Effects of exercise and training in hypoxia on antioxidant/pro-oxidant balance

(1)

ORIGINAL ARTICLE

Effects of exercise and training in hypoxia on

antioxidant/pro-oxidant balance

V Pialoux

1

, R Mounier

1

, E Ponsot

2

, E Rock

3

, A Mazur

3

, S Dufour

2

, R Richard

2

, J-P Richalet

4

,

J Coudert

1

and N Fellmann

1

1_{Laboratoire de Biologie des Activite´s Physiques et Sportives, Faculte´ de Me´decine, Clermont-Ferrand, France;}2_{Service de Physiologie}

Clinique et des Explorations Fonctionnelles et Respiratoires et de l’Exercice, Hoˆpital Civil, Strasbourg, France;3_{Unite´ Maladies}

Me´taboliques et Micronutriments, INRA Clermont-Ferrand/Theix, Saint-Gene`s Champanelle, France and4ARPE, Laboratoire ‘Re´ponses cellulaires et fonctionnelles a` l’hypoxie’, Universite´ Paris XIII, Bobigny, France

Objective: The aim was to investigate the effects of acute exercise under hypoxic condition and the repetition of such exercise in a ‘living low-training high’ training on the antioxidant/prooxidant balance.

Design: Randomized, repeated measures design. Setting: Faculte´ de Me´decine, Clermont-Ferrand, France.

Subjects: Fourteen runners were randomly divided into two groups. A 6-week endurance training protocol integrated two running sessions per week at the second ventilatory threshold into the usual training.

Intervention: A 6-week endurance training protocol integrated two running sessions per week at the second ventilatory threshold into the usual training. The first hypoxic group (HG, n ¼ 8) carried out these sessions under hypoxia (3000 m simulated altitude) and the second normoxic group (NG, n ¼ 6) in normoxia. In control period, the runners were submitted to two incremental cycling tests performed in normoxia and under hypoxia (simulated altitude of 3000 m). Plasma levels of advanced oxidation protein products (AOPP), malondialdehydes (MDA) and lipid oxidizability, ferric-reducing antioxidant power (FRAP), lipid-soluble antioxidants (a-tocopherol and b-carotene) normalized for triacyglycerols and cholesterol were measured before and after the two incremental tests and at rest before and after training.

Results: No significant changes of MDA and AOPP level were observed after normoxic exercise, whereas hypoxic exercise induced a 56% rise of MDA and a 44% rise of AOPP. Plasma level of MDA and arterial oxygen hemoglobin desaturations after the acute both exercises were highly correlated (r ¼ 0.73). a-Tocopherol normalized for cholesterol and triacyglycerols increased only after hypoxic exercise (10–12%, Po0.01). After training, FRAP resting values (21%, Po0.05) and a-tocopherol/ triacyglycerols ratio (24%, Po0.05) were diminished for HG, whereas NG values remained unchanged.

Conclusions: Intense exercise and hypoxia exposure may have a cumulative effect on oxidative stress. As a consequence, the repetition of such exercise characterizing the ‘living low-training high’ model has weakened the antioxidant capacities of the athletes.

Sponsorship: International Olympic Committee and the Direction Re´gionale de la Jeunesse et des Sports de la Re´gion Auvergne.

European Journal of Clinical Nutrition (2006) 60, 1345–1354. doi:10.1038/sj.ejcn.1602462; published online 21 June 2006

Keywords: hypoxia; endurance training; oxidized lipid; FRAP; AOPP; a-tocopherol

Introduction

Endurance athletes often incorporate altitude exposure into their training regimen in order to improve their oxygen

transport capacity and their subsequent sea-level aerobic performance. However, the efficiency of such training is controversial (Levine and Stray-Gundersen, 1992; Bailey et al., 1998). Since the beginning of the 1990s, new methods combining endurance training and intermittent hypoxia exposure have been proposed: ‘living high-training low’ (Levine et al., 1991) and ‘living low-training high’ (Emonson et al., 1997). The former method is based on sea-level endurance training and a simulated altitude exposure during sleep in hypoxic rooms whereas the latter method consists

Received 7 December 2005; revised 30 January 2006; accepted 5 May 2006; published online 21 June 2006

Correspondence: Dr V Pialoux, Faculte´ de Me´decine. Laboratoire de Physiologie-Biologie du Sport, 28 place Henri Dunant, Clermont-Ferrand 63000, France.

E-mail: Physio.sport@u-clermont1.fr

(2)

in performing endurance sessions at simulated altitude by breathing a hypoxic gas and living at sea-level. Acute physical exercise is known to increase reactive oxygen species (ROS) (Ji, 1996). During maximal exercise, the oxygen flux (O2flux) through the exercising muscles could

increase 100-fold above the resting values. In mitochondria, the increase of O2 flux can lead to an overproduction of

superoxide anion (O2K) and oxygen-derived intermediates

(Sen, 1995) that contribute to oxidative damage on biologi-cal molecules. As a consequence, a significant increase in lipid peroxidation markers (Clarkson and Thompson, 2000), protein (Sen et al., 1997) and DNA oxidation-derived molecules (Alessio, 1993) are usually described after physical exercise.

Surprisingly, in hypoxic conditions, free radical generation is also observed despite a decrease in mitochondrial O2flux.

In resting conditions, oxidative stress markers have been detected in humans under hypobaric hypoxia (Bailey et al., 2001a; Jefferson et al., 2004), at simulated altitude in hypobaric chambers (Joanny et al., 2001) and after breathing gas mixture reduced in O2content (Wing et al., 2003). Firstly,

the increase in catecholamine production under hypoxic conditions (Mazzeo et al., 1998) could be responsible for oxidative stress. Additionally, in hypoxia, the reduction of mitochondria potential redox owing to an accumulation of reducing equivalent which cannot be transferred to oxygen by cytochrome oxidase (Kehrer and Lund, 1994) also enhances ROS production. Finally, after the return in normoxia, the hypoxia/reoxygenation cycle similar to ischemia/reperfusion in flooding the anoxic tissue with oxygen, can exacerbate free radical generation (Askew, 2002).

Consequently, exercise associated with hypoxia exposure would worsen oxidative damage in comparison to normoxic exercise. During a submaximal exercise, an overproduction of DNA strand break under hypoxic condition has been reported by Moller et al. (2001) whereas Bailey et al. (2001b) have found an increase of malondialdehyde (MDA), a lipid peroxidation marker, after strenuous exercise performed under normobaric hypoxia conditions (FiO2¼ 0.16) as

compared with normoxia.

The magnitude of oxidative damages depends on the ability of antioxidant defenses to neutralize free radicals (Sen, 2001). However, literature data are controversial. Some studies have shown that training could improve endogenous enzymatic antioxidant defenses (Robertson et al., 1991; Miyazaki et al., 2001). On the contrary, Clarkson and Thompson (2000) reported after endurance training that acute exercise increased oxidative stress markers. It has been suggested that the non-enzymatic antioxidant mobilizations against the subsequent exercise-induced oxidative stress could not counterbalance the antioxidant enzymatic activity increase (Pincemail et al., 1988; Bailey et al., 2001b).

Very little is known about the effect of exercise training associated with hypoxia. As suggested by Moller et al. (2001), our hypothesis was that the cumulated effects of exercise and

hypoxia on overgeneration of free radicals could outmatch the body antioxidant capacities and consequently the non-enzymatic antioxidant reserves could be partially depleted.

Therefore, the aim of our study was to investigate whether endurance training with high-intensity exercise sessions under normobaric hypoxia (3000 m simulated altitude) weakened the antioxidant status. To answer this issue, we investigated to what extent acute hypoxia during an intense exercise increased the oxidative stress markers and whether endurance training characterized by repeated intense exercise sessions under hypoxia would affect the antioxidant status.

Materials and methods

Subjects

The goal of ‘living low-training high’ protocol was to improve the running performance of trained athletes. Four-teen endurance running male athletes, recruited in local teams, participated in the study which was approved by the Ethics Committee on Human Research for Medical Sciences of Strasbourg Hospital. They received a medical examination with an echocardiography, and gave informed written consent. Means (_{7s.d.) for age, height, body mass and body} fat mass were 30.876.8 years, 18174 cm, 73.477.6 kg and 11.473.4%, respectively. They had a normoxic maximal oxygen uptake (VO2max) of 61.873.7 ml min1kg1 on

treadmill. The inclusion criteria were to train at least five times per week, to have personal 10 km-run record of o35 min and to be less than 45 years old.

Experimental design

The study was divided into three periods (Figure 1): 1. The control period (2 weeks before the endurance

training), which included anthropometric measurements (height, weight, fat mass), two maximal tests on ergocycle under both normoxia and hypoxia conditions, and performance tests.

2 days 2 days 1 week

1 2 3 4 5 6

Normoxia cycling testHypoxia cycling testPerformance tests Performance test

Training (6 weeks) 5 days

Post-training Pre-training

Figure 1 Experimental design. The arrows indicate the blood

withdrawals.

(3)

2. The endurance training protocol associated or not with intermittent hypoxia exposure during training which lasted 6 weeks.

3. The post-training period (5 days after the end of the training) including performance tests.

Control period

Forty-eight hours separated each maximal test. The first was performed on ergocycle (Ergometer Me´difit 1000 S, Belgium) in normoxia (FiO2 (inspired fraction of O2) ¼ 0.21) and the

second under normobaric hypoxia (FiO2¼ 0.14), simulating

an altitude of 3000 m. The normobaric hypoxic gas was obtained through an integrated system of N2 enrichment

(Altitrainer 200, SMTec, Genova, Switzerland).

Before exercising, the subjects remained in resting condi-tions for 4 min breathing either ambient air or hypoxic gas mixture. The work load of the first stage was 80 W and was increased by 40 W every 2 min. The exercises were stopped when the subjects were no longer able to sustain the work at the required rate.

For each test, VO2, ventilation (VE) (Me´disoft Ergocard,

Belgium), ear arterial hemoglobin oxygen saturation (SaO2)

(Oxypleth, novametrix-medical system Inc., Wallingford, USA), heart rate and electrocardiogram were continuously recorded. About 2 ml of blood were collected from an antecubital vein after 1 min of recovery to measure lactate concentration (ABL 700 series, Radiometer, Denmark).

Performance tests

Before and after the training period, the performance tests consisted of two running maximal tests on treadmill (Gymrol 2500, Saint Etienne, France): the first in normoxia (FiO2¼ 0.21) and the second under normobaric hypoxia

(FiO2¼ 0.14) simulating an altitude of 3000 m. The speed of

the first stage was 10 km h1and was increased by 1 km h1 every 2 min. The tests were stopped when the subjects were no longer able to sustain the speed. The VO2, VE (Me´disoft

Ergocard, Belgium), SaO2 (Oxypleth, novametrix-medical

system Inc, USA), heart rate and electrocardiogram were continuously recorded during both performance tests. The speeds corresponding to the second ventilatory threshold (SVT) defined as the Respiratory Compensation Point (Wasserman, 1987) were also determined in normoxia and under hypoxia. These speeds were used to adapt the exercise intensity during the training sessions.

Training protocol

The training program was determined with the help of the coach. Participants were randomly assigned into normoxic (n ¼ 6) and hypoxic (n ¼ 8) groups. Both groups underwent a 6-week endurance training program. Each week, the athletes performed two controlled training sessions performed at SVT on treadmill (Gymrol 2500, Saint Etienne, France) and two

additional sessions. The controlled sessions were run at SVT under normobaric hypoxia (FiO2¼ 0.164) for hypoxic group

(HG) (15.470.7 km h1; 92.173.7% of VO2max assessed in

hypoxia) and in normoxia for normoxic group (NG) (16.871.7 km h1; 85.477.2% of VO2max assessed in

hypoxia). Both groups exercised for the same duration and at the same relative intensity (i.e. the speed corresponding to the SVT assessed in normoxia for NG and under hypoxia for HG). The duration of SVT sessions was for 50 min with 3278 min at SVT. The two additional sessions were run at 80% SVT in normoxic conditions for both groups and lasted 60 min. SaO2 and heart rate were continuously monitored

during each session. NG values allowed to differentiate hypoxia from training effects.

Biochemical analysis

Blood was collected in sitting position from an antecubital vein in three 5 ml ethylenediaminetetraacetic acid tubes for biochemical (plasma) and hematological (total blood) ana-lysis. The plasma was obtained by centrifugation of the samples at 1000 g for 10 min at 41C. Plasma was separated into aliquots and frozen at 801C for biochemical analyses. Before (at rest) and immediately after the two ergocycle tests in control period, and at rest before and 5 days after the last training session (pre-training and post-training samples) were measured:

1. Oxidative stress markers: plasma AOPP and plasma-oxidized lipids measured as MDA.

2. Pro-oxidant status markers: plasma lipid oxidizability. 3. Antioxidant markers: plasmatic ferric-reducing

antioxi-dant power (FRAP), lipid-soluble antioxiantioxi-dants (a-toco-pherol, and b-carotene).

4. Blood lipids: plasmatic triacyglycerols (TGs) and total cholesterol (CH).

5. Hemoglobin concentration ([Hb]) and hematocrit (Ht).

Reagent

All reagents were purchased from Sigma Chemical Co. (Sigma-Aldrich, St Quentin Fallavier, France).

Plasma AOPP

AOPP were determined in the plasma using the semiauto-mated method described by Witko-Sarsat et al. (1996). Briefly, AOPP were measured by spectrophotometry on a microplate reader (Benchmark, Bio-Rad Laboratories, Her-cules, CA, USA) using Microplate Manager software (Bio-Rad laboratories, Hercules, CA, USA) and were calibrated with chloramine-T solutions that in a presence of potassium iodide absorb at 340 nm. In test wells, 200 ml of plasma diluted 1/5 in phosphate-buffered saline (PBS) mixed with 20 ml of acetic acid was placed on a 96-well microliter plate (Nunc, Roskilde, Denmark). In standard wells, 20 ml of acetic acid was added to 200 ml chloramine-T solution

(4)

(0–200 mmol l1_{) mixed with 10 ml of 1.16}

M potassium iodide. The absorbance of the reaction mixture was immediately read at 340 nm on the microplate reader against a blank containing 200 ml of PBS, 10 ml of potassium iodide and 20 ml of acetic acid. AOPP concentrations were expressed as micromoles per liter of chloramine-T equivalents.

Plasma MDA

Concentrations of basal plasma MDA were determined as thiobarbituric-reactive substances by a modified method of Ohkawa et al. (1979) as previously described (Rayssiguier et al., 1993). The pink chromogen was extracted with n-butanol (4 ml), and its absorbance was measured at 532 nm by spectrophotometry (Perkin Elmer Cetus, Norwalk, CT, USA). MDA concentration was calculated using 1,1,3,3-tetraethoxypropane as standard. Plasma lipid were subjected to oxidation by incubation with FeSO4 (2 mmol/l)-ascorbate

(50 mmol/l) for 30 min in a 371C water-bath in an oxygen-free environment. The amount of MDA newly formed was measured as above. The difference in MDA concentration between before (basal level) and after in vitro-induced oxidation as been referred as plasma lipid peroxidizability.

Ferric reducing antioxidant power

Plasma FRAP concentrations were assessed according the manual Benzie and Strain (1996) method and were measured by spectrophotometry (Uvikon). FRAP concentration was calculated using an aqueous solution of known Fe2 þ concentration (FeSO4, 7H2O2) as standard at a wavelength

of 593 nm.

Lipid-soluble antioxidants

a-Tocopherol and b-carotene were analyzed by reversed-phase HPLC (Lyan et al., 2001) on a Waters (Milford, MA, USA) apparatus equipped with a 600 pump, a 710 automatic injector, and a 996 diode-array detector and controlled by MILLENNIUM 2.1 (Millipore Waters Chromatography, Millipore, France).

Tocopherol acetate and echinenone (Sigma Chemical Co., St Louis, MO, USA) were added to samples as internal standards (for a-tocopherol and b-carotene, respectively). They were then extracted twice with hexane, after ethanol precipitation of the proteins. This extract was evaporated to dryness under N2 and dissolved in

methanol-dichloro-methane (65:35, v/v). The samples were eluted on a Nucleosil 150 4.6-mm C18column (Interchim, Montlucon,

France) coupled with a Vydac-TP 250 4.6-mm (Interchim) with acetonitrile, methanol, dichloromethane, and ultra-pure water (70:15:10:5, v/v) as the mobile phase (2 ml/min); detection was performed at 292 and 450 nm for a-tocopherol and b-carotene, respectively. Identification was based on co-elution with authentic standards and ultraviolet light-spectrum comparisons. The sample loop size was 80 ml.

Quantification involved internal standardization and dose– response curves established with authentic standards.

TGs and total CH

TGs and CH were assayed in plasma by using enzymatic colorimetric methods with commercial kits (Biomerieux, Craponne, France). The concentrations were measured spectrophotometrically (Hitachi U 2001; Hitachi Instru-ments, Inc, San Jose, CA, USA) at 505 and 500 nm for TG and CH, respectively.

Hematological analysis

Hb and Ht were determined on total blood using a Pentra 120 Analyser (ABX, Montpellier, France). The relative changes in plasma volume (% PV) were calculated according to the following equation: DPV(%) ¼ 100[HbPre/HbPostt

(1HtPost 102)/(1HtPre 102)]100 (Dill and Costill,

1974) where the symbol Pre and Post represent Hb and Ht values before and after maximal exercise, respectively. In pre-training tests, the post-exercise plasma values were corrected for relative changes in PV using the following equation: Corrected values ¼ not corrected values [100DPV(%)]/100.

Statistics

Statistical analyses were performed using Statview software (5.0, SAS Institute Inc., USA). The values were expressed as means7s.d. The responses to maximal exercises were assessed using a two-factor (state (rest vs exercise) condi-tion (normoxia vs hypoxia)) repeated-measure analysis of variance (ANOVA) followed by a paired Student’s t-test. The responses to the type of training (normoxia vs hypoxia) were analyzed using a two-way (group: normoxic vs hypoxia and time: before vs after training) ANOVA. The values of the two groups were a posteriori analyzed with a Bonferroni test in order to compare the intermittent hypoxic exposure effect during training. Z-Fischer’s test correlations were also used to analyze the relationship between two qualitative variables. The level of statistical significance was set at Po0.05 for all analyses.

Results

Maximal cycling exercise under normobaric hypoxia and normoxia conditions in control period

Bioenergetic responses. Table 1 shows the data obtained during both normoxic and hypoxic cycling exercises. As compared to normoxia, hypoxic conditions decreased VO2max (15%, Po0.0001), HRmax (3%, P ¼ 0.002), SaO2

(13%, Po0.0001), power output (17%, Po0.0001) and time duration (10%, Po0.0001). On the contrary, [La] was increased ( þ 16%, P ¼ 0.02).

(5)

Hematological responses. In both conditions, acute exercise increased Hb (from 15.1_{70.6 to 16.371.0 g/dl, Po0.001 in} normoxia and from 15.171.0 to 16.370.6 g/dl, Po0.001 under hypoxia) and Ht (from 44.472.0 to 47.872.9%, Po0.001 on normoxia and from 44.472.8 to 48.271.7%, Po0.001 under hypoxia).

Acute exercise reduced plasma volume by the same magnitude in both conditions: 13.077.6% (Po0.001) in normoxia and 12.7710.3% (Po0.001) under hypoxia. Biochemical responses. All the parameters (AOPP, MDA, FRAP, TGs, CH and lipid-soluble antioxidants) measured at rest were not significantly different before the two maximal cycling tests.

Plasma AOPP

No significant changes in AOPP were observed after normoxic exercise whereas hypoxic exercise induced a

significant increase as compared to the basal level with (26%, P ¼ 0.04) or without ( þ 44%, P ¼ 0.001) plasmatic volume corrections (Table 2). Moreover, the post-exercise hypoxic values were significantly higher than the corre-sponding normoxic ones ( þ 41%, P ¼ 0.04 without correc-tion and þ 50%, P ¼ 0.04 with correction for plasma changes).

Plasma MDA

As for plasma AOPP, no significant changes in basal MDA were observed after normoxic exercise whereas hypoxic exercise induced a significant rise of 56% of the basal level (Table 2). The post-exercise concentrations were significantly higher in hypoxia than in normoxia ( þ 52%, P ¼ 0.01). The plasma lipid peroxidizability increased (from 0.0570.20 to 0.91_{70.29 mmol l}1_{, P ¼ 0.03) during normoxic exercise}

whereas the hypoxic exercise did not induce significant change (from 0.0170.35 to 0.2570.23 mmol l1) (Figure 2). The same results were obtained when the values were corrected for plasma changes: for normoxic exercise from 0.05_{70.20 to 0.8970.34 mmol l}1_{(P ¼ 0.03) and for hypoxic}

exercise from 0.0370.35 to 0.0370.57 mmol l1_.

Plasma FRAP

No significant changes of FRAP were observed after both types of exercise with or without plasma volume corrections (Table 2).

Plasma TGs and CH

Normoxic exercise increased TG ( þ 30%, Po0.001) and CH ( þ 14%, P ¼ 0.005) concentrations whereas no changes were observed after hypoxic exercise (Table 2). When the values were corrected for DPV(%), only TG values remained significant ( þ 16%, P ¼ 0.02) after normoxic exercise.

Table 2 Effects of acute hypoxia during cycling exercise on AOPP, MDA, FRAP, selected lipids-soluble non-enzymatic antioxidant markers, TG and CH

before training

Normoxia Hypoxia

Rest Post-exercise Post-exercise (corrected) Rest Post-exercise Post-exercise (corrected)

APPO (mmol l1₎ ₁₄₈₇₈₄ ₁₉₆₇₄₃ ₁₆₀₇₃₅ ₁₉₂₇₆₂ _277769**w 243759*w MDA (mmol l1₎ _1.64_70.40 _1.73_70.46 _1.42_70.59 _1.69_70.49 _2.63_70.67**ww 2.3170.58*w FRAP (mmol l1₎ ₇₈₂₇₁₇₂ ₈₅₁₇₂₁₂ ₇₀₃₇₁₆₁ ₇₂₇₇₁₇₃ ₇₈₉₇₁₅₇ ₆₉₂₇₁₀₂ TG (mmol l1) 0.8770.29 1.1470.39*** 1.0170.28* 1.0370.42 1.0570.41 0.8870.28 CH (mmol l1₎ _5.48_70.87 _6.25_71.10** _5.53_70.92 _5.73_71.21 _6.03_71.15 _5.46_71.21 a-Tocopherol (mmol l1) 24.573.4 30.575.0*** 26.474.3* 26.973.4 31.774.8** 27.874.9

a-Tocopherol/TG (mmol mol1_TG) _31.3_710.8 _29.3_78.8 _29.3_78.8 _30.3_712.5 _33.3_710.4*w

33.3710.4*w

a-Tocopherol/CH (mmol mol1_CH) _4.60_71.00 _4.94_70.88 _4.94_70.88 _4.80_70.74 _5.38_71.00**w

5.3871.00**w

b-Carotene (mmol l1₎ _0.8470.31 _0.8270.35 _0.8070.35 _0.8070.40 _0.8470.37 _0.8370.41

b-Carotene/TG (mmol mol1_TG) _0.97_70.29 _0.72_70.47 _0.72_70.47 _0.78_70.41 _0.80_70.32 _0.80_70.32

Abbreviations: AOPP, advanced oxidation protein products; CH, cholesterol; FRAP, ferric-reducing antioxidant power; MDA, malondialdehydes; TG, triacyglycerols. Values are means7s.d. of 14 subjects. Post-exercise values were corrected for plasma volume shifts evaluated according to Dill and Costill’s equation (Dill and Costill, 1974). Significance of differences: *Po0.05, **Po0.01 and ***Po0.001 as compared with their corresponding resting values andw_Po0.05,ww_{Po0.01 as compared}

with their corresponding normoxic values.

Table 1 Bioenergetic responses to maximal cycling exercise performed

under normoxia and normobaric hypoxia conditions during the control period

Parameters Normoxic exercise Hypoxic exercise

FIO2(%) 0.21 0.14

Time duration (min) 34.771.7 31.871.8***

VO2(ml STPD min1) 42967603 36357623***

Respiratory exchange ratio 1.0670.05 1.1270.10*

VE (l BTPS min1) 141726 131716

Heart rate (b min1₎ ₁₇₂₇₁₂ ₁₆₇_711**

Lactate (mmol l1) 9.673.6 11.474.3*

SaO2(%) 95.271.6 82.674.1***

Power output (W) 357750 298742***

VE, ventilation; VO2, maximal oxygen uptake; SaO2, ear arterial hemoglobin

oxygen saturation, BTPS, body temperature and pressure, saturated; STPD, standard temperature and pressure, dry.

Values are means7s.d. for 14 subjects. Significance of differences: *Po0.05, **Po0.01 and ***Po0.001 compared with normoxic exercise.

(6)

Lipids-soluble antioxidants

The maximal hypoxic and normoxic cycling tests induced significant increases of the same magnitude for plasmatic concentrations in a-tocopherol (24.4%, Po0.001 in normox-ia and 17.8%, Po0.002 under hypoxnormox-ia) (Table 2). When a-tocopherol values were corrected for plasma volume shift, post-exercise values remained significantly elevated in normoxia ( þ 10.5%, P ¼ 0.03) but not under hypoxia. Nevertheless, when a-tocopherol values were expressed relative to the concentrations of TG and CH, only hypoxic exercise induced significant changes ( þ 12% in a-tocopher-ol/CH, P ¼ 0.01 and þ 10% in a-tocopherol/TG, P ¼ 0.04) whereas the normoxic exercise values remained unchanged. Moreover, the post-exercise concentrations of a-tocopherol/ CH and a-tocopherol/TG were significantly higher in hypoxia than in normoxia ( þ 9%, P ¼ 0.02 and 14%,

P¼ 0.02, respectively). b-Carotene did not significantly change regardless the conditions.

Correlations

In post-exercise, a-tocopherol and TG concentrations were positively correlated (r ¼ 0.56, P ¼ 0.02 in normoxia and r¼ 0.69, P ¼ 0.02 under hypoxia).

Individual variation arterial oxygen hemoglobin desatura-tion was highly correlated (r ¼ 0.74, Po0.001) with the change in lipid oxidation during exercise (difference of basal MDA after exercise and basal MDA before exercise) (Figure 3).

Effect of endurance training in intermittent hypoxia

Performance tests. VO2max in normoxia determined on

tread-mill increased in HG (64.074.2 vs 67.073.3 ml min1kg1,

−1 −0.5 0.5 1 2 1.5

*

0 Pre

Pre Post Post

Hypoxic test Normoxic test

Lpid per oxidizability (µmol.l

-1)

Post corrected Post corrected

*

Figure 2 Lipid peroxidizability responses to maximal cycling exercise under hypoxic and normoxic conditions during the control period. Values

are means7SD for 14 subjects. Post-exercise values were corrected for plasma volume shifts evaluated according to Dill and Costill’s equation (Dill and Costill, 1974). Significance of differences: *Po0.05 and compared with pre-exercise values.

y = 7.1 x + 6.6 r = 0.73 P<0.0001 0 5 10 15 20 25 -1 -0.5 0 0.5 1 1.5 2 Decrease in SaO 2 (%)

Post-exercise MDA minus pre-exercise MDA (µmol.l-1₎

Figure 3 Relation between arterial desaturation (decrease in SaO2) and MDA changes after cycling exercise under hypoxic and normoxic

conditions during the control period.

(7)

P¼ 0.02) whereas no changes were obtained in NG (59.4_{72.4 vs 60.373.3 ml min}1_kg1_). _Regardless _the

group, the training had no effect on maximal heart rate (174710 vs 17479 b min1 for HG and 186710 vs 187711 b min1 for NG) and respiratory exchange ratio (1.0470.05 vs 1.0470.06 for HG and 1.0570.07 vs 1.06_{70.10 for NG) at the end of this VO}2max test

determined in normoxia.

Biochemical responses Plasma AOPP

In both groups (HG and NG), the endurance training did not change resting plasma AOPP (Table 3).

Plasma MDA

At rest, the endurance training reduced plasma basal MDA (before vs after training) by the same magnitude for NG (19%) and HG (16%) (i.e. 17%, P ¼ 0.01 for NG and HG pooled values). Only normoxic training significantly decreased plasma lipid peroxidizability (50%, P ¼ 0.03) (Table 3).

Plasma FRAP

Six weeks of training with intermittent hypoxia exposure induced a decrease in resting antioxidant power (21%: 7957194 vs 597786 mmol l1, P ¼ 0.05) whereas FRAP values were not altered by the normoxic training (7027171 vs 7467132 mmol l1) (Figure 4). Moreover, FRAP

Table 3 Effects of training associated (HG) or not (NG) with hypoxia on AOPP, MDA, FRAP lipid peroxidizability, selected lipids-soluble non-enzymatic

antioxidant markers, TG and CH measured at rest

Training Pre-training Post-training

NG HG NG HG

AOPP (mmol l1) 137744 156735 167752 123725

MDA (mmol l1₎ _1.57_70.30 _1.63_70.49 _1.33_70.16* _1.32_70.13*

lipid peroxidizability (mmol l1) 2.1670.75 1.9570.93 1.0970.67* 1.1470.79

TG (mmol l1₎ _0.87_70.15 _0.76_70.36 _0.97_70.24 _1.02_70.42

GH (mmol l1₎ _5.45_70.68 _5.72_71.11 _6.14_71.50 _6.39_72.35

FRAP (mmol l1₎ _7027171 _7957194 _7467132 _597786*w

a-Tocopherol (mmol l1₎ _24.7_73.8 _24.3_74.0 _29.7_76.2 _27.9_75.4

a-Tocopherol/TG (mmol mol1TG) 28.7574.98 37.91714.16 31.9979.54 28.7578.53*w

b-Carotene (mmol l1₎ _0.83_70.26 _0.98_70.40 _1.17_70.62 _1.29_70.83

b-Carotene/TG (mmol mol1TG) 0.9570.57 1.2970.76 1.2170.59 1.2670.81

Abbreviations: AOPP, advanced oxidation protein products; CH, cholesterol; FRAP, ferric-reducing antioxidant power; HG, hypoxic group; MDA, malondialdehydes; NG, normoxic group; TG, triacyglycerols.

Values are means7s.d. of 14 subjects. Significance of differences: *Po0.05 as compared with their corresponding values determined before training andw

Po0.05 when HG was compared with NG values at the same period.

500 550 600 650 700 750 800 850 900 950 1000 FRAP (mmol.l -1)

Pre-training Post-training Pre-training Post-training

HG (n=8) NG (n=6)

∗

Figure 4 Resting FRAP values before and after training under hypoxic (HG) and normoxic (NG) conditions. Values are means7SD for six

subjects for NG and eight subjects for HG. Significance of differences between pre- and post-training values: *Po0.05.

(8)

was significantly lower (P ¼ 0.04) in HG than in NG after the training whereas we did not find any differences before the training.

Plasma TGs and CH

In resting condition, no changes were observed after 6 weeks of training in TG and CH concentrations for NG and HG (Table 3).

Lipids-soluble antioxidants

After training, a-tocopherol/TGs ratio was significantly reduced (24%, P ¼ 0.04) in HG whereas NG values remained unchanged (Table 3). Regardless the training conditions (hypoxia or normoxia), the training had no effect on resting b-carotene/TG ratio.

Correlation

The TGs and a-tocopherol individual training variations were significantly correlated (r ¼ 0.56, P ¼ 0.05) when the values of both groups were pooled.

Discussion

In accordance with our hypothesis, our data give evidence that the association of two oxidative stimuli (exercise and hypoxia) emphasized the ROS generation and the subse-quent lipoperoxidation and protein oxidation markers. As a consequence, the repetition of such exercises performed under hypoxia during endurance training weakened the antioxidant capacities.

Effects of maximal exercise under normobaric hypoxia and normoxia conditions

Our results did not show basal MDA and AOPP changes after maximal exercise in normoxic conditions despite the number of studies reporting a rise of markers lipid oxidation (Sen, 1995) or protein oxidation (Radak et al., 2003) in the plasma after strenuous exercise. This discrepancy may be due to the higher training level of our subjects who might improve their pro-oxidant/antioxidant balance. Indeed, Yagi (1992) reported that blood lipid peroxide decreased in response to an acute exercise with increasing training duration. Similarly, Toskulkao and Glinsukon (1996) failed to find in long-distance runners, a plasma MDA increase after an exhaustive cycling exercise. It has been demon-strated that the high-level antioxidant enzymatic activities reached in responses to intense training (Miyazaki et al., 2001) could remove harmful effect of ROS. In contrast, the maximal exercise performed under hypoxia induced a significant increase in plasma MDA ( þ 56%) and AOPP ( þ 44%). Cellular generation of ROS under hypoxia was reported previously in animals at rest (Hitka et al., 2003), and in humans at high altitude (Simon-Schnass, 1996; Jefferson et al., 2004) and in hypobaric chambers (Joanny et al., 2001).

Our results are also in line with Bailey et al. (2001b) who reported increased oxidative stress also evaluated by MDA during a maximal exercise performed under hypoxia.

Therefore, it would be hypothesized that the cumulative effect of exercise and hypoxia resulted in ROS overproduc-tion in the muscle which exceeded the endogenous enzy-matic antioxidant defense of the subjects. The increase in plasma basal MDA and AOPP may reflect the cellular oxidation from the muscular compartment during the exercise performed in hypoxia condition.

In normoxia, it has been assumed that the ‘electron leakage’ from the electron transport chain proportionally increases with the O2 flux. This assumption implied a

constant ratio between superoxide anions and molecular oxygen. Therefore the subsequent ROS may increase pro-portionally with oxygen consumption (Alessio, 1993; Kanter, 1994). Nevertheless, our data clearly indicated that the increase in mitochondrial O2 flux was not the only source

of oxidative stress during exercise as the maximal oxygen consumption was lower under hypoxia. On the other hand, an important finding was the strong correlation between arterial desaturation and lipid peroxidation products (r ¼ 0.74). Our results are consistent with the finding of Bailey et al. (2001b) who reported a significant positive correlation between arterial desaturation and lipid hydro-peroxides after an exercise performed under hypoxia (FiO2¼ 0.16). Thus, it is possible that arterial desaturation

induced decrease in mitochondrial O2flux. In mitochondria,

the lower oxygen molecule available to accept electrons from oxidative phosphorylation enhances reducing equivalents and resulting in reduction of molecular oxygen to anion superoxide (Kehrer and Lund, 1994; Mohanraj et al., 1998). Moreover, as our blood samplings were performed after the hypoxic exposure, the reoxygenation of hypoxic tissue may have enhanced the O2 flux in the mitochondria and the

subsequent free radical production, as described after ischemia reperfusion (Askew, 2002).

Surprisingly, the plasma a-tocopherol/TG ratio increased after hypoxic exercise and a-tocopherol concentration as well, but not TG. Such transient plasmatic vitamin E rise has been also described following normoxic (Pincemail et al., 1988) and hypoxic exercises (Bailey et al., 2001b). Kawai et al. (1994) have reported that after strenuous endurance ex-ercise, a-tocopherol changes were correlated to blood lipid variations. We also found significant relationships between a-tocopherol and TG concentrations in the plasma after the normoxic (r ¼ 0.56) and hypoxic (r ¼ 0.69) exercises. Thus, under hypoxic conditions, the increase in a-tocopherol/TG ratio could be explained by their higher mobilization from tissue in order to reduce the overgenerated ROS.

After exercise, the lipid peroxidizability increased in normoxia whereas it did not move under hypoxia. The lipid peroxidizability represents plasma potential to produce oxidized lipid as MDA after an in vitro-induced oxidation. Therefore, a decrease in plasmatic antioxidant capacity during normoxic exercise could explain the lipid peroxidizability

(9)

rise. Nevertheless, this assumption is not in line with FRAP and a-tocopherol values that were the same after normoxic and hypoxic exercises (Table 2). Another possibility is that the available quantity of substrates (i.e. the blood lipids) might limit the in vitro oxidation. This hypothesis is supported by the increase in TGs and CH after acute exercise in normoxic conditions (Table 2). Consequently, the higher plasma lipid peroxidizability after the normoxic exercise would be mainly explained by higher values of oxidizable lipids before the in vitro oxidation.

Consequences of repetitive exercise in hypoxia during training on antioxidant/pro-oxidant balance

In spite of the small number of subjects, the repetition of 12 intense exercises associated with hypoxia, similar to the one performed in control period induced a decrease in plasma FRAP (20%) and a-tocopherol/TG ratio (24%) whereas the training without hypoxia exposure had no effect on the antioxidant capacities (Figure 4 and Table 3).

By mobilizing the antioxidant defenses in view to neutralize ROS overgeneration during each training session performed under hypoxia, this ‘living low-training high’ model may have induced a decrease in antioxidant capa-cities. The improvement of antioxidant defenses that are usually observed after an endurance training performed at sea-level (Ohno et al., 1988; Robertson et al., 1991; Di Massimo et al., 2004) could be due to an increase in antioxidant enzyme efficacy (Ohno et al., 1988; Miyazaki et al., 2001) and to a less extent to a plasma rise of antioxidant nutrient such as a-tocopherol (Robertson et al., 1991). Nevertheless, from studies in human (Margaritis et al., 1997) and rats (Moran et al., 2004), it seems that the upregulation of antioxidant enzymes activities following an aerobic training could depend on the pre-training level and on the duration the training. Indeed, highly trained athletes seem to exhibit antioxidant enzymes activities close to their maximal level (Dernbach et al., 1993) that could justify the lack of antioxidant adaptations in NG.

On the other hand, the HG results are consistent with several studies showing that antioxidant efficiency was reduced after training at moderate altitude (3000 m) (Vasankari et al., 1997; Chao et al., 1999; Subudhi et al., 2001). Pre-training tests give evidence that strenuous exercise performed under hypoxia magnified oxidative stress as compared to the same normoxic exercise. The repetition of such stimuli during training could reduce the antioxidant pool especially if the non-enzymatic antioxidant dietary cannot match their utilization (Sen, 2001). It has been recently described that a-tocopherol and/or b-carotene depletion weakened antioxidant status (Winklhofer-Roob et al., 2003). Thus, these results suggest that our subjects may have insufficient intakes in antioxidant nutrient. The most common dietary strategy of endurance athletes (Burke et al., 2003) consists in increasing carbohydrate intakes and reducing lipids that are the most suppliers in lipids-soluble

antioxidants. We recently found in an another group of elite runners that vitamin A and E intakes were fivefold lower than dietary allowances recommended for endurance athletes (unpublished data).

To conclude, this study provides strong evidence that 6 weeks of ‘living low-training high’ model worsened anti-oxidant capacities. The exposure of two independent stimuli (exhausting exercise and hypoxia) may have a cumulative effect on oxidative stress. As a consequence, the repetition of antioxidant demand during such training sessions outmatched the antioxidant intakes. Therefore, further researches are necessary to investigate whether an antio-xidant supplementation could maintain antioxidant capacities in highly trained athletes submitted to hypoxia during intense training sessions.

Acknowledgements

We thank the subjects for their contribution. We also thank Eric Clottes for reviewing the manuscript and Clark Ellice and English review of the manuscript.

This study was funded by the ‘International Olympic Committee’ and the ‘Direction Re´gionale de la Jeunesse et des Sports de la Re´gion Auvergne’.

References

Alessio HM (1993). Exercise-induced oxidative stress. Med Sci Sports Exerc 25, 218–224.

Askew EW (2002). Work at high altitude and oxidative stress: antioxidant nutrients. Toxicology 180, 107–119.

Bailey DM, Davies B, Romer L, Castell L, Newsholme E, Gandy G (1998). Implications of moderate altitude training for sea-level endurance in elite distance runners. Eur J Appl Physiol Occup Physiol 78, 360–368.

Bailey DM, Davies B, Young IS (2001b). Intermittent hypoxic

training: implications for lipid peroxidation induced by

acute normoxic exercise in active men. Clin Sci (London) 101, 465–475.

Bailey DM, Davies B, Young IS, Hullin DA, Seddon PS (2001a). A potential role for free radical-mediated skeletal muscle soreness in the pathophysiology of acute mountain sickness. Aviat Space Environ Med 72, 513–521.

Benzie IF, Strain JJ (1996). The ferric reducing ability of plasma (FRAP) as a measure of ‘antioxidant power’: the FRAP assay. Anal Biochem 239, 70–76.

Burke LM, Slater G, Broad EM, Haukka J, Modulon S, Hopkins WG (2003). Eating patterns and meal frequency of elite Australian athletes. Int J Sport Nutr Exerc Metab 13, 521–538.

Chao WH, Askew EW, Roberts DE, Wood SM, Perkins JB (1999). Oxidative stress in humans during work at moderate altitude. J Nutr 129, 2009–2012.

Clarkson PM, Thompson HS (2000). Antioxidants: what role do they play in physical activity and health? Am J Clin Nutr 72 (2 Suppl), 637–646.

Dernbach AR, Sherman WM, Simonsen JC, Flowers KM, Lamb DR (1993). No evidence of oxidant stress during high-intensity rowing training. J Appl Physiol 74, 2140–2145.

Di Massimo C, Scarpelli P, Penco M, Tozzi-Ciancarelli MG (2004). Possible involvement of plasma antioxidant defences in

(10)

training-associated decrease of platelet responsiveness in humans. Eur J Appl Physiol 91, 406–412.

Dill DB, Costill DL (1974). Calculation of percentage changes in volumes of blood, plasma, and red cells in dehydration. J Appl Physiol 37, 247–248.

Emonson DL, Aminuddin AH, Wight RL, Scroop GC, Gore CJ (1997).

Training-induced increases in sea level VO2max and endurance are

not enhanced by acute hypobaric exposure. Eur J Appl Physiol Occup Physiol 76, 8–12.

Hitka P, Vizek M, Wilhelm J (2003). Hypoxia and reoxygenation

increase H2O2production in rats. Exp Lung Res 29, 585–592.

Jefferson JA, Simoni J, Escudero E, Hurtado ME, Swenson ER, Wesson DE et al. (2004). Increased oxidative stress following acute and chronic high altitude exposure. High Alt Med Biol 5, 61–69. Ji LL (1996). Exercise, oxidative stress, and antioxidants. Am J Sports

Med 24, S20–24.

Joanny P, Steinberg J, Robach P, Richalet JP, Gortan C, Gardette B et al. (2001). Operation everest III (Comex’97): the effect of simulated sever hypobaric hypoxia on lipid peroxidation and antioxidant defence systems in human blood at rest and after maximal exercise. Resuscitation 49, 307–314.

Kanter MM (1994). Free radicals, exercise, and antioxidant supple-mentation. Int J Sport Nutr 4, 205–220.

Kawai Y, Iwane H, Takanami Y, Shimomitsu T, Katsumura T, Fujinami J (1994). Vitamin E is mobilized in relation to lipolysis after strenuous endurance exercise. Med Sci Sports Exerc 26, S27. Kehrer JP, Lund LG (1994). Cellular reducing equivalents and

oxidative stress. Free Radical Biol Med 17, 65–75.

Levine BD, Stray-Gundersen J (1992). A practical approach to altitude training: where to live and train for optimal performance enhancement. Int J Sports Med 13, S209–S212.

Levine BD, Stray-Gundersen J, Duhaime G, Schell PG, Friedman DB (1991). ‘Living high-training low’: the effect of altitude acclima-tization/normoxic training in trained runners. Med Sci Sports Exerc 23, S25.

Lyan B, Azais-Braesco V, Cardinault N, Tyssandier V, Borel P, Alexandre-Gouabau MC et al. (2001). Simple metho for clinical determination of 13 carotenoids in human plasma using an isocratic high-performance liquid chromatographic method. J Chromatogr B Biomed Sci Appl 751, 297–303.

Margaritis I, Tessier F, Richard MJ, Marconnet P (1997). No evidence of oxidative stress after a triathlon race in highly trained competitors. Int J Sports Med 18, 186–190.

Mazzeo RS, Child A, Butterfield GE, Mawson JT, Zamudio S, Moore LG (1998). Catecholamine response during 12 days of high-altitude exposure (4300 m) in women. J Appl Physiol 84, 1151–1157. Miyazaki H, Oh-ishi S, Ookawara T, Kizaki T, Toshinai K, Ha S et al. (2001). Strenuous endurance training in humans reduces oxida-tive stress following exhausting exercise. Eur J Appl Physiol 84, 1–6. Mohanraj P, Merola AJ, Wright VP, Clanton TL (1998). Antioxidants protect rat diaphragmatic muscle function under hypoxic condi-tions. J Appl Physiol 84, 1960–1966.

Moller P, Loft S, Lundby C, Olsen NV (2001). Acute hypoxia and hypoxic exercise induce DNA strand breaks and oxidative DNA damage in humans. FASEB J 15, 1181–1186.

Moran M, Delgado J, Gonzalez B, Manso R, Megias A (2004). Responses of rat myocardial antioxidant defences and heat shock protein. HSP72 induced by 12 and 24-week treadmill training. Acta Physiol Scand 180, 157–166.

Ohkawa H, Ohishi N, Yagi K (1979). Assay for lipid peroxides in animal tissues by thiobarbituric acid reaction. Anal Biochem 95, 351–358.

Ohno H, Yahata T, Sato Y, Yamamura K, Taniguchi N (1988). Physical training and fasting erythrocyte activities of free radical scaven-ging enzyme systems in sedentary men. Eur J Appl Physiol Occup Physiol 57, 173–176.

Pincemail J, Deby C, Camus G, Pirnay F, Bouchez R, Massaux L et al. (1988). Tocopherol mobilization during intensive exercise. Eur J Appl Physiol Occup Physiol 57, 189–191.

Radak Z, Ogonovszky H, Dubecz J, Pavlik G, Sasvari M,

Pucsok J et al. (2003). Super-marathon race increases serum and urinary nitrotyrosine and carbonyl levels. Eur J Clin Invest 33, 726–730.

Rayssiguier Y, Gueux E, Bussiere L, Mazur A (1993). Copper deficiency increases the susceptibility of lipoproteins and tissues to peroxidation in rats. J Nutr 123, 1343–1348.

Robertson JD, Maughan RJ, Duthie GG, Morrice PC (1991). Increased blood antioxidant systems of runners in response to training load. Clin Sci (London) 80, 611–618.

Sen CK (1995). Oxidants and antioxidants in exercise. J Appl Physiol 79, 675–686.

Sen CK (2001). Antioxidants in exercise nutrition. Sports Med 31, 891–908.

Sen CK, Khanna S, Reznick AZ, Roy S, Packer L (1997). Glutathione regulation of tumor necrosis factor-alpha-induced NF-kappa B activation in skeletal muscle-derived L6 cells. Biochem Biophys Res Commun 28, 645–649.

Simon-Schnass I (1996). Oxidative stress at high altitude and effects of vitamin E. In: Marriott BM, Carlson SJ (eds). Nutritional Needs in Cold and in High Altitude Environments. National Academy Press: Washington DC, pp 393–418.

Subudhi AW, Davis SL, Kipp RW, Askew EW (2001). Antioxidant status and oxidative stress in elite alpine ski racers. Int J Sport Nutr Exerc Metab 1, 32–41.

Toskulkao C, Glinsukon T (1996). Endurance exercise and muscle damage: relationship to lipid peroxidation and scavenging enzymes in short and long distance runners. Jpn Phys Fitness Sports Med 46, 63–70.

Vasankari TJ, Kujala UM, Rusko H, Sarna S, Ahotupa M (1997). The effect of endurance exercise at moderate altitude on serum lipid peroxidation and antioxidative functions in humans. Eur J Appl Physiol Occup Physiol 75, 396–399.

Wasserman K (1987). Determinants and detection of anaerobic threshold and consequences of exercise above it. Circulation 76, VI29–VI39.

Wing SL, Askew EW, Luetkemeier MJ, Ryujin DT, Kamimori GH, Grissom CK (2003). Lack of effect of Rhodiola or oxygenated water supplementation on hypoxemia and oxidative stress. Wilderness Environ Med 14, 9–16.

Winklhofer-Roob BM, Khoschsorur G, Meinitzer A, Maritschnegg M, Hiller D; Wuga S, Wonisch W et al. (2003). Effects of vitamin E depletion/repletion on vitamin E status and oxidative stress in healthy volunteers. Clin Nutr 22, S33.

Witko-Sarsat V, Friedlander M, Capeillere-Blandin C, Nguyen-Khoa T, Nguyen AT, Zingraff J et al. (1996). Advanced oxidation protein products as a novel marker of oxidative stress in uremia. Kidney Int 49, 1304–1313.

Yagi K (1992). Lipids peroxides and exercise. Med Sport Sci 37, 40–42.