Download Effects of Treadmill Exercise and High-Fat

Clinical Science (1995) 89, 447-452 (printed in Great Britain)
447
Effects of treadmill exercise and high-fat feeding on
muscle degeneration in mdx mice at the time of weaning
Asghar MOKHTARIANI, Jean Pascal LEFAUCHEUR2, Patrick C. EVEN' and Alain SEBILLP
'Laboratoire de Neurobiologie des Regulations, CNRS URA 1860, College de France, and
Laboratoire de Physiologie, DRED EA 278, Faculte de Medecine Saint-Antoine, Paris, France
(Received 31 October 1994/15 May 1995; accepted I June 1995)
1. Dystrophin-deficient hindlimb muscles of mdx
mice undergo necrosis at the time of weaning when
the motor activity of the mice greatly increases and
muscle energy metabolism becomes more dependent
on insulin and carbohydrates.
2. We have attempted to determine if the onset of
myofibre necrosis in mdx mice at the time of
weaning is related to the development of motor
activity and/or the change in diet.
3. Fourteen-day-old mdx mice were divided into two
groups after weaning. One group was trained to run
on a treadmill and the other group was kept on a
high-fat diet. Muscle necrosis was assessed histologically in the soleus and extensor digitorum longus
muscles of mice in both experiments.
4. Keeping mice on a high-fat milk diet from the
time of weaning up to 42 days of age did not
influence the occurrence of necrosis in the soleus and
extensor digitorum longus muscles of the mdx pups.
In contrast, treadmill exercise greatly increased necrosis in both muscles.
5. We conclude that an increase in motor activity
exacerbates the degeneration of hindlimb muscles of
mdx mice at the time of weaning.
INTRODUCTION
The mdx mouse, a mutant of the C57BL/10 strain
[1], presents a genetic deficiency of dystrophin [2],
like patients with Duchenne muscular dystrophy
(DMD) [3, 4]. As a result of this lack of dystrophin,
a large subsarcolemmal protein of muscle cells
[5, 6], mdx muscles undergo necrosis. In contrast to
DMD muscle, in which an extensive fibrosis develops early, necrosis in mdx muscles is followed by a
successful regeneration [7-10].
The onset of myofibre necrosis occurs in the
hindlimb muscles of mdx mice at approximately 3
weeks of age [7, 10, 11], i.e. at the time during
which two simultaneous events take place: (1) the
development of exploratory behaviour that greatly
increases motor activity, and therefore mechanical
as well as metabolic stress to skeletal muscles; (2) a
weaning-induced switch from a high-fat (milk) to a
low-fat (cube diet) food regimen, associated with the
progressive dependence of muscle metabolism on
glucose and insulin [12].
Dystrophin is thought to stabilize myofibre membranes [13, 14], and may minimize the damage induced by contractile activity in normal muscles. The
increased sarcolemmal permeability of dystrophindeficient muscle [15] could lead to an excessive
influx of calcium [16], and thus activation of
various proteases [17] that contribute to myofibre
degeneration. Consequently, mdx muscle may suffer
from an increased susceptibility to exercise-induced
injury [18-20], even if the basis for this hypothesis
is still not well established [21, 22].
On the other hand, the intracellular free calcium
accumulation could also be due to alterations in
intracellular organization of enzyme complexes
[23, 24], thus affecting the control of mitochondrial
function and energy metabolism, and resulting in a
reduction in active ion transport. This hypothesis is
supported by a reduced metabolic rate observed in
myofibres of adult mdx mice [25], a high glycogen
concentration [26, 27] and a decreased oxidation of
glucose, free fatty acids and various intermediates of
the tricarboxylic acid cycle [23].
In this study, we have attempted to determine if
the onset of myofibre necrosis in mdx mice at the
time of weaning is related to the enhancement of
motor activity or to the switch from a high-fat
(milk) to a low-fat (cube diet) food regimen.
METHODS
Animals
Mdx mice from the inbred colony of the Faculte
de Medecine Saint-Antoine (Paris, France) were
housed in large plastic cages (20cm x 27 em) in a
room kept at constant temperature (21°C) with a
natural night-day light cycle. They were fed a
commercial cube diet (A03 DAR) and water was
available ad libitum, except when indicated. Control
Key words: diet, dystrophin, eccentric contraction, mdx, myofibre type, necrosis, regeneration.
.. .
.
Abbreviations: OMO, Ouchenne muscular dystrophy; EOL, extensor digitorum longus; SOL, soleus; TA, tlblalls anterior.
Correspondence: Dr Asghar Mokhtarian, Laboratoire de Neurobiologie des Regulations, CNRS URA 1860, College de France, II place M. Berthelot, 75231 Paris cedex OS,
France.
448
A. Mokhtarian et al.
Table I. Surface area (mean ±SEM) occupied by necrotic and
regenerating myofibres, expressed as a percentage of whole crosssectional area, in soleus (SOL) and extensor digitorum longus
(EDL) muscles of 25-day~ld sedentaryand exercised mdx miceand
age-matched C57BL/IO mice. Significant difference between sedentary
and exercised mdx groups: ***p < 0.0002.
mice were from an inbred colony of the wild strain
C57BL/IO.
Treadmill running
The treadmill runnmg experiments were conducted on three groups of 14-day-old mice: (1) a
group of four sedentary mdx mice placed on a
motionless treadmill for 30 min every day, (2) a
group of four exercised mdx mice and (3) a group of
four exercised C57BL/1O mice.
The exercised mice ran for 30 min every day
(11.00 to 11.30 hours) for 7 days [28] on an uphill
home-made motorized treadmill (speed 4 m/min,
slope 15/~). The pups were always handled with
disposable gloves, and the separation was strictly
limited to the duration of the exercise to prevent
dismissal of the pups by the parents. During the
first 2 days, the pups were encouraged to run by
gently pushing them with the finger. Thereafter, all
the pups ran spontaneously. Electric shocks were
never used.
SOL
EDL
Sedentary mdx
Exercised mdx
Exercised CS7BL/10
20.8±ll
4.1 ± 1.0
67.5 ± 6.4'**
28.4 ± 2.8***
0.6±0.2
0.3±0.1
The high-fat diet consisted of 'first-age' human
synthetic milk (Enfalac, Mead-Johnson). The highfat feeding experiment was paired with a control
experiment and conducted on two litters of four
mdx pups. The first litter was given, up to 42 days
of age, only human synthetic milk [weight (g) per
100g: carbohydrates, 57.5; lipids, 30.4; proteins,
12.1]. The second litter was given for the same time
period the standard low-fat laboratory cube diet
[weight (g) per 100 g: carbohydrates, 63.6; lipids, 6.4;
proteins, 30].
independent observer, using manual planimetry. The
difference between these two measurements gave the
surface occupied by necrotic and regenerating myofibres and was expressed as a percentage of whole
cross-section area.
To determine the prevalence of necrosis in a
specific type of myofibre, cross-sections were also
processed for myosin-A'TPase activity [30]. The pH
of the preincubation medium was either 4.53 or
10.40, and all final incubations with ATP were
performed at pH 9.4 [31]. The number of type
I-lIB, IIA and IIC myofibres in both EDL muscles
of each mouse was assessed after preincubation at
pH 4.53; type I (slow oxidative) and lIB (fast
glycolytic) myofibres remained unstained, type IIA
(fast oxidative and glycolytic) myofibres were darkly
stained and type IIC (intermediate) myofibres exhibited distinct pale staining. The number of type I
and II myofibres in both SOL muscles of each
mouse was assessed after preincubation at pH 10.40;
type I myofibres remained unstained, and type IIA
and lIB myofibres were darkly stained.
The data are reported as means ± SEM. Statistical
differences were assessed using the Mann-Whitney
test.
Histological studies
RESULTS
The soleus (SOL) and the extensor digitorum
longus (EDL) muscles were removed bilaterally
under 3.5% chloral hydrate anaesthesia (0.35 ml per
animal intraperitoneally). In the high-fat feeding
experiments the muscles were removed when the
pups were 42 days of age. In the treadmill experiments they were removed when the mice were 25
days old, i.e. 4 days after the end of exercise, in
order to allow for optimal exercise-induced myofibre
necrosis [29]. The excised muscles were mounted in
a piece of cork with tragacanth gum, and frozen in
isopentane chilled by liquid nitrogen. Muscles were
cut on a cryostat all along their length. To ensure
the greatest cross-sectional area, quantification was
performed on whole-muscle cross-sections (10 /lm
thick) taken from the mid-point of the muscle body
[18].
The whole cross-sectional area and the total
surface area occupied by surviving normal myofibres (identified by the presence of peripheral
nuclei) were measured on microphotographs of
haernatoxylin-eosin-stained cross-sections by an
Effect of treadmill running
High-fat diet
Necrotic and regenerating myofibres were very
rare in SOL and EDL muscles of 25-day-old exercised control (C57BL/IO) pups (Table 1). On the
other hand, areas of necrotic and regenerating myofibres were found in SOL and to a lesser extent in
EDL muscles of 25-day-old sedentary mdx pups
(Table 1). Treadmill running enhanced mdx muscle
necrosis and resulted in a significant extension of
regenerating areas (P < 0.0002) in both SOLand
EDL muscles of 25-day-old exercised mdx mice
(Table 1). This extension may have been minimized
by the presence of early regenerating myofibres
(basophilic fibres with central nuclei), which are
small in diameter and were particularly observed in
foci of degeneration in exercised mdx EDL muscles.
Histological features in muscles of sedentary and
exercised mdx mice are presented in Fig. 1. Foci of
inflammatory cells were seen in SOL but not in
EDL muscles. We saw no localization in muscle
damage along the length of exercised muscles (Fig.
Exercise and high-fat feeding in mdx mice
449
Fig. I. CrosHection of soleus (SOL: a, b) and extensor digitorum longus (EDL: c, tI) muscles in 25-day-old sedentary (a, c) and treadmill-exercised (b,
tI) mdx mice. Haematoxylin-eosin staining ( x 250).
2), therefore, as others [18], we analysed a single
section taken from the mid-point of each muscle.
All types of muscle fibres seemed to be identically
affected by necrosis after treadmill running as characterized by ATPase reactions (Fig. 3). Although the
ratio of type I to type II fibres in SOL muscles after
pH lOA preincubation was decreased in exercised
mdx pups (- 25°0)' this difference did not reach
significance (Table 2). In EDL muscles, the ratio of
type IIA to type lIB fibres following pH 4.53
preincubation was not significantly modified (Table
2). Type IIC fibres were also discernible in EDL
muscles owing to their pale staining. In sedentary
mdx muscles, these fibres exhibited peripheral nuclei
and were probably immature, growing post-natal
myofibres. After treadmill running, type IIC fibres
exhibited central nuclei and were regenerated myofibres. In both groups, their ratio to the total
number of myofibres was similar (0.37 ±0.03 in
exercised mice versus 0.35 ± 0.03 in sedentary mice).
Effect of high-fat feeding
Maintenance of mdx pups on a high-fat milk diet
from weaning up to 42 days of age did not influence
muscle weight gain or the occurrence of necrosis in
SOL and EDL muscles of the pups. The extent of
necrotic and regenerating areas was similar (high-fat
diet, 43.2%; cube diet, 42.7%,
whatever the diet.
III
both muscles)
DISCUSSION
In this study, the onset of mdx muscle necrosis
was not modified by keeping mdx pups on a highfat diet after weaning. In contrast, muscle necrosis
was enhanced in mdx mice by treadmill exercise at
the time of weaning.
In mdx mice, hindlimb muscle necrosis occurs at
the time of weaning, i.e. at 3 weeks of age. By this
time, the pups exhibit an increase in motor activity
(e.g. exploratory behaviour, seeking food) and a
switch from a high-fat (maternal milk) to a low-fat
(cube diet) food regimen. Weaning therefore
increases muscle dependence on carbohydrate, and
exercise increases cell energy requirements. Both
processes could result in a metabolic stress for mdx
muscle cells, which are known to suffer from a
defective energy metabolism [25]. Muscle necrosis
could also result from an increased influx of calcium
into mdx myofibre sarcoplasm, due to either an
altered ion channel function [32, 33] or membrane
breakings resulting from the absence of dystrophin
[15, 17]. The lack of beneficial effects of a food
regimen of higher fat content than the standard
cube diet (30% versus 6%) suggests that a defective
adaptation of muscle energy metabolism to the low-
450
A. Mokhtarian et al.
Fig. 2. Cross-section taken in the middle part of a soleus muscle in 25-day-old treadmill-exercised mdx mice. Distance between each section = 200/1 m. Total
distance between sections (a) to (e) = BOO 11m. Haematoxylin-eosin staining (x 100).
fat cube diet is not the leading cause of muscle cell
necrosis at the time of weaning. However. the
human synthetic milk used in this study contained
the same proportion of carbohydrate as the
standard cube diet. To determine then the exact role
played by the insulin-dependent metabolic processes
at the time of weaning, further experiments using
diets of higher fat content and lower carbohydrate
content are planned.
The paradigm of mdx muscle susceptibility to
exercise-induced damage only results from studies
on adult mdx regenerated muscles. Most studies on
isolated adult muscles in a recording chamber have
reported a susceptibility to eccentric activity of mdx
muscles. An increased percentage of damaged myo-
fibres resulting from eccentric, isometric contractions, or passive lengthenings, was shown in diaphragm and EDL muscles of 90- to l l O-day-old
mdx mice [14]. Furthermore, an increased force
drop was induced by contractions with stretch in
EDL and SOL muscles of mdx mice of unknown
age [20]. However, the maximal tetanic force after
lengthening contractions during tetanic contraction
was altered in EDL but not in SOL muscles of 180to 400-day-old mdx mice [34]. The effects of
increased motor activity in mdx mice are more
controversial. Long-term overload by removing the
synergist tibialis anterior (T A) led to a progressive
EDL weakness in 60- to 240-day-old mdx mice [19].
Susceptibility to necrosis was increased in the T A
Exercise and high-fat feeding in mdx mice
451
Fig. 3. Cross-section of soleus (SOL: a, b) and extensor digitorum longus (EDL: c, d) muscles in 25-day-old sedentary (a, b) and treadmill-exercised (b,
d) mdx mice. ATPase activity staining following preincubation at pH 10.40 (a, b) and at pH 4.53 (c, d) ( x 250).
Table 2. Ratios (mean± SEM) of type I to type II myofibres in
soleus (SOL) muscle and of type IIA to type liB myofibres in
extensor digitorum longus (EDL) muscles of 25-day-old sedentary
and exercised mdx mice. Statistical comparisons between groups are not
significant.
SOL
EDL
Sedentary mdx
Exercised mdx
0,34±0.08
1.72±0.23
0.25 ±0.05
1.61±O.16
muscle following lengthening (eccentric) contractions
induced by stimulation of the sciatic nerve in 100day-old mdx mice [18]. However, force loss in the
TA muscle after lengthening contractions induced
by stimulation of the peroneal nerve was similar in
110- to 180-day-old mdx mice and age-matched
C57BL/10 mice [21]. In addition, endurance swimming has been claimed to have beneficial effects on
the contractile properties of EDL and SOL muscles
in 35- to 140-day-old mdx mice [22].
Thus, there is some discrepancy regarding the
effects of exercise on mdx muscle, even concerning
eccentric activity [18, 21], but our study was in
some respects different from these previous studies.
First, we assessed the extent of myofibre necrosis in
young mdx mice at the first round of muscle
degeneration, not in adult mdx mice in which
muscle regeneration had already occurred. Second,
our exercise protocol did not investigate the effects
of a specific type of contraction (eccentric or concentric, as with electrical stimulation procedures
[18,21]), and is different from non-weight-bearing
endurance training with few eccentric contractions,
such as swimming [22]. To our knowledge,
unweaned 15-day-old mdx pups have never been
exercised using an uphill treadmill. Since no significant necrosis was found in exercised C57BL/1O
pups, our exercise protocol was adapted to young
animals which run spontaneously without collapsing. Compared with sedentary mdx muscles, treadmill running increased the mean surface area occupied by necrotic (SOL) or regenerating (EDL)
myofibres. The prejudicial effect of exercise mimics
or exacerbates the spontaneous degeneration, which
is more evolved in SOL muscles than in EDL
muscles in sedentary young mdx mice. Muscle regeneration was allowed to take place in EDL but
not in SOL muscles, which exhibited constantly
necrotic processes. This predominant involvement in
SOL, which is a postural muscle, could be due
either to the permanent contractile activity of this
muscle or to the presence of some type I myofibres.
In our study, the ratio of type I to type II myofibres
in exercised mdx SOL muscles was decreased. Type
II (fast glycolytic) myofibres were previously shown
to have an increased susceptibility to damage
induced by eccentric contraction compared with
type I (slow oxidative) myofibres [34]. During
452
A. Mokhtarian et al.
uphill exercise, SOL muscles probably undergo
fewer eccentric contractions than during lengthening
or downhill running, but are subjected to more
intense motor activity than EDL muscles. Therefore,
in young mdx mice, the importance of the exerciseinduced injury may be related to the increase in
motor activity rather than to the type of exercise.
In EDL muscle of sedentary 25-day-old mdx mice,
we observed 37% of myofibres to be of intermediate
type (type IIC). These myofibres were mostly immature, growing and peripherally nucleated [35].
In exercised mdx EDL muscles, the percentage
of type IIC myofibres was similar, but these cells
were regenerating centronucleated myofibres. Thus,
immature, growing intermediate myofibres disappeared from EDL muscles after exercise as a result
of either an increased susceptibility of these fibres to
necrosis or an accelerated maturation in type II
myofibres.
In conclusion, we suggest that the onset of necrosis in the hindlimb muscles of mdx mice at the
time of weaning is mainly related to or exacerbated
by the increase in motor activity. Although we
should be careful in drawing conclusions regarding
DMD pathophysiology on the basis of such experiments in small quadrupedal animals, these results
suggest that dystrophin-deficient muscle necrosis
could be partially controlled by restraining intensive muscle exercise. However, contraction-induced
muscle damage cannot be avoided in incessantly
working muscles.
ACKNOWLEDGMENTS
We thank Dr S. Thornton for his help with the
English, Mrs M. J. Meile and J. Chandellier for
technical assistance and Mrs N. Ouvrard and Mr P.
Casanovas for animal care. This work was supported by the Association Francoise contre fes
Myopathies (A.M., P.C.E., A.S.) and the Ministere de
fa Recherche et de fa Technofogie, Paris, France
(J.-P.L.).
REFERENCES
I. Bulfield G, Siller WG, Wight PAL. Moore KJ. X-chromosome linked muscular
dystrophy (mdx) in the mouse. Proc Natl Acad Sci USA 1984; 81: 1189-92.
2. Sicinski P, Geng Y, Ryder-Cook AS, Barnard EA, Darlinson MG, Barnard PJ.
The molecular basis of muscular dystrophy in the mdx mouse: a point
mutation. Science (Washington DC) 1989; 244: IS78.'l0.
3. Koenig M, Monaco AP, Kunkel LM. The complete sequence of dystrophin
predicts a rod-shaped cytoskeletal protein. Cell 1988; 53: 219-28.
4. Hoffman EP, Kunkel LM. Dystrophin abnormalitles in DuchennefBecker
muscular dystrophy. Neuron 1989; 2: 1019-29.
5. Ahn AH, Kunkel LM. The structural and functional diversity of dystrophin.
Nature Genet 1993; 3: 283-91.
6. Ervasti JM, Campbell KP. Dystrophin and the membrane skeleton.
Curr Opin Cell Bioi 1993; 5: 82-7.
7. Dangain J, Vrbova G. Muscle development in mdx mutant mice. Muscle Nerve
1964; 7: 7(){)-4.
8. Anderson JE, Ovalle WK, Bressler BH. Electron microscopic and
autoradiographic characterization of hindlimb muscle regeneration in the mdx
mouse. Anat Rec 1987; 219: 243-57.
9. Anderson JE, Bressler BH, Ovalle WK. Functional regeneration in the hindlimb
skeletal muscle of the mdx mouse. J Muscle Res Cell Motil 1988; 9: 499-515.
10. Coulton GR, Morgan JE, Partridge TA, Sloper [C. The mdx mouse skeletal
muscle myopathy, I. A histopathological, morphometrical and biochemical
investigation. Neuropathol Appl Neurobiol 1988; 14: 53-70.
II. Torres LFB, Duchen LW, The mutant mdx: inherited myopathy in the mouse.
Morphological studies of nerves, muscles and end-plates. Brain 1987; 110:
269-99.
12. Girard J, Ferre P, Pegorier JP, Duee PH. Adaptations of glucose and fatty acid
metabolism during perinatal period and suckling-weaning transition.
Am J Physiol 1992; 72: 507~2.
13. Beam KG. Duchenne muscular dystrophy, localising the gene product. Nature
(London) 1988; 333: 798-9.
14. Petrof BJ, Shrager JB, Stedman HH, Kelly AM, Sweeney HL. Dystrophin
protects the sarcolemma from stresses developed during muscle contraction.
Proc Natl Acad Sci USA 1993; 90: 3710-4.
15. Menke A,Jockusch H. Decreased osmotic stability of dystrophin-Iess muscle
cells from the mdx mouse. Nature (London) 1991; l49: 69-71.
16. Turner PR, Westwood T, Regen CM, Steinhardt RA. Increased protein
degradation results from elevated free calcium levels found in muscle from
mdx mice. Nature (London) 1988; 335: 735--8.
17. Turner PR, Fong P, Denetclaw WF, Steinhardt RA. Increased calcium influx in
dystrophic muscle. J Cell Bioi 1991; 115: 1701-12.
18. Weller B, Karpati G, Carpenter S. Dystrophin-deficient mdx muscle fibers are
preferentially vulnerable to necrosis induced by experimental lengthening
contractions. J Neurol Sci 1990; 100: 9-13.
19. Dick J, Vrbova G. Progressive deterioration of muscles in mdx mice induced
by overload. Clin Sci 1993; 84: 145-50.
20. Moens P, Baatsen PHWW, Manichal G. Increased susceptibility of EDL muscles
from mdx mice to damage induced by contractions with stretch.
J Muscle Res Cell Motil 1993; 14: 446--51.
21. Sacco P, Jones DA, Dick JRT, Vrbova G. Contractile properties and
susceptibility to exercise-induced damage of normal and mdx mouse tibialis
anterior muscle. Clin Sci 1992; 82: 227-36.
22. Hayes A, Lynch GS, Williams DA. The effects of endurance exercise on
dystrophic mdx mice. I. Contractile and histochemical properties of intact
muscles. Proc RSoc Lond Bioi 1993; 253: 19-25.
23. Chinet AE, Even Pc. Decrouy A. Dystrophin-dependent efficiency of metabolic
pathways in mouse skeletal muscles. Experientia 1994; 50: 602-5.
24. Even PC, Decrouy A, Chinet A. Defective regulation of energy metabolism in
mdx mouse skeletal muscle. Biochem J 1994; 304: 649-54.
25. Decrouy A, Even PC, Chinet A. Decreased rates of Ca' +-dependent heat
production in slow- and fast-twitch muscles from the dystrophic (mdx) mouse,
Experientia 1993; 49: 843-9.
26. Cullen MJ, Jaros E. Ultrastructure of the skeletal muscle in the X
chromosome-linked dystrophic (mdx) mouse: comparison with Duchenne
muscular dystrophy. Acta Neuropathol 1988; 77: 69-a1.
27. MacLennan PA, McArdle A, Edwards RHT. Acute effects of phorbol esters on
the protein-synthetic rate and carbohydrate metabolism of normal and mdx
mouse muscles. Biochem J 1991; 275: 477-a3.
28. Fowler MF, Abresch RT, Larson DB, Sharman RB, Entrikin RK. High-repetitive
submaximal treadmill exercise training: effect on normal and dystrophic mice.
Arch Phys Med Rehabil 1990; 71: 552-7.
29. McNeil PL, Khakee R. Disruptions of muscle fiber plasma membranes: role in
exercise-induced damage. Am J Pathol 1992; 140: 1097-109.
30. Dubowitz V. Muscle biopsy: a practical approach. London: Bailiere Tindall,
1985.
31. Brooke MH, Kaiser KK. Muscle fiber types: how many and what kind?
Arch Neurol 1970; 23: 369-379.
32. Fong P, Turner PR, Denetclaw WF, Steinhardt RA. Increased activity of
calcium leak channels in myotubes of Duchenne human and mdx mouse origin.
Science (Washington DC) 1990; 250: 673~.
33. Franco AJ, Lansman JB. Calcium entry through stretch-inactivated ion channels
in mdx myotubes. Nature (London) 1990; 344: 670-3.
H. Head 51, Williams DA, Stephenson DG. Abnormalities in structure and
function of limb skeletal muscle fibres of dystrophic mdx mice.
Proc RSoc Lond Bioi 1992; 248: 163-9.
35. Pastoret C, Sebille A. Fibres of intermediate type IC and 2C are found
continuously in mdx soleus muscle up to 52 weeks. Histochemistry 1993; 100:
271~.