Content uploaded by Fergal Grace
Author content
All content in this area was uploaded by Fergal Grace on Nov 10, 2015
Content may be subject to copyright.
Clinical Science (2006) 110, 133–141 (Printed in Great Britain) doi:10.1042/CS20050135 133
Manipulation of systemic oxygen flux by acute
exercise and normobaric hypoxia: implications
for reactive oxygen species generation
Gareth W. DAVISON
∗
, Rhian M. MORGAN†, Natalie HISCOCK‡,JuanM.GARCIA†,
Fergal GRACE†, Natalie BOISSEAU‡,BruceDAVIES†, Linda CASTELL‡,
Jane M
CENENY§, Ian S. YOUNG§, David HULLIN, Tony ASHTON¶
and Damian M. BAILEY †
∗
School of Health Sciences, University of Ulster Jordanstown, Newtownabbey, County Antrim BT37 OQB, U.K., †School of
Applied Sciences, University of Glamorgan, Pontypridd CF37 1DL, Wales, U.K., ‡Department of Biochemistry, Oxford
University, South Parks Road, Oxford OX1 3QU, U.K., §Department of Medicine, Queen’s University Belfast, Mulhouse,
Grosvenor Road, Belfast BT12 6BJ, U.K., Department of Clinical Biochemistry, Royal Glamorgan Hospital, Llantrisant
CF72 8XR, Wales, U.K., and ¶Department of Sport Science, Bedford Faculty, De Montfort University, Bedford MK40 2BZ, U.K.
ABSTRACT
Maximal exercise in normoxia results in oxidative stress due to an increase in free radical
production. However, the effect of a single bout of moderate aerobic exercise performed in either
relative or absolute normobaric hypoxia on free radical production and lipid peroxidation remains
unknown. To examine this, we randomly matched {according to their normobaric normoxic
˙
V
O
2peak
[peak
˙
VO
2
(oxygen uptake)]} and assigned 30 male subjects to a normoxia (n = 10),
a hypoxia relative (n = 10) or a hypoxia absolute (n = 10) group. Each group was required to
exercise on a cycle ergometer at 55 % of
˙
VO
2peak
for 2 h double-blinded to either a normoxic or
hypoxic condition [F
i
O
2
(inspired fraction of O
2
) = 0.21 and 0.16 respectively]. ESR (electron
spin resonance) spectroscopy in conjunction with ex vivo spin trapping was utilized for the
direct detection of free radical species. The main findings show that moderate intensity exercise
increased plasma-volume-corrected free radical and lipid hydroperoxide concentration (pooled
rest compared with exercise data, P < 0.05); however, there were no selective differences between
groups (state × group interaction, P > 0.05). The delta change in free radical concentration was
moderately correlated with systemic
˙
V
O
2
(r
2
= 0.48, P < 0.05). The hyperfine coupling constants
recorded from the ESR spectra [a
N
= 13.8 Gauss, and a
H
β
= 1.9 Gauss; where 1 Gauss = 10
−4
T
(telsa)] are suggestive of oxygen-centred free radical species formed via the decomposition of lipid
hydroperoxides. Peripheral leucocyte and neutrophil cells and total CK (creatine kinase) activity
all increased following sustained exercise (pooled rest compared with exerc ise data, P < 0.05), but
no selective differences were observed between groups (state × group interaction, P > 0.05). We
conclude that a single bout of moderate aerobic exercise increases secondary free radical species.
There is also evidence of exercise-induced muscle damage, possibly caused by the increase in free
radical generation.
Key words: antioxidant, electron spin resonance (ESR), exercise, free radical, hypoxia, oxidative stress.
Abbreviations: CV, coefficient of variation; ESR, electron spin resonance; F
i
o
2
, inspired fraction of oxygen; HR, heart rate; LH, lipid
hydroperoxide; MDA, malondialdehyde; PBN, α-phenyl-tert-butylnitrone; Po
2
, partial pressure of oxygen; PUFA, polyunsaturated
fatty acid; ROS, reactive oxygen species; Sao
2
, arterial oxygen saturation;
˙
Vo
2
, oxygen uptake;
˙
Vo
2max
, maximum
˙
Vo
2
;
˙
Vo
2peak
, peak
˙
Vo
2
.
Correspondence: Dr Gareth W. Davison (email gw.davison@ulster.ac.uk).
C
2006 The Biochemical Society
134 G. W. Davison and others
INTRODUCTION
During aerobic exercise, oxygen utilization is increased
which may lead to the incomplete reduction of oxygen
molecules in one or more mitochondrial complexes,
resulting in an extensive increase in oxidative stress
[1,2]. Oxidative stress can occur during aerobic exercise
not only via the mitochondria as a primary source, but
from other mechanisms such as substrate auto-oxidation,
xanthine oxidase activity, neutrophil activation, nitric
oxide synthesis and metal-catalysed reactions [3].
Paradoxically, a growing body of evidence suggests
that exercise performed in hypoxia may also stimulate
oxidative stress due primarily to a decrease in mito-
chondrial respiration and a build-up of reducing equival-
ents that cannot be transferred to molecular oxygen
at the level of cytochrome oxidase [4]. In addition to
this concept, known as reductive stress, other potential
sources of free radical generation in hypoxia include
increased nitric oxide production, xanthine oxidase and
phospholipase A
2
activation, and increased availability of
free Fe
2+
and Cu
2+
ions [2,5]. The influence of exercise
and hypoxia on lipid peroxidation has been explored by
a number of investigators [4,6,7], and work by Bailey
et al. [8] has shown an increase in LH (lipid hydro-
peroxide) and MDA (malondialdehyde) production
following maximal exhaustive exercise in normobaric
hypoxia. Other research by Bailey et al. [9] has shown that
4 weeks of exercise training in intermittent hypoxia can
attenuate the exercise-induced increase in these putative
biomarkers of lipid peroxidation more effectively than
normoxic training, suggesting that hypoxic exercise train-
ing may have influential molecular adaptive proper-
ties. These studies, however, used indirect indices of free-
radical-induced molecular damage, thus the claim that
hypoxia generates free radical species has not been con-
firmed in exercising humans. ESR (electron spin reson-
ance) spectroscopy, which is the most direct method of
measuring free radical molecules in conjunction with
the spin-trapping technique, has largely been used to
Table 1
Subject characteristics
Values are means
+
−
S.D. All groups were equally matched for the above
characteristics.
Group
Normoxia Hypoxia (relative) Hypoxia (absolute)
n
10 10 10
Age (years) 21
+
−
121
+
−
121
+
−
2
Height (m) 1.76
+
−
0.5 1.73
+
−
0.4 1.78
+
−
0.4
Body mass (kg) 75
+
−
13 72.4
+
−
8 73.1
+
−
9
Body fat (%) 17.4
+
−
415
+
−
6 15.8
+
−
5
˙
V
O
2peak
46.7
+
−
5 48.1
+
−
6 48.6
+
−
9
(ml · kg
−1
· min
−1
)
confirm the presence of free radical species in venous
blood of humans exercising to exhaustion in normoxia
[10–12]. To our knowledge there are currently no studies
that have used ESR spectroscopy to directly determine
the pro-oxidant effects of moderate aerobic exercise per-
formed in normobaric hypoxia. Therefore the purpose
of the present study was to quantify the degree of free-
radical-mediated oxidative stress in a single bout of
moderate normobaric hypoxic exercise. Furthermore, we
hypothesize that exercising at the same absolute workload
in hypoxia would result in an increased free radical
response compared with exercise in relative hypoxia.
METHODS
Subjects
Thirty apparently healthy male volunteers were recruited
from a student population to participate in the present
study. The subject characteristics are shown in Table 1.
Subjects were free of any diseases or ailments as assessed
by a medical history questionnaire prior to experimental
exercise. All subjects were non-smokers, and any sub-
jects taking antioxidant supplements were excluded.
Volunteers provided written informed consent prior to
participation, and the Local Medical Research Ethics
Committee (Bro Taff, Cardiff, South Wales, U.K.)
granted ethical approval.
Experimental design
Subjects were instructed to refrain from exercise and
alcohol for 48 h before all tests, and to maintain their
usual dietary pattern up until the last meal consumed
12 h before experimental exercise, where a standardized
meal-replacement drink (Wake-up cereal; Retail Brand)
was issued to all subjects. A quantity (0.32 g/kg of total
body weight) of cereal powder was mixed with 2.7 ml
of semi-skimmed milk per kg of total body weight by
the same investigator. Dietary composition and caloric
intake in the 72 h before the exercise test was recorded by
means of a food diary and was assessed using a commercial
nutritional assessment package (Nutri-check; Health
Options Limited). All subjects attended a familiarization
session 1 week prior to the commencement of the
incremental test protocol.
Incremental test protocol
On arrival at the laboratory, body mass and height of the
subjects were determined according to standard methods.
Body fat was measured using Harpenden skin fold cal-
lipers (British Indicators) and the equations as described
by Durnin and Womersley [13]. All incremental exercise
tests were performed between 09.00 and 17:00 hours at
the University of Glamorgan under the supervision of the
same investigators to minimize inter-analytical subject
variation. Each subject in a randomized double-blind
C
2006 The Biochemical Society
Exercise-induced oxidative stress 135
placebo-controlled fashion performed two incremental
cycling tests to volitional exhaustion, one in normobaric
normoxia [F
i
o
2
(inspired fraction of oxygen) = 0.21]
and the other in normobaric hypoxia (F
i
o
2
= 0.16).
Each test was separated by 7 days. The ergometer cycl-
ing test consisted of a 5 min warm up period at 0.5 kg
at 80 rev./min. The subject was then required to main-
tain 80 rev./min while 0.4 kg was applied to the basket
every 2 min (power output increase of 32 W/stage) until
volitional exhaustion. This protocol was chosen as it
had previously been validated to elicit
˙
Vo
2peak
[peak
˙
Vo
2
(oxygen uptake)] in a hypoxic environment [14].
Pre- and post-exercise blood lactate was measured using
an automated electrochemical analyser (Analox PGM7
Champion).
˙
Vo
2
was monitored during the last 60 s of
each stage and at the point of volitional exhaustion using
the off-line gas analysis Douglas bag system. HR (heart
rate) was recorded continuously throughout exercise
using a three-lead ECG system (Life pulse; HME). The
relationship between
˙
Vo
2
, power output (maximum
workload) and HR was subsequently determined for
each subject and was used to assess the individual exercise
intensity level for the 2 h exercise protocol.
Experimental protocol
All exercise tests were performed between 08.00 and
13.00 hours at the University of Glamorgan. Subjects
were randomly matched according to their normobaric
normoxia
˙
Vo
2peak
scores and r andomly assigned to one
of three groups prior to performing a single bout of
moderate exercise: group 1 (normoxia), subjects (n = 10)
performed 2 h of cycling exercise in normobaric nor-
moxia (F
i
o
2
= 0.21) at a workload corresponding to 55 %
of the pre-determined
˙
Vo
2peak
in normobaric normoxia;
group 2 [hypoxia (relative)], subjects (n = 10) performed
2 h of cycling exercise in normobaric hypoxia (F
i
o
2
=
0.16) at a workload corresponding to 55 % of the pre-
determined
˙
Vo
2peak
in normobaric hypoxia; and group 3
[hypoxia (absolute)], subjects (n = 10) performed 2 h of
cycling exercise in normobaric hypoxia (F
i
o
2
= 0.16) at a
workload corresponding to 55 % of the pre-determined
˙
Vo
2peak
in normobaric normoxia.
˙
Vo
2
was monitored
during the last 60 s of each time period. HR and RPE
(rate of perceived exertion) [15] were measured at selected
intervals throughout. Sao
2
(arterial oxygen saturation)
was determined using a finger pulse oximeter (Nonin
Model 8800) with a reported accuracy of
+
−
1%.
Blood sampling
Venous blood was collected following a 12 h overnight
fast from a prominent forearm vein after supine rest and
during the last minute of experimental exercise using
an intravenous cannula (Venflon IV cannula; Becton-
Dickinson). Following blood collection, EDTA vacutain-
ers were placed on ice, whilst blood in the serum separa-
tion tubes (SST) was allowed to clot at room temperature.
After centrifugation at 996 g at 4
◦
C for 10 min, the serum/
plasma aliquots were stored at − 80
◦
C. ESR–PBN (α-
phenyl-tert-butylnitrone) adduct analysis was completed
on the experimental day, whereas the remaining aliquots
were assayed within 6–8 weeks of collection. Post-
exercise blood samples were corrected for the possibility
of an exercise-induced haemoconcentration using the
equations as described by Dill and Costill [16].
Blood biochemistry
Oxidative stress
Free radicals were measured using methodology pub-
lished previously [11]. Briefly, toluene extracts of PBN-
trapped species were analysed at room temperature on
a Bruker EMX X-band ESR spectrometer with a 110
TM
cavity. All samples were vacuum degassed using liquid ni-
trogen and a turbo molecular pump (West Technologies).
ESR spectral peak height (average of six peaks) was used
as a measure of PBN spin adduct concentration present in
the biological sample. PBN contains a hydrogen atom β,
which forms due its interaction with an unpaired electron
in a manner that causes each of the three nitrogen lines to
split into doublets, resulting in a six line spectrum. The
distance between the two lines of the first doublet is
termed the β-hydrogen split (a
H
β
). The distance between
the first and second doublets (nitrogen component) is
termed a
N
. Nitrone values for a
N
and a
H
β
are called hyper-
fine coupling constants and aid in the identification of
the type of free radical species present in the biological
sample. Results are expressed in arbitrary units. Intra-
assay CV (coefficient of variation) at 1795 arbitrary units
was 5.2%.
MDA was measured by HPLC in EDTA plasma
using a modified method as described by Young and
Trimble [17]. Intra- and inter-assay CV at 0.56 µmol/l
was 6.2 and 9.1 % respectively. LH was measured spectro-
photometrically in serum using a modified method as
described by Wolff [18]. Intra- and inter-assay CV at
0.57 µmol/l was 4.6 and 6% respectively.
Antioxidant status
Ascorbic acid was measured in EDTA plasma with 5 %
metaphosphoric acid using the fl uorimetric method as
described by Vuilleumier and Keck [19]. The interassay
CV at a concentration of 51.1 µmol/l was 0.72 %. The
HPLC method as described by Thurnham et al. [20] was
used to simultaneously determine plasma lipid soluble
antioxidant status. Intra- and inter-assay CV were both
< 5%.
Muscle damage
Total CK (creatine kinase) was determined using a diag-
nostic kit and measured on a Vitros 750 analyser (Ortho-
Clinical Diagnostics), with an intra- and inter-assay CV
of 4.4 %. Myoglobin was analysed using an automated
C
2006 The Biochemical Society
136 G. W. Davison and others
Table 2
Physiological responses to incremental exercise in
acute hypoxia
Values are means
+
−
S.D.
∗
P
< 0.05 compared with the normoxia group. [La
−
]
B
,
corrected whole-blood lactate.
Parameter Normoxia (
F
i
O
2
= 0.20) Hypoxia (
F
i
O
2
= 0.16)
n
30 30
˙
V
O
2peak
(ml · kg
−1
· min
−1
) 47.8
+
−
730
+
−
8
∗
˙
V
O
2peak
(litres/min) 3.49
+
−
0.4 2.18
+
−
0.5
∗
Maximum workload (kg) 3.5
+
−
0.4 3.32
+
−
0.3
Maximum HR (beats/min) 195
+
−
9 193
+
−
8
[La
−
]
B
(mmol/l) 7.5
+
−
27.7
+
−
2
S
aO
2
(%) 96
+
−
388
+
−
5
∗
chemiluminescene immunoassay (ACS 180; Bayer-
Chiron Immunodiagnostics), with an intra- and inter-
assay CV of 3.5 % and 1.3% respectively.
White blood cells
Whole blood leucocytes and neutrophils cells were ana-
lysed using a COULTER
®
GEN.S
TM
automated haema-
tology analyser (Coulter).
Statistical analysis
Statistical analysis was performed using the SPSS social
statistics package (version 9.0). Prospective power calcu-
lations were calculated as described by Altman [21].
Retrospective power calculations were performed using
SPSS. Data were analysed using parametric statistics
following mathematical confirmation of a normal distri-
bution by Shapiro–Wilks tests. Independent sample
Student t tests were used to compare physiological
responses to incremental exercise between groups. A
one-way ANOVA was used to compare subject charac-
teristics. Experimental resting and exercise data were
analysed using two-way split plot [A × (B)] mixed
ANOVA, which incorporated one between [group:
normoxia compared with hypoxia (relative) compared
with hypoxia (absolute)] and one within (state: rest com-
pared with exercise) subjects factor. Following a signi-
ficant interaction effect (state × group), within-subject
factors were analysed using Bonferroni-corrected paired
samples Student t test. Between-subject differences were
analysed using a one-way ANOVA with a posteriori
Tukey HSD (honestly significant difference) test. The α
value was set at P < 0.05, and all values are reported as
means
+
−
S.D.
RESULTS
Dietary status
There was no significant difference in caloric intake
and macronutrient composition between groups, and all
values were within recommended UK daily range.
Preliminary tests
There was a difference in relative and absolute
˙
Vo
2peak
and Sao
2
between groups (P < 0.05) as shown in Table 2.
Experimental tests
As shown in Table 3,
˙
Vo
2
and HR increased (P < 0.05)
during s ubmaximal exercise. Sao
2
was lower in the
hypoxic groups (P < 0.05 compared with normoxia).
Oxidative stress indices
Figure 1 shows that exercise itself, but not selectively in
hypoxia, increased the concentration of free radicals
in venous blood (pooled rest compared with exercise data,
P = 0.004; state × group interaction, P = 0.12). Typical
Table 3
Physiological responses to sustained exercise in acute hypoxia
Values are means
+
−
S.D. for subjects in the normoxia (
n
= 10), hypoxia relative (
n
= 10) and hypoxia absolute (
n
= 10) groups. Main effect for state indicates a
difference (
P
< 0.05) between 5 min of exercise compared with 110 min of exercise (pooled normoxia and hypoxia values). Main effect for group indicates a difference
(
P
< 0.05) between normoxia and hypoxia groups.
Normoxia Hypoxia (relative) Hypoxia (absolute)
Parameter Time ... 5 min 100 min 5 min 100 min 5 min 100 min
˙
V
O
2
(litres/min) 1.6
+
−
02 1.9
+
−
0.3 1.5
+
−
0.3 1.9
+
−
0.2 1.4
+
−
0.2 2
+
−
0.3
Main effect for state
HR (beats/min) 112
+
−
8 137
+
−
16 111
+
−
12 142
+
−
10 123
+
−
19 151
+
−
23
Main effect for state
Normoxia Hypoxia (relative) Hypoxia (absolute)
Parameter Time ... 0 min 90 min 0 min 90 min 0 min 90 min
S
aO
2
(%) 97
+
−
196
+
−
193
+
−
390
+
−
392
+
−
387
+
−
4
Main effect for group
C
2006 The Biochemical Society
Exercise-induced oxidative stress 137
Figure 1
Rest and exercise PBN adduct concentration in the
normoxia and hypoxia groups
Main effect for state (pooled rest compared with exercise data,
P
= 0.004).
Figure 2
Typical rest (A) and (hypoxia) exercise (B) ESR
spectra of PBN adducts in serum
ESR spectra of the PBN adduct before and following
aerobic exercise are shown in Figure 2. For all ESR spectra
of the PBN adduct detected ex vivo in human sera, the
hyperfine coupling constants were measured at a
N
of
13.8 Gauss and a
H
β
of 1.9 Gauss [where 1 Gauss = 10
−4
T
(telsa)]. Based on these coupling constants, the free
radical species were identified as being secondary oxygen-
centred lipid radicals. A positive correlation was observed
between the delta (exercise−rest) PBN adduct and delta
˙
Vo
2
(r = 0.48, P < 0.05). Exercise increased circulating
LH concentration (pooled rest compared with exercise
data, P = 0.01), but not selectively between groups
(state × group interaction, P = 0.27). This latter finding
supports the supposition that the free radicals detected
are lipid in origin. In contrast, no between- or within-
group differences were observed for MDA concentration
(state × group interaction, P = 0.71).
Antioxidants
As shown in Table 4, exercise selectively decreased venous
lycopene concentration by 18.5% in the normoxia group
only (P = 0.05). Both α-andβ-carotene concentrations
increased following exercise (pooled rest compared with
exercise data, P = 0.004 and 0.01 for α-andβ-carotene
respectively). No between- or within-group differences
were observed for any other antioxidant parameter
(state × group interaction, P = 0.06, 0.43 and 0.97 for
ascorbic acid, α-tocopherol and retinol respectively).
White blood cells and muscle
damage markers
Peripheral leucocyte and neutrophil cell number
increased following sustained exercise (pooled rest com-
pared with exercise data, P = 0.000; state × group inter-
action, P = 0.72 and 0.34 for leucocytes and neutrophils
respectively); however, there were no differences between
normoxia and hypoxia groups (Table 5). Total CK activity
increased following exercise (pooled rest compared
with exercise data, P = 0.02; state × group interaction,
P = 0.50) as shown in Figure 3. No differences were
observed within or between groups for myoglobin activ-
ity (state × group interaction, P = 0.13) as shown in
Table 4.
DISCUSSION
Recent literature postulates that exercise performed in
hypoxia causes oxidative stress [8,9]. The present s tudy
used ESR spectroscopy to measure free radicals directly,
and demonstrates for the first time that a single bout of
moderate aerobic exercise in hypoxia (of an absolute and
relative nature) does not selectively (against normoxia)
increase free radical concentration. This finding is at
variance with previous reports that have quantified free
radical generation by measuring the by-products of mol-
ecular oxidation, and may be due, in part, to a limited
statistical power. In addition, other hypoxia studies have
largely used short-term exhaustive protocols [i.e. at or
near
˙
Vo
2max
(maximum
˙
Vo
2
)] to generate an oxidative
stress [8,9], whereas the exercise intensity was much lower
in the present study. Nevertheless, pooled group data
would suggest that 2 h of exercise at 55 %
˙
Vo
2peak
gener-
ates free radical species within the systemic circu-
lation to levels beyond the capacity of the antioxidant
defence system.
The hyperfine coupling constants of all ex vivo PBN-
trapped free radicals were consistently the same and are
C
2006 The Biochemical Society
138 G. W. Davison and others
Table 4
Effect of acute hypoxia on oxidative stress, muscle damage and antioxidant indices
Values are means
+
−
S.D.;
n
= 10 per group. Interaction effect (state × group) indicates a difference (
P
< 0.05) within group as a function of time; †within group
difference (
P
< 0.05). Main effect for state indicates a difference (
P
< 0.05) between rest compared with exercise (pooled normoxia and hypoxia values).
Metabolite Group Rest Exercise Main effect Interaction effect
LH (µmol/l) Normoxia 0.59
+
−
0.06 0.7
+
−
0.16 State –
Hypoxia (relative) 0.62
+
−
0.16 0.62
+
−
0.07
Hypoxia (absolute) 0.57
+
−
0.07 0.65
+
−
0.07
MDA (µmol/l) Normoxia 0.48
+
−
0.1 0.5
+
−
0.2 – –
Hypoxia (relative) 0.56
+
−
0.2 0.55
+
−
0.2
Hypoxia (absolute) 0.65
+
−
0.2 0.68
+
−
0.2
Myoglobin (units/l) Normoxia 36
+
−
637
+
−
14 – –
Hypoxia (relative) 43
+
−
16 46
+
−
23
Hypoxia (absolute) 43
+
−
15 59
+
−
26
Ascorbic acid (µmol/l) Normoxia 60.8
+
−
10 60.9
+
−
16 – –
Hypoxia (relative) 62.2
+
−
10 61.5
+
−
8
Hypoxia (absolute) 51
+
−
18 61.6
+
−
1
α-Tocopherol (µmol/l) Normoxia 19.7
+
−
2 19.9
+
−
4– –
Hypoxia (relative) 17.7
+
−
3 19.1
+
−
3
Hypoxia (absolute) 18.2
+
−
1 18.8
+
−
3
α-Carotene (µmol/l) Normoxia 0.02
+
−
0.01 0.02
+
−
0.01 State –
Hypoxia (relative) 0.01
+
−
0.01 0.02
+
−
0.01
Hypoxia (absolute) 0.02
+
−
0.01 0.02
+
−
0.01
β-Carotene (µmol/l) Normoxia 0.06
+
−
0.04 0.09
+
−
0.08 State –
Hypoxia (relative) 0.04
+
−
0.02 0.05
+
−
0.02
Hypoxia (absolute) 0.07
+
−
0.05 0.07
+
−
0.05
Lycopene (µmol/l) Normoxia 0.32
+
−
0.08 0.27
+
−
0.06† – State × group
Hypoxia (relative) 0.27
+
−
0.10 0.30
+
−
0.19
Hypoxia (absolute) 0.35
+
−
0.12 0.37
+
−
0.13
Retinol (µmol/l) Normoxia 0.74
+
−
0.11 0.73
+
−
0.22 – –
Hypoxia (relative) 0.80
+
−
0.15 0.80
+
−
0.15
Hypoxia (absolute) 0.76
+
−
0.14 0.73
+
−
0.15
Table 5
Effect of acute hypoxia on white blood cell
numbers
Values are means
+
−
S.D.;
n
= 10 per group. Main effect for state indicates a
difference (
P
< 0.05) between rest compared with exercise (pooled normoxia and
hypoxia values).
Blood cells Group Rest Exercise Main effect
Leucocytes Normoxia 5.6
+
−
1.2 12.7
+
−
5.4 State
(10
3
cells/µl) Hypoxia (relative) 5.6
+
−
1.0 10.7
+
−
4.5
Hypoxia (absolute) 5.7
+
−
1.3 12.6
+
−
7.5
Neutrophils Normoxia 2.8
+
−
0.7 8.7
+
−
4.5 State
(10
3
cells/µl) Hypoxia (relative) 2.8
+
−
0.8 6.8
+
−
4
Hypoxia (absolute) 3.1
+
−
1.0 9.8
+
−
5.0
suggestive of secondary oxygen-centred lipid-derived
alkoxyl radicals [10,11]. ESR spectra from the auto-
oxidation of α-linolenic acid show similar hyper-
fine coupling constants (a
N
= 13.8 Gauss, and a
H
β
=
1.9 Gauss) to those of human blood in the present study,
supporting our observation that the free radical species
detected are lipid in origin [11]. Coupled with the fact
Figure 3
Rest and exercise CK concentration in the
normoxia and hypoxia groups
Main effect for state (pooled rest compared with exercise data,
P
= 0.02). CPK,
CK.
that LH concentration increased following exercise, it
is reasonable to suggest that the primary source of the
alkoxyl radical observed in the present study is via initial
primary oxygen-centred radical attack and subsequent
C
2006 The Biochemical Society
Exercise-induced oxidative stress 139
PUFA (polyunsaturated fatty acid) decomposition [22].
Lipid peroxidation is a self-perpetuating chain reaction,
which generates oxidation by-products and free radical
intermediates. In support, Ashton et al. [10] suggest
that the alkoxyl radical results from lipid peroxidation
of cellular membranes by primary free radical attack
resulting in increased levels of plasma LH post-exercise.
Direct free radical detection, in addition to a rise in LH
concentration in the present study, provides evidence
that the damage inflicted to cellular membranes during
aerobic exercise may well be free radical mediated. In
the present study, we suggest that either muscle mito-
chondria or extracellular leucocytes had a role to play
in the generation of primary free radicals and lipid
peroxidation during aerobic exercise. The mitochondrial
electron transport chain within a muscle fibre has long
been considered the major site of ROS ( reactive oxygen
species) production at rest and during exercise [23].
Studies on isolated mitochondria suggest that between 2–
5% of total electron flux through the cytochrome chain
may undergo one electron univalent reduction with the
formation of superoxide and H
2
O
2
[24,25]. Given
the positive association between
˙
Vo
2
and the PBN adduct
and the significant rise in
˙
Vo
2
from rest to post-exer-
cise in the present study, it is postulated that aerobic
exercise increased tissue oxygen flux, causing an increase
in electron flux within the mitochondrial respiratory
chain with rapid formation of primary oxygen-centred
radical species (e.g. superoxide anion) [26]. As supported
by the work of Davies et al. [27], rodents performing
endurance exercise produce ESR-detected semiquinone
radicals derived from an inner mitochondrial membrane
‘leakiness’. More recent evidence using a microdialysis
probe has demonstrated a rise in the reduction of cyto-
chrome c, suggesting an increase in superoxide pro-
duction within the interstitial space of contracting mouse
muscle [28]. These data provide evidence that free radicals
are formed in exercising skeletal tissue and supports the
hypothesis that mitochondria may induce the formation
of free radical species. Although muscle mitochondria are
generally regarded as the main site of intracellular free
radical generation during exercise, it must be taken into
account that other tissues, such as liver, have the potential
to produce free radical species [27]. It is implied that
the observed increase in PBN adduct concentration after
exercise is due to the primary radical species attack-
ing either intracellular or extracellular PUFAs, which
would decompose LH in the presence of iron, yielding a
rapid rise in alkoxyl radical formation and allowing sub-
sequent detection via ESR spectroscopy [11]. Evidence
is provided for the presence of free iron to aid in such a
reaction by Jenkins et al. [29], who have demonstrated
that iron becomes more loosely bound to transferrin
during exercise.
LH concentration throughout was shown to increase
after exercise and, to our knowledge, this is the first
evidence of an increase in LH after 2 h of moderate
aerobicexerciseat55%
˙
Vo
2peak
. Since there is evidence
of membrane peroxidation, perhaps the integrity of the
muscle cell membrane was compromised, particularly as
an increased release of the glycolytic protein CK was
observed. The presence of systemic CK may suggest in-
creased sarcolemmal permeability, which during pro-
longed exercise predominately rises from sarcolemmal
membrane rupture [30], possibly as a result of free radicals
produced during exercise [31].
Although aerobic exercise in hypoxia did not increase
free radical generation, other studies with hypoxia have
demonstrated an increase in lipid peroxidation. The inter-
pretation of these findings may have some relevance to
the overall increase in oxidative stress in the present
investigation, as it can be observed from Figure 1 that
hypoxia plus exercise causes the greatest contribution
to free radical generation. Simon-Schnass and Pabst [32]
observed a significant increase in lipid peroxidation in
hypobaric hypoxia, and Simon-Schnass [4] suggests
further that hypoxic cells are particularly susceptible to
oxidative stress due to an accumulation of electrons that
cannot be transferred to oxygen at the level of mito-
chondrial cytochrome oxidase, but transferred to other
low-molecular-mass molecules which, in turn, induce
radical chain reactions. Notwithstanding the fact that
overall
˙
Vo
2
increased in the present study, there was a
selective decrease in Sao
2
between normoxia and both
hypoxia groups. This may suggest a decrease in diffusive
oxygen delivery to active skeletal myocytes, causing a
decrease in the
˙
V
max
of cytochrome oxidase and an
increase in univalent reduction [33]. Work by Bailey
and co-workers [8] has shown an increase in LH con-
centration following exhaustive exercise in normobaric
hypoxia (F
i
o
2
= 0.16). This increase in lipid peroxidation
was associated with a decrease in Sao
2
(LH, r =−0.61,
P < 0.05; MDA, r =−0.50, P < 0.05) as opposed to
˙
Vo
2max
. Bailey et al. [9] suggest that an exercise-induced
mass action effect of mitochondrial oxygen flux is not
the exclusive mediator of ROS production and postulate
that a decrease in mitochondrial Po
2
(partial pressure of
oxygen) may be a contributor to ROS generation.
It is clear from Figure 1 that the main effect observed
for free radical concentration is largely due to a change
from rest to exercise in the hypoxia subgroups. The lack
of a selective difference between normoxia and relative
and absolute hypoxia exercise may in part be due to the
consequence of limited statistical power. Retrospective
calculation of power (power = 0.42) shows that, for a
possible interaction effect to occur, any future study using
the same methodological design would need approxi-
mately double the existing number of subjects. Further-
more, although it is known that a decrease in intracellular
Po
2
can influence free radical exchange [34], perhaps a
more mechanistic reason for the lack of a hypoxic effect
may be due to an insufficient decrease in intracellular Po
2
C
2006 The Biochemical Society
140 G. W. Davison and others
as a consequence of using a 16 % F
i
o
2
at 55 %
˙
Vo
2peak.
It can also be observed in Table 3 that
˙
Vo
2
and work
rate were similar across all groups, and perhaps this also
contributed to the lack of an interaction effect.
Activated vascular leucocyte and neutrophil cells are
known to produce ROS [35] and both were increased fol-
lowing exercise in the present study. As
˙
Vo
2
increased
during exercise, it is possible that phagocytosis activated
membrane-bound NADPH oxidase, which reduced mol-
ecular oxygen to superoxide [36]. As erythrocyte mem-
branes are rich in PUFAs and highly susceptible to
oxidative damage [37], it is possible that the increase in
LH and lipid alkoxyl radicals originate from this extra-
cellular source.
The lack of change in MDA observed as a function
of exercise may possibly be due to a number of factors.
Because MDA levels were not measured continuously
during exercise, we cannot exclude the possibility of
MDA redistribution between plasma and exercising tis-
sue [11]. Moreover the lack of change in MDA may be
as a consequence of free radical and lipid peroxidation
termination by available circulating antioxidants. An
increased mobilization of important carotenoids as a
function of exercise and a selective decrease in lycopene
would suggest an increase in plasma antioxidant activity
and perhaps a stabilization of lipid-derived free radicals,
thereby inhibiting the production of MDA. It is also
conceivable that perhaps MDA was taken from an inap-
propriate sampling site at an inappropriate time following
exercise. Jenkins [38] suggest that the failure of some
investigators to detect a rise in MDA that was actually
produced include potential factors such as a failure to
look for a marker at the right time or place.
In conclusion, the present study has measured and
quantified oxidative stress in human blood following a
single bout of moderate aerobic exercise and demon-
strates that exercise performed at 55 % of
˙
Vo
2peak
can in-
crease (pooled group data) free radical concentration as
measured by ESR spectroscopy. Exercise performed in
absolute and relative hypoxia did not selectively (against
normoxia) increase oxidative stress, due in part to
limited statistical power. In addition to an exercise-
induced increase in the PBN adduct, exercise of this
duration and intensity increased LH and CK production,
which suggests cell membrane damage and increased
sarcolemmal permeability.
REFERENCES
1 Rokitzki, L., Logemann, E., Sagredos, A. N., Murphy, M.,
Wetzel-roth, W. and Keul, J. (1994) Lipid peroxidation and
antioxidative vitamins under extreme endurance stress.
Acta Physiol. Scand. 151, 149–158
2 Mohanraj, P., Merla, A. J., Wright, V. P. and Clanton, T. L.
(1998) Antioxidants protect rat diaphragmatic muscle
function under hypoxic conditions. J. Appl. Physiol. 84,
1960–1966
3 Sen, C. K., Roy, S. and Packer, L. (2000) Exercise-induced
oxidative stress and antioxidant nutrients. In Nutrition
in Sport (Maughan, R. J., ed.), pp. 292–317, Blackwell
Science, Oxford
4 Simon-Schnass, I. (1994) Risk of oxidative stress during
exercise at high altitude. In Exercise and Oxygen Toxicity
(Sen, C. K., Packer, L. and Hanninan, O., eds.),
pp. 191–210, Elsevier, Amsterdam
5 Kanter, M. M. (1998) Free radicals, exercise and antioxidant
supplementation. Proc. Nutr. Soc. 57, 9–13
6 Radak, Z., Lee, K., Choi, W. et al. (1994) Oxidative stress
induced by intermittent exposure at a stimulated altitude of
4000 m decreases mitochondrial superoxide dismutase
content in soleus muscle of rats. Eur. J. Appl. Physiol. 69,
392–395
7 Vasankari, T. J., Kujala, U. M., Rusko, H., Sarna, S. and
Ahotupa, M. (1997) The effect of endurance exercise at
moderate altitude on serum lipid peroxidation and
antioxidant functions in humans. Eur. J. Appl. Physiol. 75,
396–399
8 Bailey, D. M., Davies, B. and Young, I. S. (2000) Evidence
for reactive oxidant generation during acute physical
exercise and normobaric hypoxia in man. J. Physiol.
528P,99
9 Bailey, D. M., Davies, B. and Young, I. S. (2001)
Intermittent hypoxic training: implications for lipid
peroxidation induced by acute normoxic exercise in active
men. Clin. Sci. 101, 465–475
10 Ashton, T., Rowlands, C. C., Jones, E. et al. (1998)
Electron spin resonance spectroscopic detection of
oxygen-centred radicals in human serum following
exhaustive exercise. Eur. J. Appl. Physiol. 77, 498–502
11 Davison, G. W., George, L., Jackson, S. K. et al. (2002)
Exercise, free radicals, and lipid peroxidation in type 1
diabetes mellitus. Free Radical Biol. Med. 33,
1543–1551
12 Bailey, D. M., Davies, B., Young, I. S. et al. (2003) EPR
spectroscopic detection of free radical outflow from an
isolated muscle bed in exercising humans. J. Appl. Physiol.
95, 1714–1718
13 Durnin, J. V. G. A. and Womersley, J. (1974) Body
fat assessed from total body density and its estimation
from skinfold thickness: measurements on 481 men
and women aged from 16 to 72 years. Br. J. Nutr. 32,
77–79
14 Engfred, K., Kjaer, M., Secher, N. H. et al. (1994)
Hypoxia and training-induced adaptation of hormonal
responses to exercise in humans. Eur. J. Appl. Physiol. 68,
303–309
15 Borg, G. A. V. (1973) Perceived exertion. A note on history
and methods. Med. Sci. Sports Exercise 5, 90–93
16 Dill, D. B. and Costill, D. L. (1974) Calculation of the
percentage changes in volumes of blood, plasma and red
cells in dehydration. J. Appl. Physiol. 37, 247–248
17 Young, I. S. and Trimble, E. R. (1991) Measurement of
malondialdehyde in plasma by high performance liquid
chromatography with fluorimetric detection.
Ann. Clin. Biochem. 28, 504–508
18 Wolff, S. P. (1994) Ferrous ion oxidation in presence of
ferric ion indicator xylenol orange for measurement
of hypoperoxides. Methods Enzymol. 233, 183–189
19 Vuilleumier, J. P. and Keck, E. (1989) Fluorimetric assay of
vitamin C in biological materials using a centrifugal
analyser with fluorescence attachment. J. Micronutr. Anal.
5, 25–34
20 Thurnham, D. I., Smith, E. and Flora, P. S. (1988)
Concurrent liquid-chromatographic assay of retinol,
α-tocopherol, β-carotene, α-carotene, lycopene, and
β-cryptoxanthin in plasma with tocopherol acetate as
internal standard. Clin. Chem. 34, 377–381
21 Altman, D. G. (1982) Statistics and ethics in medical
research. III How large a sample size? Br. Med. J. 281,
1336–1338
22 Duthie, G. G. (1993) Lipid peroxidation. Eur. J. Clin. Nutr.
47, 759–764
23 Di Meo, S. and Venditti, P. (2001) Mitochondria in
exercise-induced oxidative stress. Biol. Signals Recept. 10,
125–140
C
2006 The Biochemical Society
Exercise-induced oxidative stress 141
24 Boveris, A. and Chance, B. (1973) The mitochondrial
generation of hydrogen peroxide: general properties
and effect of hyperbaric oxygen. Biochem. J. 134,
707–716
25 Cadenas, E., Boveris, A., Ragan, C. I. and Stoppani,
A. O. M. (1977) Production of superoxide radicals and
hydrogen peroxide by NADH-ubiquinone reductase
and ubiquinol-cytochrome c reductase from beef-heart
mitochondria. Arch. Biochem. Biophys. 180, 248–257
26 Sj
¨
odin, B., Westing, Y. H. and Apple, F. S. (1990)
Biochemical mechanisms for oxygen free radical formation
during exercise. Sports Med. 10, 236–254
27 Davies, K. J. A., Quintanilha, A. T., Brooks, G. A. and
Packer, L. (1982) Free radicals and tissue damage produced
by exercise. Biochem. Biophy. Res. Commun. 107,
1198–1205
28 McArdle, A., Pattwell, D., Vasilaki, A., Griffiths, R. D. and
Jackson, M. J. (2001) Contractile activity-induced oxidative
stress: cellular origin and adaptive responses. Am. J.
Physiol. Cell Physiol. 280, C621–C627
29 Jenkins, R. R., Krause, K. and Schofield, L. S. (1993)
Influence of exercise on clearance of oxidant stress
products and loosely bound iron. Med. Sci. Sports Exercise
25, 213–217
30 Hikida, R. S., Staron, R. S., Hagerman, F. C., Sherman,
W. M. and Costill, D. L. (1983) Muscle fibre necrosis
associated with human marathon runners. J. Neurol. Sci.
59, 185–203
31 Ebbeling, C. B. and Clarkson, P. M. (1989) Exercise-
induced muscle damage and adaptation. Sports Med. 7,
207–234
32 Simon-Schnass, I. and Pabst, H. (1988) Influence of
vitamin E on physical performance. Int. J. Vitam. Nutr.
Res. 58, 49–54
33 Duranteau, J., Chandel, N. S., Kulisz, A., Shao, Z. and
Schumacher, P. T. (1998) Intracellular signalling by reactive
oxygen species during hypoxia in cardiomyocytes.
J. Biol. Chem. 273, 11619–11624
34 Bailey, D. M., Young, I. S., McEneny, J. et al. (2004)
Regulation of free radical outflow from an isolated muscle
bed in exercising humans. Am. J. Physiol. Heart
Circ. Physiol. 287, H1689–H1699
35 Santos-Silva, A., Rebelo, M. I., Castro, E. M. B. et al.
(2001) Leukocyte activation, erythrocyte damage, lipid
profile and oxidative stress imposed by high composition
physical exercise in adolescents. Clin. Chim. Acta 306,
119–126
36 Suzuki, K., Sato, H., Kikuchi, T. et al. (1996) Capacity
of circulating neutrophils to produce reactive oxygen
species after exhaustive exercise. J. Appl. Physiol. 81,
1213–1222
37 Clemens, M. R. and Waller, H. D. (1987) Lipid
peroxidation in erythrocytes. Chem. Phys. Lipids 45,
251–268
38 Jenkins, R. R. (2000) Exercise and oxidative stress
methodology: a critique. Am. J. Clin. Nutr. 72, 670S–674S
Received 21 April 2005/6 September 2005; accepted 3 October 2005
Published as Immediate Publication 3 October 2005, doi:10.1042/CS20050135
C
2006 The Biochemical Society
