an interaction effect was indicated, Bonferroni-corrected (when appropriate) paired samples t-tests were applied to each level of the within-subjects factor of interest. A Friedman test followed by Bonferonni-corrected Wilcoxon matched pairs signed ranks tests were incorporated as the nonparametric equivalents. Differences in outflow were assessed by using a one-factor repeated-measures ANOVA and Bonferonni-corrected paired-samples t-tests or appropriate nonparametric equivalents. The relationship between two dependent variables was analyzed using a Pearson product moment or Spearman’s rank correlation. Significance for all two-tailed tests was established at P < 0.05, and data are expressed as means ± SD.
Blood Gas, Hemodynamics, and General Metabolic Responses
Incremental KE exercise increased O2 delivery and V̇o2 and was associated with an increase in LBF, whereas peripheral O2 extraction did not change across exercise intensities (Table 1). A general hemodilution effect (i.e., increase in PV) was apparent due in part to the infusion of saline (∼20 ml for each exercise intensity) during the measurement of LBF. However, metabolic parameters were not corrected for this artifact because PV shifts were independent of exercise intensity and sample site. Exercise also resulted in a net outflow of H+ and lactate.
Exercise resulted in a progressive increase in the venoarterial concentration difference for norepinephrine, thus increasing net outflow and spillover (Table 2). In contrast, exercise did not influence arterial or venous concentrations of epinephrine or dopamine.
With the exception of slightly elevated pooled arterial and venous values for derived LDL at 100% WRMAX, incremental exercise did not influence the venoarterial concentration difference of remaining lipids or lipoproteins (Table 3).
Exercise was associated with a general increase in the venoarterial concentration difference for myoglobin and LDH that was only apparent for total CPK at 100% WRMAX (Table 4). However, no progressive increase in outflow for any of the related parameters was observed during incremental exercise.
There were no apparent changes in the arterial or venous concentrations of ascorbic acid and selected lipid-soluble antioxidants (Table 5).
Oxidative Stress Biomarkers
A positive venoarterial concentration difference and hence net outflow of LH, PBN adducts, and uric acid was observed during exercise (Table 6). Typical EPR spectra of PBN adducts detected ex vivo are illustrated in Fig. 2 with control and simulation data.
The dominant signal exhibited hyperfine coupling constants of aN = ∼13.7G and aβH = ∼1.9G, consistent with published values for an O2-centered alkoxyl species (PBN-LO·) using similar extraction solvents (18). Visual inspection of spectra identified the likely presence of a second adduct, albeit of low signal intensity. Computer simulation tentatively confirmed that this adduct, detectable across all exercise intensities and sample sites, contributed to ∼10% of the total signal recovered. Coupling constants were determined as aN = 14.0 G and aβH = 4.0 G, consistent with the trapping of a carbon-centered alkyl (PBN-LC·) species (18). A more comprehensive examination of these species was beyond the scope of the present investigation, although alternative approaches, including high-performance liquid chromatography mass spectrometry, electron nuclear double-resonance spectroscopy, and ab-initio-density functional theory calculations are currently being applied in an attempt to further resolve species identification.
An exercise-induced increase in single-leg V̇o2 was associated with a progressive increase in the net outflow of LH and PBN adducts that was most prominent between 25% and 70% WRMAX as indicated by the departure from linearity at the highest exercise intensity (Fig. 3A). The latter was significant for PBN only (P < 0.05 vs. predicted increase based on a linear function). Accordingly, a positive venoarterial concentration difference for radicals was only apparent from 25% to 70% WRMAX when LH and PBN-spin adducts were expressed relative to changes in the individual components of muscle V̇o2, specifically O2 extraction (Fig. 3B) and LBF (Fig. 3C). These changes were inverse to previously reported changes in iPo2 during identical KE exercise (49, 50).
Changes in the outflow of LH and PBN adducts were no longer apparent when expressed relative to the exchange of related parameters that also exhibited positive venoarterial concentration differences. These included H+ (Fig. 4A), norepinephrine (Fig. 4B), myoglobin and LDH (Fig. 4, C and D despite inherent variability), and uric acid (Fig. 4E). Finally, a disproportional increase in PBN-adduct outflow was observed from 25% to 70% WRMAX when expressed relative to LH outflow (Fig. 4F).
Extending previous observations during submaximal exercise (4, 8), the current data collected during incremental exercise to maximum identified a progressive increase in the venoarterial concentration difference and hence net outflow of LH and lipid-derived PBN adducts measured by EPR spectroscopy subsequently identified as a mixture of alkoxyl-alkyl species. However, the exchange of a comprehensive selection of nonenzymatic aqueous and lipid-phase antioxidants was unremarkable. Furthermore, expression of radical outflow to related parameters tentatively suggested that in addition to potential increases in mitochondrial electron flux, a decrease in iPo2 and increases in H+ generation, norepinephrine auto-oxidation, peroxidation of damaged tissue, and xanthine oxidase activation may each have contributed to exercise-induced free radical generation.
It is important to emphasize that whereas EPR is considered the most direct, specific, and sensitive analytic technique for the molecular detection and subsequent identification of free radical species (4), the spin-trapping approach employed in the present study still relies on the ex vivo detection of resonance-stabilized reactants formed clearly downstream of the primary oxidant production pathway that we assume reflects events initiated in vivo (4, 8, 11). Nonetheless, the complementary increase in LH, one of the major initial reactants of lipid peroxidation (7), provided additional convincing evidence that our experimental paradigm induced peroxidative damage.
Supraphysiological changes in perfusate O2 concentration has been shown to influence the trapping efficiency and stability of PBN adducts detected in vitro (43). Thus it is plausible that the exchange of oxidative stress markers detected ex vivo in the present study was merely a consequence of differences in the ambient Po2 of arterial and venous blood. However, this is unlikely for a number of reasons. First, we did not identify any incremental change in the arteriovenous Po2 difference despite clear changes in LH and PBN adduct concentration. Second, follow-up studies have since identified similar PBN adduct concentrations in arterial and venous blood despite marked differences in resting Po2 (arterial: ≈95 Torr vs. venous: ≈25 Torr). Finally, sample aeration of an in vitro Fenton system (cumene hydroperoxide + Fe2+) to physiologically relevant ambient Po2 values did not influence spectral characteristics of what we consider to be comparable PBN adducts (D. M. Bailey, unpublished findings).
Mitochondrial Mechanisms: Increased O2 Flux vs. Decreased Intracellular Po2
Mitochondrial flavoproteins, iron-sulfur clusters, and ubisemiquinone are thermodynamically capable of reducing O2 to the superoxide anion (O2−·) accounting for an estimated 1–2% of overall O2 consumption in vitro (58). Thus a mass action effect initiated by a systemic increase in pulmonary V̇o2 has traditionally been considered the major mechanism associated with mitochondrial free radical generation in exercising humans (5, 52). However, the “flux concept” is clearly incompatible with in vitro evidence identifying a decrease in mitochondrial O2−· generation during the state 4 (basal respiration) to state 3 (ADP stimulated such as that encountered during KE exercise) transition subsequent to a decrease in mitochondrial membrane potential (53). Additional mechanisms capable of catalyzing proton leak across the inner mitochondrial membrane thus providing functional protection against O2−· may involve activation of uncoupling protein 3, adenine nucleotide translocase, and mitochondrial permeability transition pores (27, 28, 34).
These inconsistencies warranted the first comprehensive examination of the relationship between radical exchange and the individual convective components that contribute to local muscle V̇o2. Our findings highlight a tentative disassociation or apparent “uncoupling” between V̇o2 and radical outflow, with more marked increases observed between the low-to-moderate exercise intensity domains. A subsequent examination of what the muscle bed per se was generating as a function of independent changes in O2 extraction and LBF confirmed this bias and revealed that radical release responded inversely with remarkable precision to existing data collected in our laboratory for exercise-induced changes in iPo2 (50). Though not directly measured, we are confident that the iPo2 response to our strategically chosen exercise paradigm would have been qualitatively similar. This raises the intriguing possibility that exercise-induced free radical generation is more intimately regulated by increased mitochondrial redox subsequent to a decrease in (extra) mitochondrial Po2 per se rather than increased electron flux as traditionally assumed (24, 38, 52). Whereas related mechanisms await investigation in vivo, a lower mitochondrial Po2 has been shown to decrease the VMAX of cytochrome oxidase for O2, thus increasing the lifetime of reduced mitochondrial electron carriers such as ubisemiquinone and amplifying O2−· generation by complex III (22, 26). In the present study, we would have anticipated a higher mitochondrial redox for a given electron flux during the 25–70% WRMAX transition, as was seen here. Whether the apparent disassociation between V̇o2 and radical outflow was merely a threshold phenomenon subsequent to anaerobiosis also deserves attention. The inclusion of additional exercise intensity increments in future studies would facilitate a more comprehensive examination of radical exchange energetics.
Lipid Metabolism and Radical Species
Several lines of evidence suggest that the species recovered in the present study may have been lipid derived. Adducts were clearly soluble in nonaqueous solvents, and comparable coupling constants have been detected following in vitro oxidation of linoleic and α-linolenic polyunsaturated fatty acids (PUFA) (8). Furthermore, EPR spectral signals displayed consistently lower signal intensities associated with the high-field lines (Fig. 2), characteristic of spin inhibition, which may indicate the binding of a longer length biopolymer radical such as a PUFA carbon side chain (45). Finally, similar PBN adducts and supporting biomarkers of lipid peroxidation have been shown to increase markedly following ingestion of a fatty meal subsequent to lipoprotein triacylglycerol enrichment (3). It is important to emphasize that this was not a conflicting factor in the present study based on the dietary restrictions imposed.
These observations justified a complimentary examination of lipid-lipoprotein exchange across the working leg. Contrary to previous observations during two-legged cycling exercise (6), there was no evidence for increased lipid uptake and the lack of glycerol outflow, although not definitive, strongly argues against any increase in intramuscular lipolysis. We are therefore confident that radical outflow occurred independently of altered lipid-lipoprotein substrate delivery.
It has been suggested that LO· detected by using this technique may have evolved during the metal-catalyzed Fe2+-reductive decomposition of extracellular LH formed subsequent to primary radical-mediated damage to membrane phospholipids (4, 8, 42). Despite the fact that transferrin, ferritin, and albumin limit the transit pool of iron to below 5–10 μmol/l (33), it is not unreasonable to suggest that muscle contraction may have overloaded conservation mechanisms and increased the extracellular availability of catalytic iron. We (10) have recently demonstrated an increase in the concentration of PBN adducts and ascorbyl radical following in vitro addition of a specific iron promoter (e.g., EDTA stimulating catalytic activity of iron and reducing that of copper) and marked reduction following the addition of the metal-chelating agent diethylenetriaminepentaacetic acid to serum. Identical coupling constants, similar to those reported by Mergner et al. (42), have also been obtained during in vitro incubation of PBN with a Fenton system containing cumene hydroperoxides and Fe2+ (D. M. Bailey, unpublished observations).
Previous studies have demonstrated an exercise-induced increase in oxidative stress and bleomycin-detectable iron (23, 36) with effective protection conferred following iron chelator prophylaxis (55). As demonstrated in previous studies (13, 42), the increase in sarcolemmal membrane permeability reported in the present study may have initiated iron release from ferritin. Additional mechanisms associated with KE exercise that may have altered catalytic iron availability include activation of polymorphonuclear leukocytes and decrease in pH shown to increase iron release from ferritin and myoglobin (13, 33), oxidant-mediated damage to RBC/heme proteins rendered structurally “susceptible” due to perturbation of osmotic homeostasis, and compression of large muscle group (quadriceps femoris) on capillaries (54). Furthermore, extracellulary generated H2O2 is capable of reacting with nonferritin iron intracellularly to generate OH−· (40).
The minor signal assigned to LC· may be a short-chain radical such as the ethyl or pentyl species formed during β-scission of LO· (30) and is consistent with previous observations (57). This species is thermodynamically capable (ΔEO′ ≈ +1,300 mV vs. PUFA-H/PUFA-OOH) of initiating lipid peroxidation (17). However, it is of interest to note that LH and PBN-adduct exchange were not stoichiometrically related as indicated by a disproportional increase in PBN adduct relative to LH outflow from the low-to-moderate exercise intensity domains.
Although the lactate ion associated with metabolic acidosis can directly scavenge initiating species (31), including O2−· and hydroxyl radicals (·OH), the concomitant generation of H+ can initiate oxidation. An increase in H+ can decrease Ca2+ uptake by the sarcoplasmic reticulum (20), convert O2−· to the more reactive hydroperoxyl radical, disassociate protein-bound iron, and accelerate dismutation to H2O2, that, via a Haber-Weiss or Fe2+-catalyzed Fenton reaction, can yield the highly toxic ·OH (12, 16) and secondary LO·/LC· (45). Furthermore, differential changes in the EPR signal intensities of LO·, LC· and ascorbyl radical detected in untreated human serum following exercise and ex vivo addition of a metal chelator/catalyst and H+ donor/scavenger confirms the interactive significance of iron and H+ as oxidative catalysts for the reductive cleavage of organic peroxides and present EPR detection of PBN-LO· and -LC· (4, 10).
Additionally, a reduction in pH has been shown to increase the ability of heme proteins to initiate oxidative stress subsequent to protonation of the oxoferryl heme to form the highly reactive Fe4+-OH· (47). Destablization of the ferryl species can promote lipid peroxidation via decomposition of hydroperoxides to form peroxyl (LOO·) and LO· (44) species such as that presumed in the present study.
Isolated quadriceps exercise typically activates a small muscle mass (ca 2.5 kg) that by consequence, resulted in a moderate increase in norepinephrine spillover. Catecholamine-mediated activation of β-adrenergic receptors could increase mitochondrial O2 flux and compound iPo2-mediated O2−· leakage, as the previously documented reduction in plasma markers of oxidative stress following prophylactic β-blockade in exercising subjects would suggest (46). An incremental increase in radical outflow was no longer apparent in the present study when expressed relative to norepinephrine outflow. Auto-oxidation of epinephrine to adrenochrome can yield epinephrine semiquinone and increase O2−· (14). However, it is unlikely that this contributed to radical outflow in the present study because we did not observe a concomitant increase in epinephrine outflow.
Exchange measurements across the muscle bed minimized the interpretive limitations associated with intracellular myofiber proteins as quantitative markers of exercise-induced tissue damage (41, 60) and thus provide compelling evidence for compromised sarcolemmal membrane permeability. Though not a primary focus, we could not detect any evidence of cardiac troponin I outflow (data not shown) which, based on its diagnostic specificity as a cardiac protein (1), combined with the limited central demands imposed by KE exercise, firmly suggests that molecular damage was confined to the skeletal and not myocardial tissue bed as previously suggested (7, 9).
However, the present findings did not unequivocally establish whether free radicals were a direct cause or merely a consequence of myocellular enzyme outflow, because damaged tissue peroxidizes more rapidly than healthy, structurally intact tissue (32). We support a minimal role for mechanical trauma as an initiating event because the exercise modality is exclusively concentric by nature, and no reports of transient myalgia have been recorded in our laboratory either during or following KE exercise. We therefore suggest that tissue damage was initiated principally by chemical events, namely the intra/extracellular generation of free radicals, capable of initiating and propagating secondary oxidative injury, distal to the contracting tissue bed. The liberation of extracellular iron and LH by damaged skeletal tissue may therefore prove the unifying mechanism responsible for secondary LO· generation in the blood of exercising subjects. However, it is important to emphasize that unlike the PBN-adduct data, normalization of LH outflow relative to myoglobin and LDH outflow was highly variable (Fig. 3D). Future studies are clearly warranted to define the potential contributory role of tissue damage in the global exercise-induced oxidative stress response.
In summary and in support of our original hypothesis, this study has provided direct evidence for an incremental increase in free radical outflow across a functionally isolated and energetically active skeletal muscle bed that was associated with an increase in sarcolemmal membrane permeability. EPR spectroscopic examination of spin adducts detected ex vivo identified LO· as the dominant species though a minor signal consistent with LC· was also apparent. Our data tentatively suggest that these secondary species were formed distal to the locus of generation during the metal-catalyzed reductive decomposition of LH liberated by oxidatively damaged skeletal muscle. As anticipated, the kinetics of radical outflow were similar to predicted decreases in iPo2 and not, as traditionally thought, simply due to an increase in muscle O2 flux. A complementary examination of related redox-reactive parameters suggests that H+ generation, norepinephrine auto-oxidation, and xanthine oxidase activation may have also contributed to the exercise-induced oxidative stress response.
We express our gratitude to the enthusiastic subjects who participated in this study and gratefully acknowledge the technical expertise provided by Dr. Richardson’s and Prof. Young’s respective research teams. We also extend our appreciation to Professors Bruce Davies, Peter Wagner, John West, and Joe McCord, and Dr. Damien Murphy for stimulating discussions, and Drs. Kemp and Hullin for technical assistance.
We remember the late Dr. Casey Kindig.
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 by the American Physiological Society