Clinical Science (1995) 89, 581-590 (Printed in Great Britain)
581
Calf muscle mitochondria1 and glycogenolytic ATP
synthesis in patients with claudication due to peripheral
vascular disease analysed using 31Pmagnetic resonance
spectroscopy
G . J. KEMP*, L. J. HANDS", G. RAMASWAMlf, D. J. TAYLOR*, A. NICOLAIDESS, A. AMATOg
and G . K. RADDA*
*MRC 6iochemical and Clinical Magnetic Resonance Unit, Oxford Radcliffe Hospital Trust.
Oxford, U.K., ?Nufield Department of Surgery, University of Oxford Medical School, Oxford
Radclire Hospital Trust, Oxford, U.K., $Department of Academic Vascular Surgery, St Mary's
Hospital Medical School, London, U.K., and @igma-Tau lndustrie Farmaceutiche Riunite s.p.a.,
Rome, Italy
(Received 9 May/4 July 1995; accepted I September 1995)
1. We set out to define abnormalities of oxidative
ATP synthesis, cellular proton efflux and the
efficiency of ATP usage in gastrocnemius muscle of
patients with claudication due to peripheral vascular
disease, using data obtained by 31Pmagnetic resonance spectroscopy during aerobic exercise and
recovery.
2. Eleven patients with moderate claudication were
studied and results were compared with 25 agematched control subjects. Changes in pH and phosphocreatine concentration during recovery were used
to calculate the maximum rate of oxidative ATP
synthesis (QmJ and the capacity of net proton
efflux. Changes in pH and phosphocreatine concentration were used to estimate rates of non-oxidative
and (indirectly) oxidative ATP synthesis throughout
exercise, taking account of abnormalities in proton
efflux during exercise.
3. In patients with claudication, slow post-exercise
phosphocreatine recovery showed a 42 9% decrease
in Qmax.,
and the slow ADP recovery was consistent
with this. pH recovery was slow, showing a 77f9%
decrease in the capacity for proton efflux. Both
abnormalities are compatible with a substantial
reduction in muscle blood flow.
4. During exercise, increased phosphocreatine depletion and intracellular acidification were a consequence
of impaired oxidative ATP synthesis and the consequent increase in non-oxidative ATP synthesis, compounded by reduced proton efflux. The acidification
prevented an increase in ADP concentration which
could otherwise partially compensate for the oxidative defect. All these abnormalities are compatible
with a reduced muscle blood flow.
5. In addition, initial-exercise changes in pH and
phosphocreatine concentration implied a 44f 5%
reduction in 'effective muscle mass', necessitating an
ATP turnover (per litre of muscle water) twice as
high for given power output as in control muscle.
Some of this is probably due to a localized loss of
muscle fibres, but the rest appears to reflect reduced
metabolic efficiency of the muscle. This is not a
direct consequence of reduced blood flow, and may be
related to change in muscle fibre type.
INTRODUCTION
The effects of chronic partial ischaemia on skeletal muscle are surprisingly little understood.
Classically, chronic peripheral vascular disease is
associated with an increase in the muscle content of
enzymes of aerobic and anaerobic metabolism [l],
an increased proportion of small oxidative fibres [2]
and an increase in the number and size of muscle
mitochondria [3]. During exercise, muscle blood
flow is reduced [4], oxygen consumption is decreased [4], oxygen extraction is increased [ S ] and
lactate/pyruvate efflux is increased [4, 61. Studies of
calf muscle in patients with claudication using 31P
magnetic resonance spectroscopy (MRS) showed
increased changes of pH and phosphocreatine (PCr)
concentration during exercise [7]. These findings are
in accordance with studies using needle biopsy,
which demonstrated increased accumulation of lactic acid and increased depletion of cellular ATP and
PCr during exercise in patients with claudication
Key words: arterial disease, ATP synthesis, bioenergetics, glycogenolysis. mitochondria, occlusive oxidative metabolism, "P magnetic resonance spectroscopy, skeletal muscle,
vascular disease.
Abbreviations: EMM, effective muscle mass; K,,, creatine kinase equilibrium constant; K,, [ADP] for half-maximal oxidative ATP synthesis; I,,,, apparent maximum rate of
glycogenolytic ATP synthesis; MRS, magnetic resonance spectroscopy; Pi, inorganic phosphate; PCr. phosphocreatine: Q,,
apparent maximum rate of oxidative ATP synthesis.
Correspondence: Miss L. J. Hands, MRC Biochemical and Clinical Magnetic Resonance Unit, Oxford Radcliffe Hospital Trust, Oxford OX3 9DU, U.K.
582
G. J. Kemp et al
compared with control subjects at comparable
workloads [4].
There are several possible explanations for these
observations. Impaired oxidative ATP synthesis or
an increase in the requirement for ATP (resulting
from reduced metabolic efficiency [S]) would both
tend to increase glycogenolytic ATP production.
Increased glycogenolysis would result in increased
cellular acidification, and this would be exacerbated
by any reduction in proton efflux. 31P MRS studies
have shown slow recovery of pH after exercise,
suggesting a reduced capacity for proton efflux, and
slow recovery of [ADP] and [PCr] which must
result from a defect of mitochondrial oxidation [7].
These abnormalities could be consequences of
reduced vascular perfusion, which would decrease
both oxygen delivery (thus impairing mitochondrial
oxidation) and the washout of acid from the muscle
extracellular space (hence inhibiting cellular proton
efflux). Changes in metabolic efficiency (i.e. in work
done per ATP hydrolysed) might result from
changes in muscle fibre type composition [9-1 13.
A key question is to what extent these several
factors (reduced oxygen supply, impaired acid washout and possible reduction in metabolic efficiency)
contribute to the abnormal exercise response in
patients with claudication. To address this, the
factors contributing to pathophysiology of muscle
MRS abnormalities can now be analysed quantitatively [12-181. In this paper we will show that the
abnormal skeletal muscle response to exercise in
patients with claudication is due to a substantial
reduction in proton efflux and the capacity for
mitochondrial ATP synthesis, as well as a decrease
in muscle mass and/or metabolic efficiency, with a
consequent stimulation of glycogenolysis.
METH0DS
Subjects
Studies were performed on 1 1 patients (six male,
five female) with calf claudication (defined as calf
muscle pain associated with walking) due to peripheral vascular disease, aged 43-82 years (mean
68 years). Maximum walking distance was established on a treadmill (4km/h, zero incline) and
ranged from 55-226 m (median 135 m). Anklebrachial pressure indices in the affected leg ranged
from 0.36-0.83 (median 0.63). Only one patient had
never smoked; another was still smoking, but the
remainder had all relinquished the habit. Two
patients had type I1 diabetes mellitus and six were
on treatment for hypertension (none with 8adrenoceptor-blocking agents). Results were compared with those of 25 healthy control subjects (12
male, 13 female), aged 52-74 years (mean 62 years).
All subjects gave informed consent according to a
protocol approved by the local hospital ethics
committee.
MRS methods
The study used a 2.0T superconducting magnet
(Oxford Magnet Technology, Eynsham, Oxford,
U.K.) interfaced to a Bruker spectrometer (Bruker,
Coventry, U.K.). Subjects lay with the affected calf
(or in control subjects, the right calf) overlying a
6 cm diameter surface coil. Spectra were acquired
using a 2 s interpulse delay at rest (64 scans) and
during exercise and recovery. In control subjects,
the muscle was exercised by plantar flexion at
OSHz, lifting a weight of 10% of lean body mass
(calculated from body weight and skinfold thickness
[19]) through a distance of 7cm. After four 32-scan
spectra (1.25 min) had been acquired, the weight was
incremented by 2% of lean body mass for each
subsequent spectral acquisition (1.25 min). The subjects exercised until stopped by claudication pain (or
in control subjects, fatigue). The muscle was then
studied for 13min during recovery (four 8-scan
spectra, four 16-scan spectra, three 32-scan spectra
then two 64-scan spectra). In nine of the patients,
the exercise protocol had to be modified according
to clinical assessment of exercise tolerance, the
starting load being set at 5% of lean body mass
rather than 10%; in six of these the load could be
increased to 10% after two spectral acquisitions and
in four of these at least one normal increment of 2%
lean body mass could be applied after four spectral
acquisitions. The analysis takes account of these
variations in the absolute power output.
Relative concentrations of inorganic phosphate
(Pi), PCr and ATP were obtained by a time-domain
fitting routine (VARPRO; R. de Beer, Delft, The
Netherlands) and were corrected for magnetic saturation. Absolute concentrations were obtained in
normal muscle by assuming a normal [ATP] of
8.2mmol/l (i.e. mmol per litre of muscle water) [20].
In patients, biopsy studies have shown a reduction
in resting [ATP] of about 16% [4, 211, and the
calculation of metabolite concentrations in patients
was modified accordingly. Intracellular pH was calculated from the chemical shift of the Pi peak
relative to PCr (measured in ppm using the spectrometer's built-in peak-picking facility), as pH = 6.75 +
log,,((ppm - 3.27)/(5.69- ppm)}. Free cytosolic
[ADP] was calculated from pH and [PCr] using
the creatine kinase equilibrium constant [22] of
K , , = 1.66 x lo9mol/l, assuming a normal total creatine content of 42.5mmol/l [20], as
[ADP] =([total creatine]/[PCr] - 1)
x CADPII(K,,[H 'I)
This assumption may perhaps be in error for the
muscle of patients, but the analysis is not very
sensitive to such uncertainties.
Analysis of exercise
During exercise, ATP is produced by net hydrolysis of PCr, by glycogenolysis to lactic acid and by
oxidative phosphorylation; protons are produced by
Calf muscle bioenergetics in peripheral vascular disease
glycogenolysis, buffered by passive processes and as
a consequence of PCr breakdown, and also leave
the cell by several membrane transport processes [8,
12, 131. The contributions of these are calculated as
follows (see Appendix for details). First we consider
the start of exercise. The cytosolic buffer capacity,
based on measurements in rat muscle in uitro [23]
and 31P MRS studies of human muscle in uiuo [13],
was taken as 20 slykes plus the calculated contribution of [Pi] [13, 161. Using this, measurements of
pH and [PCr] from rest to the first exercise spectrum are used to calculate the initial rates of
glycogenolytic ATP synthesis and PCr depletion [8,
121. As the contribution of oxidation in the early
stages of exercise can be neglected [12], their sum
gives the initial rate of ATP synthesis. We define the
effective muscle mass (EMM) [24] as the ratio of
initial mechanical power output to initial rate of
ATP synthesis (for convenience, EMM is expressed
in arbitrary units: see Appendix). EMM is proportional to the metabolic efficiency, i.e. the work
achieved by hydrolysis of a given amount of ATP,
and also to the real muscle mass (see Discussion)
Next we consider ATP synthesis rates during later
exercise. At each point, changes in pH and [PCr]
are used to calculate the glycogenolytic ATP synthesis rate (allowing for proton efflux by using the
recovery analysis: see below), and when added to
the rate of net PCr hydrolysis this gives the nonoxidative ATP synthesis rate. The total ATP synthesis rate is calculated from the mechanical power
output using the EMM. From these, the oxidative
ATP synthesis rate is calculated by difference [8, 12,
131. The flux-generating step of glycogenolysis is
glycogen phosphorylase, whose activity in uiuo is
largely that of the phosphorylated form, phosphorylase a. This is strongly dependent on the concentration of Pi, its substrate and activator, and its
activity is half-maximal at 26mmol/l [25]. Thus the
apparent maximum rate of glycogenolysis (L,,,,),
calculated from [Pi] and the glycogenolytic ATP
synthesis rate at each point, is an approximate
measure of the activity of phosphorylase a in uiuo
during exercise [18].
To facilitate comparisons where exercise
durations differ widely, it is convenient to calculate
cumulative ATP synthesis over the whole of exercise
(oxidative and non-oxidative), and divide by the
total work done (in arbitrary units), to estimate the
cumulative ATP cost of work [24]. This has two
components, which can be estimated separately. The
cumulative non-oxidative cost of work tends to be
increased by a reduction in either the oxidative
contribution to ATP synthesis or in the EMM. The
cumulative oxidative cost of work tends to be
increased by a fall in EMM. The effects of changes
in oxidative contribution and in EMM can be
separated by using the work per unit EMM to
calculate the cumulative non-oxidative cost of work
per unit EMM and the cumulative oxidative cost of
work per unit EMM [24].
583
Analysis of recovery
During recovery, PCr is resynthesized as a consequence of oxidative ATP synthesis, and this is
accompanied by the net generation of protons [13,
15, 16, 261. Protons are lost from the cell by several
processes [12, 13, 161. Analysis of recovery yields
information about mitochondrial ATP synthesis and
net proton efflux.
First, the initial rate of PCr resynthesis (measured
here by linear regression using the end-exercise and
first two recovery spectra) is an estimate of the endexercise rate of oxidative ATP synthesis [12, 14, 271;
the component due to ‘basal’ ATP synthesis is
normally small enough to ignore [27], although it
might be significantly increased in these patients due
to increased energy requirement in damaged, ‘leaky’
cells. The initial rate of PCr resynthesis has a
hyperbolic (Michaelis-Menten) relationship to its
driving force, the cytosolic free [ADP] [26], as
expected from experiments on mitochondrial control
in uiuo in exercising muscle [28, 291 as well as in
uitro [30]. Thus, the end-exercise [ADP] and initial
PCr resynthesis rate can be used to calculate the
maximum oxidative ATP synthesis rate (Q,,,,),
which reflects both intrinsic mitochondrial capacity
and the delivery of oxygen [l5]. We assume a
normal value (30pmol/l) for the K , for ADP [26,
28-30]. Two other indices of mitochondrial function
[8], the half-times of ADP and PCr recovery, are
calculated from the slope of semilogarithmic plots
[20]. Both the PCr recovery half-time and the ADP
recovery half-time are increased in the presence of a
defect in Q,,,.[8,
121.
Second, the proton efflux rate is calculated from
changes in pH and [PCr] [l2, 13, 16). This has an
approximately linear pH-dependence [12, 13, 161,
and so the proton efflux capacity is estimated as a
first-order proton efflux rate constant, calculated as
the linear regression slope of eflux against pH. A
modification of this method calculates the rate of
end-exercise proton efflux required to make the
calculated end-exercise rate of oxidative ATP synthesis equal to the initial post-exercise rate of PCr
resynthesis. This and the end-exercise pH yield the
value of the proton efflux rate constant which is
used to estimate rates of proton efflux in exercise
[12, 131.
Presentation and statistics
Results are presented as means k SEM. Statistical
significance of differences of patients from control
subjects was determined by Student’s unpaired
t-test.
RESULTS
Resting muscle
In resting muscle, cytosolic pH was increased, and
both calculated [Pi] and [PCr] were normal (Table
G. J. Kemp et
584
Table 1. "P MRS data in resting muscle (meanfSEM). *P is the
statistical significance of difference from control (Student's f-test).
PH
[Pi] (mmol/l)
[PCr] (mmol/l)
Calculated [ADP] (pmol/l)
PCr/(PCr Pi)
+
Patients
(n=ll)
Control subjects
(n = 25)
P*
7.05+ 0.01
3.9k0.3
34+ I
14+2
0.90+0.01
7.02+O.oO
3.9 0.2
33+ I
14+ I
o.ooo1
0.90,O.OI
0.8
+
0.9
0.4
0.8
1). Note that, in biopsy studies, [PCr] at rest has
been reported as unchanged [4] or reduced by 19%
P I .
al
when account was taken of the reduced EMM, but
the overall non-oxidative cost of work per unit
EMM was still increased in patients with claudication (Table 2). The oxidative costs of work tell the
same story. The overall oxidative cost of work was
not different from its control value. This is because
a decreased EMM tends to increase, and a reduced
oxidative contribution to decrease, this quantity
[24]. When account was taken of the reduced
EMM, the overall oxidative cost of work per unit
EMM was decreased in patients with claudication
(Table 2); this reflects the oxidative defect alone.
Thus, defects in both EMM and in oxidative capacity (and therefore contribution) are responsible for
the increased changes in pH and [PCr] seen in
patients with claudication.
Exercise
The initial ATP synthesis rate was not significantly different in the two groups. However, as the
initial power was reduced by 41 f 6 % in patients
with claudication, the EMM was reduced by
44f 5% in patients with claudication compared with
control subjects (Table 2). Exercise duration was
reduced by more than 50%. The later changes in pH
and [PCr] were larger in patients with claudication
(Fig. l), and so was the calculated rate of nonoxidative ATP synthesis (Fig. 2a), but by the end of
exercise pH and [PCr] did not differ from control
(Table 2). End-exercise [ADP] was marginally lower
in patients with claudication (Table 2). As [PCr]
was smaller in patients than in control subjects at
each time point (Fig. lb), so [Pi] was higher (as the
sum of [Pi] and [PCr] does not change significantly
during exercise [8]). Taking account of the
increased glycogenolytic ATP synthesis rate and
[Pi] at each point, the apparent maximum rate of
glycogenolytic ATP synthesis
was normal at
the start of exercise, and then increased (Fig. 2b)
(presumably due to more rapid net conversion of
phosphorylase b to a), although by the end of
exercise it did not differ from control (Table 2). The
oxidative ATP synthesis rate (shown in Fig. 2c as a
fraction of total ATP synthesis rate) decreased
slowly during exercise in control subjects, and much
more sharply in patients with claudication; by the
end of exercise it was reduced by 50% in patients
with claudication compared with control (Table 2).
A cumulative analysis of ATP synthesis and
mechanical work provides a way of asking to what
extent the observed abnormalities in pH and [PCr]
during exercise are due to the reduction in EMM or
to the reduction in the oxidative contribution to
ATP synthesis [24]. Taking account of the reduced
work, the cumulative non-oxidative cost of work
was increased in patients with claudication (Table
2), most of the abnormality being in the glycogenolytic component. This is because both a decreased
EMM and a reduced oxidative contribution tend to
increase this quantity [24]. The difference between
patients and control subjects was therefore reduced
Recovery from exercise
During recovery from exercise (Table 3), pH
recovery was substantially slower in patients with
claudication (Fig. 3a). Q,,,, was decreased by 42+
9% in the patient group compared with the control
group, and the PCr and ADP recovery half-times
were prolonged approximately 5-fold (Figs. 3b and
3c). The rates of oxidative ATP synthesis (Fig. 4a)
and proton efflux (Fig. 4b) remained lower in
patients with claudication throughout the first
minute of recovery, before the rates become too
small to measure. The proton efflux rate constant
was reduced by 77+9% (Table 3).
DISCUSSlON
Increased changes in pH and [PCr] during exercise in patients with claudication might be due to
reduced net proton efflux during exercise, reduced
muscle mass and/or metabolic efficiency, or reduced
oxidative ATP synthesis rate [8]. These are discussed below.
Proton efflux
Although proton efflux rates cannot be assessed
directly during exercise, the present analysis suggests
that the rate of net proton efflux during exercise was
substantially decreased in patients with claudication
(Fig. 4b). The defect in the capacity for proton emux
is best estimated by the 75% reduction of the rate
constant (Table 3). This contributes substantially to
the increased acidification during exercise, and must
be taken into account when estimating rates of
glycogenolysis during exercise [12, 13, 161.
There are few published data to compare with
this analysis. Proton efflux can in principle be due
to the sodium-proton antiporter, sodium-dependent
chloride/bicarbonate exchange or lactate/proton cotransport [ 161. Previously only lactate efflux has
been measured in patients with claudication. A
comparison of lactate efflux from exercising calf
muscle with measured muscle lactate concentrations
Calf muscle bioenergetics in peripheral vascular disease
585
Table 2 I‘P MRS data in exercising muscle (mun&SEM). *P is the statistical significance of difference from control (Student’s
1-test).
Initial exercise
Initial ATP turnover rate (F) (mmolmin-’
Initial power output (%lean body mass)
EMM (arbitrary units)
Initial lmU.
(mmolmin-ll-l)
1-I)
End of exercise
PH
PCr/(PCr Pi)
Calculated [ADP] (pmol/l)
Oxidative fraction (Q/F)
lmS,
(mmol min - I I- I)
+
Overall exercise
Duration (min)
Overall cost of work (arbitrary units)
Non-oxidative
Oxidative
Cost per unit EMM (arbitrary units)
Non-oxidative
Oxidative
6.7
6.6
1
1 1
0
2
4
6
10
8
12
14
16
18
14
16
18
Exercise time (min)
-5
1.0
(b)
0.9
0.8 -
6 0.7$0.6 -
2 0.5 0.40.31
0
2
4
6
10
8
I2
Exercise time (min)
Patients
Control subiects
20f3
5.9f 0.6
0.32k0.03
7f2
23f I
IOfO
0.46 f0.02
8f2
6.75 0.08
0.40 f 0.05
50f3
0.26 fO.05
5Of8
6.78f0.03
0.37 k0.02
70f7
0.51f0.03
54f6
0.05
4.6f 0.5
11.9fO.7
o.Oo0 I
I.9f 0.3
I .5 f 0.2
0.7 f 0.I
1.6 f O . l
o.Oo01
0.53f 0.07
0.48f 0.06
0.29 f 0.03
0.70 f 0.03
0.001
0.002
p*
0.9
o.ooo1
0.002
0.8
0.7
0.5
o.Oo01
0.7
0.7
[4] suggests that the rate constant of net lactate
efflux is reduced by 50-75% in patients with claudication. Direct measurements of [lactate] during
recovery from exercise in the stimulated rat leg
imply that no more than 30% of net proton efflux
can be accounted for by lactate/proton co-transport
[161, which is presumably dependent on blood flow.
It may be asked what fraction of calculated
proton ‘efflux’ could be accounted for by oxidation
of accumulated lactate. Comparison of the initial
rates of PCr resynthesis (a measure of oxidative
ATP synthesis, regardless of substrate) and net
proton efflux given in Table 3 shows that even if all
oxidative ATP synthesis resulted from oxidation of
retained lactate, this would only account for 10% of
net proton eflux in both patients and control
subjects.
The difference in pH in resting muscle (Table 1)
implies a small alkaline shift in the aggregate setpoint of all the mechanisms of net proton efflux,
including the sodium-proton antiporter [161. However, this would have a negligible effect on proton
efflux at any pH that was significantly more acid.
Glycogenolytic ATP synthesis rate
:0
i
4
6
8
I0
I2
I4
I6
(8
Exercise time (min)
Fig. 1. Metabolite changes during exercise. Time course of (a) cytosolic
pH, (b) PCr/(PCr Pi), a measure of PCr concentration, and (c) calculated
+
free cytosolic [ADP] during exercise in patients with claudication ( 0 )and
control subjects ( 0 )In. later exercise the number of subjects contributing
to each point decreases as exercise ceased early, particularly in the patient
group. [Pi] is proportional to {I - PCr/(PCr Pi)}. The calculation of free
cytosolic [ADP] assumes a normal creatine content (see Methods). Values are
given as mean f SEM.
+
The rate of glycogenolysis (the major component
of non-oxidative ATP synthesis) was increased in
patients with claudication, consistent with reports of
increased blood and muscle lactate concentrations
[4,61. The results in Fig. 3(a) can be compared with
the finding that lactate output during bicycle exercise is increased about 5-fold [4]. Glycogenolysis
during exercise is stimulated because of the calciumdependent conversion of phosphorylase b to phosphorylase a catalysed by phosphorylase b kinase, as
G. J. Kemp et al.
586
in normal muscle [25, 331, but no data exist for
patients with claudication. In patients with claudication, the glycogen content of resting muscle is
normal [l, 41.
Effective muscle mass (EMM)
z
0
1
4
6
8
10
12
14
18
16
Exercise time (min)
0
2
4
6
8
10
I2
14
16
18
I2
14
16
18
Exercise time (min)
40
0.0
2
4
6
8
10
Exercise time (min)
Fig. 2. ATP synthesis rates during exercise. ( a ) Time course of the
non-oxidative ATP synthesis rate (i.e. mainly glycogenolytic ATP synthesis,
plus also the rate of PCr depletion) during exercise in patients with
claudication (0) and control subjects (0).
(b) Time course of the apparent
maximum rate of glycogenolysis ( t m wduring
)
exercise in patients with
). This is calculated from the rate
claudication (0) and control subjects (0
of glycogenolytic ATP synthesis and the [P,] at each time point, and is in
principle proportional t o the activity of glycogen phosphorylase a at each
point. (c) Time course of the rate of oxidative ATP synthesis expressed as a
fraction of total ATP synthesis rate during exercise in patients with claudiValues are given as mean +SEM.
cation (0) and control subjects (0).
well as the stimulation of phosphorylase a activity
by increasing concentrations of its substrate Pi [25,
31, 321. Further analysis [18] of the rates of glycogenolysis measured in the present experiments
makes use of the known Pi-dependence of phosphorylase activity [25] and the relatively low
activity of phosphorylase b under physiological conditions [25, 331. The results suggest (Fig. 2b) that
the activity of phosphorylase a increases more
rapidly in patients with claudication, and is higher
at comparable power outputs. As the total activity
of phosphorylase is also normal in patients with
claudication [lo], this implies a greater conversion
of phosphorylase b to a. The reported maximum
activity of phosphorylase in normal human leg
muscle amounts to approximately 65 mmol/l ATP/
min (recalculated from published data [25]). In both
groups, the end-exercise values of L,,,, were close
to this (Table 2), suggesting near-complete activation of phosphorylase (at least in control muscle).
Biopsy measurements show a peak value of 50-65x
for the active fraction i.e. a/(a + b) of phosphorylase
EMM is calculated as the ratio of absolute
mechanical power output to the corresponding rate
of ATP synthesis per litre of muscle water [24]. In
the present study where exercise is dynamic, mechanical power is measured directly from the force
applied, the distance moved and the rate of contraction (but expressed here, for simplicity, in nominal
units of percentage of lean body mass: see Table 2).
In isometric exercise [141, an equivalent calculation
could be performed using force (since power output
is formally zero in isometric exercise).
The initial rate of ATP synthesis was similar in
patients with claudication and control subjects, but
as the initial power output was lower in patients,
the EMM was reduced by 44+5% (Table 2). This is
proportional to the metabolic efficiency of the
muscle and to its actual mass. Mechanical power
was standardized according to lean body mass, and
our recent work shows that this is proportional to
the cross-sectional area of the gastrocnemius muscle
as assessed by 'H MR imaging: in 13 normal
subjects, gastrocnemius area at its maximum point
(cm2) was proportional to lean body mass (kg) over
an approximately 2-fold range with slope = 3.2,
intercept not significantly different from zero and
r=0.78 (D. J. Taylor, unpublished work). Thus, in
normal subjects, lean body mass is a convenient and
plausible surrogate for calf muscle mass, although it
is possible that standardizing power output on a
reliable direct measure of muscle mass would further
reduce the variation in exercise response.
However, we cannot exclude a contribution from
localized muscle atrophy in the patients as we did
not measure individual muscle cross-sectional area,
or maximum voluntary contraction, which correlates with this in normal subjects [14] but perhaps
not necessarily in disease. There appear to be no
published quantitative data on muscle volume in
patients with claudication. In patients with rest pain
due to advanced peripheral vascular disease, the
total number of muscle fibres in the gastrocnemius
may be reduced by as much as half [34], which
could explain a decrease in EMM of the size
observed in the present work (Table 2). However, in
the present patients, with much less severe disease,
any such reduction must presumably be smaller
than half.
Thus, although we cannot be certain, it is likely
that there is at least a contribution from a reduction
in metabolic efficiency. This is perhaps associated
with changes in fibre type, and perhaps also with
altered ATP requirements in damaged muscle fibres
(e.g. increased need for ion pumping). Reported
findings on fibre type change in patients with
Calf muscle bioenergetics i n peripheral vascular disease
Table 3. IlP MRS data from muscle during recovery from exercise (mean -ISEM). *P is the statistical significance of difference
from control (Student's t-test). ?To increase precision, this is an average calculated over the first four recovery spectra
(t=0.13- 1.13min). As proton efflux rate decreases with time during recovery these values are somewhat lower than the true initial
rates (first point in Fig. 4b).
Mitochondria1 function
PCr tlp (min)
Calculated ADP t1p (min)
Initial PCr resynthesis rate (mmol min-' I - # )
Qmu. (mmolmin-l1-l)
0
I
3
2
I
2
3
8
4
8
6
2.0 f 0.7
I.O f 0.3
0.49 0.04
0.27 0.01
23+ I
36+3
0.002
0.003
0.000I
0.0002
12+ I
39f7
om I
+
11f2
10
9
II
p*
+
5+ I
9*3
4
5
6
7
Recovery time (min)
5
Control subiects
21 f 3
Proton efflux
Initial rate (mmol min-' I-')?
Rate constant (mmolmin-'I-l)/pH unit
0
Patients
0.0
0.5
0.0001
I.o
1.5
2.0
2.5
3.0
2.5
3.0
Recovery time (min)
7
9
10
II
0.0
0.5
I.o
1.5
2.0
Recovery time (min)
Recovery time (min)
Fig. 4. Oxidative ATP synthesis and proton efflux during recovery
from exercise. Time course of ( 0 ) the rate of PCr resynthesis (a measure of
oxidative ATP production) and (b) the rate of proton efflux during recovery
from exercise in patients with claudication ( 0 )and control subjects (0).
Values are given as mean If:SEM.
0.4
B 0.2
Y
0
1
2
3
4
5
6
7
8
9
1
0
1
1
Recovery time (min)
Fig. 3. Metabolite changes during recovery from exercise. Time course
of (a) cytosolic pH. (b) PCr concentration (expressed in terms of fractional
recovery) and (c) calculated free cytosolic [ADP] (also expressed as a fraction)
during recovery from exercise in patients with claudication ( 0 )and control
The calculation of free cytosolic [ADP] assumes a normal
subjects (0).
creatine content (see Methods). Values are given as meankSEM.
claudication are variable and confusing. In gastrocnemius there are reports of a reduced [9] or
normal [lo] relative abundance of type I fibres, and
of both reduced type IIa and increased type IIb
fibre numbers [lo], and vice versa [ll].
Oxidative metabolism
During recovery from exercise, we find a 42f9%
defect in the maximum rate of oxidative ATP
synthesis [8, 151 (Table 3), and the prolongation of
the recovery half-times of PCr and ADP (Table 3)
is consistent with this. This is clearly due to the
reduction in vascular supply of oxygen and perhaps
substrate. Any increase in the intrinsic capacity
for oxidative ATP synthesis can evidently do little
to compensate for this. There are reports of
48-64% increases in the content per mass of muscle
of the mitochondrial enzymes citrate synthase,
cytochrome-c oxidase and 3-hydroxyacyl coenzyme
A dehydrogenase [35], of 20-85% increases in the
content per mass of muscle of mitochondrial protein
[35] and also increases of approximately 10% in
mitochondrial volume per mass of muscle [3]. However, the activity of the mitochondrial succinate
oxidation per mass of muscle is variously reported
as increased by 22% [l] or decreased by 26% [lo].
Next we consider the oxidative contribution to
exercise. A reduction in muscle mass or efficiency
would increase the required rate of ATP synthesis
expressed per litre of muscle water, while reduction
588
G. J. Kemp et al
in oxidative capacity tends to decrease the contribution of oxidative ATP synthesis. We find that
oxidative ATP synthesis rates decreased dramatically in patients with claudication, contributing
much less than in control muscle to total ATP
synthesis (Figs. 2c and 3c). This is partly because of
the decrease in Qmax,,and partly because of the
decrease in the mitochondrial driving force, [ADP].
This analysis is consistent with the finding that rates
of oxygen consumption in exercising quadriceps in
patients with claudication are reduced by about a
third [4]. Direct measurements have shown that in
gastrocnemius during exercise, the partial pressure
of 0, in muscle falls by about 70% in patients,
compared with 30% in control subjects [21].
In a purely oxidative exercise, an increase in the
'error signal' [ADP] would be an inevitable response to impairment of mitochondria1 function, but
in mixed oxidative and glycogenolytic exercise this
response depends on the balance between the
increased rates of glycogenolysis and PCr depletion
[lS]. An increase in [ADP] would have permitted a
more normal rate of oxidative ATP synthesis, with
smaller changes in pH and [PCr]. This would
require an adaptation such as in mitochondrial
myopathy, where an up-regulation of proton efflux
(evidenced by rapid pH recovery) results in reduced
acidification during exercise, despite increased glycogenolytic ATP synthesis [13]. In patients with claudication this is impossible because of the impairment of proton efflux.
In summary, analysis of 3 1 P MRS data from
gastrocnemius muscle during incremental aerobic
exercise suggests a 40% decrease in maximum rate
of oxidative ATP synthesis Qmax. and a 40%
decrease in muscle mass and/or metabolic efficiency
in patients with claudication. Rates of oxidative
ATP synthesis were substantially reduced, while
glycogenolytic ATP synthesis was stimulated. Thus
the defects in metabolic efficiency and in Q,,,, both
contributed to the increased PCr depletion and
intracellular acidosis observed in later exercise. Substantial metabolic defects are therefore present in
the calf muscle of patients with only mild to moderate claudication.
metabolic proton generation rate; U , 'unaccountedfor' ATP synthesis rate; V , initial post-exercise PCr
resynthesis rate; /I,
cytosolic buffer capacity; 1, proton eMux rate constant; KPi, [Pi] for half-maximal
glycogenolytic ATP synthesis; m, oxidative proton
1.
production; $, l/[l+ 10'pH-6.75)
Exercise: principles
During exercise [S, 12, 131, ATP is produced by
net hydrolysis of PCr (D, mmol min- ' 1- '), by glycogenolysis to lactic acid (L, mmol min - 1 - ') and
by oxidative phosphorylation (Q, mmol min- I-').
The total ATP synthesis rate, which is proportional
to mechanical power, is given by
'
'
F=Q+D+L
(1)
Protons are produced (at P , mmolmin-' I-') by
glycogenolysis: to a first approximation the small
production by oxidation can be neglected, so that
P 4 2 / 3 ) L . This must equal the rate at which protons are buffered by passive processes and by net
PCr breakdown, plus the rate of net proton eflux
( E , mmolmin-'l-'), so that
in which fl
P=$D-fldpH/dt + E
is the cytosolic buffer capacity.
(2)
Start of exercise
In the first exercise spectrum (where oxidation
and proton efflux can, to a first approximation, be
ignored [ 12]), the glycogenolytic rate and total rate
of ATP synthesis can be obtained as
+
Initial F = D (3/2)(1+bD- fldpH/dt)
(3)
(eqn. 3 is equivalent in practice to the initial
ATPase rate measured by the very early rate of PCr
depletion [17]). We define effective muscle mass (in
arbitrary units) as EMM =(initial power)/(initial F )
where power is expressed as percentage lean body
mass. This is proportional to metabolic efficiency
and to real muscle mass.
Later phases of exercise
ACKNOWLEDGMENTS
The studies described in this paper were supported by Sigma-Tau Industrie Farmaceutiche
Riunite s.p.a., Rome, Italy.
APPENDIX: MATHEMATICAL DETAILS OF
ANALYSIS
Symbols used
The following symbols are used in the Appendix:
F , total ATP synthesis rate; Q, oxidative ATP
synthesis rate; L, glycogenolytic ATP synthesis rate;
D, PCr depletion rate; E , proton eflux rate; P ,
In the later stages of mixed exercise [12, 131, we
define the rate of ATP synthesis 'unaccounted for'
by non-oxidative means without allowing for proton
efflux as
U = F- D-(3/2)($D-pdpH/dt)
(4)
in which F is obtained from eqn. 3 as the product
(initial F ) x (relative power). This quantity U contains the oxidative ATP synthesis rate and the
component of glycogenolysis which gives rise to
protons that are removed by net proton efflux. The
contributions of oxidative and glycogenolytic ATP
synthesis can be separated by assuming a value of
the proton eflux rate. Thus, from eqns. 1, 2 and 4,
Calf muscle bioenergetics in peripheral vascular disease
oxidative ATP synthesis rate can be estimated
approximately as Q z U-(3/2)E, or more exactly
[8, 12, 131 as
(3/2)mI [ u - (3/2)E1
(5)
where m=0.16/[1 + 10(6.'-PH'
1, which is the small
number of protons produced per mol of oxidative
ATP generated. From eqns. 1 and 5, glycogenolytic
rate can then be calculated as
Q = { 1/[1-
L = F - D - Q = F - D - { 1/[ 1 -( 3/2)m]}[U -(3/2)E]
(6)
We describe below how we estimate the rate of
proton efflux E in these expressions. (Note that if
glucose were the major source of lactic acid rather
than glycogen, the factor 312 in the above equations
would be replaced by unity; this might occur in
later exercise, but if the proton efflux correction is
obtained as described below, this has a rather small
effect on the results).
During exercise, the apparent glycogenolytic
capacity (an approximate measure of phosphorylase
a activity) is calculated as L,,,. = L( 1 Kpi/[Pi]),
where Kpi= 26 mmol/l, the [Pi] at which phosphorylase activity is half-maximal [25]. The overall nonoxidative cost of work [24] is calculated as the ratio
of the time-integral of L+D to the time-integral of
power W (expressed for convenience relative to the
starting power). The overall oxidative cost of work
is calculated analogously as the ratio of the timeintegral of Q to the time-integral of W . The nonoxidative and oxidative costs of work per unit of
EMM are derived from these by dividing by the
initial exercise value of F [24].
+
This analysis of exercise depends partly on a
correction for the amount of proton efflux E during
exercise (eqns. 5 and 6). We obtain this as
E = -3.ApH on the assumption of a linear relationship between proton efflux and pH. Such a relationship is seen during recovery, in which PCr repletion
releases protons which are extruded from the cell [IS,
12, 13, 16, 261. The rate constant i, can be estimated
in two ways.
To estimate i. from the kinetics of recovery, the
rate of change of pH and the rate of PCr resynthesis
( V , m m o l m i n ~ ' l - ' ) at the start of recovery are
used to estimate the end-exercise rate of proton
efflux as
+ m)I/ + /?dpH/dt
(7)
from which the effective proton efflux rate constant
[12] is given by
i.= - E / ( ApH at end of exercise)
buffers as the pH rises; the calculation also takes
account of the small amount of net proton generation resulting from the oxidative generation of
co,.
Alternatively, since the initial rate of PCr resynthesis ( V ) is an easily-measured estimate of endexercise oxidative ATP synthesis rate, the endexercise proton efflux rate is calculated as
E =(2/3)U - [(2 - 3m)/3] V
(9)
which is used to calculate ias above. We used the
first of these methods (eqns. 7 and 8) to define the
defect of proton efflux, whose most visible effect is
the slow rate of PCr recovery (Fig. 3 4 . However,
the most reliable fact about the later phase of
exercise is that end-exercise Q should be equal to
the initial V [12, 14]), and so we have used the
second method (eqns. 8 and 9) to obtain a proton
efflux correction for the analysis of exercise in the
present work.
Mitochondria1 function in recovery from exercise
Also from recovery data, the maximum rate of
oxidative ATP synthesis (QmaX,)
is calculated from
the initial V and the end-exercise [ADP] as
Qmax. = V(1+ KnAADPI)
(10)
making use of the hyperbolic relationship between
oxidation rate and [ADP] and assuming a normal
K , of 30pmol/l [26, 28, 29). This evidently depends
on the end-exercise value of calculated [ADP],
which depends in turn on the assumed value of total
creatine; however, at the high values of [ADP]
observed in the present work this makes relatively
little difference.
REFERENCES
Proton efflux in recovery from exercise
E =(I)
589
(8)
In this analysis, E is estimated as the rate at which
protons are 'generated' by PCr resynthesis, plus the
rate at which they are 'liberated' from cellular
I. Holm J, Bjorntorp P, Scherston T. Metabolic activity in human skeletal muscle.
Effect of peripheral arterial insufficiency. Eur J Clin Invest 1972; 2: 321-5.
2. Sjostrom M. Anguist KA, Rais 0. Intermittent claudication and muscle fibre
fine structure: correlation between clinical and morphological data. Ultrastruct
Pathol 1980; I: 309-26.
3. Anquist KA, Sjostrom M. Intermittent claudication and muscle fibre fine
structure: morphometric data on mitochondria1 volumes. Ultrastruct Pathol
1980; I: 461-70.
4. Pernow B. Saltin B. Wahren J, Cronestrand R, Ekestrom S. Leg blood flow
and muscle metabolism in occlusive arterial disease of the leg before and after
reconstructive surgery. Clin Sci 1975; 49: 265-75.
5. Carlson L, Pernow B. Studies on the peripheral circulation and metabolism in
man: II. Oxygen utilization and lactate-pyruvate formation in the legs at rest
and during exercise in patients with arteriosclerosis obliterans. Acta Med
Scand 1962; 171: 311-23.
6. Hlavova A, Linhart J, Prerovsky I, Ganz V. Lactate and pyruvate changes in
the leg during and after exercise in normal subjects and in patients with
femoral artery occlusion. Clin Sci 1968; 34: 397-409.
7. Hands LJ, Bore PJ, Galloway G, Morris PJ, Radda GK. Muscle metabolism in
patients with peripheral vascular disease investigated by "P nuclear magnetic
resonance spectroscopy. Clin Sci 1986 71: 283-90.
8. Kemp GJ, Radda GK. Quantitative interpretation of bioenergetic data from "P
and 'H magnetic resonance spectroscopic studies of skeletal muscle: an
analytical review. Magn Reson Quart 1994; 10 43-63.
9. Mackitie J, Teravainen H. Histochemical changes in striated muscle in patients
with intermittent claudication. Arch Pathol Lab Med 1977: 101: 65B-63.
10. Henriksson 1, Nygarrd E, Anderson J, Eklof B. Enzyme activities, fibre types
and capillarization in calf muscles of patients with intermittent claudication.
Scand J Clin Lab Invest 1980; 40: 361-9.
590
G. J.
I I. Hammarsten L. Bylund-Fellenius AC. Holm J, Schersten T, Krotkiewski M.
Capillary supply and muscle fibre types in patients with intermittent
claudication: relationships between morphology and metabolism. Eur J Clin
Invest 1980; 10: 301-5.
12. Kemp GI. Thompson CH, Barnes PRJ, Radda GK. Comparisons of ATP
turnover in human muscle during ischaemic and aerobic exercise using "P
magnetic resonance spectroscopy. Magn Reson Med 1994 31: 248-58.
13. Kemp GI, Taylor Dj. Styles P. Radda GK. The production, buffering and efflux
of protons in human skeletal muscle during exercise and recovery. NMR
Biomed 1993: 6 73-83.
14. Boska M. Estimating the ATP cost of force production in the human
gastrocnemius/soleusmuscle group using I1PMRS and ' H MRI. NMR Biomed
1991; 4 173-81.
IS. Kemp GI, Thompson CH, Taylor DJ, Hands LJ, Rajagopalan 6, Radda GK.
Quantitative analysis by I ' P MRS of abnormal mitochondrial oxidation in
skeletal muscle during recovery from exercise. NMR Biomed 1993: 6 302-10.
16. Kemp GI, Thompson CH, Sanderson AL, Radda GK. pH control in rat skeletal
muscle during exercise, recovery from exercise and acute respiratory acidosis.
Magn Reson Med 1994 31: 103-9.
17. Foley JM, Meyer RA. Energy cost of twitch and tetanic contractions of rat
muscle estimated in situ by gated "P NMR. NMR Biomed 1993; 6 32-8.
18. Sanderson AL, Kemp GI. Thompson CH, Radda GK. Regulation of oxidative
and glycogenolytic ATP turnover in exercising rat skeletal muscle. Biochem
SOC Trans 1995; 823: 292.
19. Durnin JVGA, Womersley J. Body fat assessed from total body density and its
estimation from skinfold thickness; measurements in 481 men and women aged
from 16 to 72 years. Br J Nutr 1974 3 2 77-97.
20. Arnold DL, Matthews PM. Radda GK. Metabolic recovery after exercise and
the assessment of mitochondrial function in vivo in human skeletal muscle by
means of P-31 NMR. Magn Reson Med 1984; I: 307-15.
21. Bylund-Fellenius AC. Walker PM. Elander A, Holm 5. Holm J. Schersten T.
Energy metabolism in relation to oxygen partial pressure in human skeletal
muscle during exercise. Biochem J 1981: u)o: 247-55.
22. Veech RL, Lawson JWR, Cornell NW, Krebs HA. Cytosolic phosphorylation
potential. J Biol Chem 1979; 254: 6 5 3 H 7 .
Kemp et al.
23. Adam GR. Foley JM, Meyer RA. Muscle buffer capacity estimated from pH
changes during rest-tc-work transitions. J Appl Physiol 1990; 6 9 96B-72.
24. Kemp GI. Thompson CH. Taylor DJ, Radda GK. ATP production and
mechanical work in exercising skeletal muscle: a theoretical analysis applied t o
"P magnetic resonance spectroscopic studies of dialysed uraemic patients.
Magn Reson Med 1995; 33: 601-9.
25. Chasiotis D. The regulation of glycogen phosphorylase attd glycogen
breakdown in human skeletal muscle. Acta Physiol Scand 1983; 518 (Suppl.):
168.
26. Kemp GI, Taylor DJ. Radda GK. Control of phosphocreatine resynthesis
during recovery from exercise in human skeletal muscle. NMR Biomed 1993;
6 6672.
27. Thompson CH, Kemp GI. Sanderson AL, Radda GK. Skeletal muscle
mitochondrial function studied by kinetic analysis of postexercise resynthesis.
J Appl Physiol 1995; 78: 2131-9.
28. Kushmerick MI. Meyer RA, Brown TR. Regulation of oxygen consumption in
fast- and slow-twitch muscle. Am J Physiol 1992 263: C598-605.
29. Nioka S, Argov Z. Dobson GP. et al. Substrate regulation of mitochondrial
oxidative phosphorylation in hypercapnic rabbit muscle. j Appl Physiol 1992
12: 5216.
30. Gyulai L. Roth Z, Leigh IS. Chance 6. Bioenergetic studies of mitochondrial
oxidative phosphorylation using "Phosphorus NMR. J Biol Chem 1985; 260:
3947-54.
31. Griffiths JR. A fresh look at glycogenolysis in skeletal muscle. Biosci Rep 1981;
I: 595610.
32. Chasiotis D, Sahlin K, Hultman E. Regulation of glycogenolysis in human
muscle a t rest and during exercise. J Appl Physiol 1982 53: 70CIS.
33. Ren jM, Hultman E. Phosphorylase activity in needle biopsy samples - factors
influencing transformation. Acta Physiol Scand 1987; 133: 109-14.
34. Hedberg B. Angquist KA. Henriksson-Larsen K, Sjostrom M. Fibre loss and
distribution in skeletal muscle from patients with severe peripheral arterial
insufficiency. Eur J Vasc Surg 1989; 3: 315-22.
35. Elander A, Sjostrom M, Lundgren F. Schersten T, Bylund-Fellenius A t .
Biochemical and morphometric properties of mitochondrial populations in
human muscle fibres. Clin Sci 1985; 6 9 153-64.